1
DEPARTMENT OF HEALTH AND HUMAN
SERVICES
FOOD AND DRUG
ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION
AND RESEARCH
BIOLOGICAL RESPONSE MODIFIERS
ADVISORY COMMITTEE
MEETING #37
Thursday, March 18,
2004
8:30 a.m.
Hilton Hotel
Silver Spring,
Maryland
2
PARTICIPANTS
Mahendra S. Rao, M.D., Ph.D., Chair
Gail Dapolito, Executive Secretary
MEMBERS
Jonathan S. Allan, D.V.M.
Bruce R. Blazar, M.D.
David M. Harlan, M.D.
Katherine A. High, M.D.
Joanne Kurtzberg, M.D.
Alison F. Lawton
James J. Mul, Ph.D.
Thomas H. Murray, Ph.D.
Anastasios A. Tsiatis, Ph.D
CONSULTANTS
Jeffrey S. Borer, M.D.
Susanna Cunningham, Ph.D.
Jeremy N. Ruskin, M.D.
Michael E. Schneider, M.D.
Michael Simons, M.D.
INDUSTRY REPRESENTATIVE
John F. Neylan, M.D.
GUEST HEALTH CANADA REPRESENTATIVE
Norman Viner, M.D.
GUEST SPEAKERS
Stephen Epstein, M.D.
Silviu Itescu, M.D.
Robert J. Lederman, M.D.
Philippe Menasch, M.D.
Emerson C. Perin, M.D., F.A.C.C.
Doris A. Taylor, Ph.D.
NIH PARTICIPANTS
Richard O. Cannon, M.D.
Stephen M. Rose, Ph.D.
FDA PARTICIPANTS
Jesse L. Goodman, M.D., M.P.H.
Philip Noguchi, M.D.
Dwaine Rieves, M.D
Stephen Grant, M.D.
Richard McFarland, Ph.D., M.D.
Donald Nick Jensen, D.V.M., M.S.E.E.
3
C O N T E N T S
PAGE
Call to Order
Mahendra Rao, M.D., Ph.D., Chair 5
Conflict of Interest Statement
Gail Dapolito, Executive
Secretary 5
Introduction of Committee 9
FDA Opening Remarks
Presentation of Certificate of
Appreciation
to Retiring Member
Jesse Goodman, M.D., M.P.H. 14
Philip Noguchi, M.D. 16
Open Committee
Discussion
Cellular Therapies for Cardiac Disease
FDA Introduction and Perspectives
Dwaine Rieves, M.D. 18
Guest Presentations
Overview Cardiomyopathy and Ischemic
Heart
Disease
Emerson Perin, M.D., Ph.D. 35
Q&A 65
Clinical Experience of Autologous
Myoblast
Transplantation
Philippe Menasch, M.D. 85
Q&A 115
Bone Marrow Cell Therapy for
Angiogenesis:
Present and Future
Steven Epstein, M.D. 128
Q&A 148
Cellular Therapies for Cardiac Disease
Richard McFarland, Ph.D., M.D. 159
Guest Presentations
Myoblasts: The First Generation Cells for
Cardiac Repair: What Have We Learned?
Doris Taylor, Ph.D. 169
Q&A 202
Preclinical Models - Hematopoietic and
Mesenchymal Cell Therapies for Cardiac
Diseases
Silviu Itescu, M.D. 219
Q&A 245
4
C O N T E N T S
(Continued)
From Mouse to Man: Is it a Logical Step for
Cardiac Repair?
Doris Taylor, Ph.D. 257
Q&A 275
Cardiac Catheters for Delivery of Cell
Suspension
Donald Nick Jensen, D.V.M.,
M.S.E.E. 292
Transcatheter Myocardial Cell
Delivery: Questions
and Considerations from the Trenches
Robert Lederman, M.D. 307
Q&A 333
Open Public Hearing 343
5
1 P R O C E E D I N G S
2 Call to Order
3 DR. RAO: Good
morning. Welcome to the
4
37th meeting of the Biological Response Modifiers
5
Advisory Committee.
6 Today's topic, as you all know, is related
7
to use of cells in cardiovascular disorders, and we
8
have a pretty full schedule for the next couple of
9
days, but before we can start the meeting, we have
10
to have a few sort of committee stuff that needs to
11
be gotten through, so I will turn the mike over to
12
Gail, so that she can make the mandatory
13
announcements.
14 Conflict of Interest Statement
15 MS. DAPOLITO:
Good morning.
16 The following announcement addresses
17
conflict of interest issues associated with this
18
meeting of the Biological Response Modifiers
19
Advisory Committee on March 18 and 19, 2004.
20 Pursuant to the authority granted under
21
the Committee Charter, the Associate Commissioner
22
for External Relations, FDA, appointed Drs. Jeffrey
23
Borer and Susanna Cunningham as temporary voting
24
members.
25 In addition, the Director of FDA's Center
6
1
for Biologics Evaluation and Research, appointed
2
Drs. Jeremy Ruskin, Michael Schneider, and Michael
3
Simons as temporary voting members.
4 Based on the agenda, it was determined
5
that there are no specific products considered for
6
approval at this meeting. The
committee
7
participants were screened for their financial
8
interests. To determine if any
conflicts of
9
interest existed, the agency reviewed the agenda
10
and all relevant financial interests reported by
11
the meeting participants.
12 The Food and Drug Administration prepared
13
general matters waivers for participants who
14
required a waiver under 18 U.S.C. 208.
Because
15
general topics impact on many entities, it is not
16
prudent to recite all potential conflicts of
17
interest as they apply to each member.
18 FDA acknowledges that there may be
19
potential conflicts of interest, but because of the
20
general nature of the discussions before the
21
committee, these potential conflicts are mitigated.
22 We note for the record that Dr. John
23
Neylan is participating in this meeting as a
24
non-voting industry representative acting on behalf
25
of regulated industry. Dr.
Neylan's appointment is
7
1
not subject to 18 U.S.C. 208. He
is employed by
2
Wyeth Research and thus has a financial interest in
3
his employer.
4 With regards to FDA's invited guest
5
speakers and guests, the agency determined that
6
their services are essential.
The following
7
disclosures will assist the public in objectively
8
evaluating presentations and/or comments made by
9
the participants.
10 Dr. Stephen Epstein is the Executive
11
Director, Cardiovascular Research Institute,
12
Washington Hospital Center. He
receives research
13
support, is a consultant to and has financial
14
interests with, firms that could be affected by the
15
committee discussions.
16 Dr. Philippe Menasch is employed at the
17
George Pompidou Hospital in Paris, France. He has
18
an association with a firm that could be affected
19
by the committee discussions.
20 Dr. Emerson Perin is employed by the Texas
21
Heart Institute. He receives
consultant fees from,
22
and is a scientific advisor to, firms that could be
23
affected by the committee discussions.
24 Dr. Doris Taylor is employed by the
25
University of Minnesota, Center for Cardiovascular
8
1
Repair. She receives consultant
fees from a firm
2
that could be affected by the committee
3
discussions.
4 Dr. Norman Viner is employed by the
5
Biologics and Radiopharmaceuticals Evaluation
6
Centre, Biologics and Genetic Therapies
7
Directorate, Health Canada, in Ottawa, Canada.
8 FDA participants are aware of the need to
9
exclude themselves from the discussions involving
10
specific products or firms for which they have not
11
been screened for conflicts of interest. Their
12
exclusion will be noted for the public record.
13 With respect to all other meeting
14
participants, we ask in the interest of fairness
15
that you state your name, affiliation, and address
16
any current or financial involvement with any firm
17
whose product you wish to comment upon.
18 Waivers are available by written request
19
under the Freedom of Information Act.
20 Thank you, Dr. Rao.
21 DR. RAO: Now
you know why I always have
22
Gail read that statement.
23 Before we start any committee work, I
24
would like to welcome two new members to the
25
committee, Dr. Murray and Dr. James Mul. We
9
1
generally introduce everyone on the committee
2
first, and we generally go in alphabetical order,
3
but this time I will try and start with the new
4
members, so that they can tell us a little bit
5
about themselves before we have the others
6
introduce themselves.
7 Introduction of Committee
8 DR. MULE: I am
Dr. Jim Mul. I am
9
currently the Associate Center Director for the H.
10
Lee Moffitt Cancer Center in Tampa.
I oversee all
11
translational research at the Center including all
12
cell-based therapies for the treatment of cancer as
13
it applies to the clinical treatment of patients
14
with advance tumors.
15 Prior to being in Tampa since September of
16
last year, I was at the University of Michigan
17
Cancer Center for 10 years, and prior to that, the
18
NCI for another 10 years, and I am delighted to be
19
here.
20 DR. MURRAY:
Good morning. I am Tom
21
Murray. I am President of the
Hastings Center,
22
which is celebrating its 35th years as the world's
23
first research institute devoted to ethics in
24
medicine and the life sciences.
25 I spent 15 years as professor at medical
10
1
schools including 12 at Case Western Reserve
2
University School of Medicine.
My interests are
3
fairly broad. I write a lot
about ethics and
4
ethics in the life science and science policy.
5 Thank you. I
am delighted also to be
6
here.
7 DR. RAO: If we
can go down the table, Dr.
8
Tsiatis.
9 DR. TSIATIS:
Hi. I am Butch Tsiatis. I
10 am
from the Department of Statistics at North
11
Carolina State University.
12 DR. BORER: My
name is Jeff Borer. I am a
13
cardiologist. I work at Weill
Medical College of
14
Cornell University in New York City.
I run a
15 division
and an institute at Cornell and, relevant
16
to this meeting, I am the Chairman of the
17
Cardiorenal Drugs Advisory Committee of the FDA.
18 DR. CUNNINGHAM:
Good morning. My name is
19
Susanna Cunningham. I am a professor
in the School
20
of Nursing at the University of Washington in
21
Seattle, and I am the consumer representative for
22
the Cardiovascular Renal Advisory Committee.
23 DR. SCHNEIDER:
I am Michael Schneider. I
24
co-direct the Center for Cardiovascular Development
25
at Baylor College of Medicine, and our interests
11
1
are in the molecular genetics of cardiac muscle
2
formation, cardiac growth, cardiac cell apoptosis
3
and its relation to heart failure, and, relevant to
4
this meeting, cardiac progenitor cells of different
5
kinds.
6 DR. SIMONS:
Hi. I am Michael Simons. I
7
am Chief of Cardiology at Dartmouth Medical School.
8 I
work in the area of vascular biology, gene and
9
cell therapy.
10 DR. RUSKIN:
Good morning. I am Jeremy
11
Ruskin. I am a cardiologist and
12
electrophysiologist, and I direct the
Cardiac
13 Arrhythmia
Service at Massachusetts General
14
Hospital.
15 DR. NEYLAN:
Good morning. I am John
16
Neylan. I am a nephrologist and
an organ
17
transplanter by training. Currently, I am Vice
18
President of Clinical Research and Development at
19
Wyeth, and I serve as a industry representative to
20
the committee.
21 DR. KURTZBERG:
Hi. I am Joanne
22
Kurtzberg. I am a pediatric
oncologist. I direct
23
the Pediatric Bone Marrow and Stem Cell Transplant
24
Program at Duke University and the Carolinas Cord
25
Blood Bank at Duke.
12
1 DR. ALLAN:
Hi. I am Jon Allan. I am a
2
virologist at the Southwest Foundation for
3
Biomedical Research. My area is
nonhuman primate
4
models for AIDS pathogenesis.
5 DR. CANNON:
Good morning. I am Richard
6
Cannon. I am at the National
Heart, Lung, and
7
Blood Institute. I am Clinical
Director of NHLBI,
8
and I am representing NHLBI at this meeting.
9 DR. ROSE: Good
morning. I am Stephen
10
Rose. I am Deputy Director for
the Recombinant DNA
11
Program in the Office of Biotechnology Activities
12
in the NIH.
13 DR. JENSEN:
Good morning. My name is
14
Nick Jensen. I am a reviewer in
the Center for
15
Devices and Radiological Health.
I am a
16
veterinarian and an engineer.
17 DR. McFARLAND:
Good morning. I am
18
Richard McFarland. I am a
reviewer in the
19
Pharm/Tox Branch in the Center for Biologics in the
20
Office of Cellular, Tissue and Gene Therapies.
21 DR. RIEVES:
Good morning. My name is
22
Dwaine Rieves. I am a medical
officer in FDA's
23
Center for Biologics Evaluation and Research.
24 DR. GOODMAN:
Good morning. I am Jesse
25
Goodman. I am the Center
Director of the Center
13
1
for Biologics. I would just like
to join in
2
welcoming especially the new members.
My
3
background is as an infectious disease physician in
4
academic medicine for many years.
5 DR. NOGUCHI: I
am Phil Noguchi, Acting
6
Director of the Office of Cellular, Tissue and Gene
7
Therapies in CBER.
8 DR. RAO: Thank
you, everyone.
9 We are very fortunate in having some
10
really leaders in the field come and present some
11
of the data which will be the basis of where we can
12
address some of the questions that have been raised
13
by the FDA.
14 I am going to ask them to just briefly
15
introduce themselves, as well.
16 DR. EPSTEIN: I
am Steve Epstein, a
17
cardiologist. I am head of the
Cardiovascular
18
Research Institute at the Washington Hospital
19
Center. We are involved in
vascular biology, gene,
20
and cell therapy.
21 DR. MENASCHE:
I am Philippe Menasch. I
22
am cardiac surgeon at the Hospital European George
23
Pompidou in Paris, France.
24 DR. PERIN:
Good morning. I am Emerson
25
Perin. I am an interventional
cardiologist and
14
1
Director of Interventional Cardiology at Texas
2
Heart Institute in Houston.
3 DR. TAYLOR:
Hi. I am Doris Taylor. I am
4 a
scientist. I just moved from Duke
University to
5
the University of Minnesota to head the Center for
6
Cardiovascular Repair.
7 DR. ITESCU:
Hi. I am Silviu Itescu. I
8
am Director of Transplantation Immunology at
9
Columbia Presbyterian, New York.
10 DR. RAO: I
would also like to welcome Dr.
11
Viner who is from Health Canada.
Health Canada has
12
been following a lot of what the FDA has been doing
13
and it is nice to have them there.
14 I would like to invite Dr. Goodman to make
15 a
statement.
16 FDA Opening Remarks
17
Presentation of Certificate of Appreciation
18 to Retiring Member
19 DR. GOODMAN:
My main purpose is to thank
20
Joanne Kurtzberg for I guess about four years of
21
service to the BRMAC. We really
appreciate that
22
tremendously. She has also
interacted with CBER
23
before that.
24 One of the reasons I really wanted to come
25
by this morning. Joanne is
rotating off this
15
1
committee. I know from
interactions both within
2
this committee and outside, and from all the
3
leadership and staff within CBER, just what a
4
tremendous advisor and asset Joanne has been for
5
FDA and for your various fields here.
6 Of course, she has mostly contributed very
7
extensively in her areas of hematopoietic stem
8
cells, et cetera, but she has also been a very
9
important thinker and discussant and contributor on
10
the whole range of other cellular therapies and
11
even gene therapy.
12 Please join me in thanking Joanne for her
13
service over these years. Also,
we like to say,
14 particularly CBER, that we are a family and that
15
nobody ever leaves it, and that we, just like a
16
family, we will keep asking for favors in the
17
future and probably causing grief in return.
18 Thanks so much, Joanne. We have a plaque
19
for her, of course.
20 [Applause.]
21 DR. GOODMAN: I
guess I will just turn it
22
over to Phil to just give a brief introduction for
23
the meeting, but just to say that, as I mentioned a
24
little while back about the islet cell therapies,
25
we, at FDA, are extremely excited about cellular
16
1
therapies and their potential, and I think nowhere
2
is some of that potential clearer, but also perhaps
3
more difficult to evaluate and help move forward
4
than in the area of cardiovascular disease whether
5
it is for ischemic disease or heart muscle disease
6
or trauma, et cetera, some of the uses where there
7
have been some very promising reports.
8 So, we think this is a very timely
9
meeting. It is very important to
get input about
10
how to go forward with efficient development of
11
those products, how to address some of the clinical
12
and safety issues, and how to hopefully make this
13
field positioned to realize its successes in the
14
most efficient manner and also help FDA get that
15
right to the extent that we all can based on
16
incomplete information.
17 Again, we really look forward to this. I
18
apologize, my usual schedule means I will be in and
19
out, but I really appreciate it.
20 Phil.
21 DR. NOGUCHI:
Thank you, Jesse, and, of
22
course, Dr. Kurtzberg, our sincere thanks for the
23
many years of service. Jesse is
absolutely right,
24
don't be surprised if the next meeting, you get a
25
funny call early in the morning.
17
1 This is one of our, in a way, continuing
2
series of dealing with things that seem really
3
wonderful and amazing when they come up, where
4
there is a lot of hope and there is perhaps a
5
little bit of hype, but what we have always found
6
over the years, and here I would like to just
7
acknowledge Dr. Rose in the Office of Biotechnology
8
Activities and the Recombinant DNA Committee, what
9
we have learned from them is that one of the best
10
ways that we have of really dealing with things
11
controversial and where there is both hope and
12
there is some trepidation about whether or not this
13
is actually going to work or not, is to bring
14
everyone together, put them in the same room.
15 Our continuing--and this really goes back
16
at least 25 years through the RAC and many years
17
for the BRMAC--is that when you get reasonable
18
people together who may have differing opinions
19
about things, but are presented the facts and the
20
realities, as well as the unknowns, we all
21
basically pretty much come out with the same
22
conclusion, and then we can make significant
23
progress in making these therapies not just
24
experimental, but a reality.
25 With that, what I would really like to
18
1
do, because we have such a full schedule, is now
2
turn it over to Dr. Rieves for the introduction.
3 DR. RAO: As
Dr. Rieves comes up to the
4
mike, I just want to remind people of a few simple
5
rules. Remember that when you
want to ask a
6
question, make sure that you are recognized. Use
7
the button. You will see that
the light comes on.
8
When you are done, just hit the button again to
9
switch it off, because otherwise, there is sort of
10 a
feedback loop and noise. Make sure you
identify
11
yourself when you ask questions.
12 Cellular Therapies for Cardiac Disease
13 FDA Introduction and Perspectives
14 DR. RIEVES:
Good morning. My name is
15
Dwaine Rieves. I am a medical
officer within FDA's
16
Center for Biologics Evaluation and Research. This
17
morning I am going to present a brief overview of
18
FDA's perspective on cellular products used in the
19
treatment of cardiac diseases.
20 As will be covered in a subsequent
21
presentation, certain cellular products, when
22
either perfused into the heart or directly injected
23
into heart muscle, are thought to be capable of
24
regenerating heart tissue and/or augmenting heart
25
function.
19
1 Consequently, these products may have
2
special utility in the treatment of heart failure
3
and certain other cardiac diseases.
Today and
4
tomorrow, we will discuss issues in the early
5
clinical development of these products.
6 [Slide.]
7 This talk is divided into three major
8
sections. First, I will cite the purpose in
9
convening this advisory committee.
Secondly, I
10
will provide a regulatory background on FDA's
11
understanding and activities within the realm of
12
clinical development of these products.
Finally, I
13
will introduce the major questions we have proposed
14
for discussion.
15 [Slide.]
16 Unlike many advisory committees where the
17
topics center around assessment of data associated
18
with a specific product or data related to a
19
specific regulatory concern, our purpose in
20
convening this committee is not to obtain
21
definitive regulatory advice, instead, FDA has
22
convened this committee to listen to, and learn
23
from, the voiced thoughts and perspectives with the
24
understanding that this information will enhance
25
our ability to promote the safe clinical
20
1
development of these products.
2 As you are aware, the clinical development
3
of cellular products is in its infancy and many
4
questions surround the very early stages of product
5
development. Consequently, our purpose today and
6
tomorrow is to stimulate a solid scientific
7
discussion of the major facets associated with the
8
very early clinical development of these products.
9 As noted here, we will focus upon three
10
major areas: manufacturing
aspects of the cellular
11
product, preclinical testing of the products, and
12
finally, items related to the early clinical
13 studies.
14 [Slide.]
15 What are the cellular products we will be
16
discussing? These products may
be broadly grouped
17
into two categories.
18 Firstly, those manufactured without
19
ex-vivo culture methodology, that is, the cells are
20
harvested from humans, processed, and then
21
delivered to a recipient without maintaining the
22
cells in culture for a period of time.
23 In general, these cells consist of bone
24
marrow mononuclear cells and certain peripheral
25
blood mononuclear cells, hematopoietic progenitor
21
1
cells that are variously referred to as stem cells,
2
cells thought to be capable of assuming phenotypic
3
characteristics of non-hematopoietic cells.
4 The second category consists of cells
5
that, following harvesting, are maintained in ex
6
vivo culture for a period of time before final
7 processing
and administration.
8 In general, these cells consist of those
9
derived from skeletal muscle tissue, cells
10
frequently referred to as myoblasts, and certain
11
bone marrow stromal cells, cells also referred to
12
as mesenchymal cells. Whether
these cultured cells
13
should be regarded as forms of stem cells is more
14
questionable than that for the hematopoietic
15
progenitor cells.
16 Lastly, as the slide notes, most of the
17
cellular products we will be discussing today and
18
tomorrow are of autologous origin.
19 [Slide.]
20 The many questions surrounding the
21
scientific basis for cellular product development
22
illustrate the very nascent nature of the field.
23
As we are probably all aware, there is almost no
24
precedent for the clinical development of products
25
intended to regenerate and/or augment disease
22
1
tissue.
2 The scientific data surrounding this field
3
are relatively new, such that the data are limited
4
in depth and the extent of replication.
Hence we
5
come to the table of clinical development with many
6 hypothetical
considerations and some, but
7
relatively limited background supportive data.
8 [Slide.]
9 Given these limitations, our discussions
10
today and tomorrow assume a scientific focus in
11
which certain insights and perspectives are
12
presented, and you, the committee members, will be
13
asked to share your thoughts.
Three points are
14
cited here.
15 First, we acknowledge that these thoughts
16
are all tentative and susceptible to revision based
17
on accumulating data.
18 Secondly, we are not requesting any
19
definitive assessment of data, and we note that the
20
data presented here today are within the public
21
arena, and have not undergone FDA vetting.
22 Finally, I reiterate an earlier comment,
23
that no specific cellular product discussed here is
24
under review with respect to regulatory
25
decisionmaking.
23
1 [Slide.]
2 This slide illustrates the
3
interconnectedness of clinical research and
4
regulatory paradigms. The
connecting link between
5
the two fields is the science.
Clinical research
6
generates the scientific background for clinical
7
development of cellular products and the scientific
8
background forms the major basis for our regulatory
9
paradigms.
10 [Slide.]
11 FDA is charged with many responsibilities,
12
but as cited here, two are especially relevant to
13
this discussion. Specifically, FDA's mission is to
14
promote and protect the public health by optimizing
15
pre-market product development and ensuring
16
sufficient post-marketing product monitoring.
17 The key word in these two statements is
18
"product." A notation that whereas we frequently
19
hear the terms transplant, graft, and procedure, we
20
need to think in terms of a cellular product, a
21
product that is manufactured, labeled, and
22
potentially marketed.
23 [Slide.]
24 A little over 10 years ago, FDA clarified
25
the regulatory basis for oversight of clinical
24
1
development programs for cellular products. In
2
general, this regulatory framework is the same as
3
that for the drugs and biologic products we
4
commonly recognize as marketed products.
5 Hence, the commonly cited biologic
6
product, drug, and device regulations applied to
7
the clinical development of these cellular
8
products, and the clinical studies must be
9
conducted under the purview of submission of a
10
investigational new drug application.
11 The last bullet on this slide reminds us
12
that clinical development programs may be divided
13
into early and late stages, with the late stages
14
focused upon the ascertainment of data definitive
15
to safety and efficacy, and the early stage, what
16
we are talking about today and tomorrow, focused
17
upon the ascertainment of exploratory safety and
18
bioactivity data.
19 That is, we hope to examine the nature and
20
extent of background data necessary to introduce
21
the cellular products into small, sample size,
22
Phase I clinical studies.
23 [Slide.]
24 As previously noted, the keystone
25
consideration in early clinical development is
25
1
safety. Specifically, we need to
ensure that the
2
tripod of product development is solid. That tripod
3
consists of manufacturing control and testing
4
information, sufficient preclinical testing
5
information, especially information that may inform
6
the design of a clinical study, and finally, the
7
clinical study itself.
8 The next few slides will cite each of
9
these three components.
10 [Slide.]
11 Cellular products must be manufactured in
12
some manner, that is, the cells must be harvested
13
and processed prior to administration to a
14
recipient. Manufacturing aspects
may be divided
15
among four major areas, three being shown on this
16
slide.
17 The top bullet notes that documents should
18
describe the cell source and reagents used in the
19
manufacturing process, such as growth factors,
20
sera, salt solutions and additives.
We need to be
21
confident that all the reagents used in the
22
manufacturing are of clinical or pharmaceutical
23
grade, or that if they are not pharmaceutical
24
grade, they are sufficient for human use.
25 One may envision many potential concerns
26
1
with these materials, such as the use of sera that
2
may contain infections agents, or the use of only
3
partially purified reagents that contain harmful
4
excipients.
5 Secondly, documents should describe the
6
procedures used in manufacturing, specifically
7
describing how cells are aseptically harvested,
8
isolated, and potentially selected.
9 For example, a distinct population of
10
cells may be selected based upon the presence of
11
certain cell surface markers, such as the CD34
12
antigen with the selection process involving
13
incubation with an antibody to CD34.
14 As we know, many investigational
15
antibodies have been developed to target cell
16
surface antigens, and we need to be confident that
17
these selection techniques are performed in a
18
reproducible and safe manner.
19 Additionally, documents should describe
20
the storage and tracking of the cellular products,
21
this being of special concern because certain
22
cellular products may be patient-specific products.
23 For example, measures must be in place to
24
ensure that for autologous products, the cellular
25
product is returned to the correct donor. Of
27
1
course, the cellular product needs to be labeled as
2
one for investigational use only.
3 The bullet at the bottom of this slide
4
emphasizes the importance of testing the cellular
5
product, an especially important concern since
6
cellular products cannot be sterilized in the same
7
manner as one might sterilize a drug product or a
8
device. Notable aspects of
testing include tests
9
for sterility, endotoxin, viability, enumeration,
10
or cell counting.
11 [Slide.]
12 The fourth component of manufacturing
13
information is product characterization as
14
highlighted here. When one
speaks of product
15
characterization, we are generally talking about
16
cellular phenotype and/or functional
17
characterization and the characteristics of the
18
product's final formulation.
19 For example, a product containing solely
20
CD34 positive cells in saline with no preservatives
21
or media. Product characterization is especially
22
important from a clinical perspective, because
23
failure to consistently manufacture a product makes
24
the clinical data virtually uninterpretable.
25 As noted here, the major aspects of
28
1
product characterization consist of a description
2
of identity, purity, and potency of the final
3
cellular product.
4 [Slide.]
5
Pre-clinical testing is the
second major
6
component of product development, and the major
7
aspects of this testing are cited here.
The top
8
bullet notes that consistent with the science, the
9
extent and depth of preclinical testing necessary
10
to support a clinical study is an evolving paradigm
11
and is a major topic for discussion at this
12
meeting. However, we generally
take the stance
13
that this preclinical testing paradigm should be
14
consistent with that used for other biological
15
products.
16 The last bullet notes another important
17
aspect of preclinical testing, the testing of the
18
product administration procedure.
19 This is especially important because many
20
cellular products involve injection directly into
21
heart muscle either through the epicardial surface
22
or the endocardial surface.
These techniques
23
represent inherent safety concerns that may be best
24
evaluated in animals prior to their use in humans.
25 As noted, all available catheters, whether
29
1
marketed or not, are regarded as investigational
2
with respect to administration of cellular
3
products.
4 [Slide.]
5 This slide highlights three aspects of
6
preclinical testing that will be the focus of the
7
preclinical questions tomorrow.
8 Firstly, the choice of the relevant
9
species is central to designing preclinical studies
10
with the major choices being between large animals,
11
such as pigs, versus small animals, such as mice,
12
as well as the choice between immunocompetent
13
animals where, for autologous products, the
14
cellular products would be the animal cells, not
15
human cells, or immunocompromised animals, where
16
the actual human cellular product may be tested.
17 Secondly, designing preclinical studies
18
raise questions of the choice of model, that is, a
19
disease model, such as ischemic heart disease
20
induced in the pig versus a healthy animal.
21 Lastly, preclinical concerns relate to
22
testing of the administration procedure itself,
23
such items as the impact of the catheter materials
24
upon cells, the potential for occlusion of
25
catheters by the cellular product, and the safety
30
1
concerns associated with manipulation of the
2
catheters in the heart.
3 [Slide.]
4 The third component of the clinical
5
development program for cellular products is the
6
clinical study. There are many
aspects of clinical
7
study design that could be discussed, but at this
8
meeting, we are focusing upon two, the first shown
9
here, that is, adverse event detection.
10 This slide highlights two aspects of
11
clinical study design that are frequently
12
engineered to optimize adverse event detection, the
13
evaluation plan with attention to the duration of
14
clinical follow-up, the frequency of evaluations,
15
and the extent or nature of these evaluations.
16 Secondly, the clinical study safety
17
monitoring plan may be optimized through the use of
18
close scrutiny of each study subject based upon the
19
sequential, not simultaneous, enrollment and
20
treatment of the subjects, as well as the
21
prespecifications of the types and numbers of
22
adverse events that should prompt interruption of
23
the study, that is, the study stopping rules.
24 Tomorrow, the committee will be asked to
25
discuss potential adverse events in these early
31
1
clinical studies, both the nature of the events and
2
ways to optimize the safety of the studies.
3 [Slide.]
4
This slide illustrates an
additional
5
clinical study design item that we will bring to
6
the committee, that is, a discussion of the
7
analysis of adverse events.
8 Exploratory clinical studies are, by their
9
nature, small sample size studies in which it is
10
often difficult or impossible to distinguish
11
treatment-related events from adverse events that
12
might occur in the natural history of the disease,
13
potential study design mechanisms that might help,
14
but certainly not resolve this issue are cited in
15
the bullets, design features that incorporate
16
randomization of subjects among groups, such that
17
comparisons may be made, the use of controls,
18
especially placebo controls, to make comparisons,
19
the use of masking or blinding to help lessen the
20
bias associated with concomitant therapies or
21
clinical care.
22 Tomorrow, the committee will be asked to
23
discuss mechanisms that might aid in adverse event
24
attribution.
25 [Slide.]
32
1 In this presentation, we have covered
2
three major topics. Firstly, we
have noted that
3
the focus of this meeting is upon a discussion of
4
the scientific aspects of early cellular product
5
development.
6 Secondly, we have noted the regulatory
7
precedent for the cellular products.
8 Finally, we come to the questions.
9 [Slide.]
10 This slide highlights the four major areas
11
of tomorrow's questions.
Specifically, questions
12
related to manufacturing, we will request a
13
discussion of the extent of safety testing and
14
characterization that should be performed prior to
15
the release of a cellular product for
16
administration to humans.
17 The second and third discussion areas are
18
especially critical and may consume the bulk of our
19
time, that is, the extent and nature of preclinical
20
testing necessary to support the introduction of a
21
cellular product into humans, testing that involves
22
questions related to the product itself, as well as
23 the delivery mechanism, the catheter.
24 Finally, we will pose clinical questions
25
centered around adverse event detection and
33
1
analysis with a discussion of the pros and cons
2
associated with the use of controls in these
3
studies.
4 [Slide.]
5 Our agenda is summarized on this slide.
6
As you can see, today, we have a series of invited
7
presentations by FDA staff and leading
8
investigators in the field, as well as the
9
opportunity for public presentations.
10 Tomorrow, we will have another opportunity
11
for public presentations followed by a discussion
12
of the questions.
13 [Slide.]
14 In closing, listed here are some documents
15
that are especially pertinent to our discussions.
16
All these documents are available at www.fda.gov
17
under the CBER sites, specifically the guidance
18 section.
19 The first document is entitled "Draft
20
Guidance for CMC Reviewers: Human Somatic Cell
21
Therapy Investigational New Drug Applications."
22
This document describes the types of information
23
FDA reviewers will examine following the submission
24
of an IND. Consequently, it
provides a very clear
25
description of the types of manufacturing
34
1
information that needs to be submitted with an IND
2
application.
3 The second document is from the
4
International Conference on Harmonization of
5
Regulatory Practices, and it is entitled "
6
Preclinical Safety Evaluation of
7
Biotechnology-derived Pharmaceutics," the S6
8
document.
9 This document is cited because it contains
10 a
paradigm that one may apply to cellular products.
11 Finally, the last bullet cites one of the
12
most useful guidances to sponsors and
13
investigators, the ICH Guideline on Good Clinical
14
Practice.
15 This guideline provides detailed
16
information on how to design and conduct a clinical
17
study, information presented in a simple to read,
18
yet relatively comprehensive format.
19 This concludes my presentation and I thank
20
you for your attention.
21 [Applause.]
22 DR. RAO:
Before we continue with the rest
23
of the presentations, I would like to just welcome
24
Dr. Harlan and ask him to introduce himself.
25 DR. HARLAN: I
apologize for being late,
35
1
but I am David Harlan, NIDDK. I
study
2
transplantation of islets and immunotherapies.
3 DR. RAO: Our
first speaker will be Dr.
4
Perin, whom you already were introduced to.
5 Guest Presentations
6
Overview Cardiomyopathy and Ischemic Heart Disease
7 DR. PERIN: I
want to thank you for the
8
invitation to be here to present to you today,
9
especially Dr. Grant, who has helped me put this
10
together in a way.
11 So, what I want to do here this morning,
12
the task that has been laid before me is that of in
13 a
way setting the stage or giving you a general
14
idea of the kinds of patients that we are treating.
15 Obviously, this is fundamental if we are
16
thinking about doing clinical trials.
It is very
17
important to understand the nature of the disease
18
in which these kind of therapies will frequently be
19
applied.
20 What I plan to do is talk about the
21
following topics. First, we will
start from the
22
beginning, define what heart failure is, look at
23
the scope of heart failure, talk a little bit about
24
the pathophysiology, look at some prognostic
25
markers, talk about the treatment to some extent
36
1
and that is important in terms of monitoring, and
2
then really work our way towards end stage heart
3
failure because that is where I think the focus of
4
most of the future clinical trials will likely be
5
initially, and finally, talk about adverse events,
6
which I think is a major concern, and the
7
monitoring of there adverse events.
8 Now, I know many of you are not
9
cardiologists, so hopefully, I can go from a level
10
where we are not getting too complicated, but not
11
too simple.
12 Starting with the definition of what heart
13
failure is. Firstly, heart
failure is a clinical
14
syndrome very simply defined by certain symptoms
15
and certain signs that come together.
These
16
symptoms are fatigue, shortness of breath, and
17
congestion, and these are translated on a physical
18
exam by being able to hear a third heart sound, the
19
patient manifesting peripheral edema, and jugular
20
venous distention.
21 If we start looking at this problem and
22
have a broad overview of this, first, I want to
23
show you a graph from the HOPE trial.
This is a
24
trial that was conducted in thousands of patients,
25
as you can see here, over 9,000 patients. It was a
37
1
study primarily of ramipril and vitamin E in
2
patients with hypertension over a long period of
3
time, involved a five-year follow-up.
4 But it is just very interesting, as we
5
start out looking at heart failure, to look at this
6
patient population, and here we have over 500 days,
7
so here is about a year out, and if we look at this
8
population, who is not primarily designated as
9
particularly sick or harboring heart failure, that
10
identified the patients that did have heart failure
11
and we look at their survival, you will see the
12
mortality.
13 It separates from the beginning, and when
14
we get out to about a year, you have got a 10
15
percent mortality in the group that has heart
16
failure compared to less than 4 percent mortality
17 in the general population.
So, you can see that
18
the problem that we are dealing with seems to be
19
very serious.
20 If we go here and let's just look at the
21
placebo arms of some very large heart failure
22
trials, these are trials pretty much aimed at
23
evaluating different forms of therapy now in heart
24
failure patients, and looking at different severity
25
of heart failure patients, for example, in the
38
1
V-HeFT trial, inclusion criteria might be an
2
ejection fraction less than 40 percent.
3 If we look at PRAISE, which evaluated
4
amlodipine in more severe heart failure, an
5
ejection fraction was less than 30 percent,
6
comparing this with Class III and Class IV
7
patients, very sick patients.
8 So, you can see here if we look at just
9
the placebo arms of all these trials, a very
10
striking mortality as we go along.
If we look at 1
11
year here, this will vary from 10 percent down to
12
around 30 percent.
13 If we go out to 2 years in the very sick
14
patients, we see that half of the patients are
15
dead. So, heart failure,
depending on the
16
presentation, carries a very ominous prognosis.
17 It is a very broad problem, 5 million
18
Americans are living with heart failure now,
19
550,000 new cases are diagnosed each year.
20 From 1979 to 2000, heart failure deaths
21
increased by 148 percent. Now,
what is
22
interesting, over this period of time, we have
23
actually gotten a lot better at treating heart
24
failure, and we do treat it. I will get into this a
25
little later, and I will show you the modern treat
39
1
of heart failure and how much better we are doing,
2
but at the same time that we are treating heart
3
failure better, we are also treating the patients
4
that have coronary disease, which is a very
5
dominant problem in this country and around the
6
world, we are treating those
patients better, too,
7
so what happens is we are getting more patients
8
with heart disease that normally would have died
9
earlier, to live longer, and as we are able to
10
bypass and stent and do all these revascularization
11
procedures and come up with better treatments, we
12
are getting people that go further down the road,
13
that otherwise would have succumbed a long time
14
ago.
15 So, despite our improvements in treatment
16
of coronary disease, we are dealing with an
17
increasing amount of heart failure deaths.
18
In individuals diagnosed with
heart
19
failure, cardiac death occurs at 6 to 9 times the
20
rate in the general population.
If you are more
21
than 40 years old, you have a 1 in 5 chance of
22
developing heart failure, and 22 percent of men and
23
46 percent of women that have heart attacks will be
24
disabled within 6 years with heart failure.
25 So, as you can imagine, the high
40
1
prevalence and multiple complications have an
2
implication in terms of health costs.
If we look
3
at the costs, and these numbers vary, and it
4
depends on what you are looking at and what year
5
you are looking at, but this is a very significant
6
financial burden on the country, over 5 percent of
7
the total health care costs.
8 You can see that most of the cost involved
9
is really involved in inpatient care, and as I will
10
show you hopefully, that really translates to the
11
sickest portions of these patients, that as you get
12
sicker with heart failure, you start coming into
13
the hospital more, and that is what runs up the
14
cost of treating these patients.
It is interesting
15
that transplant is just a little sliver out of the
16
pie here.
17 So, let's look at the causes of heart
18
failure, and I am not going to get into all the
19
little minor details, but let's look at the major
20
causes of what brings on heart failure.
21 Seventy-five percent of people that go on
22
to develop heart failure had hypertension
23
previously. Valvular heart
disease is a big
24
contributor and also heart failure engenders
25
valvular heart disease, mitral regurgitation
41
1
further contributes to the problem.
2 Coronary artery disease, you are all
3
familiar with this, the number one problem in this
4 country, and this is
really what we are going to
5
focus majorly on in terms of causing heart failure
6
and the specific kind of heart failure that this
7
engenders.
8 In cardiomyopathy, there is many different
9
kinds of things that get a heart to perform poorly,
10
all the way from an idiopathic cardiomyopathy to
11
such things as iron overload, et cetera, which are
12
not as common.
13 Now, what I want to talk about here is
14
really systolic heart failure.
There is something
15
called diastolic heart failure, and that really has
16 a
lot to do with compliance problems of the
17
ventricle, and in these patients, we are going to
18
see a normal ejection fraction.
19 So, this is really a different
animal and
20
it is really not what we are focusing on, so what I
21
am going to be talking about today is systolic
22
heart failure, and as I will show you, with the
23
hallmark being a low left ventricular ejection
24
fraction.
25 This is just to give you a practical
42
1
example. This is an angiogram
from one of the
2
patients that we treated with stem cell therapy in
3
Brazil, who all had an ejection fraction that
4
averaged about 20 percent. This
patient has an
5
ejection fraction of 10 percent.
6 You can see the coronaries are calcified.
7
This is a catheter in the left ventricle. This
8
heart is supposed to be pumping this contrast we
9
just put into the aorta. As you
can see, it is not
10
doing that very well at all.
Only 10 percent of
11
what is in here gets out with each beat.
12
So, you can tell this is a
dilated big
13
heart that just doesn't contract well.
That is the
14
picture of severe heart failure right there, and
15
this is what I want to talk about.
16 Now, when we talk about heart failure, I
17
think everybody is aware of the classification.
18
There is Class I, II, III, IV, which are commonly
19
used, but it is important to acknowledge this.
20
Class I involves no limitation of physical
21
activity, Class II slight limitations, Class III
22
marked limitations, you can't walk up a flight of
23
stairs without getting short of breath, and Class
24
IV, you have symptoms at rest.
25 If we look at this, if we put Class III
43
1
and Class IV together, you see the division is
2
about a third for each of these pieces of the pie
3
here.
4 Now, if somebody comes in with Class IV
5
heart failure, they are very short of breath at
6
rest, you can give them some diuretics and they
7
will feel better. They are not
Class IV anymore,
8
they are Class III.
9 So, it is interesting, there has been a
10
want in development of a little different way of
11
looking at heart failure, and a staging or
12
classification put out by joint AHA and ACC shows
13
four different stages, and really looks at heart
14
failure more like a disease like cancer.
15 So, where we can identify patients that
16
are at high risk of developing it, we can screen
17
patients, and then we can start treating patients
18
before they really manifest symptoms of the
19
disease.
20 Again, this is a progressive disease and
21
we are going to end up with people that are
22
refractory even to all kinds of treatment. I am
23
going to go over this a little bit more in detail a
24
little later.
25 So, in defining what heart failure is, I
44
1
hope I have given you a general idea of the scope
2
of the problem, just talk a little bit about what
3
causes it because it is important to understand
4
that to be able to know how we treat it and how we
5
monitor these patients.
6 Usually, we are talking about ischemic
7
heart disease and we are dealing with a myocardial
8
insult, which is usually a heart attack, so that
9
heart attack causes damage to the heart muscle, and
10
that is going to result in dysfunction of that
11
heart muscle.
12 Well, the body is going to try to
13
compensate this dysfunction and especially in two
14
major ways. One is neurohumoral
activation, so we
15
will talk a little bit about this in more detail,
16
but essentially, these compensatory mechanisms are
17
going to make the heart change its shape and its
18
size. It is something we call
remodeling. It
19
involves hypertrophy of the myocytes and then it
20
involves fibrosis and dilatation.
21 So, these mechanisms that the body helps,
22
to try to help to reverse what is going on,
23
actually wind up causing toxicity, hemodynamic
24
alterations that all lead to remodeling, and
25
remodeling really is the hallmark.
45
1 You saw that big heart. Well, remodeling
2
is how you get from a normal small heart, which you
3
have, to a big boggy heart that doesn't contract.
4
That is the problem of heart failure.
5 This was very simply put by Doug Mann in a
6
nice editorial a few years ago.
Basically, here is
7 the
heart over time, as we have an index event, and
8
basically, remodeling occurs, the heart gets
9
bigger, the ejection fraction goes down as time
10
goes by and symptoms occur as time progresses, as
11
well.
12 So, I have told you we have a myocardial
13
insult. This leads to LV dysfunction and
14
remodeling, and this really instigates a
15
neurohumoral response. In
return, this is going to
16
have an impact on remodeling again.
17 So, what are these neurohumoral things
18
that happen? Well, first of all,
most importantly,
19
is the renin- angiotensin-aldosterone system, and
20
there are several points in which the body
21
upregulates the system and ultimately, it acts on
22 the AT-1 receptor, which will cause
23
vasoconstriction, proteinuria, again LV remodeling.
24 As you can identify, here are several
25
sites in which medications, the mainstay of some of
46
1
the therapy for heart failure works, namely ACE
2
inhibitors that work at this point, ARBs that work
3
at this point, beta blockers have a role in
4
inhibiting renin, as well. So,
some of the
5
mainstay of therapy is actually directed at one of
6
these mechanisms of compensation.
7 On the other side, we have sympathetic
8
activation. We have increased sympathetic activity
9
that again leads to myocardial toxicity and
10 arrhythmias,
and then on the other side, with the
11
sympathetic outflow, we get vasoconstriction. This
12
impacts negatively on the kidney, sodium retention,
13
more vasoconstriction, and progression of the
14
disease.
15 Just to get a slightly little bit more
16
complicated, just to mention that it is really not
17
all that simple, there are other things involved,
18
and we have cytokines, TNF-alpha, IL-6,
19
inflammation that actually progresses with the
20 progression of heart failure.
21 Endothelin is a potent vasoconstrictor.
22
All these things lead to apoptosis and unfavorable
23
effects upon the myocyte, but then lead to LV
24
remodeling, which I have told you is one of the
25
mainstays of reasons for heart failure.
47
1 Now, natruretic peptides are important,
2
as well. It's another compensatory mechanism that
3
the body has. I am sure you are
familiar with
4
these BNP, it's a B-type natruretic protein that
5
actually comes from the ventricle, the A types
6
comes from the atrium. We will
just focus on the B
7
type.
8 What this does, basically, in response to
9
elevated pressure inside the heart, we secrete BNP.
10
This suppresses the renin-angiotensin-aldosterone
11
system and suppresses endothelin.
It helps with
12
peripheral vascular resistances, decreases
13
vasodilatation, and it increases natruresis.
14 So, if we go on to understand now that
15
there is an interplay between LV dysfunction and
16
remodeling, and that basically, this will lead to
17
low ejection fraction, and that is what we see in
18
the patients.
19 On the other hand, as a result of this, we
20
will start getting a constellation of symptoms, and
21
it is the combination of having a low ejection
22
fraction and symptoms that defines heart failure.
23 Let's look a little bit at the prognostic
24
markers. I just talked a little bit about BNP.
25
Well, it is very interesting. If
we divide BNP in
48
1
quartiles here, depending on the amount of BNP that
2
you have circulating, your survival will go down.
3
It is a prognostic marker, as well as a treatment.
4
Norepinephrine, the same way.
So, these are
5
markers of prognosis.
6 It is very interesting. These are levels
7
of BNP, and if you can decrease them, decrease to a
8
less degree, or here, we have an increase. So,
9
depending on which direction your BNP goes, your
10
survival varies as well, and that is an important
11
concept.
12 Let's look at another different kind of
13
marker. Exercise capacity, peak oxygen consumption.
14
In the transplant world, this is very important.
15
Here you see the number 14, so a peak oxygen
16 consumption greater than 14 or less than 14 has
17
very different prognostic indicators and in many
18
centers, this serves as a marker threshold for one
19
of the criteria for entering the patient into a
20
transplant program.
21
You can see here a
difference in mortality
22
from 53 percent mortality over two years in
23
patients that have an NVO2 of less than 14, to that
24
of 11 with greater than 14, so this is another
25
important number in patients with heart failure.
49
1 Then, if we look overall and look at
2
symptoms and hospitalizations, here is a New York
3
Heart Class I to IV, and this is fairly intuitive,
4
but as we get more symptomatic, we have an impact
5
on survival, and as we are getting more
6
symptomatic, we have an increase in
7
rehospitalization.
8 What about ejection fraction? I just
9
talked about ejection fraction, and you can see
10
here, similarly to NVO2, ejection fraction can
11
divide prognostically how patients will do. Here
12
we see more than 20 percent, less than 20 percent.
13
Here you see a two-year survival, 54 percent, so
14 half
the people dying that have an ejection
15
fraction less than 20 percent.
At one year, that
16
is a little over 20 percent.
17 The same thing, this is a large randomized
18
clinical trial, ejection fraction less than 40
19 percent. Over time,
people die more frequently.
20 Now, let's add a little arrhythmia to
21
this. Looking at different
levels, the first two
22
are greater than 30 percent ejection fraction, here
23
less than 30 percent, so that stratifies that out,
24
but then if we just add the amount of extra
25
ventricular beats to this, and if we have less than
50
1
10 per hour, more than 10 per hour, and then with a
2
poorly contractile ventricle, your survival goes
3
down as we add extra ventricular beats.
4 One attempt that has been made to sort of
5
graph this problem, because now I have shown you
6
many different prognostic markers and different
7
things we can use to classify these patients to
8
decide what to do and how to follow them.
9 One of them is a heart failure survival
10
score. There is an invasive
model, there is a
11
non-invasive model. So, things
like cause of heart
12
failure, resting heart rate, EF, mean blood
13
pressure, if there is a conduction delay
14
electrically in the heart, oxygen consumption, and
15
serum sodium can enter into a risk classification.
16 Here, you just basically have a graph that
17
shows according to low, medium, and high, your
18
survival will vary according to the risk.
19 In our little schema here, that leads
20
symptoms and low ejection fraction to heart
21
failure, what are really the things, though, that
22
are driving mortality? They are
going to be pump
23
failure, on the one hand, and arrhythmia, on the
24
other, because sudden death, as I talked to you
25
about before, is a very prominent problem in people
51
1
that have heart failure.
2 So, it is the combination of these three
3
things that will pretty much drive patients to a
4
lethal exit.
5 Let's talk a little bit about treatment
6
now. What are the goals of
treatment of heart
7
failure? You want to delay the
progression or
8
reverse remodeling, which you can do in some
9
patients, and delay the progression and reverse
10
myocardial dysfunction.
11 You want to reduce mortality, relieve the
12
symptoms, improve functional capacity, and reduce
13
disability, also decrease the intensity of medical
14
care and hopefully reduce economic cost.
15 I have shown you we go from initial
16
injury, initial infarct, we suffer remodeling, we
17
get a remodeled heart that now has a low ejection
18
fraction, and over this course of time, we have a
19
worsening of symptoms, so how are we going to
20
impact this in terms of treatment?
21 Well, the two mainstays are neurohumoral
22
blockade, we have kind of gone over some of the
23
things that we can do, and we will look at those,
24
and the other is revascularization.
So, many times
25
with the use of medication or with the use of
52
1
revascularization, we can reverse some of this
2 remodeling
in some patients, and in some patients
3
we don't.
4 One thing that is very important in terms
5
of being able to recover patients that have
6
remodeled hearts, and that are in this road of
7
heart failure, is identification of viable
8
myocardium.
9 Myocardial viability has clearly been
10
shown to influence the prognosis of people that are
11
undergoing revascularization procedures, so if you
12
have a viable myocardium, you are going to do
13
better. You have a chance of
improving more than
14
someone who doesn't.
15 Just to shift gears for just a second
16
here, these are electromechanical maps.
These are
17
representations of the left ventricle. This is
18
from a patient in our Brazil stem cell study.
19 This is an electrical map, this is a
20
mechanical map. Let's just look
at the electrical
21
map because I just talked to you about viability.
22
Very simply, if your cells are alive, they have an
23
electrical signal that is high.
If you have a big
24
scar with no cells, you have no electricity, you
25
have a low electrical signal.
53
1 We put it on
a little color scale. Red is
2
dead or red is very little voltage.
Purple is
3
high. Here, you see on this
electromechanical map,
4
an area of myocardial viability.
Again, just as it
5
is important to understand viability when you are
6
vascularizing patients that have heart failure,
7
that have coronary disease, it is also going to be
8
important, in my view, to understand myocardial
9
viability when we are applying some of these
10 therapies, and I think there will be differences in
11
bone marrow therapies and myoblast therapy, but
12
that is something to keep in mind.
13 I just wanted to show you an example of
14
the very common things that we deal with, so this
15
is not some esoteric difficult patient to find. We
16
come across people like this all the time in the
17
hospital every day.
18 This is a patient who was 41 years old, he
19
had bypass, he stopped up all his vein grafts and
20
his memory artery, and he had ejection fraction of
21
20 percent, very similar to the one that I showed
22
you, and Class IV congestive heart failure.
23 This gentleman was really delightful. He
24
was actually a pilot for a major airline, and
25
because of his bypass, he had to be put off the
54
1
flying, and he was actually in charge of all the
2
simulators, and he was the guy that graded all the
3
pilots when they had to come in and do the
4
simulation testing.
5 Basically, here, we have a 41-year-old
6
guy, very active man who has gone bypass, he has
7
lost his graft, he obviously has very aggressive
8
disease, and why I hear the talk about why some
9
people have more aggressive coronary disease than
10
others.
11 You see this is his right coronary, it is
12
completely blocked up, X's mean that you can't see
13 anything on angiography, so this kind of
fills from
14
the other side by collaterals, see these little
15
twigs down here.
16 Then, the circumflex is completely
17
occluded. This is a floating
marginal branch.
18
This is supposed to be connected, but this is
19
totally occluded, as well. The
only artery he has
20
got left is the one down the front of his heart,
21
but this is very much infarcted, and has a very
22
significant blockage here, as well as the takeoff
23
of this.
24 So, this patient, there is really nothing
25
to do, and we are faced with this a lot every day.
55
1
This patient, as I have shown you these curves of
2
mortality, this patient at our hospital wound up
3
going for an LVAD type procedure and died, and that
4
is what we see again and again, so this is a very
5
serious problem.
6 So, looking of an overview of treatment of
7
heart failure, let's see, we have medical-based
8
therapy, on one hand, we have device-based therapy,
9
on the other.
10 On the medical side, we need neurohumoral
11
blockade, we can have a hemodynamic approach and
12
also antiarrhythmic approach, so we are going to
13
use these drugs, ACE inhibitors, aldosterones,
14
diuretics, beta blockers, and then antiarrhythmics,
15
such as amiodarone, and then we are going to use
16
more potent i.v. inotropes that improve
17
hemodynamics, and asaratide [ph], which is
18
basically similar to BNP, it is like giving the
19
patient BNP.
20 On the other hand, we are going to have a
21
device-based approach using resynchronization
22
therapy. It really hasn't shown
a benefit in
23
survival, but in combined endpoints.
We are going
24
to put defibrillators into people, and I will show
25
you how that has improved survival.
56
1 Then, we will have ventricular assist
2
devices, and when all this fails, we have an option
3
of heart transplant, that is very little available
4
actually, and as you saw, it is a very little
5
sliver of what we are able to do.
6 But as you cumulatively add these
7
therapies, you are able to impact on survival and
8
make patients live longer. Here,
you see sort of
9
adding digoxin and diuretic, adding an ACE
10
inhibitor, and then adding a beta blocker, we get
11
progressive improvement. So,
this is pretty well
12
established in terms of medical therapy.
13 When we look at defibrillators, here is a
14
curve. This is from the MADA-2. This is primary
15
prevention, defibrillator in patients, previous MI,
16
LVF less than 30 percent, a very significant
17
survival difference in the patients that get a
18
defibrillator, so treating the arrhythmias is also
19
important.
20 Back to our schema of the different
21
classification of stages of heart failure. You see
22
that we can gradually, we start with ACE inhibitors
23
and gradually add different medications, but
24
everybody kind of goes up these stairs and ends up
25
here at the top, and that is why we have increasing
57
1
mortality from heart failure, because we are
2
getting people to get to this point where before
3
they really didn't reach that stage.
4 Then, we get to a stage of basically
5
refractory symptoms, so they have been bypassed,
6
they have had stents, everything has been done for
7
them, and they have that bad heart, it doesn't pump
8
well, they have a lot of symptoms, they can't
9
breathe very well. Many of them
have angina. I
10
want to want to give you a little bit of my own
11
perspective on that.
12 If we look at current trends, this was
13
published last week in JACC, very interesting.
14
Heart failure treatment--this is the survival
15
curves--heart failure treatment in 1994 to 1997.
16
Here is a survival curve. We
have improved the
17
treatment of heart failure.
18 1999 to 2001, gee, we are doing a lot
19
better, and this is comparable actually to
20
transplant from 1993 to 2000, and it really raises
21
the question if transplant, with the modern
22
management in medical management of heart failure,
23
how important is it and what the role of transplant
24
really is.
25 Really, there is a gap between a very
58
1
invasive transplant or LVAD and the medical
2
therapy, there really is, and we are here to talk
3
about stem cell therapy. There
is a gap of
4
something that could be done that is not quite as
5
invasive and traumatic as an LVAD or transplant,
6
and that can improve the patient significantly
7
since we are doing so well with medical therapy.
8 I want to talk to you a little bit about
9
my perspective on end-stage ischemic heart disease.
10
Basically, as I have told you, we have improved the
11
medical management, so we have longer survival, we
12
have improved the vascularization treatments of
13
coronary disease, we have improved the survival
14
following a heart attack, and that is why we have
15
more patients, and now we are using widely
16
defibrillators, and that is why people are living
17
longer.
18 So, this is sort of my understanding of
19
this end-stage patient. You
progress with coronary
20
disease until you get to the Stage III and Stage
21
IV, Class III/Class IV heart failure.
22 If we look at these patients, sometimes
23
there will be a little surprise, because some
24
patients really just have shortness of breath, so
25
this is a variable. This may
occupy the whole
59
1
square or angina may occupy the whole square.
2 So, some patients predominantly have heart
3
failure, and these patients that predominantly have
4
heart failure probably weren't very good at forming
5
collaterals when they had heart attacks and
6
developed a lot of scar tissue, and have a very low
7
ejection fraction. These are the
sickest patients
8 and the patients that are going to have a
very high
9
mortality.
10 On the other hand, but also in the Class
11
III or Class IV, and sometimes we pool these people
12
together in trials and that is why I am making this
13
distinction, some people have angina more than they
14
have heart failure. These
probably have a much
15
better collateral formation when they had these
16
events, so their ejection fraction is a little more
17
preserved.
18 I have had many patients that have lived
19
on one artery. Their whole heart
is beating okay.
20
That one artery feeds everything by collaterals,
21
but they are in really bad shape.
I mean it's an
22
illusion that they are doing okay, but they do have
23 a
preserved ejection fraction, and their
24
manifestation is a lot of chest pain.
25 So, symptoms can vary from one side to the
60
1
other and some patients have a balance here, and I
2
think we need to keep this in mind when we are
3
designing these trials.
4 So, there is a predominant angina, and
5
this is the kind of patient that got, let's say,
6
these TMR type procedures. That
is the kind of
7
population you are dealing with.
The predominant
8
aspect is disabling angina, preserved EF, 100- to
9
200,000 new cases per year, and constitute about 5
10
percent of the patients undergoing angiography at
11
tertiary referral centers. This
has been studied
12
in this particular case at the Cleveland Clinic.
13 One year mortality is still very high.
14
Then, that other group, predominantly heart failure
15
symptoms, very low EF, myocardial ischemia, though,
16
is still present, but with more scar.
No option
17
really for any kind of revascularization. One year
18
mortality, 20 to 50 percent. I
have shown you one
19
curve where it is up to 80 percent, I mean it can
20
be really bad.
21 Here, we have ICD therapy trials. If we
22
look at secondary prevention trials, very sick
23
patients in this study, treated with amiodarone,
24
you see here one year mortality 44 percent. I mean
25
heart failure can be worse than cancer.
61
1 Here is the REMATCH trial. This is an
2
LVAD. This is the impact of
LVAD, and there is an
3
impact of survival, but again you are dealing, in
4
this case, with Class IV patients that are
5
unresponsive to medical therapy, so these very sick
6
patients, but again an invasive, costly, not widely
7
available kind of therapy, but it does have an
8
impact on failure.
9 I want to finish now talking a little bit
10
then, hopefully, I have given you an overview of
11
the problems with heart failure, and how are we
12
going to look at adverse events.
13 Well, what are the things that are going
14
to drive the adverse events here, are going to be
15
arrhythmia, ejection fraction, and symptoms, and I
16
think if we focus here, we can pretty much decide
17
what we need to look at in these patients over time
18
as we use new therapy towards these patients.
19 Let's look at low ejection fraction, how
20
are we going to monitor that?
Well, we need to
21
look at cardiac function, cardiac size, and the
22
perfusion status of the ventricle.
We can do that
23
very simply, if you take a simplistic approach,
24
with echocardiography.
25 I empirically have placed this here based
62
1 on my own limited
experience here, but I read in
2
the document that you wanted some more practical
3
advice, so I will give you my own sort of practical
4
feel for what I would do.
5 If we did echocardiogram on these
6
patients, we could do it monthly for the first
7
three months and then at six months follow-up. We
8
can do SPECT, we know that we don't need it too
9
early, and that is a very simple way of doing it,
10
three to six months. Clinical
visits, which will
11
be very frequent, and I will talk about that, and
12
BNP can be done for that, as well.
13 Now, we can get fancy and use alternative
14
imaging strategies, we can use MRI,
15
electromechanical mapping, PET, depending on the
16
institution, and depending on what we are really
17
looking for and want to find.
18 Cardiac arrhythmias, it is important to
19
monitor cardiac rhythm. Holter
monitoring is very
20
simple, probably should be done after the
21
procedure, one, three, six months later. Q-T
22
interval when the patient comes in for his clinic
23
visit is a strong predictor of survival, just a
24
plain-old, good-old 12-lead EKG, and that should
25
always be looked at.
63
1 In the patients I guess that are getting
2
myoblast therapy, there may be a little bit more
3
concern about this, and this is really not my area
4
of expertise, but these patients, many of them
5
already entering with an AICD, that have sort of a
6
built-in little computer that is already monitoring
7
their rhythm as it is. If they
don't, you might
8
want to consider event monitoring.
9 For symptoms, well, clinical visits
10
biweekly for 8 weeks, monthly up to 6 months. We
11
are going to look at heart class, we are going to
12
look at EKG, CBC, CRP, look for inflammation.
13
Exercise capacity, ramp treadmills, as you know, if
14
you put a patient that has end-stage heart failure
15
on a graded treadmill test, every time the
16
treadmill bumps up and goes a little faster, he
17
just may not be able to exercise at that point.
18 So, the advantage of a ramp treadmill
19
protocol is that you have a gradual continuous
20
increase, so these people that really can't do very
21
much at all, they will be able to tolerate the
22
exercise and probably get further than they could
23
in any other kind of exercise test.
24 There is a very simple way of evaluating
25
an exercise test, a 6-minute walk test.
You just
64
1
define a distance, walk the patient walk for 6
2
minutes, see how fast he can go.
You can do that
3
at a clinic visit, and it is very simple to do.
4
So, you can do something like this at one, three,
5
and six months.
6
Rehospitalization. We look at the
7
rehospitalization rates. It is
important to look
8
at the use of i.v. medications that are used to
9
control symptoms, because this is, as you saw, the
10
biggest part of the pie in terms of costs, and is a
11
real problem in the end-stage patients.
12 Quality of life, it is important to assess
13
quality of life, for example, SF36, Minnesota
14
Questionnaire.
15 Just some suggestions. I want to wrap
16 this
up and saying I hope I have given you a
17
general idea and scope of this problem.
We deal
18
with a very, very serious problem, which is heart
19
failure, specifically, that which is ischemic heart
20
failure and specifically, end-stage ischemic heart
21
failure.
22 I hope I have given you a flavor of this
23
and set the stage for the discussions.
24 Thank you very much.
25 [Applause.]
65
1 DR. RAO: Thank
you, Dr. Perin.
2 There is time for questions, and we can
3
open it up to the committee.
4 Q&A
5 DR. SCHNEIDER:
Emerson, one of the things
6
that you did very nicely was lay out the clinical
7
spectrum for people who may not be familiar with it
8
in this context.
9 I wanted to follow up on that point
10
because work presented at international meetings
11
recently by the Frankfurt group of Andreas Sire and
12
Stephanie Dimler suggests that bone marrow derived
13
cells and circulating progenitor cells from
14
patients with established heart failure may be
15
deficient relative to the performance of bone
16
marrow derived and circulating progenitor cells
17
from patients with an acute infarct.
18 So, while it is not quite an apples and
19
oranges comparison to envision cardiac cell
20
grafting immediately post infarction or in the
21
first week post infarction in patients without
22
severe ventricular dysfunction versus patients,
23
let's say, two to four months out with mild or no
24
ventricular dysfunction versus the end-stage heart
25
failure patients who have been a focus in your talk
66
1
this morning, it does seem to me that that clinical
2
heterogeneity introduces a couple of problems.
3 I am curious to know how you have worked
4
those through in your own work.
One of them is
5
because what we are discussing today and tomorrow,
6
is autologous cell therapy, I believe that there is
7 a
serious issue of patient-to-patient cell
8
heterogeneity which has been relatively little
9
discussed in the field except in these still
10
unpublished or perhaps one paper has come out in a
11
secondary journal from Stephanie and Andreas about
12
the defects.
13
So, one question is what kinds
of
14
standards should a proposed production center be
15
required to meet in terms of their ability to
16
generate cells that perform in accordance with some
17
standard when there is patient-to-patient variation
18
of this kind.
19 Secondly, if you are envisioning putting
20
cells of different kinds into a so severely an
21
ischemic background as the 41-year-old former pilot
22
that you mentioned, doesn't it become important to
23
clearly distinguish, as the prefatory remarks did,
24
between mechanisms of action for proposed donor
25
cells that are aimed at regeneration specifically
67
1
versus benefits that are achieved through entire
2
different mechanisms, such as angiogenesis?
3 If you put new cells into an ischemic
4
background, they will surely die, and if the goal
5
is to achieve angiogenesis in a background where
6
the native coronary circulation has failed and the
7
graft has failed, then, it seems to me we need a
8
clearer resolution of the problem of which cells do
9
which things well, and really fine-tune much better
10
than the field has to date, you know, which are the
11
cells that we want where the spectrum is normal
12
vasculature, insufficient muscle cells versus the
13
hypothetical ischemic patient that you described
14
where revascularization is the major goal.
15 DR. PERIN:
Well, that's fantastic.
16 [Laughter.]
17 DR. PERIN: I
think the basic answer to
18
your question is I don't know, but, you know, these
19
are all very good points, starting with the cell
20
type, we really don't know.
21 Actually, we have submitted a manuscript
22
in which we have had the pathology of one or our
23
patients in our study in Brazil who received
24
autologous bone marrow, died 11 months later, and I
25
really can't preempt I guess our publication, but I
68
1
think we will be seeing some evidence of myogenesis
2
and angiogenesis from autologous bone marrow cells,
3 but we really don't know what we are getting when
4
we are putting, let's say, autologous bone marrow,
5
and even in that patient that has, let's say he has
6
predominantly ischemia, if we want to
7
revascularize, can we get a predominantly
8
angiogenic effect, so we really don't know, and we
9
need to define that.
10 Mononuclear fraction of the bone marrow is
11 a
very simple approach, the one that we have taken,
12
and it seems to initially, and we haven't really
13
done efficacy studies and we are continuing on, but
14
there is a suggestion that it does, so I think that
15
we need to take every step that we take should be
16
put one foot in front of the other, and if the
17
mononuclear cell fraction works, I think we can go
18
from there and keep investigating that.
19 Now, the average age in our trial was
20
about 58, and you mentioned the problem--
21 DR. RAO: Can I
interrupt? These are
22 really important questions, but they discuss
data
23
which was not presented in the talk right now. I
24
would like to at least focus the questions
25
initially on the issues that relate to the
69
1
presentation right now.
2 We should really come back to these
3
questions tomorrow when we discuss exactly these
4
sorts of issues.
5 Do you think that that would be okay with
6
you, Dr. Schneider?
7 DR. SCHNEIDER:
We will certainly return
8
to them tomorrow, but I was discussing issues that
9
were raised in this talk, which was clinical
10
heterogeneity.
11 DR. RAO: Let's
then focus, not on the
12
cells per se, and the choice of cells, because none
13
of the presentation was related to the production
14
facility or how the cells would be, or the quality
15
would be, or how you would choose the mechanism,
16
but maybe how do you choose patients for a trial or
17
is there some reasonable way of selecting patients,
18
that there would be consensus on.
19 DR. PERIN:
Okay. So, we will get back to
20
your first question and really, that is something
21
that actually, we are working on trying to
22
understand, is there a thumbprint or is there a
23
profile in the study by Dimler and their colleagues
24
looking at the characteristics of cells in certain
25
patients, and obviously, they may not be the same
70
1
in a diabetic, in a severe heart failure, we don't
2
know, so there is another important we don't know.
3 Age obviously is a very important thing,
4
so harvesting cells from a 75-year-old may be very
5
different than doing that in a 55-year-old, so
6
these are all questions that need to be answered.
7 DR. RAO: Dr.
Mul.
8 DR. MULE:
Given the slides you showed of
9
the steps toward progression of heart failure, and
10
given the current interventions along that pathway,
11
from your perspective, where would you see
12
cell-based therapy intervention falling into that
13
step toward complete heart failure?
14 DR. PERIN:
Right now, at close to the
15
last few steps, I think ethically, we are propelled
16
to really study the problem in the patients that
17
really don't have a proven conventional option for
18
treatment. In brief, I would say
in the patients
19
who can't be revascularized, because really medical
20
therapy, we are going to apply to everyone, so then
21
we are left with revascularization.
22 Well, can we revascularize? Well, we do,
23
and we do it again and again, and there is a point
24
where you are out of revascularization options, and
25 I
think that is one place we are
initially now,
71
1
then, you could think about applying this kind of
2
treatment.
3 DR. HARLAN:
Building upon what Dr. Rieves
4
mentioned when he gave his introductory comments, I
5
want to just congratulate you on, it seems like our
6
task is to weigh the risk-benefit, and you have
7
outlined very clearly the risk, and I accept that
8
it is severe, and I also want to congratulate you
9
on mentioning the JACC paper that was just
10
published, that showed how dangerous it is to look
11
at historical controls, because we are making such
12
rapid progress.
13 My question is along those lines, not in
14
this field, I just read in the journal, the
15
Washington Post, about the great advance that has
16
been made in super-high statin therapies, and I
17
wonder if you could comment on that study, that
18
these super-physiologic statin doses seem to have a
19
major impact on mortality.
20 DR. PERIN: I
really don't have an
21
expertise in a lot of things, and that is not one
22
of them, so it is really hard for me to comment on
23
that. I know that it looks like
giving people HDL
24
in the future may be a very exciting thing, and we
25 may be able to finally find our liquid
plumber kind
72
1
of solution for people.
2 Then, again, statins are just--more and
3
more if you study statins, you have probably come
4
to the conclusion it should be in the water pretty
5
soon, I mean the patient benefit is on every single
6
aspect of cardiovascular disease.
7 DR. RAO: Dr.
Kurtzberg.
8 DR. KURTZBERG:
You mentioned some
9
practice-based methods to evaluate outcomes and
10
function in these patients, but I think the
11
challenge is to determine what the cells are doing,
12
you know, are they differentiating into other kinds
13
of cells, are they mediating inflammation, are they
14
mediating angiogenesis, and I don't see how you can
15
sort that out by clinical-based study.
16 Do you know of other technologies that are
17
on the horizon that may help
with that, that are
18
non-invasive, or would you consider serial biopsies
19
in patients like this to answer those questions?
20 DR. PERIN:
That is a good question. I
21
don't know that serial biopsies would be a very
22
efficient way of evaluating that.
You would have
23
to have a very precise way of being able to
24
identify where you put the cells and be able to go
25
exactly to that same spot.
73
1
We do have that
technology. Dr. Lederman
2
is going to follow me eventually here.
The MRI
3
field, I think is very promising in that regard in
4
terms of labeling and following cells.
5 Now, I really don't know that even
6
labeling a cell, even if it died, if the label
7
stays there, you still see the
label, so I think
8
that we have to even go a step further and be able
9
to prove the functionality of the cell that is
10
alive and was implanted.
11 That can be done on an experimental basis,
12
so we figure ways out to do that, but this is a
13
very intriguing problem and a very difficult
14
problem to evaluate. I
think you have put your
15
finger on something that is going to be hard to
16
know.
17 DR. DINSMORE:
Jonathan Dinsmore from
18
GenVec.
19 I just had a question on your angina heart
20
failure continuum. I was
confused because most
21
heart failure patients present without angina, with
22
symptoms of fatigue, so what percentage of heart
23
failure patients actually experience angina?
24 DR. PERIN: If
we are talking about
25
ischemic heart failure, we are not talking about
74
1
other kinds of heart failure, actually, idiopathic
2
heart failure, you kind of get the same remodeling
3
and everything except you didn't have that infarct
4
in the beginning, but you go through the same sort
5
of pathophysiologic processes.
6 So, we are talking about ischemic heart
7
failure. People that have ischemic heart failure
8
have coronary disease. Coronary
disease is
9
narrowing of your coronary arteries.
10 Depending on what your response is, you
11
will or will not have angina, but angina is one of
12
the manifestations of coronary disease, and it is
13
really not a good thing to base a lot on, because
14 the
expression of angina is very variable.
15 It depends on your pain threshold. I mean
16
if you are a diabetic, you may not have as much
17
pain. It is a subjective thing
subject to
18
interpretation by the actual patient, so it is
19
something that is very difficult to evaluate, and
20
that is why I put the continuum, because it is all
21
there and you really shouldn't take a patient
22
population based on angina or based on shortness of
23
breath.
24
I think you have got to
bring both of
25
these things together to understand they are sort
75
1
of in the spectrum of a similar underlying
2
pathophysiologic process.
3 DR. SIMONS: I
would like to come back to
4
the issues of the differences among the patients
5
having these kind of therapies.
We have learned
6
from a number of trials of growth factor therapies
7
that there is a very large difference in how the
8
patients respond.
9 This issue that there are different
10
subgroups that we are not defining is fairly
11
critical to the field. You
mentioned one or two
12
biomarkers, but there seemed to be a general
13
association of markers as opposed to really
14
identifying which patients respond in which manner.
15 What would you suggest as a way of trying
16
to sort of stratify these patient groups? Not
17
suggest ejection fraction, that is probably in a
18
way sort of crude measure, but in terms of
19
biological responses.
20 DR. PERIN: If
we look at the trials of
21
devices, I think that probably a common way to look
22
at these patients is exercise capacity.
23 I think that probably is one of the
24
unifying parameters that we cannot only use at
25
entry, but also you are able to follow as a patient
76
1
goes along, and if he has a response to therapy, he
2
will have a positive response in terms of what he
3
is able to do in terms of function.
4 That has a very practical translation into
5
quality of life and people feeling better. I would
6
say in a broad sense, that exercise capacity, peak
7
oxygen consumption might be something that I might
8
consider an important thing to follow in these
9
patients, and not just ejection fraction, which is
10 dependent on a lot of things, how much loading the
11
ventricle has that day, the amount of mitral
12
regurgitation, et cetera, so there is a lot of
13
things that will make that extremely variable.
14 DR. RAO: As an
extension of that, it's a
15
very general question. Is there
any problem with
16
many of these studies which are in high-risk
17
patients enrolling people for the placebo arm of
18
the trial? Not in cell therapy,
but maybe when you
19
do devices or you do assists, has this been
20
historically a problem for the cardiovascular
21
field?
22 DR. PERIN:
Well, it has been done as you
23
can see, so I have showed you a bunch of studies
24
where it has been done, and it can be done.
25 Personally, the way I like to see it is I
77
1
want to offer patients that get in the placebo arm
2
some kind of a treatment, so in our future upcoming
3
study, what I am going to do is I will tell a
4
patient you are going to get randomized to maybe
5
not getting treatment, but if you don't get that
6
treatment at an X period of time, six months, you
7
will cross over to get the treatment.
8 I think that is a humane way of doing it,
9
in which these patients are very ill and desperate
10
to get something to help, so again, if you can
11
cross over, sometimes these placebo patients at
12
some point after you have achieved your assessment,
13
then that makes it a more palatable or fair way to
14
do things maybe.
15 DR. RAO: Dr.
Cunningham.
16 DR. CUNNINGHAM:
I just wonder, in your
17
data, if you see any difference
by either
18
socioeconomic status or by gender, or by any way of
19
culture, dividing populations, whether it would be
20
race or ethnicity or any other factor like that?
21 DR. PERIN: You
mean in our own--
22
DR. CUNNINGHAM: Yes, reading the JACC
23
data, was there anything by gender, for instance,
24
or by subpopulation?
25 DR. PERIN:
Females, there are some
78
1 differences in the female population in which
there
2
are some differences. There is the catch-up
3
phenomenon in the end, but socioeconomic
4
differences, I am not aware that it would have an
5
impact on that, as well, but maybe gender
6
differences, yes.
7 DR. RAO: One
question to sort of follow
8
on Dr. Simons' question, in at least the way I
9
understood it, it is really kind of difficult to
10
stratify patients or to extrapolate from one class
11
of patients to the other.
Historically, that has
12
always been a problem.
13 Again, it's a general feeling when one
14
conducts studies in the cardiovascular field, is
15
there some consensus that
everybody says that,
16 well, if you measure by ejection fraction, and we
17
take patients, which is what it seemed like a lot
18
of studies have done, that that is a reasonable
19
criteria that you can extrapolate from one
20
classification of that kind to the next, or one
21
cannot? Just as a general
statement.
22 DR. PERIN: It
has been done, and it is a
23
general way of separating--there is definitely a
24
correlation with your ejection fraction and your
25
survival, so it is probably not the most refined
79
1
way of dividing patients, and it depends where you
2
make the cutoff, so if you make a fairly high
3
cutoff, let's say, patients that had ejection
4
fraction less than 40 percent, then, you are
5
including most of the population of patients that
6
have heart failure, so it's a general way to divide
7
things.
8 If you start decreasing that number of
9
that cutoff, then, you are really selecting out
10
more I think subpopulations we were talking about,
11
maybe some different kind of subpopulations of
12
patients with heart failure.
13 DR. RAO: Dr.
Borer.
14 DR. BORER: Dr.
Rao, a few minutes ago you
15
made a point, and I would like to restate it in
16
another way, because what Dr. Perin did, as I see
17
it, is very well present an overview as an outline,
18
was a scaffold upon which we can conduct subsequent
19
more specific discussions.
20 I think that right now we are getting into
21 a
series of questions that are way beyond the data
22
that exist, and you couldn't expect Dr. Perin to
23
respond to them in a meaningful way because the
24
data don't exist.
25 In specific response to your question,
80
1
which was a very fundamental one, I think we are at
2 a
point now with this form of therapy where if we
3
could define any group in which we saw a response
4
which seemed credible, which was statistically
5
valid, we would then have a series of hypotheses
6
that would have been generated that would allow one
7
to move further, but I think that is the level we
8
are at.
9 The idea of defining a general population
10
in which to test therapy the way we do with drugs,
11
we are not there yet, so I think the specific
12
questions have to come a little later in this
13
forum.
14 DR. RAO: I
just wanted to get it clear to
15
people that that was the case, but your point is
16
very well taken.
17 Dr. Neylan.
18 DR. NEYLAN:
Thank you.
19 That was a very nice clinical overview,
20
and I wanted to ask you from your perspective as a
21
clinician, there are obviously many parameters
22
whose relief or improvement would be significant in
23
terms of the lives of individual patients, and many
24
of these could be utilized as endpoints for proof
25
of concept.
81
1 But ultimately, what do you believe is the
2
most relevant clinical endpoint for defining
3
registration criteria for this form of therapy, is
4
it patient mortality or something else?
5 DR. PERIN: I
don't know if we are going
6
to be impacting patient mortality.
That is a very
7
difficult question. I would go
back and what I had
8
said earlier, and use an endpoint, I would use
9
something like the LV02 as an endpoint.
10 I think that is a little bit more
11
palpable, and obviously, looking at mortality, this
12
is such an initial incipient field in which we have
13
barely treated any patient, so to think about
14
looking at mortality, which involves a much larger
15
number of patients, I think that is probably
16
getting ahead of ourselves a little bit.
17 We need to first verify if this is
18
efficacious and if there is some objective
19
improvement in these patients, and one of those
20
objective ways of doing that would be something
21
like exercise capacity, like I mentioned.
22 DR. RAO: Dr.
Ruskin.
23 DR. RUSKIN:
Just two quick comments on
24
Dr. Perin's very nice presentation.
25 One is that we have learned from drug and
82
1
device trials that both ejection fraction and heart
2
failure classification are critically important
3
predictors, but that they are not necessarily fully
4
interactive, that is, they are independent, so
5
using both, I think in any classification with
6
regard to these kinds of interventions would be
7
critical because the outcomes are very, very
8
different in Class III and IV even with the same
9
EF.
10 The other relates to a question
that Dr.
11
Rao raised about recruitment and controls. I think
12
that given the excitement in this area, but the
13
unknown issues that have already been raised, doing
14
trials that have adequate controls perhaps is more
15
important here than anywhere else one can imagine
16
given the severity of the illness that we are
17
dealing with and the kinds of outcomes that Dr.
18
Perin has described.
19 As someone who recruits for device trials,
20
though, I can tell you that it is not easy, and
21
randomizing patients to acceptable controls in this
22
kind of illness is going to be a huge challenge,
23
but I think it is important for this group to
24
emphasize that there is no place where this could
25
be more important, otherwise, we will never get an
83
1
answer, and I think that mortality ultimately will
2
have to be a critical part of any trial that is
3
done.
4 DR. RAO: Go
ahead, Dr. Borer.
5 DR. BORER: I
agree completely with Jeremy
6
that controls are essential in this kind of
7
research and really in any clinical research, but I
8
think again to put this whole area in context, and
9
in response to Dr. Neylan's point and question, we
10
are at the point now of looking at physiological
11
variables and what we would call in drug
12
development "surrogates," to see whether cardiac
13
performance, cardiac perfusion, this, that, and the
14
other thing, is affected in one way or another, so
15
that one could extrapolate to the point where it
16
would be legitimate to define hypotheses about
17
clinical outcome.
18 We are not there yet, and the clinical
19
outcome, just to put it in context from the drug
20
world, is perfectly legitimate in the view of most
21
people who deal with this area and these agents to
22
think of a therapy as being approvable if it makes
23
people feel better, but doesn't make them live
24
longer.
25 If it makes people feel better, even if it
84
1
makes them live a little bit shorter, as long as
2
you know how much shorter that is, and if it makes
3
people live longer while not making them feel too
4
much worse.
5 I don't think we are at a point yet again
6
to define what the outcomes variables should be. I
7
think we are at the point of defining physiological
8
and pathophysiological surrogates, and that is what
9
is being done in the studies to date, and then we
10
can decide what the outcomes are, clinically
11
important for registration.
12 DR. RAO: I
guess that leads us to the
13
fact that many of these things should be discussed
14
tomorrow, just like you pointed
out.
15 If there are no critical questions
16
remaining, I will thank Dr. Perin.
17 [Applause.]
18 DR. RAO: We
are going to take a short
19
break.
20 [Break.]
21 DR. RAO: We
are really extremely
22
fortunate in having Dr. Menasch here to present
23
his findings, and I look forward to a really
24
interesting talk.
25 Clinical Experience of Autologous
85
1 Myoblast Transplantation
2 DR. MENASCHE:
Good morning. First of
3
all, I would like really to thank you for the
4
privilege of this invitation and this unique
5
opportunity of sharing some data on the clinical
6
myoblast transplantation.
7 What I would like to do in this talk is
8
first to briefly touch on the preclinical data
9
which have paved the way for these early clinical
10
trials, and then, as requested by Dr. Grant, to
11
focus on the various aspects of the clinical
12
experience which has accumulated so far before
13
drawing some perspectives which may have clinical
14
relevance in the near future.
15 Now, I think just to make things clear,
16
that the basic assumption is fairly
17
straightforward, and the objective of this therapy
18
is really to try to repopulate areas of dead
19
myocardium with new contractile cells with the hope
20
that these areas can regain some function, and
21 given the close relationship between function
and
22
survival, which has been already mentioned, the
23
ultimate hope is obviously that it can have a
24
significant impact on clinical outcomes.
25 The reason why we initially started with
86
1
the skeletal myoblasts are actually listed here.
2
These cells are not really stem cells, they are
3
better termed precursor cells for muscular fibers
4
in that they are very committed to their skeletal
5
muscle phenotype as you will see.
6 The first advantage of the myoblasts is
7
that they can be very easily retrieved from the
8
patient himself, thus overcoming any problem
9 associated with
rejection and immunosuppressive
10
therapies.
11 These cells feature a very great expansion
12
potential which is important given the relationship
13
which exists between the number of cells which are
14 injected
and the ultimate functional outcome.
15 As I have just said, they are pretty well
16
committed to their myogenic lineage, and the risk
17
of tumor development is virtually negligible.
18
Finally, they are pretty resistant to ischemia, and
19
although unfortunately, many of them die shortly
20
after the injections, fortunately, some of them
21
will survive and may positively affect function.
22 So, this is type of animal model which has
23
been used initially in rodents.
You see here the
24
heart and the needle injecting the cells. I just
25
would like to mention that it took us seven years,
87
1
seven years of preclinical work before I did
2
operate on the first patient June 15, 2000.
3 During the seven years, we moved from the
4
rodent models to the large animal models, which I
5
think is absolutely necessary before arriving to
6 clinical trials.
7 Just to summarize the bulk of this data,
8
we can say, number one, that when you inject
9
skeletal myoblasts into an infarcted area, they
10
retain the possibility of differentiating into
11
typical myotubes. Here is a
typical myotube,
12
elongated structure, and this is a sheep heart and
13
this is a human heart.
14 This is an autopsy specimen. One patient
15
of our Phase I trial died 18 months after his
16
surgery from stroke, and we had permission for the
17
autopsy. You will appreciate the
striking
18
similarity of these two slides.
Here you find in
19
this human heart, a typical myotube embedded in
20
scar tissue.
21 At closer magnification, you can
22
appreciate the typical cross-striations, and I
23
think two observations are important to be made at
24
this point. Number one, these
cells really remain
25
committed to their skeletal muscle phenotype. In
88
1
other words, there is virtually no evidence that
2
they can ever turn to cardiomyocytes.
They will
3
not become cardiac cells.
4 Number two, they remain electrically
5 insulated from the surrounding myocardium, which
6
obviously raises major mechanistic questions
7
regarding the underlying mechanisms by which they
8
can improve function, but the fact is that there is
9
no real evidence that they develop connections with
10
the neighboring cardiomyocytes.
11 Nevertheless, when you subject them to
12
strong depolarizing currents, they show excitable
13
properties, and you see here, this is a fluorescent
14
myotube which has been grafted in a myocardial
15
scar. This is an in vivo study
and definitely they
16
can respond to currents by generating action
17
potentials followed by contractions.
18 This translates into an improvement in
19
function, both regional function here in the sheep
20
model, and global function, the LV ejection
21
fraction. This improvement, as
you can see, seems
22
to be sustained over time until one year in our rat
23
studies, and basically, these kinds of observations
24
have been made by several other investigators
25
already past 10 years.
89
1 So, there is a fairly good consistency
2
showing that these myoblasts can, to some extent,
3
improve function at least in animal models, and
4
obviously, the gap with the humans is a wide one.
5 So, if we now move to the clinical
6
experience, so far there are 44 patients who have
7
been included in early Phase I trials, and 34
8
patients currently included in our ongoing
9
randomized, multi-centered Phase II study .
10 This list is by far not exhaustive. I
11
have not tabulated anecdotal case or me-too cases.
12 I
have just kept those studies which have been
13
published in peer-reviewed journals.
14 Basically, the inclusion criteria have
15
been fairly straightforward across all these
16
studies. Patients with low
ejection fractions,
17
usually below 35 percent, patients with a history
18
of myocardial infarct, and obviously, patients
19
requiring concomitant coronary bypass surgery since
20
for ethical reasons, it is difficult to open the
21
chest just for injecting a product we don't really
22
know whether it is effective or not.
23 If we try to summarize the main results,
24
we can say, number one, that multiple epicardial
25
injections look to be safe. I
have never seen any
90
1
bleeding from the needle holes, and overall, this
2
experience has been shared by the other surgeons
3
who have practiced the operation.
4 Number two, it is possible--and we will
5
come back on that--that the procedure increases the
6
risk of arrhythmia postoperatively, at least in the
7
early post-op period.
8 Number three, I will be extremely careful
9
and cautious about that, there are some data
10
suggesting that maybe function can improve, but it
11
is clear that until we have the results of the
12
ongoing randomized, placebo-controlled study, we
13
cannot make any meaningful conclusion.
14 This is the list of the studies and of the
15
patients. I have just added the
last one a few
16
days ago. Professor Siminiak presented at the
17
American College of Cardiology another series of 10
18
patients who got the cells through a percutaneous
19
catheter using the coronary sinus route. I will
20
come back on that catheter in a few minutes, but I
21
will rather concentrate on the surgical
22
implantations listed here.
23 Dr. Smits also injected cells through a
24
catheter using the interventricular approach
25
similar to the one alluded to by Dr. Perin.
91
1 This goes back to the inclusion criteria
2
which have previously been mentioned.
I think it
3
is important to look at all words, because as you
4
will see, differences in definition may really be
5
confounders in the interpretation of the results.
6 It is important to look at akinetic areas
7
that is really dead myocardium, not simply
8
ipokinetic or dyskinetic, really akinetic
9
myocardium, which are not amenable to
10
revascularization and obviously, it is also
11
important that the bypass surgery be done in other
12
areas.
13 For example, you will see that in one
14
study, the area which was transplanted with cells
15
was also revascularized, so when the authors
16
conclude that cell therapy improves function, it is
17
clearly meaningless since the same area has got
18
simultaneous revascularization.
19 For those of you who are not familiar with
20
the procedure, I just would like briefly to show
21
you this three-step operation.
It starts with a
22
muscular biopsy. We take it at
the thigh. It is a
23
very simple procedure under local anesthesia.
24 We remove a chunk of muscle, which is then
25
cut into small pieces, put in this sheeping medium
92
1
and sent to the cell culture lab where a multiple
2
tri-cell factory is being designed to allow for
3
large-scale cell production.
4 Then, there are regular morphological
5
controls. Obviously, the key point is to inject the
6
cells before they reach confluence.
What you would
7
like to do is that confluence occurs in vivo
8
following the engraftment, not before, so it is
9
important to check the morphological state of the
10
cells on a regular basis.
11 This is how human myoblasts look like
12
during the cell culture process, and this is how
13
the cells look like when they are back in the
14
operating room.
15 Then, with the curved needle, we inject
16
the cells all across the infarcted area including
17
the borders. It's a
time-consuming, I would say
18
10, 12, 15 minute procedure, rather tedious and
19
boring procedure, by the way, where you have to
20
mentally construct the grids and then go with the
21
needle from side to side, so we are working on the
22
multiple shot device, but it is more tricky than we
23
initially thought.
24 So, right now we have the requirement for
25
these multiple injections all across.
This is
93
1
another view of the injections.
2 So, if we start by feasibility, I think it
3
is quite well established that this technique is
4
perfectly feasible. In other
words, it does
5
demonstrate that provided you
have the appropriate
6
techniques, you can take a small piece of muscle
7
which contains, say, 3- 4 million skeletal
8
myoblasts initially and expand it over two to three
9
weeks until approximately 1 billion cells.
10 These are the results of our cultures
11
during the Phase I trial, during which the target
12
numbers which have been prespecified have
13
consistently been obtained and even overshoot it.
14 You will note that you can get up to 90
15
percent of skeletal myoblasts in that--and this is
16
an important point--you really end up with a pretty
17
well defined cell therapy product.
You really know
18
what you are injecting.
19 Importantly, what we have seen is that
20
heart failure does not prevent skeletal myoblasts
21
to differentiate into myotubes,
and this was a
22
question because when we did preclinical rounds, I
23
got pieces of tissue from orthopedic colleagues,
24
but often these patient were young, and the
25
question was are the myoblasts from this Class
94
1
III/IV heart failure patients going to
2
differentiate normally, and the answer is yes, so
3
far we have had no failure.
4 The only thing is that it may take a
5
little bit more time for some patients until we get
6
the target number of cells, but at the end of the
7
day, it has always been possible to achieve the
8
prespecified target number of cells in myoblasts.
9 What about safety now? These are the
10
different adverse events we were concerned with by
11
the time we started the trial, and fortunately, I
12
must say that none of them has occurred except--and
13
we are going to discuss that--possibly the
14
arrhythmias, but it is important to emphasize that,
15
for example, there was never any particular
16
bleeding from these multiple puncture sites.
17 There was no unusual complication in the
18
postoperative course of these patients, and when
19
the cells were injected in newt immunocompromised
20
mice, there was never any evidence for tumor
21
formation.
22 Obviously, before we started the study, we
23
had to go through a lot of regulatory constraints,
24
indeed, what I did is to discuss with the French
25
FDA and ask them what was approved or not, and the
95
1
game was not so easy because as previously
2
mentioned, there was no precedent.
3 So, they told us, well, this is what you
4
are allowed to do. This is the
kind of culture
5
medium, ancillary product additives which are
6
permitted for human use, so we immediately from the
7
onset designed our cell culture in accordance to
8
the prespecified instructions, and obviously, it
9
was timesaving because when we came back with the
10
process, there was nothing else than to accept it.
11 Well, what about the V-tachs? In the
12
initial series we had 4 patients with sustained
13
episodes of ventricular tachycardia.
14 All of them occurred during the early
15
post-op period, the early three first week,
16
postoperative weeks, and there was virtually no
17
recurrence later on because these patients had a
18
defibrillator put on and only one of them
19
experienced firing of the defibrillator one year
20
later, so it really appears to be a relatively
21
early post-op event.
22 Now, there are different mechanisms which
23
could account for these arrhythmias, in particular,
24
the differences in electrical membrane properties
25
between the grafted cells and the neighboring
96
1
cardiomyocytes. Obviously, other
mechanisms can
2
also be considered, but we really favor the first
3
one because we did an EP study in which we looked
4
at the different membrane properties of the cells.
5 Here, you see a typical action potential
6
of a muscular fiber and here of a cardiomyocyte.
7
Now, if you graft skeletal myoblasts back into a
8
muscle, these cells retain a typical skeletal
9
muscle phenotype, and this is also true for
10 myotubes
which grow in culture.
11 The question is how does it look like when
12
you graft the skeletal myoblasts into the heart.
13
Well, definitely it remains very similar to what it
14
was initially and different from the action
15
potential of the cardiomyocyte.
16 If you expressed it graphically, you would
17
see that the action potential duration is quite
18
different between the cardiomyocyte and the
19
myotube, and this heterogeneity might account for
20
some of these arrhythmias.
21 Now, having said that, the picture is
22
probably more complex and the reason, as you know,
23
and it has been mentioned by Dr. Perin in his talk,
24
is that heart failure by itself predisposes
25
patients to arrhythmias.
97
1 So, I think that as long as we don't have
2
the results of the randomized trial in which all
3
patients have been instrumented with a
4
defibrillator, it will be difficult to conclusively
5
establish a causal relationship between grafting of
6
cells and the occurrence of arrhythmia.
7 I can also tell you that we currently have
8
randomized 34 patients in the Phase II trial and
9
the incidence of arrhythmia has been strikingly
10
low, much lower than in the initial study we had
11
done, so things are probably less clear than they
12
were initially, and once again we have to wait for
13
the results of the randomized trial before we can
14
definitely say yes, there is no relationship
15
between myoblast transplantation and arrhythmia.
16 Anyway, these patients or most of them
17
would require at one point a defibrillator, so it
18
was not a big issue for us to implant those
19
defibrillators in all the Phase II patients.
20 Now, what about efficacy? Now, we have to
21
be extremely careful in the interpretation of the
22
results which are presented because of the
23
multiplicity of the confounding factors.
24 The culture conditions, for example, the
25
Spanish group has used a culture medium which
98
1
contains the patient's own serum, and the
2
conclusion is we had no arrhythmia, so if you use
3
the patient's own serum instead of fetal calf
4
serum, you prevent arrhythmia.
5 I think it is really a simplistic
6
conclusion based on 12 patients, but it can
7
introduce an additional bias.
There is currently
8
no evidence that fetal calf serum is really
9
responsible for the arrhythmias.
10 Dosing has been extremely different and
11
variable from one study to the other, as well as
12
the kinetics of the grafted
area.
13 Once again, any kinetic area is different
14
from a dyskinetic area, which features a
15
paradoxical motion, and, for example, in the U.S.
16
trial, some patients were included who had
17
hypokinesia, which we know can improve just because
18
of the revascularization even if revascularization
19
is not targeted at this particular area.
20 The same for bypasses. In the Spanish
21
study, for example, the cell grafted areas were
22
also bypassed, which makes the interpretation of
23
results impossible.
24 Type of surgery has also been different.
25
In the U.S. study, for example, some patients had
99
1
additional reconstructions of the left ventricle in
2
addition to the bypass surgery, which make things
3
still more complicated.
4
Finally, the method of
outcome assessment,
5
in some studies, the assessment has been
6
centralized at one side, in others, each center has
7
made its own assessment, which obviously makes big
8
differences.
9 This is just to illustrate the variability
10
in the number of cells which have been injected. I
11
don't have the figures for the initial surgical
12
study from Professor Siminiak, but as you can see,
13
there is a wide variability.
14 The U.S. study of Dr. Dib was, as you
15
know, was a dose escalating study accounting for
16
this variability in the numbers.
Dosing is
17
probably important. This is one
study among others
18
showing that there seems to be a tight relationship
19
between the number of injected cells and the
20
functional outcomes.
21 This is the reason why, in our early Phase
22 I
trial, we have targeted a high number of cells,
23
800 million. In the Phase II, we have two arms with
24
two different doses of cells, but the number
25
probably makes a big difference given the high rate
100
1
of early cell death.
2 The characteristics of the grafted
3
segments, as I previously mentioned, have also been
4
different from one study to the other, as well as
5
the method for assessing viability, usually,
6
dobutamine echocardiography, occasionally MRI or
7
PET scan.
8 Same variability in the characteristics
9
of injections, but you see that you can go up to
10
almost 60 injections without any concern related to
11
bleeding, and obviously, the number of injections
12
depends on the extent of the area of infarction.
13 It is also important to look at the cell
14
concentration. We extensively
studied that before
15 I
started doing patients. You have to
find a
16
tradeoff because if you use a large needle, then,
17
you can have large holes and some bleeding
18
problems.
19 If you use a too small needle, you will
20
eliminate the bleeding problems, but the cells may
21
be packed and damaged through their passage, so we
22
ended with a 27-gauge needle which gave an
23
acceptable rate of cell viability.
24 The concentration of cells is important,
25
and probably still more important when you are
101
1
using a long catheter. We are
using a short needle
2
with directly the serum hooked to the needle, but
3
if you are using a long catheter, concentration may
4
make a big difference.
5 Finally, revascularization is occasionally
6
being done in the same area as the area where cells
7
were put in, which completely confuses the results.
8 This is, for example, the Spanish study,
9
what you see is that, what they call the untreated
10 segments, that it is segments which had just
11
bypassed, the wall motion score went from 1.2 to
12
1.1 and 1, but really, this is almost normal
13
motion, so obviously, it makes it easier to
14
demonstrate that in the other segments which have
15
bypass surgery and cells, the improvement was
16
greater.
17 This is a summary of our data from the
18
Phase I trial. We had an
improvement in the
19
functional status and an increase in ejection
20
fraction. These results are
meaningless because
21
these patients had associated bypass surgery.
22 So, we rather looked at the number of
23
scarred segments, and I remind you these were
24
akinetic segments without viability on dobutaminic
25
echocardiography without any possibility for
102
1
revascularization. So, we looked
at the changes in
2
the contractions of these segments which have been
3
grafted with cells.
4 So, initially, obviously, there was no
5
motion since it was one of the inclusion criteria,
6
and afterwards we had, at two different time
7
points, approximately 60 percent of segments
8
regaining some function.
9 I am not saying that these segments were
10
normally contracting, they were not.
There was a
11
slight and modest improvement.
This was a blinded
12
assessment, in other words, we blinded the dates of
13
the echo tapes and asked independent
14
echocardiographers to review them and to grade
15
them. There was a modest
improvement, not normal
16
contraction, but it was sufficient to push us to
17
move forward to the Phase II study.
18 I just show you a couple of examples.
19
This is a flat exterior wall, no motion at all, and
20
this is the same wall with the systolic thickening
21
following myoblast transplantation.
This is the
22
MRI study which does not project on the screen. I
23
have it on the computer, but not on the screen.
24 You see here the interior infarct which
25
has been grafted, and you can appreciate an
103
1
improvement in wall motion in the postoperative
2
period. This is an exterior
infarct. You see the
3
thin wall here, which has been grafted, and this is
4
the post-op pattern with a thickening of the wall.
5 I add intentionally that these patients
6
also had bypasses in the left system.
I don't like
7
the slides where you see pre-transplantation,
8
post-transplantation, just omitting that in
9
addition, there was either bypass surgery or
10
balloon angioplasty.
11 This is another example of an interior
12
infarct pre-transplantation and bypass to the
13
posterior descending coronary artery and the
14
post-op, with an improvement in the wall motion.
15 So, now, can it be due to the
16
revascularization of the PDA? It
is unlikely, but
17
it cannot be eliminated.
18 So, basically, this is the design of the
19
MAGIC, the Phase II trial which has been initiated
20
now in Europe, in different countries in Europe.
21
It is targeted to include 300 patients in different
22
countries, and to emphasize what Dr. Ruskin was
23
mentioning earlier, it is a placebo-controlled
24
study. In other words, patients
following
25
randomization have a muscular biopsy and they have
104
1
eventually injection of a placebo solution in
2
addition to their bypass surgery.
3 There are three arms, one control and two
4
treated groups, one having 400 million, the other
5
having 800 million cells. The
production of cells,
6
and this is probably important, has been
7
centralized in two sites, one in Paris and one in
8
Boston, and it is exactly the same technology which
9
is used in the two sites.
10 The primary endpoint is the improvement in
11
the contractility of the segments which have been
12
grafted with cells in the core lab and in a blinded
13
fashion. In addition to that, we
are obviously
14
looking at major adverse cardiovascular events at
15
the one-year follow-up time.
16 I would like to move on now before
17
finishing to some clinically relevant perspectives
18
which may have really clinical implications in the
19
near future.
20 First of all, so far we have been talking
21
primarily of ischemia cardiomyopathy, but as
22
mentioned by Dr. Perin, there are other causes of
23
heart failure in particular non-ischemic, globally
24
dilated cardiomyopathy.
25 So, we have been interested in assessing
105
1
myoblast transplantation in this particular
2
context, and use a particular genetic strain of
3
hamsters which develop a non-ischemic dilated
4
cardiomyopathy, and randomize the animals to
5
receive either autologous skeletal myoblasts,
6
because phenotypically, these myoblasts are free
7 from the disease, or
culture medium.
8 To make a long story short, you see that
9
there is a definite improvement in function which
10
correlates with a major engraftment of cells in
11
this non-ischemic myocardium. I think it just
12
brings another piece of evidence that maybe
13
something good is occurring.
14 The second problem is cell death.
15
Regardless of the cell type, cell death is
16
extremely high, 80, 90 percent of cells are dying
17
shortly after the injections for a variety of
18
causes, in particular, apoptosis, but also
19
ischemia. It makes sense since
we are injecting
20
cells in scar areas which receive very little
21
vascularization. So, even if
myoblasts are fairly
22
resistant, they die nevertheless.
23 So, now there are several studies
24
suggesting that the co-induction of angiogenesis
25
may be an effective means of improving survival of
106
1
the cells, and ultimately, of improving function of
2
the animals.
3 This is a study comparing transplantation
4
of fetal cardiomyocytes, injection of fibroblast
5
growth factor, or a combination of both. As you
6
can see, function is improved when you combine the
7
two therapies.
8 Recently, we have duplicated this study
9
except that we used myoblasts and another growth
10
factor, and we found exactly similar results.
11 So, there are different ways of inducing
12
angiogenesis, and I know Dr. Epstein is going to
13
discuss that, but the point I wanted to make, this
14
is, you know, the difference in cell survival
15
between myoblasts alone and myoblasts plus an
16
angiogenic growth factor.
17 The point I would like to make is that
18
probably in the future, you will have to deal with
19
proposal of studies trying to combine cell
20
transplantation with some form of angiogenesis just
21
to optimize cell survival and potentiate the
22
benefits of the intervention.
23 A third point regards cycling. This is
24
the muscular biopsy of the patient who died. I
25
previously talked about this patient who died from
107
1 a
stroke. Initially, in this biopsy, and
this is
2
not unexpected, you find fast skeletal myosin and
3
slow type myosin. You don't find
fibers, virtually
4
no fibers co-expressing fast and slow.
5 When we looked at the heart of this dead
6
patient, we found approximately 30 percent of cells
7
co-expressing fast and slow myosin, which means
8
that although once again these myotubes remain
9
myotubes and do not turn to cardiac cells, it seems
10
that some of them may incur some phenotypic changes
11
in response to their new myocardial environment and
12
start expressing slow myosin, which as you know is
13 a
fatigue-resistant myosin.
14 So, this is important and should be put in
15
parallel with this study showing that if you
16
co-culture myoblasts in cardiac cells,
17
cardiomyocytes, this is the green myoblast, this is
18
an antibody against a cardiac troponin and against
19
another cardiac marker, some of the myoblasts, as
20
you can appreciate here, will express some cardiac
21
markers.
22 Now, what is shown here is that if you
23
stop the beating of the co-culture fetal cells,
24
there is no myoblasts which can acquire cardiac
25
cell characteristics.
Conversely, if you subject
108
1
the preparation to a cyclic stretch, then, the
2
stretch makes some of these myoblasts able to
3
express the cardiac markers.
4 In other words, it is quite possible that
5
in vivo, the cyclic contraction of the neighboring
6 cardiomyocytes
leads to the expression of some
7
cardiac markers and leaves the slow myosin in the
8
grafted cells, and the practical implication could
9
be that maybe combining cell transplantation with
10
ventricular stimulation, resynchronization could
11
actually improve the extent by which the grafted
12
cells express slow myosin and become fatigue
13
resistant.
14 So, once again, because this is a
15
clinically used modality, biventricular
16
resynchronization, in the future, we may have to
17
deal with studies trying to combine ventricular
18
resynchronization with cell transplantation.
19 Finally, a few words about the routes of
20
delivery, I have talked about epicardial
21
injections, we are also looking at the scaffolds,
22
which are just put on top of the infarcted area.
23
We are currently working on polyurethane, as well
24
as collagen patches. Obviously,
it is less
25
traumatic and maybe it could reduce a little bit
109
1
the extent of cell death. This
is a pattern after
2 a
couple of weeks.
3 Now, the catheters, I know that this issue
4
will be discussed this afternoon.
I just want to
5
say that from a surgical perspective, I am amazed
6
by the fact that many clinical studies have been
7
initiated in spite of the fact that we had few data
8
on cell retention, functionality, cell viability is
9
not the only issue.
10 It is not because you see myoblasts, that
11
they are going to turn into myotubes.
You really
12
have also to assess the functionality, the ability
13
for these cells to become myotubes, long-term
14
engraftment, as well as the possible interactions
15
between catheter materials and the cells.
16 Most of the studies published so far in
17
the preclinical setting have dealt with technical
18
feasibility rather than functional efficacy, and
19
the various routes have not really been compared.
20 Having said that, we are very interested
21
in the percutaneous routes, and in our group, we
22
have interventional cardiologists working in that.
23 I
must say that we have been primarily interested
24
by the transvenous cell injection through the
25
coronary sinus, and this is the summary of the
110
1
study which was presented last week at the ACC,
2
which is a functional study that is a sheep model
3
of myocardial infarction in which we injected cells
4
through this catheter.
5 You can appreciate that it allows a real
6
delivery of the myoblasts, which turn into
7
myotubes, and this correlated with a significant
8
improvement in function. So,
this is a not a
9
feasibility study, this is a true efficacy study,
10
which is encouraging at least with regard to this
11
particular catheter.
12 So, these are maybe the challenges of
13
the future, in the setting of bone marrow, the
14
famous MAPS, the mesenchymous adult report in
15
cells, which feature distant advantages, possible
16 disadvantages. We are currently working on the
17
cells, cardiac progenitors, and I am sure Michael
18
Schneider will have a lot of things to say about
19
that. Also, possible embryonic
stem cells.
20 What is also important now is to compare
21
cells between them, and not exclusively with
22
controls. This is true, for
example, with the bone
23
marrow.
24 I just would like to show you a recent
25
study that we have done comparing skeletal
111
1
myoblasts, CD133 progenitors or culture medium in a
2
randomized study, and the result is that there is
3
virtually no difference between the CD133 and the
4
myoblasts in terms of function.
5 If you look at histology, it is easy to
6
find the myotube. It is
co-expressed, you know,
7
specific markers like myosin heavy chain. It has
8
been extremely difficult to identify the CD133. We
9
have to rely on PCR to find some of them, which
10
means that probably very few are still present
11
after one month.
12 So, I think it is important to compare the
13
cells, and to some extent, given the amount of data
14
which have accumulated over years, skeletal
15
myoblasts may provide a sort of benchmark for
16
testing other cell types.
17 A similar study is being done in Doris
18
Taylor's lab showing basically that there was no
19
difference between skeletal myoblasts and
20
mesenchymous cells.
21 Once again, we are back to the question
22
which was raised by Michael Schneider.
This is the
23
setting of chronic heart failure, and in this
24
particular setting, current evidence will rather
25
favor skeletal myoblasts.
112
1 This is a completely different setting
2
from acute MI in which bone marrow cells seem to
3
generate impressive results, but they are different
4
patient populations, and it is quite possible that
5
the acute stage of the MI, the bone marrow cells
6
receive appropriate signals which allow them to
7
improve function. The setting
may be quite
8
different where you are dealing with heart failure
9
patients and old scars for which apparently,
10
skeletal myoblasts look more suitable for improving
11
function.
12 So, I just would like to close by two
13 general
slides summarizing a little bit what we
14
have learned from our 10-year experience in the
15
field.
16 Regarding preclinical issues, it is clear
17
that screening experiments have to be done in
18
rodents, but I think it is critically important to
19
validate that in large animal models before
20
arriving to clinical trials.
21 We have a good example of that with the
22
combination of bone marrow cells and JCSF. You are
23
aware of the initial study by Orlig's group showing
24 a
regeneration of mouse myocardium by combining
25
JCSF and bone marrow cell transplantation.
113
1 This mouse study could now be duplicated
2
by two independent groups including Orlig's group
3
in primates, and then we have the Lancet paper last
4
week showing that there was a higher rate of
5
restenosis in patients receiving these two
6
therapies.
7 So, this jump from the mouse to the
man
8
without an intervening large animal model seems to
9
be maybe a little questionable.
10 It is also important that this preclinical
11
study be designed just like clinical studies with
12
appropriate controls and blinded assessment, but
13
having said that, we must be aware that all these
14
models have serious limitations at what point we
15
are not able really to model the very complex
16
situation of heart failure patients with a
17
long-standing coronary artery disease.
A good
18
example is that in our preclinical work, we have
19
never seen any arrhythmia in any of the animals.
20 Regarding clinical issues, it is important
21 to have a well characterized cell therapy
product.
22 I
am not sure that once the feasibility has been
23
demonstrated in small pilot trials, it is necessary
24
to multiply this 10-patient studies, because the
25
amount of information that can be collected from
114
1
these small studies is indeed limited once
2
visibility has been established, and I think it is
3
rather important to move on to larger clinical
4
trials focusing in efficacy and safety, and
5
allowing to draw more meaningful conclusions,
6
safety, the arrhythmias with the myoblasts and
7
possibly instant restenosis with the bone marrow,
8
and efficacy obviously is left ventricular function
9
and major inverse cardiovascular events.
10 So, we are really now at very early stage,
11
as you know, in the field. We
have some evidence
12
that the myoblasts, among others, may improve
13
function, but we still have a lot of basic
14
questions to answer, and in the meantime, I don't
15
think we can make any progress without the
16
implementation of well-designed clinical trials
17
more or less resembling those which have been
18
designed for drugs with appropriate controls,
19
randomization, blinded assessment, and so on,
20
because this is the only way really to know whether
21
hosts will be matched or not.
22 I would like to acknowledge obviously all
23 those
who have participated in this endeavor with a
24
special thanks for those who really did the work.
25 Thank you very much.
115
1 [Applause.]
2
DR. RAO: We will be open for questions.
3 Go ahead, Dr. Borer.
4 Q&A
5 DR. BORER:
First of all, Dr. Menasch, I
6
have to tell you I think that was one of the most
7
exciting talks I have heard in a long time. That
8
was really wonderful.
9 I have some specific questions. I will
10
only give a couple of them, so that everybody else
11
can talk, and then maybe ask a few more later.
12 During your many years of preclinical
13
studies, I am sure you made efforts to determine
14
whether there were aspects of the preparation that
15
could increase the plasticity of the myoblasts, so
16
that they would manifest themselves more as
17
cardiomyocytes than as myocytes.
18 If you did, number one, did you find
19
anything that altered the character, that increased
20
plasticity, because if it did, that suggests that
21
maybe the current preparation isn't the end of the
22
line, maybe one could do better.
23 With that in mind, you mentioned that you
24
saw increased evidence of differentiation into
25
cardiomyocytes or more cardiomyocyte
116
1
characteristics in the beating setting.
2 So, I wonder--I am sure you thought about
3
it--but I wonder if you did culture any of the
4
cells, rather than on flat plates, on flexor cell
5
plates where periodic stress was applied in the
6
culture phase, so that you could perhaps generate
7
some of these cardiomyocyte characteristics before
8
injection.
9 That is one set of questions, and a second
10
question I would like to just put in here, because
11
it's a one-word answer, if these cells are
12
electrically isolated, as you mentioned, how is it
13
that they were caused to be in concert with the
14
rest of the heart?
15 DR. MENASCHE:
Regarding your first
16
question, to be honest, we have not found any trick
17
during the cell culture process which could really
18
increase the transdifferentiation of these
19
myoblasts into cardiomyocytes, and really, I don't
20 think--I am thinking of the works of Chuck
Murray,
21
for example--I think no one has really shown that
22
changes in the culture conditions could really make
23
them turn into cardiomyocytes.
24 The only evidence that can acquire some
25
cardiac-like characteristics is this expression of
117
1
slow type myosin. We are
currently exploring the
2
possibility maybe of increasing the expression of
3 these
slow myosin isoform by implantation
4
stimulation of the cells, but these are experiments
5
which now are going to be done, and this was the
6
reason why I was mentioning ventricular stimulation
7
as a potential additive in the future to the
8
clinical trials.
9 But to summarize, no, we have not found
10
any particular intervention, although maybe we have
11
not found the right one, which could increase the
12
proportion of cardiac-like skeletal myoblasts.
13
Now, the mechanisms, I don't know; from scratch, I
14
don't know.
15 There are different possibilities. One is
16 a
limitation of remodeling. I am not sure
if there
17
is a predominant mechanism because in our patients,
18
we have had some evidence of improved systolic
19
function, but we have never seen a reduction in
20
left ventricular diastolic dimensions.
21 Another possibility is that GAB junctions
22
are not the only ways for electrical impulses to
23
travel across the heart, and as you have seen on
24
the film, these cells retain excitable properties.
25
In order words, if you excite them, they will
118
1
contract.
2 So, it is not completely impossible that
3
in areas where physically, they are very close to
4
the neighboring cardiomyocytes, they may be
5
directly excited by electrotonic currents, and
6
there is a third hypothesis we are currently
7
exploring, and which is the paracrine hypothesis.
8 It is quite possible, and it has been
9
shown for bone marrow, for example, that these
10
cells secrete various growth factors or cytokines,
11
and so on, that can positively affect the function
12
of the host cardiomyocytes.
13 For example, in our studies we have found
14
that myoblasts and myotubes from patients secrete
15
very large amount of IGF-1, which has important
16
effects on tissue regeneration.
So, maybe it has
17
nothing to do with their countertype properties,
18
but rather with the fact their behavior, small
19
factors releasing good factors for the heart.
20 So, we are currently playing with all
21
these hypotheses, but I don't have any definite
22
answer.
23 DR. RAO: So,
it is pretty clear, like Dr.
24
Schneider pointed out earlier, that mechanism is an
25
issue that is still not clear.
119
1 DR. MENASCHE:
No, it is not clear at all.
2
The only thing is I don't think that you could
3
infer from the lack of connexin 43 expression, that
4
improvement in function is not possible. I think
5
both should be dissociated.
6 DR. RAO: Dr.
Kurtzberg.
7 DR. KURTZBERG:
You mentioned studies with
8
bone marrow derived AC133 cells.
I wonder what the
9
rationale behind the selection was and why you
10
thought there would be an advantage to using
11
selected cells over whole bone marrow.
12 DR. MENASCHE:
The reason is that we first
13
did a large animal study with whole bone marrow in
14
the sheep model of myocardial infarction, once
15
again, a chronic infarct. So, we
injected the
16
whole bone marrow and we didn't find anything, no
17
improvement in function, no limitation in
18
remodeling, no evidence for transdifferentiation of
19
cells.
20 So, we said, well, maybe the whole bone
21
marrow is not the appropriate medium for this
22
particular setting, let's try to purify the cells,
23
and we went to the CD133. The
results were
24 slightly
better in that. There was some
25
improvement in function compared with controls, but
120
1
the improvement was not greater than that we got
2
with the myoblast.
3 Currently, we are comparing now myoblasts
4
with AMAPCs, so we have tried to pick the different
5
populations, in a stepwise approach, test all of
6
them.
7 DR. RAO: Dr.
Cannon.
8 DR. CANNON:
Thank you for your talk, it
9
was most interesting. I am
Richard Cannon from
10
NHLBI.
11 My question is in any of your preclinical
12
animal work, did you ever inject your myoblast
13
culture preparations or cell suspensions into the
14
circulation to see where they might end up and what
15
toxicity they might cause.
16 This may not be an issue with an
17
intra-operative injection into scar, but I would
18
imagine with a catheter-based approach, it is
19
conceivable that despite the operator's best
20
efforts, some of these cells might be injected into
21
systemic circulation.
22 Do you have any data on where the cells
23
end up, do they lodge in the brain or the kidneys,
24
do they cause any toxicity or injury to other
25
tissues?
121
1 DR. MENASCHE:
Well, in the preclinical
2
studies we have done, we have not found evidence
3
for, first of all, all the injections were direct
4
intramyocardial injections, so it may be difficult
5
to find them in the brain or in the liver.
6 We have not found them disseminated
7
throughout the body, but I must say that maybe if
8
we had done more extensive studies, autopsy studies
9
of the brain or the lungs, or any other organ,
10
maybe we could have found some of them.
11 We have never injected intentionally the
12
cells intravenously just to see what was happening.
13 I
don't believe it is a real issue because even
14
when you are injecting them intraoperatively in
15
humans, it is clear that some of them are escaping
16
through the lymphatic system or in the venous
17
system, and so far we have never seen any evidence
18
for unexpected or unusual complications.
19 But I agree with you that if you are
20
expecting some leakage of the cells in the systemic
21
circulation, this is probably a point that should
22
be addressed more extensively than we have done.
23 DR. RAO: I
have a practical question.
24
Did you, when you looked at the cells, ever look at
25
BRD incorporation to see whether cells continue to
122
1
divide at any time?
2 DR. MENASCHE:
Yes, absolutely, including
3
in the human trial. In the human
trial, we always
4
keep aliquot, initially, we kept aliquots of cells
5
and just let them grow, and this is why we have
6
been able to show that these cells were
7
differentiating into myotubes including in heart
8
failure patients.
9 DR. RAO: Did
you ever take your samples,
10
and look at freeze/thaw? You
know, you grow them
11
in cell culture, can you freeze these cells and do
12
they behave the same way when you send them to
13
another site like you are planning in the Phase II
14
trial, for example?
15 DR. MENASCHE:
Absolutely, we have done
16
that and we have validated that after thawing, they
17
retain their ability to differentiate into
18
myotubes.
19 DR. RAO: Have
they been done in any
20
transplant paradigm, or has it only been done by
21
looking at they are forming myotubes in culture?
22 DR. MENASCHE:
Both. We have several
23
preclinical studies in rats and in sheep, in which
24
we have used cryopreserved and thawed cells with
25
apparently functional outcomes similar to those we
123
1
had with fresh primary myoblasts.
2 This is the reason why actually we got
3
permission to freeze them should they become
4
necessary for logistical reasons in the Phase II.
5 DR. BORER: You
mentioned that about 95
6
percent of the cells that you inject have CD56
7
characteristics. That suggests
that there is some
8 alteration
in some of the cells or perhaps a
9
different cell line is growing in parallel, in the
10
cultures that you are using, so I wonder, number
11
one, how many passages do you use before
12
administering the cells, and, number two, is the
13
reproduction error rate increased with passage in
14
any meaningful way, and does it make any
15
difference?
16 Obviously, a lot of these cells that you
17
inject are nonviable. When you
inject them, they
18
don't survive. I don't know
which ones are
19
surviving and which ones aren't. But it seems to me
20
that the number of passages employed may affect the
21
ultimate outcome of the injection, and I wonder if
22
you have some data on that from your preclinical
23
work.
24 DR. MENASCHE:
We used three to four
25
passages, but it has been shown by Chuck Murray
124
1
that if you multiply passaging, you may end up with
2 a
population of differentiation of effective cells,
3
in which case you might end up with some unexpected
4
overgrowth without any functional benefit.
5 So, it is probably important not to
6
multiply passaging too much.
With three to four
7
passages, we have been able to reach the target
8
numbers of cells, 400- or 800 million of cells. We
9
don't go beyond that.
10 DR. TSIATIS:
In your randomized clinical
11
trial, do you actually have formal stopping rules
12
for either safety or efficacy, and, if so, what are
13
they, or what is the general philosophy for
14
monitoring?
15 DR. MENASCHE:
It is primarily based on
16
the judgments of the DMSB given the type of
17
surgical population we are dealing with. It is
18
difficult to have stopping rules, just as you can
19
have with drugs, for example, for each adverse
20
event is reviewed by the DSMB, and based on that,
21
they would decide whether the study has to be
22
stopped or not.
23 DR. RAO: Dr.
Neylan.
24 DR. NEYLAN:
Thank you.
25 I have another preclinical question. I
125
1
was wondering if you had an opportunity to compare
2
the morphology and functionality of the myoblasts
3
when these are injected either into the akinetic
4
areas or perhaps into an area resected, and thus
5
undergoing a normal reparative milieu, and whether
6
perhaps under that milieu, there might be a
7
different behavior of these cells or expression.
8 DR. MENASCHE:
Really, basically, our
9
model has been the model of, you know, coronary
10
ligation creating myocardial infarction, so you
11
really end up with an akinetic
scar. I cannot
12
answer this question.
13 DR. NEYLAN:
You never had the chance to
14
maybe resect that, maybe adhere to the natural
15 surgical tendency of cutting things out.
16 DR. MENASCHE:
I try to refrain from that.
17 I
have cardiologists as bodyguards, so it would
18
just refrain you from doing that.
19 DR. MULE: You
had mentioned that the vast
20
majority of cells that are injected will die, and
21
clearly, there is room for improvement with perhaps
22
increasing angiogenesis, and so forth.
23 About the kinetics of myotube formation in
24
the ischemic areas, is it a dynamic process, in
25
other words, once you inject the cells, myotubes
126
1
will form over a given period of time, and then no
2
more tubes will form, or additional tubes are
3
generated over a prolonged period of time, and do
4
those tubes, when they are formed, remain viable
5
for the extension of the observation period?
6 DR. MENASCHE:
It is difficult to tell
7
you. The kinetic studies
indicate that although a
8
substantial number of cells die, the remaining ones
9
obviously proliferate in different shape over a
10
period which seems to extend, say,
two to three
11
weeks.
12 At least in patients when we have seen
13
improvements, we have never seen improvements
14
before one month, and in animals, it is difficult
15
to see any improvement before two weeks.
16 Now, afterwards, the longest follow-up we
17
have is 14 months in animals, but we cannot know
18
whether the myotubes that we found at the end of
19
the experiments were present since the beginning or
20
whether they have been continuously regenerating.
21 The interesting observation, however, I
22
don't know whether it really answers your question,
23
is that these myotubes harbor new myoblasts, so
24
when you look at them with electron microscopy, you
25
clearly see, on their basal lamina of these
127
1
myotubes, newly formed myoblast cells, so they are
2
able to regenerate their own pool of precursor
3
cells.
4 Now, whether these cells participate in
5
the formation of new myotubes, I don't know.
6 DR. RAO: The
last comment, Dr. Noguchi,
7
and then we move on.
8 DR. NOGUCHI: I
am sorry to have prolonged
9
this, but it is just fascinating.
Of course, FDA
10
always loves these controls, but to follow up on
11
Dr. Mul's question, is it a question of liability,
12
do the cells have to be alive, or you have a
13
myotube has some structure, and then I just recall
14
there is, in tumor biology, an old effect called
15
the reverse effect where if you have a few viable
16
cells with a lot of dead cells, you can actually
17
get tumors developing from one cell where normally,
18
you might need a million or 10 million.
19 I am just wondering if you have done any
20
mixtures of dead and live cells to really see how
21
much is viability, how much is surrounding stuff.
22 DR. MENASCHE:
No, we have not done that
23
intentionally. We have just
completed a study in
24
which we have looked more carefully at the patterns
25
not only of cell death, but
also of cell
128
1
proliferation.
2 So, we know that we have this mix of dead
3
cells and living cells, but we have not done an
4
intentional mixing of them to see whether there was
5
any tumor formation.
6 Regarding oncogenicity, we have learned a
7
lot from our colleagues working in the field of
8
dystrophic myopathies, and it really seems that
9
these cells have a very low tumor-retaining
10
potential.
11 In the newt mice in which we have injected
12
our human myoblasts, we have never seen any tumor
13
in spite of the fact that many of these cells
14
expectedly died.
15 DR. RAO: Thank
you, Doctor.
16 We will move on to our next speaker, Dr.
17
Epstein.
18
Bone Marrow Cell Therapy for Angiogenesis:
19 Present and Future
20 DR. EPSTEIN:
It is really an honor to
21
have been asked to speak to this very august group.
22 I wanted to emphasize because I really do
23
think it is important in this, an ever-growing
24 field
to make sure that disclosure is presented,
25
and I have a number of potential conflicts of
129
1
interest, which I hope in no way influences what I
2
will be talking about to you for the next 20 or 30
3
minutes.
4 I will talking about bone marrow cells and
5
angiogenesis. I wanted to start
out and make a
6
careful distinction. Dr.
Schneider brought this up
7
in his earlier questions, but basically, what we
8
are considering today is really the use of bone
9
marrow cells, stem cells, progenitor cells for
10
myogenesis, but also for angiogenesis.
11 It is critically important to understand
12
that these are very distinct targets with
13
undoubtedly different mechanisms and certainly
14
very, very different issues, and therefore will
15
have a profound impact on how the FDA I think
16
judges whether or not a particular proposal is
17
meritorious.
18 For example--and you have heard this very
19
eloquently discussed by Dr. Menasch--for
20
myogenesis, the transdifferentiation of adult
21
progenitor cells or skeletal myoblasts is a
22
critically important issue. Maybe
you don't need
23
transdifferentiation into cardiac myocytes to
24
improve myocardial contractility, but nonetheless,
25
it is a very important issue to consider.
130
1 Also, if you think about how many
cells
2
are present in a large myocardial scar, the issue
3
of adequate numbers of cells to replace the scar to
4
cause a significant biologic effect has to be
5
considered, and you have heard a very eloquent
6
presentation and some demonstration relating to
7
this.
8 Now, the issues relating to angiogenesis
9
are, as I indicated a moment ago, different.
10
Transdifferentiation is really not an important
11 factor, because what has been recognized most
12
recently is that cytocrine secretion, exerting a
13
paracrine effect can induce proliferation and
14
remodeling of existing vessels.
15 So, it is not necessary, although it may
16
happen, it is not necessary for the cells that you
17
are injecting to turn into blood vessels. They
18
could induce the development of already existing
19
blood vessels, and the adequate number of cells
20
relating to angiogenesis is not nearly of similar
21
concern secondary to these paracrine effects that
22
have an amplifying activity of the individual cells
23
that are injected.
24 So, these are very different issues, each
25
are very important. I think the
path to myogenesis
131
1
is going to be a longer one. I
think that there
2
are a lot of problems that still have to be solved,
3
and I think the issue of angiogenesis, we have gone
4
along that path for probably a longer period of
5
time, and my sense is that we are closer to pivotal
6
clinical trials even when one considers cell
7
therapy, but I will be focusing my remarks on
8
angiogenesis.
9 The first
thing I wanted to point out,
10
which is obvious to anyone who is involved in the
11
field now, is how complex the molecular and
12
cellular mechanisms are that are involved in
13
collateral formation.
14
This is a slide I always like
to show.
15
This is a cartoon showing different genes
16
expressed, either increased expression or decreased
17
expression, four different times, different
18
amounts, two actually wind up with a collateral.
19 So, there are multiple, multiple genes
20
that are necessary to actually form a new
21
collateral vessel. Just to
illustrate the
22
importance of interactions between different
23
angiogenic cytokines, I wanted to show you the
24
results of a study that we did a couple of years
25
ago using a rabbit ear.
132
1 So, here is the ear.
It is supplied by
2
three major vessels. This is a
laser doppler image
3
which is color coded for velocity.
Red is highest
4
velocity, green is intermediate, and blue is low
5
velocity. If you tie off two of
these three
6
vessels, you have a marked decrease in flow, and
7
the nice thing about the rabbit ear is that you
8
could observe this hourly if you wanted, and you
9
could with this laser doppler do repeated analyses
10
of the changing flow with time.
11 What you can also do is focus on a
12 particular area of interest and measure
tissue
13
perfusion and the change in tissue perfusion that
14
occurs with time.
15 Here is the tying off of these vessels,
16
resulting in a profound decrease in flow, which
17
gradually recovers over several weeks, and in this
18
particular model, it is quite interesting. It
19
plateaus off below normal flow, so this is a model
20
of chronic hypoperfusion, which makes it kind of
21
interesting.
22 In this model, we looked at what happens
23
with endogenous VEGF levels, and VEGF is a key
24
angiogenic molecule, so we measured VEGF by western
25
blot before the induction of ischemia, and there is
133
1
essentially no VEGF present, however, if we measure
2
VEGF levels throughout the course of this, and even
3
at the end, there is a low level of VEGF present.
4
So, this is further indication that we are dealing
5
with a chronically ischemic preparation that has a
6
background of VEGF present.
7 The next issue that we wanted to document,
8
we took the model during the period of chronic
9
ischemia, and we take that now as our starting
10
point for this experiment, where we had an
11
angiopoietin-1 gene within an adenoviral vector, so
12
that is the transgene, which we inject
13
intradermally in the region of hypoperfusion in the
14
ear.
15 We inject it and we see over the course of
16
time, a major increase in collateral flow and
17
tissue perfusion, but remember there is background
18
VEGF present. If we coinject with the adenovirus
19
expressing angiopoietin-1, an inhibitor of VEGF,
20
and this is a soluble VEGF receptor, so it sops up
21
and inactivates whatever VEGF is present, it
22
obliterates the collateral-forming effects of
23
angiopoietin, so it just is an example of how you
24
need multiple factors to develop your collaterals.
25 If any one of these is perturbed, you
134
1
could seriously influence the course of collateral
2
development.
3 Now, just as background for the cell
4
therapy, there have been a number of adequately
5
powered, randomized studies that have been
6
performed using individual cytokines for
7
angiogenesis, and basically, these are either basic
8
FGF or VEGF used in the coronary circulation or the
9
peripheral vasculature, either the protein was
10
injected or a gene encoding the protein were
11
injected.
12 As of the moment, there have been no
13
definitive and robust beneficial results. There is
14
trends, there is some encouragement, there is some
15
early positive results, but nothing to really get
16
excited about.
17 Of course, as I indicated, all of these
18
randomized studies to date have involved a single
19
agent to promote collateral development, and it was
20
these considerations about four or five years ago
21
that provided the impetus for developing and
22
testing a second generation of angiogenesis
23
strategies, which is the use of cell therapy, which
24
had the potential to deliver multiple
25
collaterogenic cytokines.
135
1 I just wanted to show this slide. I am
2
not an expert at all in stem cells, progenitor
3
cells, but I just wanted to indicate to you what
4
has been used in clinical trials or in late-stage
5
preclinical trials. So,
hematopoietic stem cells
6
characterized by positive CD34-133, which do
7
progress to endothelial progenitor cells and then
8
to endothelial cells, which have been shown to lead
9
to an increase in collateral flow.
10 Now, more recently, monocyte lineage cells
11
have been demonstrated. These
are characterized by
12
the lack of CD34, but having CD14 and 45 MAC-1,
13
these monocyte lineage cells have been shown, not
14
to produce endothelial cells directly, but
15
nonetheless, are capable of inducing collateral
16
formation.
17 We have used freshly aspirated bone marrow
18
cells that have been filtered
and directly
19
injected. These are autologous
into pig ischemic
20
hearts, as well as patients. You have heard about
21
monocyte-derived bone marrow cells.
Dr. Perin used
22
these in his study.
23 We have been also working now with
24
mesenchymal stem cells or stromal cells. There are
25
multiple terms that have been used to describe
136
1
these. These are CD34-negative,
45-negative cells,
2
and these have been shown to produce collaterals.
3 Now, I won't get into this in any detail,
4
but you should be aware of the fact that some of
5
these cells are believed to incorporate into
6
developing collaterals, with that being a major
7
mechanism by which they enhance the development of
8
collaterals, whereas, other interventions are not
9
believed to have that as a major mechanism, but the
10
major mechanism being the secretion of all sorts of
11
cytokines and growth factors that lead through a
12
paracrine effect to the development of either new
13
collaterals or the enhancement of existing
14
collaterals.
15 What I will be talking about, because all
16
of our recent work has been done with these
17
mesenchymal stem or stromal cells, I will be
18
talking about that for the next few minutes, and
19
these we refer to MSCs.
20 So, these MSCs, just to start or justify
21
our further studies, were cultured in vitro and
22
assayed. The conditioned medium
was assayed, and
23
here is our control cells which produce small
24
amounts of VEGF MCP-1 and FGF, but the MSCs produce
25
really quite large amounts of these angiogenic
137
1
cytokines.
2 So, we were very excited about that,
3
thinking that they could be
little factories that
4
might enhance collateral development.
So, this is
5
the mouse hind limb model. This
is laser doppler
6
imaging, as I showed you with the rabbit ear. Here
7
is the mouse's tail and the two hind limbs, and the
8
femoral artery is ligated at day zero, and this is
9
followed now every few days, and you can see there
10
is some return of function under control
11
conditions.
12 This is just injecting media that had not
13
been exposed to cells, and here is what we see with
14
media alone and with the control cell.
This is
15
mature aortic endothelial cells.
16 But then when we inject into the hind limb
17
MSCs, we see a quite marked improvement in
18
perfusion, and this can be quantitated as shown in
19
this slide. So, this was a very
exciting
20
demonstration for us, which was repeated multiple
21
times in different experiments.
22 Now, I won't belabor the number of studies
23
that have been done here. It is
in your handout
24
that was distributed, I think it is page 20 and 22,
25
but I will just go over a couple of the highlights.
138
1 There have been 7 or probably 8 published
2
studies in either chronic ischemia--and this is
3
angiogenesis studies, no myogenesis--in chronic
4
ischemia or in acute myocardial infarction.
5 The points to be made are, number one, all
6
of these studies have shown safety, they have shown
7
feasibility, and they have all showed positive
8
trends, they have been encouraging, but as Dr.
9
Menasch very carefully pointed out with his own
10
myogenesis studies, when you are dealing with such
11
small numbers of patients, none of these studies
12
was randomized, double-blinded.
There is no way
13
you could draw any conclusions regarding efficacy.
14 So, it is encouraging and it certainly
15
would indicate that additional studies are
16
necessary, but we can't make any inferences whether
17
the strategies that work in an animal model very
18
reproducibly necessarily work in humans.
19 Now, I want to point out, and I think we
20
have to be aware of this upfront, and any
21
investigator who is involved in the field has to be
22
aware of it, that there are potential problems with
23
any angiogenic strategy including cell-based
24
strategy.
25 For example, genetics. Here are some
139
1
beautiful studies done by Birgit Kantor in
2
collaboration with us. This is microscopic CT
3 imaging
of two different strains of mice. This is
4
the femur, tibia, and this is the femoral artery,
5
and the femoral artery had been ligated, and you
6
can see that the C57 black 6 mouse has an
7
extraordinary capacity to develop collaterals,
8
however, about C. mouse, same ligation site, has a
9
paucity of collaterals. Well,
clearly, the same
10
thing must relate to humans.
11 Another thing that was raised earlier is
12
the enormous variability amongst patients to
13
respond to angiogenic interventions.
These are not
14
patients, these are mice, and this is the typical
15
experiment that I showed earlier, looking at laser
16
doppler perfusion.
17 The mouse has the femoral artery ligated,
18
and there is a gradual recovery of flow in young
19
mice, however, it you look at knockout mice that
20
have high cholesterol levels, their capacity to
21
develop collaterals is significantly impaired.
22 Now, if you take an old mouse--these are
23
mice about 18 to 20 months of age--they are really
24
having trouble developing collaterals, and then if
25
you take an old mouse who has high cholesterol
140
1
levels, they are really in bad shape.
2 Now, no one has demonstrated this
3
relationship in humans, but I am certain that it
4
occurs. So, there is going to be
different
5
capacities of different individuals to develop
6
collaterals, and undoubtedly reflecting different
7
potential to respond to angiogenic interventions.
8 Now, here is something I would like Dr.
9
Schneider to look at, because he said this has not
10
been published before, but it has been published.
11
This is in I think JACC in 2002, but when we did
12
our clinical study, injecting autologous filtered,
13
freshly aspirated bone marrow cells into ischemic
14
myocardium of patients, we took an aliquot of these
15
cells and cultured them, and looked at VEGF
16
production, as well as other cytokine production,
17
and over the course of time, one sees an increase
18
in VEGF production, so these cells do have the
19 capacity
to produce different angiogenic cytokines
20
including VEGF, but that is the mean data.
21 If you look at the individual data, there
22
is marked patient-to-patient variability in the
23
capacity to express VEGF, so here is a patient who
24
really has a great capacity to produce VEGF,
25
whereas, this is a patient who hardly can produce
141
1
VEGF at all, and it is not a great stretch to think
2
that this patient may not respond as vigorously to
3
cell therapy as the patient whose cells have a
4
great capacity to produce VEGF.
5 Now, we didn't look at enough patients to
6
be able to make such correlations, but I am sure
7
that this is an issue that has to be addressed, as
8
Dr. Schneider really pointed out before.
9 Now, let's look at the cells we are
10
injecting, and this was also raised earlier, so we
11
are looking here at a HIF-1--I will just get into
12
that in a moment--but it is a transcription factor
13
that is a key modulator of the cells response to
14
ischemia, so let's take it for the moment that it
15
is a key angiogenic factor, so this is a Western
16
showing HIF levels, and the first thing I want you
17
to concentrate on is under normoxic conditions,
18
young and old, HIF is not present, it is mostly
19
absent as a matter of fact, in the absence of
20
hypoxia.
21 Now, in the young mice, if you expose
22
these cells to hypoxia over 12 hours, there is a
23
major increase in HIF protein, and then that has
24
important compensatory effects on the cells'
25
response to hypoxia, however, cells derived from
142
1
old mice have a markedly impaired ability to form
2
HIF in response to hypoxia, so there are
3
age-related changes in the capacity of cells--these
4
are MSCs--to perform in a way that we would expect
5
them to if they were going to have a potent effect
6
on collaterals.
7 So, HIF is a master switch gene in the
8
presence of hypoxia, a heterodimer is formed,
9
HIF-1-alpha, and HIF-1-beta, which attaches to the
10
promotor of many genes and turns these genes on,
11
and amongst the genes are multiple genes related to
12
angiogenesis, VEGF, VEGF receptor, FGF, et cetera.
13 Now, just to show the biologic effects of
14
what I just showed you before, that is, the
15
inability to increase HIF protein in response to
16
hypoxia, here are young and old mice, and we are
17
looking at VEGF levels. These
are cells, MSCs
18
growing in culture, and here is the intrinsic VEGF
19
production.
20 When we expose young cells to hypoxia,
21
there is a major increase in VEGF production,
22
mediated mainly by HIF-1 reduction, but old mice
23
not only have the lower levels of HIF-1, but they
24
have a lower target production of HIF-1, that is,
25
VEGF production.
143
1 So, these are real phenomenon I think that
2
we have to be aware of and begin to start thinking
3
about when we are dealing with any angiogenic
4
intervention, but certainly with the cell
5
therapies.
6 So, these considerations provided the
7
impetus to test another generation of angiogenic
8 strategies, and this relates again to one of
the
9
earlier questions of the panelists, and we are very
10
much involved in genetic manipulation of these MSCs
11
to see if we could further enhance their ability to
12
secrete angiogenic cytokines and so to improve
13
collateral flow.
14 This is a construct of Genzyme. They have
15
been very helpful in working with us in this. In
16
the absence of severe hypoxia, these two dimers of
17
HIF, HIF-1-beta and alpha, there is no heterodimer
18
formed because HIF-1-alpha is rapidly degraded.
19 So, to overexpress HIF-1-alpha, so that
20
the heterodimer can form, we transfect these cells
21
with an adenoviral vector that has the HIF-1-alpha
22
transgene and that has a deletion insertion,
23
putting on a herpes sequence VP16, which stabilizes
24
the protein under normoxic conditions, so we are
25
able to overexpress HIF-1-alpha, the heterodimer
144
1
can be formed, and the genes can be transactivated.
2 So, we then looked at the capacity of this
3
intervention to cause these MSCs to secrete
4
angiogenic cytokines, so this is VEGF.
The cells
5
now are exposed to just hypoxia, and you can see
6
there is about a doubling of the amount of VEGF
7
present, but when we transfect these cells with the
8
HIF-1-alpha, there is a huge increase in VEGF
9
production, and the same thing is true for
10
fibroblast growth factor.
11 So, this was really exciting to us because
12
we saw that we could genetically manipulate these
13
cells to make them at least in vitro more like a
14
better collateral enhancer.
15 We then went to our mouse ischemic hind
16
limb model to test this concept.
Here are our
17
control cells, mature aortic endothelial cells.
18
Here are our non-transfected MSCs, and here are our
19
transfected MSCs. So, not only
do we see an effect
20
in vitro, but we see what would have been predicted
21
from the in vitro effects in vivo.
22 I think that this is probably something
23
that we have to think about given the effects of
24 various risk factors on the ability of cells
to
25
achieve their desired effects.
145
1 I did want to point out for the panel that
2
down the line, we are not only going to be talking
3
about cells, but cell-derived products in
4
cardiovascular therapy, and I will just spend a
5
moment on this, and mention the effects of
6
conditioned medium on collateral development.
7 As I showed you before, if you grow cells
8
in culture and allow them to produce whatever
9
goodies they are producing, and then you take the
10
media and you inject that media into the ischemic
11
hind limb of mice, you get--well, that will be the
12
next slide--but here is what I showed you before,
13
so these in the media contains more VEGF, more
14
MCP-1, more FGF, and multiple other gene products
15
that we haven't tested, but when we put this
16
conditioned medium into the ischemic hind limb of
17
the mouse, here is the control again, here is our
18
control.
19 Here is the injection of the MSC
20
conditioned medium. We see that
the media alone
21
has the capacity to increase collaterals. So, this
22
undoubtedly is going to be something that you are
23
not going to see an application to this, I don't
24
think in the next few months, but in the next six
25
months or a year, I think that the cell products is
146
1
another very interesting way to use the angiogenic
2
potential that bone marrow cells have.
3 Just to show that this is biologically
4
important, we looked at the number of collateral
5
vessels the media increased, the number of
6
collateral vessels, the strength of the leg as an
7
ambulatory score, the media increased that, and
8
also the amount of atrophy that occurs in the calf
9
as a result of ischemia, and the media injection
10
decreases that, so it was a biologically relevant
11
intervention.
12 I just wanted to mention safety concerns,
13
and I was interested in Dr. Menasch's comment
14
about this. There are multiple
well-known ones
15
that are usually tracked, and I won't get into
16
this, but I would just alert you to something that
17
is more theoretical than proven, but I think you
18
have to be aware of it and at least consider it
19
when you are considering the safety of angiogenic
20
interventions.
21 That is--and this is a general rule that I
22
have come up with--whatever induces angiogenesis,
23
induces atherogenesis, and I refer to this as the
24
Janus phenomenon. Janus was a Greek god with two
25
heads, so that when he looks out one way, he is
147
1
also looking at the other. That
goodness can also
2
be badness, there are no free lunches.
3 So, I would think that any angiogenesis
4
intervention, one of the potential side effects
5
that one should look for is the acceleration of the
6
atherogenic process.
7 This is just a slide that is still in
8
development, but basically, it shows that when you
9
induce ischemia in a hind limb, you have decreased
10
PO2. This activates cytokine
release, which
11
activates bone marrow cells, splenocytes, many
12
inflammatory cells.
13 We now know that inflammatory cells are
14
critically important to the development of
15
collaterals. Macrophages have
been shown to be
16
critical. We have shown that
both CD4,
17
t-lymphocytes, and CD8 t-lymphocytes are critical
18
to collateral development, but these same factors,
19
these same inflammatory factors also have been well
20
described to lead to, much longer than
21
angiogenesis, an acceleration of atherosclerosis.
22 So, if it causes angiogenesis, think very
23
hard as to whether it might be worsening the
24
atherosclerotic process.
25 The conclusions:
single molecule-based
148
1
strategies to improve collateral flow, although
2
effective in animals, have yet to be proven
3
efficacious in patients.
Cell-based strategies
4
have great promise because of the ability of bone
5
marrow derived progenitor cells to secrete multiple
6
collaterogenic cytokines.
7 However, cell-based therapies also have
8
the potential of achieving suboptimal effects
9
because of the effects of aging and other risk
10
factors on cell function. The
optimal strategy has
11
yet to be identified, but genetic manipulation of
12
cells would appear to hold great promise, and use
13
of cell products, such as conditioned medium
14
derived from cells, will also undoubtedly be
15
explored as a therapeutic strategy in the near
16
future.
17 Thank you.
18 [Applause.]
19 DR. RAO: Thank
you, Doctor.
20 We have time for a few quick questions.
21 Q&A
22 DR. SCHNEIDER:
Steve, as an exploratory
23
tool, conditioned medium for angiogenesis makes a
24
lot of sense for the reasons that you articulated,
25
but as a therapeutic product, that would be true
149
1
if, and only if, conditioned medium contained
2
products that could not be identified or could not
3
be added combinatorily from defined factors.
4 So, it seems to me it will be especially
5
useful in those conditioned medium experiments to
6
test the effect of specific blockers and find out,
7
at a reductionist level, what the components are.
8
If it were as simple as angiopoietin and VEGF, one
9
could use angiopoietin and VEGF.
10 DR. EPSTEIN:
Right, it's a very good
11
point. My own feeling is, having
been in this
12
field now for 12--more than that--14 years, it is
13
so complex and the number of factors that are
14
involved in collateral development are not 4 or 5,
15
but they are dozens, and maybe even more than that,
16
that I personally will not waste time trying to
17
figure out what two products are enough or what
18
three products, I don't believe that, but the
19
cells, they know how to develop collaterals, I mean
20
they are doing it all the time, so I will go with
21
cell therapy, and I will allow someone else to look
22
at what combination of three factors might be
23
optimal.
24 It may be that you could find such
25
factors, but don't forget, not only do you have to
150
1
know what factors are present, but you have to know
2
in what concentrations, and so on.
3 I think the cells, they are eliminating so
4
many of the issues that if we were going to look at
5
individual cytokines, we would have to explore for
6
years, so it is a good point, but I think that the
7
practical issues, given the huge complexity of
8
this, would be overwhelming.
9 DR. SCHNEIDER:
To follow up on your
10
comment, which I would share, that it is extremely
11
likely that engineered cells will outperform naive
12
cells, I would like to ask the
participants from
13
FDA what additional hurdles are seen in the
14
consideration of gene-engineered cells to be
15
applied to these therapeutic situations.
16 DR. NOGUCHI: I
think we can answer in
17 general
that actually, we have a fairly rich
18
experience with gene-modified cells that have been
19
given to individuals for a whole variety of
20
diseases, not too many for cardiovascular, but I
21
wouldn't expect that we would have very much
22
difficulty in really being able to handle that.
23 DR. RAO: Dr.
Mul.
24 DR. MULE:
Combining your presentation
25
with Dr. Menasch's, I was sitting here wondering
151
1
what is known about, if one takes whole bone marrow
2
cells and perhaps Dr. Kurtzberg can add to this,
3
with Dr. Menasch's studies, if there are cells
4
within the marrow that can give rise to myotubes,
5 and
you combine that with a population of cells
6
that could be responsible for collaterogenesis, the
7
issue is are they the same cell or are we at a
8
period in time where we can
identify two subsets
9
within the marrow that conceivably could be
10
combined to overcome some of the issues of
11
viability that Dr. Menasch has talked about.
12 If the answer is no, we are not there yet,
13
then, it begs the question if one were to use an
14
adenovirus to introduce a gene to improve
15
collaterogenesis into the cell population, that is
16
identifiable for producing myotubes, the issue is
17
does that manipulation adversely impact the ability
18
of that cell to create myotubes.
19 DR. EPSTEIN:
Well, those are sensational
20
questions, and I have never thought of this last
21
one before, but it is certainly--you know, it is so
22
easy to do harm, and it is so hard to do good, so I
23
mean your question is very apt, I mean does the
24
very expression of the cytokines that enhance
25
angiogenesis, might it interfere with myogenic
152
1
potential, I don't think anybody has done that
2
experiment. Hopefully, the answer will be no, but
3
it certainly is an experiment that has to be done.
4 The MSCs that we are deriving from the
5
bone marrow do not differentiate into myoblasts,
6
and it would be a very interesting experiment to
7
take Dr. Menasch's approach and mix these in with
8
the skeletal myoblasts to see, because I can't
9
understand how, if you have a scar, and you are
10
injecting cells into the scar, and you do nothing
11
about the blood flow, why those cells won't turn
12
into scar. The blood supply
clearly was
13
demonstrated to be inadequate because you have got
14
scar.
15 Of course, Dr. Menasch is actually doing
16
some experiments now using the same molecule that
17
we are using to induce collateral formation, so it
18
certainly is a very, very important strategy to
19
test.
20 DR. KURTZBERG:
In answer to the other
21
question raised, I don't personally think we know
22
which subsets are important yet, or whether subsets
23
are more important than whole cell preparations. I
24
think that all should be the focus of questions
25
going forward.
153
1 Evan Snyder has an interesting model of
2
spinal cord injury and repair.
It's a rat model.
3
They ligate, take a hunk of spinal cord and then
4
inject allogeneic cells and look at repair, and
5
they see repair and re-formation and re-connection
6
of nerves, but when they went back and looked to
7
see what cells did it, it turned out they were host
8
cells that were facilitated by something that the
9
allogeneic cells brought to the table, although
10
they don't know what.
11 To me, that just points out how much we
12
don't know and how complex the process is, and how
13
much more we need to study.
14 DR. RAO: One
last question. To me, and I
15
am somewhat naive in this field, there is a
16
difference between new vessel initiation and
17
collateral formation of regrowth in terms of the
18
factors which had acquired, and so on, and it seems
19
to me in some models of cardiac ischemia, what we
20
are looking at are completely ischemic regions and
21
long term, which there is no regrowth, and if you
22
had to do anything, it would be new vessel
23
formation.
24 Would it be fair to say that we can't
25
extrapolate from the current models that you talked
154
1
about in terms of the religation and
2
revascularization, or is it reasonable to be able
3
to extrapolate from those models to what you think
4
might happen in a cardiac model?
5 DR. EPSTEIN:
Well, I think it is
6
reasonable to extrapolate because we demonstrate
7
that we are able to improve perfusion, but your
8
question is really a very interesting one, and that
9
is, there used to be a major emphasis that
10
increased perfusion is just due to angiogenesis or
11
the development of new capillaries.
12 Well, I think most people involved in the
13
field would agree at this point that capillaries
14
don't increase flow. They
facilitate the
15
distribution of flow, and what you need is an
16
increase in conductance vessels or arteriogenesis
17
to truly produce an overall increase in flow.
18 However, we have some preliminary
data to
19
suggest that both processes are real, that
20
angiogenesis is a part of arteriogenesis, and that
21
you need the development of capillaries, that the
22
development of new vessels, capillaries, can
23
remodel to form collaterals.
24 That is why I don't use the term anymore
25
of angiogenesis. I say
"collaterogenesis," because
155
1
it gets away from the mechanistic aspects, which
2
are critically important, and we still don't have
3
the answer what cytokines produce angiogenesis,
4
what are important in terms of arteriogenesis, and
5
are both important to actually optimize the
6
development of collaterals, so we still have a
7
couple of years I think to go to answer that
8
question.
9 DR. RAO: One
last question
10 DR. HARLAN:
When you showed the
11
adenoviral HIF transfected cell lines and showed
12
that those were more efficient at elaborating
13
cytokines and inducing vessel growth, it adds
14
another question, that then becomes, however,
15
potentially anyway, a less well refined product,
16
the cell-conditioned medium from those cells, and
17
in view again of what we heard when we started
18
today, as we move forward with thinking about
19
delivering products to people, you want them to be
20
defined, and I wonder if you would comment on that.
21 DR. EPSTEIN:
Well, if this field is to
22
move forward, I think that criteria is going to
23
have to be eliminated because there is no way you
24
are going to be able to define--maybe I am
25
exaggerating--the hundreds of molecules that these
156
1
cells are producing. We don't
even know how to
2
measure them.
3 But I am sure the FDA allows the infusion
4
of serum and plasma from one individual to another.
5
Do you know what is in that serum?
6 DR. HARLAN: I
am not the FDA.
7 [Laughter.]
8 DR. EPSTEIN:
So, there is a precedent for
9
not knowing what you are injecting.
To be honest,
10
we are injecting cells, and we know a few of the
11
molecules that they are secreting, but we have no
12
idea of the concentration, and whether it is going
13
to vary from one patient to another, and if a
14
patient has diabetes or has hypercholesterol, so
15
believe me, if you going to be compulsive and say
16
we have to know the concentration, when, over the
17
course of time, those molecules are up, and what is
18
their interaction, we have to stop this field right
19
now, it can't move forward, and it is too bad. I
20
mean you would like to know everything, but this is
21
not a characteristic of cell therapy.
22 DR. RAO: On
that note, we will break for
23
lunch. We broke a little late,
so we will try and
24
come back a little bit later, but not too much, so
25
we will shoot for 1:00.
157
1 [Whereupon, at 12:15 p.m., the proceedings
2
were recessed, to be resumed at 1:00 p.m.]
158
1 A F T E R N O O N
P R O C E E D I N G S
2 [1:12 p.m.]
3
DR. RAO: Good afternoon.
4 Before we begin with the talks, I would
5
like to introduce three more members of the
6
committee who have just joined us.
I like to let
7
them do it.
8 DR. HIGH: My
name is Katherine High. I
9
am on the faculty at the University of
10
Pennsylvania. I am a
hematologist with an interest
11
in gene transfer for hematological disease.
12 DR. BLAZER: My
name is Bruce Blazer. I
13
am at the University of Minnesota in the Department
14
of Bone Marrow Transplantation with an interest in
15
immunobiology.
16 DR. RAO: We
also have Dr. Grant from the
17
FDA.
18 DR. GRANT:
Hi. I am Steve Grant. I am a
19
cardiologist. I am also a
clinical reviewer within
20
the Office of Cellular Tissue and Gene Therapies.
21 DR. RAO: We
will continue with the series
22
of talks that were scheduled.
23 The next speaker is going to be Dr.
24
McFarland.
25 Cellular Therapies for Cardiac Disease
159
1 DR. McFARLAND:
Thank you, Dr. Rao, and
2
welcome back from lunch.
3 [Slide.]
4 This slide is intended to remind me to
5
answer the implicit question which may have been
6
raised, and the question is:
Isn't the FDA putting
7
the cart before the horse?
8 The answer is, well, yes, in a way. We
9
thought that it would be good to focus and give
10
people a peek at what is in the cart below those
11
flat-screen monitors, I suppose, before we spend
12
the afternoon dealing with the horse which is
13
pulling the cart, the draft horse of product
14
development being preclinical studies and product
15
characterization.
16 [Slide.]
17 As Dr. Rao said, I am Richard McFarland,
18
and I am in the Office of Cell Tissue and Gene
19
Therapy in CBER.
20 [Slide.]
21 What I am going to do, I have been charged
22
with providing a perspective, FDA perspective to
23
the preclinical and manufacturing issues of cell
24
therapies for cardiac diseases.
25 First, I am going to describe the general
160
1
framework in which the FDA conducts our
2
science-based assessment of safety of novel
3
cellular therapies prior to allowing clinical
4
trials to proceed.
5 Second, I am going to describe the goals
6
of preclinical testing, safety testing in general,
7
and then a little specific about how that applies
8
to cellular therapies for cardiac disease, and
9
finally introduce the speakers for the rest of the
10
afternoon.
11 [Slide.]
12 FDA review is product-based, and it
13
parallels prudent product development.
This is in
14
contra-distinction to the NIH grant process,
15
which is more based on diseases
and organ systems,
16
which is illustrated just in the administrative
17
structure of the two agencies, NIH being primarily
18
divided by institutes, and FDA being primarily
19
divided by products that we regulate.
20 That means that our FDA review is
21
dependent on the characteristics of a specific
22
product, and the preclinical studies are designed
23
to support the use of specific products, and the
24
clinical trial design that we review is designed to
25
be supported by manufacturing and preclinical data.
161
1 That product-based review is framed by
2 regulations. I think I am the one designated to
3
get ready for regulations from the FDA.
4 [Slide.]
5 These are selections from the
6
Investigational New Drug regulations.
I want to
7
highlight a few things, that being that regulations
8
stipulate there is adequate information about
9
pharmacological and toxicological studies, that the
10
sponsor has concluded that it is reasonably safe,
11
and that the kind, duration, and scope of those
12 required tests vary with the nature of the proposed
13
clinical investigations.
14 [Slide.]
15 A little further down in the regs, for
16
each toxicology study that is intended primarily to
17
support the safety of the
proposed clinical
18
investigation, a full tabulation of data suitable
19
for detailed review should be submitted.
20 This is kind of critical to the way that
21
we do the review in that we need to get enough data
22
to do a detailed review.
23 [Slide.]
24 Let's back up for a minute and go to
25
preclinical evaluation in general.
What are the
162
1
goals of preclinical evaluation with a perspective
2
of supporting trials?
3 One is to provide a rationale for the
4
proposed therapy, discern mechanism of action,
5
identify "at risk" patient populations, recommend
6
safe starting doses and escalation schemes for
7
humans, do a preliminary risk/benefit assessment,
8
and to identify parameters for potential clinical
9
monitoring.
10 [Slide.]
11 I will talk a little more specifically
12
about use of preclinical models for cellular
13
therapies. Preclinical models
are used to provide
14
the scientific rationale with the cellular product
15
intended for clinical use, to understand cell
16
function, trafficking, and differentiation as all
17
these factors impinge on safety, as well as
18
modeling of routes of administration.
19 [Slide.]
20 If we had an ideal animal model for
21
cardiac cell therapies, it would have a similar
22
pathophysiology to humans that would improve the
23
predictability of human risk from the models,
24
similar anatomy to humans, which would allow us to
25
use various routes of delivery including catheters
163
1
of various types, with actually the clinical
2
catheter, it would also allow us to do extensive
3
dose exploration of cells, and it would be
4
immune-tolerant to human cells, so you could
5
actually use a human cellular product.
6 [Slide.]
7 Well, such an animal model doesn't really
8
exist, so we often use syngeneic
animal models of
9
cardiac diseases because they can provide us useful
10
data for assessment of safety.
11 That would be cells from analogous cell
12
source processed in a similar way from animals,
13
autologous cells in the animal species or syngeneic
14
species, gives rise to potential processing,
15
formulation, and storage differences, and limited
16
product characterization both preclinically and
17
clinically leads to some uncertainty in addition to
18
the uncertainty that is inherent in the modes
19
themselves.
20 [Slide.]
21 Add to that, we have added complexity due
22
to innovative delivery systems, many of which have
23
not been tested for delivery of cells.
Common
24
delivery systems that we have seen, intraoperative
25
transepicardial injection usually during CABG,
164
1
catheter-mediated transendocardial injection, and
2
catheter-mediated via cardiac vein, all of which I
3
think were discussed this morning.
4 [Slide.]
5 I only want to present a short scaffold of
6
the animal models that have been published so far,
7
and our speakers this afternoon will much more
8
extensively discuss the data.
9 There have been data in small animal
10
models, often cryoinjury, occasionally coronary
11
artery ligation to give an
ischemic area damaged
12
myocardium. One of the
advantages of the small
13
animal systems is that you have them available, at
14
least in mouse and rat, to use a human cellular
15
product to give you an idea of function and safety
16
of those cells. Primarily, it has been in mouse,
17
rat, and rabbit.
18 [Slide.]
19 Large animal models, typically, dog,
20
sheep, and pig, and we have seen some of those data
21
this morning, as well, an ameroid constrictor used
22
to generate an ischemic area has been a popular
23
model of disease. They are
amenable to catheter
24
administration and more amenable
to clinical
25
monitoring modalities, however, you are stuck with
165
1
using syngeneic cells in this situation.
2 [Slide.]
3 The regulations say that the data should
4
be adequate and extensive. What
are the potential
5
sources of data to support initiation of clinical
6
trials?
7 Well, the gold standard would really be
8
preclinical studies specifically designed to
9
support a specific trial with a specific cell.
10 We also have data from other potential
11
sources: existing animal
studies that were
12
designed to answer other questions, in-vitro
13
studies, clinical trials using the "same" product.
14 [Slide.]
15 However, we use data that are published
16
and unpublished. Using published
data either from
17
animal studies or human studies as sole support for
18
initiation of clinical trials raises some
19
questions, some complexities, because often these
20
studies are not designed to answer a toxicologic
21
question, and therefore, adequate toxicology
22
endpoints may not have been incorporated into the
23
design.
24 If they were incorporated into the design,
25
but not in the publication, we need access to those
166
1
data. Published reports may
provide sufficient
2
information for independent review.
3 [Slide.]
4 There are some limitations in using
5
published studies. Protocols in
the studies need
6
to be sufficiently detailed. We
need to be able to
7
do our independent review as per our regulations.
8
We need to see specifics of the route of
9
administration.
10 We need to see catheter specifics, such as
11
identity of the catheter, flow rate, pressures,
12
effects of catheters on cells, location of
13
injection in relation to the ischemic area, and
14
protocols, either animal studies or human
15
protocols, we need the control details of the
16
"routine" monitoring and analytical plans.
17 [Slide.]
18
The data must be presented in
sufficient
19
detail.
20 In-process and lot-release data from
21
manufacturing need to be presented in sufficient
22
detail for us to know exactly what the product is,
23
and complete study reports for both animal and
24
clinical studies.
25 [Slide.]
167
1 Cellular products used in published
2
reports may not be comparable to the intended
3 clinical product. Often,
in published reports,
4
there is insufficient data to allow us to make a
5
comparability assessment, and that is either
6
because the data don't exist or due to editorial
7
constraints of the publication.
8 [Slide.]
9 So, given the limitations of the framework
10
or the window at which FDA is required to look at
11
these, and the detail that we are required to look
12
for prudent product development raises some
13 regulatory
challenges.
14 These are rather recurrent regulatory
15
challenges. This is does the
submission contain
16
sufficient information to assess risk to the
17
subjects in the proposed trial.
It is a question
18
that we ask at the end of our review.
Were
19
adequate preclinical studies performed?
If they
20
were performed, were the data submitted in
21
sufficient detail to conduct an independent review?
22 If sufficient data are present, then we
23
get to the question, are the risk to human subjects
24
reasonable and significant?
25 [Slide.]
168
1 That gives you an idea of the framework
2 with which we need to look through to the window of
3
science and what we are obligated to do as we make
4
an assessment.
5 There is discretion, there is ability to
6
be flexible within the regulations, and what we are
7
asking the committee to do over these two days is
8
to give us an idea of what the state of the science
9
is. It will be reflected in the
questions that we
10
will be discussing tomorrow afternoon.
11 What is the state of the science? What
12
is a reasonable amount of data for us to be looking
13
at?
14 This afternoon, we are going to have two
15
speakers that are focusing
primarily on cells.
16
Doris Taylor from the University of Minnesota and
17
Silviu Itescu from Columbia.
18 After the break, followed by Dr. Nick
19
Jensen from the Center for Devices and Radiologic
20
Health at the FDA, who will focus on delivery
21
devices and some of the issues that are related to
22
development and testing of delivery devices, which
23
are an integral part of our preclinical
24
development.
25 Then, Robert Lederman from the NIH will be
169
1
discussing some of his experiences from being in
2
the trenches of doing studies with devices, cardiac
3
diseases.
4 DR. RAO: Thank
you.
5 We will now have Dr. Taylor.
6 Guest Presentations
7 Myoblasts: The
First Generation Cells for
8 Cardiac Repair:
What Have We Learned
9 DR. TAYLOR:
Thank you. I have to confess
10
that if I had seen those previous slides before I
11
had signed my talk, it would be a completely
12
different talk, so bear with me.
13 I am going to be talking about some of our
14
data and some of the data from the rest of the
15
field, but I think what I really want to focus on
16 is myoblasts for cardiovascular repair and what
17
lessons we can learn from the cells that have been
18
used for the longest period of time preclinically,
19
and I think another way to think about this is gene
20
therapy revisited, are we going to do it all over
21
again.
22 I think the point that I want to make is
23
that there are a lot of lessons that we can learn
24
from the gene therapy field as we are going forward
25
with cell therapy, and I think it is important to
170
1
take those lessons away from this.
I will be glad
2
to talk about that in more detail if people have
3
questions later.
4 I like to start with this because this is
5
what I tell the people in my lab, and I think it is
6
true - make everything as simple as it is, and no
7
simpler, from Einstein.
8 Philippe will recognize this. This is an
9
image I borrowed from his Lancet manuscript in
10
2001, showing the first patient into whom myoblasts
11
were actually delivered clinically.
I actually use
12
it to illustrate what I think is the salient point
13
here, which is that most of us are dealing with
14
animals as well as in patients, infarcted
15
myocardium, where a process of events has occurred
16
that starts with inflammation, moves to cardiocyte
17
apoptosis, a remodeling and compensation process
18
that you heard about in extreme detail this
19
morning, scar expansion, decompensation, and
20
progression to failure.
21 The truth of the matter is we are trying
22
to intervene in this with either cells, genes, or
23
devices, but we don't know where in this cascade we
24
are actually intervening, nor do we know where we
25
should be intervening.
171
1 I think most of us who got into this field
2
envisioned it, first, as a field where we would
3
intervene early after an acute myocardial
4
infarction to try to prevent the slippery slope
5
here of remodeling, scar expansion, decompensation,
6
and failure, but the truth of the matter is that
7
most of the patients in whom studies have been done
8
are patients who have already progressed to some
9
degree of failure.
10 Although we initially started to begin to
11
look at prevention and repair of not only cardiac
12
injury, but also vascular injury, we are now really
13
trying to understand whether or not we can move
14
back up this scale in a reverse remodeling way, or
15
to grow new cells, and I think those are the
16
questions that are really out there in the field
17
right now.
18 The Holy Grail in this field then is that
19
transplanted cells cannot not only engraft, but
20
restore blood flow and contractility to injured
21
myocardium, and all of you know, because you have
22
seen some of the data from Philippe and are fairly
23
well versed in this field or you wouldn't be here,
24
that there is more than 15 years of preclinical
25
data in rabbit and dog, there is at least 1 to 5
172
1
years of preclinical data in pig, rat, mouse,
2
sheep, and now hamster, all of which showed in our
3
preclinical models that transplanting autologous
4
skeletal muscle derived cells was safe, effective,
5
and feasible, and therefore, Phase I, surgical and
6
intervascular studies were initiated worldwide.
7 I think the future that we will probably
8
ultimately all try to examine is what is ultimately
9
the best cell for cardiac repair.
I have been
10
asked to focus on myoblasts, so I am going to do
11
that although a little bit later this afternoon, I
12
am going to talk necessarily a little bit about
13
comparisons among cell types because I think it
14
really begins to ask questions about mechanism
15
that we have to address as we are going forward.
16 Obviously, the best cell may be some
17
autologous bone marrow-derived product.
It may be
18 a
cell plus or minus a therapeutic gene to either
19
promote angiogenesis or some other signaling
20
cascade that maybe promote cell survival.
21
Ultimately, it may be embryonic stem cells although
22 I
would submit at least in this country, we are a
23
number of years away from those cells, not only
24
because we don't understand how to regulate their
25
differentiation, but because we also don't
173
1
understand how to make them stop dividing
2
appropriately.
3 Then, obviously, you can't be from the
4
University of Minnesota without talking about
5
adult-derived stem cells.
6 So, what are the questions in the field?
7 I
am just going to put my opinions out there.
I
8
think just keep these in mind as a background as we
9
go forward.
10 Is there a best cell?
I don't think there
11
is a best cell. I think it
really depends on the
12
patients, the time after injury, the dose and a
13
number of other factors.
14 Is there just a better way to get it
15
there? One of the questions that
we keep coming
16
back to over and over and over is whether or not
17
cell therapy is just going to ultimately be another
18
local drug delivery problem, and whether or not we
19
are really going to be able to get the cells to
20
where we need them or whether they have the
21
capacity to actually migrate there or home there,
22
and I think that is going to be an important
23
phenomenon to begin to explore.
24 The other side of that is do the cells go
25
where we want them, or do the cells go where we
174
1
want them and somewhere else where we don't want
2
them, and I think we will come back to that.
3 Should we just use growth factors and
4
forget cells? I am not going to
focus on that.
5 Is there a future for biologic
devices,
6
and is the real question dose, timing, and choosing
7
the right patients for the right cell?
I would
8
submit that it probably is.
9 So, where are we in this field? Well,
10
this is a table that I copied from a review by
11
Loren Field, and the table goes on for slides, just
12
to show that in terms of preclinical myoblast
13
transplantation, there is a huge amount of data out
14
there, and what the data really begin to show is
15
that there are a lot of different cells that have
16
been used, there are a lot of different species
17
that have been used from mouse, rat, rabbit, pig,
18
dog, and sheep, that these cells have been injected
19
either into normal heart, cryo-injured heart,
20
hearts where vessels have been occluded, and that
21
surprisingly, most people didn't actually measure
22
improvements in function or in angiogenesis or in
23
survival.
24 I think that is important to consider as
25
we really try to pull together the summary of data
175
1
from myoblast cell therapy.
2 So, how do we really do this? Well, for
3
myoblasts, you basically take a chunk of muscle and
4
you grow cells in vitro, and you end up with cells
5
in a dish, and then you deliver these cells to the
6
injured myocardium and you measure the effect.
7 So, you inject them, you deliver them, and
8
you measure the effect. So, I am
going to go
9
through the different parameters here, talking
10
primarily about myoblasts and what exists in each
11
of these areas.
12 So, the cells are typically referred to as
13
myoblasts, but the bottom line is these are
14
muscle-derived cells that contain, not only
15
myoblasts, but also often more fibroblasts than
16
myoblasts, a number of cells called SP or site
17
population cells, and then a whole lot of other
18
cells that we don't necessarily know how to
19
characterize yet.
20 So, this is a very heterogeneous cell
21
population, and you do take these cells and you
22
grow them in a dish for several weeks, and what
23
that means is that all of us are exposing these
24
cells to serum-containing medium, and that what
25
Arnold Kaplan learned years ago from mesenchymal
176
1
cells is that when you are using serum, what is in
2
that serum, FDA regulations notwithstanding,
3
matters, and that the lot number of serum is going
4
to give you a different outcome in terms of the
5
numbers of these cells and what their phenotype is
6
when you are done at the end of the day.
7 In terms of injury models where myoblasts
8
have been delivered, the primary injury models are
9
either acute myocardial infarction, acute being
10 anywhere
from a week to one month, and primarily
11
that has been a cryoinjury model or coronary artery
12
ligation.
13 The question that really arises--and then
14
those have varied dramatically in size, the timing
15
after creation of this infarction to the delivery
16
of cells matters, and inflammation, so are these
17
the same as a clinically relevant injury? Well,
18
the size differs, the timing differs, and the
19
inflammation differs with regard to what is
20
actually seen in patients.
21 In terms of cells, the myoblasts,
22
primarily how they have been delivered, more than
23
90 percent of the cells preclinically and
24
clinically have been delivered by surgical
25
approach.
177
1 Some preclinical data exists in terms of
2
delivering cells via catheter in a pig, but again
3
there are open questions about dose, about where
4
those cells are delivered, about when those cells
5
are delivered, and there is virtually no data out
6
there about the vehicle in which the cells are
7
given.
8 Typically, people either say they inject
9
the cells in saline or they inject them in the cell
10
growth medium minus the serum.
11 As I said up here, lot number matters, so
12
the vehicle differs dramatically.
13 The next issue that you have to deal with,
14
if you have got cells and you grow them and you
15
inject them into an animal, is how you measure the
16
outcome, and the question of safety is obviously an
17
open one, and none of us really have addressed the
18
safety question in preclinical models.
19 We didn't really know that there
was going
20
to be a safety issue. I was
talking at lunch about
21
the fact that, you know, we had a number of animals
22
drop dead over the course of our studies when we
23
were doing these experiments early on, but we
24
assumed it was because we were doing open-chest
25
surgeries on these animals to create the infarct,
178
1
to deliver the cells, to measure cardiac function,
2
not that it could have anything to do with a
3
potential electrical effect of the cells in vivo.
4 So, we had to then go back and evaluate
5
whether or not safety was even at all compromised
6
or relevant in these animal models, so none of us
7
are really measuring safety. There have been two
8
studies reported, one in a pig model of holter
9
monitoring animals, and we just presented some data
10
at ACC in rabbit, monitoring electrical effects of
11
myoblasts, and I will show some of those data in a
12
minute.
13 The other issue in terms of measuring
14
outcome is function. Typically,
in rodents,
15
isolated heart preps have been used to measure
16
function although in some cases, sonomicrometry is
17
used. I have actually put the
methods here in
18
order of I believe their ability to actually give
19
you useful information.
20 I think the isolated heart prep is the
21
least useful because it is subject to a lot of
22
variability, it is subject to baseline drift, it is
23
subject to flow and rate factors, and it is subject
24
to ischemia in vitro.
25 Echocardiography is obviously used
179
1
clinically, as well as experimentally, but it
2
varies dramatically with the operator and the
3
orientation of the probes, so you can make echo
4
tell you just about anything you want.
5 If you really want to measure work done in
6 a
region of the heart, I submit that you have to go
7
back and do sonomicrometry in that area and use
8
crystals to actually measure the ability of that
9
region of the myocardium to move in an electrically
10
and mechanically meaningful way.
11 Then, more recently, cine MRI has really
12
come to the fore in terms of our ability to make
13
measurements in not only people, but animals, as
14
well.
15 But measuring function is pretty useless
16
unless it correlates with histology, and we begin
17
to ask questions about angiogenesis and myogenesis,
18
and we don't always correlate histology with
19
outcome, and, in fact, one of the issues that comes
20
up over and over is clinically, as well as
21
experimentally, there is a disconnect between the
22
number of cells we can find in the heart and the
23
functional improvement we see, which begins to ask
24 questions about mechanism.
25 So, in terms of myoblasts, what are these
180
1
cells? As I said, there are myoblasts or
2
fibroblasts and there are SP cells.
I propose that
3
the mechanism of repair of these cells depends on
4
the number of cell types that you have present and
5
the percentage of each.
6 I believe myoblasts know to become muscle,
7
and they are capable of myogenesis.
I believe the
8
fibroblasts not only secrete an angiogenic factor
9
FGF, but also act as a growth factor and a mitogen
10
for myoblasts.
11 That has been known for years. Judy
12
Swain's data, they actually keep myoblasts alive
13
and keep myoblasts proliferating over a fairly
14
extended period of time, FGF does.
15 SP cells, I believe are more likely to be
16
angiogenic and also possibly to fuse with other
17
cells in the myocardium. That is
based on data
18
primarily from our group showing that the more
19
immature a cell is, the more likely it is to fuse.
20 So, in terms of our animal models, I think
21
the question we have to ask is what do the patients
22
look like first, and in the myoblast trials, those
23
patients have been post infarction, usually greater
24
than one month, up to many
years. The average in
25
one of the studies was 6.7 years.
181
1 Most of those patients are in need of
2
revascularization, they have an ejection fraction
3
of less than 35 percent, and they have heart
4
failure.
5 If you look at the European experience,
6
Philippe has already talked about some of this, and
7 I
am not going to go over these in detail.
This is
8 a
slide that was given to me by Peter Smits from
9
Rotterdam.
10 You look at Spain and Poland, and
11
obviously the French study, and then the U.S.
12
studies, both Arizona heart and Bioheart Mount
13
Sinai study, and then the Bioheart study in Europe,
14
and I should say for the sake of disclosure that I
15
have had a relationship with Bioheart, so take
16
these data with a grain of salt, that all of these
17
patients are heart failure patients.
18 The average in this study is 6.7 years
19
post-infarction, so these are patients who are
20
already pretty sick, and who have significant
21
electrical abnormalities already.
22 How do the preclinical patients compare?
23
Well, for myoblast studies, these animals all have
24
acute cardiac injury, either, as I said, cryoinjury
25
or coronary artery ligation.
Very few, if any,
182
1
studies deal with occlusion reperfusion, which is
2
what happens clinically.
3 You can open up virtually any artery in
4 the heart now, but nobody is doing preclinical
5
studies where we do ischemia reperfusion. We are
6
just now moving in that direction.
7 I think the reason initially for at least
8
us, and I believe for other people, was that we
9
wanted to kill everything that was there, so that
10
anything we found was due to something we put in,
11
and so we started with cryoinjury where we applied
12 a
minus 70 degrees C probe to the surface of the
13
heart to wipe out that region of the heart, but it
14
raises questions about the inflammatory process,
15
which we are coming to understand is critical in
16
terms of potential homing of cells and the
17
potential mechanism.
18 Most of the preclinical studies, the cells
19
are delivered two to three weeks post-injury, and
20
there is a one- to three-month follow-up.
21
Clinically, this isn't exactly relevant.
22 I just told you that every clinical study
23
is at least one month post-injury and sometimes six
24
to seven years post-injury, so we are not looking
25
at the same milieu into which we put these cells.
183
1 In term of heart failure models, there are
2 a
few models out there. Dan Burkoff's
group has
3
recently published a study in dog where they used
4
microspheres to actually create a heart failure
5
model, and have gotten data with myoblasts that
6
actually look very similar to some of the data
7
gathered in earlier models.
8 None of the studies to date have really
9
used any animals with LVAD support, and yet there
10
are clinical trials beginning to move forward in
11
that, and actually, one has already been completed
12
in that context. So, we would
expect that with a
13
completely unloaded heart, we might have very
14
different phenomena.
15 In terms of cardiomyopathy, there is a
16
hamster model and mouse genetic models that have
17
begun to be used for myoblast transplantation, and
18
Magdia Koob's group has actually published
19
reasonable data in terms of a rat model of
20
adriamycin toxicity. So, we are
beginning to get a
21
plethora of models in which we can look at myoblast
22
transplantation.
23 I put up here actually the little bit of
24
physiology that I could pull together about some of
25
the different animal models just to make a few
184
1
points.
2 One is that when you start looking at
3
these different injury models in mouse, rat,
4
rabbit, dog, pig, sheep, and humans, that there
5
really are significant differences in rodents and
6
the larger animals.
7 Mouse and rat, you know, the hearts are
8
pretty darn small. Their heart
rate consequently
9
is very high. These animals have very few
10
collaterals, and they express completely different
11
contractile proteins than are expressed in the
12
majority of the heart.
13 Moreover, the action potential in the
14
electrical capacity of the mice and rat is very
15 different. There is no plateau phase in the action
16
potential in mice, and the action potential
17
duration is on the order of 10 milliseconds. In
18
humans, it is on the order of 250 milliseconds.
19 Rabbit is the first animal model where you
20
begin to get numbers and conditions that resemble
21
humans, and that is why we chose rabbit early on,
22
and I would submit that in terms of feasibility
23
studies, rabbit is a good entry level animal under
24
most conditions except where you are trying to do
25
stem cell work.
185
1 When you are trying to do stem cell work,
2
we don't have the markers for stem cells in most of
3
these other species that we do in rat, mouse, and
4
humans, so that is when mice, rats, and humans, or
5
maybe pigs, some of the human cytokines and
6
antibodies cross to pig, but not all of them, so
7
you can begin to do some of those studies in pig.
8 Nonetheless, in terms of feasibility
9
studies, I think rabbit is a good model, and you
10
move up from there.
11 So, let's look at each of these. In terms
12
of myoblast in the mouse, you can begin to track
13
the cells because you can use genetic models of
14
where the cells actually express different markers
15
that are unavailable in the animals in which you
16
inject the cells.
17 You can begin to isolate and characterize
18
stem cells including the stem cells in muscle, but
19
you can't characterize from larger species. You
20
can use immunocompromised mice for human cells, but
21
you are missing an important component, which is
22
the inflammatory component, and we are beginning to
23
understand again relates to homing and perhaps even
24
recruitment of cells.
25 What is the advantage of a rat versus a
186
1
mouse? Well, the main advantage is it is larger
2
than a mouse, so you can do a few more things, but
3
you can still track the cells and isolate stem
4
cells, and you can still have an immunocompromised
5
model.
6 The rabbit, the bottom line about
rabbits
7
is most people make antibodies in rabbits, not
8
against rabbits, so it is really hard to find the
9
tools that you need to do some of the evaluations
10
downstream, but it is still a relatively
11
inexpensive model with cardiac characteristics very
12
similar to humans.
13 The pig, obviously, the size is good, and
14
the geometry is good for catheter-based studies.
15
One of the points I want to make about delivery of
16
these cells, and I think Philippe showed it when he
17
was talking about his clinical trials, and we have
18
done the same thing in terms of our preclinical
19
studies, surgical studies, is that when you inject
20
these cells surgically, most of us have delivered
21
the cells parallel to the surface of the heart.
22 We have done that for years because we
23
really thought it was going to increase the number
24
of cells that we could get into the myocardium.
25
Yet, all of the catheter-based studies deliver the
187
1
cells perpendicular to the surface of the heart,
2
and it is not completely unexpected that geometry
3
may make a difference in terms of how these cells
4
actually function in the myocardium.
5 So, I think it is important to evaluate
6
the geometry of the cells in some of these larger
7
animal models.
8 So, how do you choose an animal model for
9
the myoblast studies? I think
feasibility and
10
costs are obviously important, whether or not you
11
are going to do high throughput studies and need to
12
track your cells. Rodent
and hamster I think are
13
best for those. Rabbit is best
in terms of
14
beginning to be physiologically relevant to humans
15
in terms of heart rate and scalability.
I will
16
show some data in a minute in terms of scalability.
17 The large animal models are obviously much
18
more physiologically relevant.
You can get a sense
19
of dose. You can use
conventional delivery methods
20
that you would use in humans.
You can do the right
21
functional assessments, and the heart size and
22
geometry is very similar to a human.
23 So, what exists for myoblasts? Well, as I
24
said, the route of administration has been
25
primarily surgical or percutaneous.
Intravenous
188
1
and intracoronary studies are just beginning
2
although we published some intracoronary data, that
3
was the first thing we published in '96.
4 Myoblasts are unlike stem cells or unlike
5
bone marrow mononuclear cells.
They are very much
6
like stromal cells. They are
big. When we put
7
myoblasts in the coronary circulation, what we
8
found is that we got profound ST elevations, and we
9
saw transient ischemia every time we injected these
10
cells.
11 So, we actually think that with large
12
cells, that the way they are actually having an
13
effect in the myocardium is creating essentially a
14
microinfarct clogging the vessels and then getting
15
out of the vessels as a result of that.
16 Mononuclear cells are much smaller, and I
17
think don't have the same effect.
18 In terms of dose, in a mouse, typically,
19
you give about 1,000 cells. Some
people go as high
20
as a million, but typically, 1,000 is enough to
21
begin to see an effect.
22 In our hands, in rabbit, the lowest dose
23
at which we see an effect is 3 x 10
7 cells. We
24
tried 10 7, 3 x 107, 108, and
3 x
108, and this is the
25
range in which we see the most effect.
189
1 In pig, it is about 3 x 10
8, and what we
2
found, you know, pig is about 10 times bigger than
3 a
rabbit, you need about 10 times as many cells.
4
Rabbit is about I think about 4,000 times the size
5
of a mouse, and we found that we need many more
6
than 4,000 more cells in a rabbit than we do in a
7
mouse. So, I don't think you can
really
8
extrapolate from mouse, but I think you can begin
9
to extrapolate at the size of rabbit and go up.
10 In terms of cell location and where
11
myoblasts have been injected, you pick a surgical
12
fellow who is doing the experiments, and you will
13
get a different location of injection virtually
14
every time, I guarantee it, and you are not going
15
to convince them otherwise that their way isn't the
16
right way to do it. It is
completely ignored in
17
most of the preclinical studies.
18 There might be mention of one injection or
19
two injections or three injections, but in terms of
20
the exact location, I couldn't tell you, I couldn't
21
find in the literature where the majority of
22 injections occur. I know in my own lab, it is not
23
consistent from study to study.
24 In terms of timing, myoblasts have been
25
injected two to four weeks
post-injury. The
190
1
vehicle has been PBS cell growth medium minus
2
serum, or it is completely ignored, there is no
3
mention of it.
4 So, this is the slide again I borrowed
5
from Philippe's work to illustrate how he did some
6
of the early injections with a bent needle again
7
parallel to the surface of the heart, and also that
8
the injections are done, not just in the center of
9
the infarct, but in the periinfarct region, as
10
well.
11
Similarly, with a
percutaneous approach,
12
and this is another slide from Peter Smits, in the
13
first patient who received cells, and in the
14
majority of cases now with percutaneous myoblast
15
delivery, cells are delivered in the periinfarct
16
region, in the normal region of myocardium, and
17
very few of the injections percentagewise actually
18
end up in the infarcted cell, and that may have an
19
effect on safety, and I will show some preclinical
20
data that support that.
21 So, the majority of injections are
22
surgical, and you can inject the cells and find
23
them in the center of the infarct.
What you get
24
surgically when you inject these cells is one or
25
two things.
191
1 On a great day when you are really lucky,
2
you get what looks like a chunk of steak in the
3
center of the heart. On a
typical day, you see
4
something that looks more like this, where you have
5
patchy regions of cells distributed through the
6
infarct rather than these large fibers that you see
7
here, and these patchy cells distributed throughout
8
the infarct are not necessarily talking to each
9
other, but they are all oriented with the
10
extracellular matrix.
11 You can see there are some small vessels
12
here, here. We often see large
vessels in the
13
infarct, as well.
14 These are preclinical data from my group,
15
but they don't look too dissimilar from what you
16
see from Pagani's paper from myoblasts in an
17
LVAD-supported human heart or, in fact, the data
18
that Philippe showed you earlier of the myoblasts
19 surrounded by scar in a patient 17 1/2 months after
20
injection.
21 I am not going to talk about stromal cells
22
because that is not my job today, but what I am
23
going to begin to talk about is delivery.
24
Assessing delivery requires that we be able to
25
track the cells. In vivo, we
have chosen SPECT or
192
1
MRI most recently, although I think PET is going to
2
be a good method, as well.
3 That has to correlate in vitro with
4
histology and appropriate markers.
If I don't make
5
any other point today, take home the fact that
6
using desmin, using phospholamban, using GATA-4,
7
using all of these markers that people claim are
8
cardiac markers, are not cardiac-specific markers.
9
You find these markers in other muscle cells, you
10
find these markers in undifferentiated progenitor
11
cells.
12 If you look in C2C12 skeletal muscle
13
cells, you can see phospholamban, you can see
14
connexin 43, you can see in some cases, in
15
progenitor cells, you see GATA-4.
You have got to
16
use markers that are specific for cardiocytes if
17
you are going to call these cells cardiac cells,
18
and the only markers that I know of right now, that
19 I
believe are specific for cardiocytes, are channel
20
markers that are actually not expressed in skeletal
21
muscle.
22 As skeletal muscle matures, it expresses
23
many cardiac-specific proteins, and as skeletal
24
muscle matures, it expresses cardiac markers, as
25
cardiac muscle matures, it expresses skeletal
193
1 markers.
So, we don't know where in that process
2
we are, so we can't really use those markers.
3 This is an image showing that we can begin
4
to visualize these cells in the heart.
These are
5
indium-111 labeled myoblasts present in a short
6
axis view by SPECT imaging of a rabbit heart
7
showing that we can actually co-deliver
8
tetrofosmin, see perfusion, see the dropoff in
9
perfusion here with the infarct, and then see the
10
indium-labeled cells in the center of the image.
11 So, we are beginning to believe that we
12
can actually track cells over time now.
This is
13
also a cine MRI of a rabbit heart, and these are
14
data that were all gathered at Duke.
This is a
15
rabbit heart, so at the level of rabbit, although
16
we can now do the same thing in a mouse, we can
17
iron label our cells in a way that we believe
18
doesn't affect proliferation or viability of the
19
cells, and begin to see them in the center of the
20
infarct region, the infarct region here being
21
contrast-enhanced in white.
22 So, we can start now to label these cells.
23
We have followed these cells out to four months in
24
this way and can still find them.
When we kill the
25
cells and then inject them, the iron label goes
194
1
away over about two to three days, so we are fairly
2
convinced that the iron is present in viable cells.
3 So, the other issues in terms of measuring
4
outcome, I think I have already made this point,
5
safety is an open question, and I think what I take
6
away from the field so far is if you don't look,
7
you won't find it, and that we didn't look, and now
8 I
think it is important that we begin to do holter
9
monitoring and other electrically relevant studies,
10
and those are going to require large animals, pig,
11
rabbit. You can't do those in
mouse and rat, not
12
at 300 to 600 beats per minute.
You are really not
13
going to be able to see a VF or a VT.
14 In terms of function, I think if the goal
15
here with myoblasts is really to find an
16
ischemia-resistant cell that is electrically
17
compatible with a healthy heart, we have got to
18
also look at electrical activity of these cells
19
over time.
20 This is again a slide that Peter Smits
21
provided showing clinical data and the number of
22
VPCs per visit in some of the early patients who
23
had cells delivered, and I modified the slide a bit
24
to show times at which patients have actually died
25
after cell delivery.
195
1 What you begin to see is that there is a
2
window of time from about a week to a month where
3
there seems to be an increased incidence of
4
electrical abnormalities. When
we have done animal
5
studies now, we see that same sort of window from
6
about 3 days to about 3 1/2
weeks, and then it
7
drops off and we don't see the incidence after that
8
period of time.
9 So, the safety may depend on the cell
10 dose.
We have found that if we just look at PVCs
11
in our animal models, that as we increase dose, we
12
increase the number of PVCs, and it may also depend
13
on location.
14 We have found that if we inject cells in
15
the center of the infarct and we measure PVCs, and
16
this is actually 10
8 cells, 107 is not functionally
17
relevant, 10 8 is, 109
is, we
found that if we inject
18
cells in the center of the infarct, we see PVCs and
19
no monomorphic VT.
20 If we begin to inject cells in the border
21
zone, we not only increase the number of PVCs we
22
see, but we start seeing runs of ventricular
23
tachycardia. If we inject cells
in both the center
24
and the periphery, we see essentially the same
25
thing, and more up-to-data were just presented at
196
1
the ACC from my lab.
2 What is interesting is we began to take
3
these cells back out of heart, what we found is
4
that their action potential duration changed, that
5
initially, the cells had an action potential
6
duration of on the order of 20 seconds, and over
7
time it increased to something on the order of 120
8
milliseconds, but it is still not compatible with
9
the surrounding heart.
10 We have also done some modeling data.
11
What we believe is that if you have these cells
12
coupled to each other, that is a good thing in the
13
center of the infarct, but that you don't want them
14
coupled to the remainder of the heart until they
15
are electrically compatible with the remainder of
16
the heart, and yet clinically, very little
17
attention has been paid to location.
18 Again, it is an issue that we didn't know
19
we were going to have to address, and now we have
20
got to go back and address. In
fact, some of the
21
locations of injections could explain why there
22
have been ventricular tachycardia in some of these
23
patients.
24 The only possibility is, you know, over
25
that window of time, we don't know if these cells
197
1
are integrating, dying, or changing their
2
phenotype, we have no idea, and I think we really
3
have to begin to elucidate that.
4
So, in a standardized model
where we know
5
how myoblasts function, we have now got to look at
6
location, dose, and route of administration.
7 I think I have already said this, so I am
8
not going to really belittle, spend time on
9
function especially other than
to say we have
10
begun to collect a lot of data now with a lot of
11
different cell types and a lot of different growth
12
factors, and what we have begun to realize is that
13
virtually anything we put into the heart, cells,
14
myoblasts, fibroblasts, bone marrow stromal cells,
15
bone marrow mononuclear cells, growth factors
16
including VEGF and other growth factors, improve
17
the mechanical properties of the scar, and change
18
diastolic performance.
19 They do that first, before they have any
20
effect on systolic performance,
usually by several
21
weeks. What we figure is that
having something
22
alive in the scar is better than having just this
23
dense collagen matrix, and it really doesn't seem
24
to matter what you have alive in the scar, if is
25
vessels, if it is muscle, if it is whatever, you
198
1
improve compliance.
2 But we don't see the corresponding
3
improvement in systolic performance, at least not
4
with fibroblasts in our hand, but we do with
5
myoblasts, we don't with VEGF, but we have now with
6
bone marrow stromal cells and bone marrow
7
mononuclear cells.
8 These are some data that just came out in
9
Circulation showing that, that if we use crystals
10
to measure regional stroke work in our sham-treated
11
animals, regional function gets worse, but in our
12
myoblast-treated animals, function goes from pretty
13
bad to better, and in our bone marrow stromal cell
14
animals, the same thing is true.
15 I think that really raises a question
16
about mechanism, but the positive outcome in our
17
hands at least is dose dependent, 10
7 no effect, 108
18
positive effect, sham continues to get worse.
19 What is interesting is this is not
just
20
improvement versus cell number.
This is log of
21
injected cells, but this is the percentage of
22
animals that actually improve.
23 So, what we found is that the percentage
24
of animals increases with cell dose, as well.
25 So, will myoblast transfer work in
199
1
patients? Philippe already has told us that it
2
will, and we have begun to believe that, in fact,
3
that there are different mechanisms of action for
4
these cells.
5 I think I will just very quickly go
6
through the last couple slides.
We believe that
7
myoblasts improve both regional and global function
8
in the heart based on our preclinical data.
9 If we use cine MRI and actually measure
10
thickness in the wall of the myocardium over time
11
and global wall thickening, so areas where cells
12
were not injected, we use contrast to define where
13
the infarct is, and this is area that has no
14
contrast in it, so the remainder of the heart
15
actually gets better, wall thickening improves in
16
the cell treated, but not in the control vehicle
17
injected animals.
18 Regional wall thickening where we actually
19
inject the cells gets better to a greater degree,
20
so we only measured this where there was a
21
transmural infarct. We didn't
measure it in
22
regions at the periphery of the infarct where there
23
can be tethering going on. So,
we use contrast and
24
only measured it in the region.
25 Diastolic volume decreased, heart weight
200
1
decreased, so global indices of failure also
2
improved.
3 Every cell we and virtually anyone has
4
injected seems to work, which either means the
5
myocardium is easier to repair than we thought or
6
we don't understand what is happening and we aren't
7
looking at the data correctly.
8 I would like to believe it is this, and I
9
am actually going to posit this in a little bit,
10
but I have a bad feeling. I
would also submit that
11
they work despite the fact that we don't know how
12
to get the cells there in large numbers, and we
13
can't always find them histologically, and that we
14
don't really know what to look for.
15 These cells may be promoting angiogenesis,
16
myogenesis, they may just be unloading the heart,
17
changing wall stress. They may
be secreting
18
paracrine factors that recruit other endogenous
19
stem cells to the area of injury, either cells from
20
the heart, if that's your fancy, of cells from the
21
bone marrow, or maybe a combination thereof, or
22
maybe they work because we are lacking long-term
23
follow-up in both animals and patients, and we
24
haven't asked the right questions.
25 So, I will just stop by saying we have
201
1
started in small animals, we have moved to large
2
animals, and we have moved to patients, and these
3
are the cells that we knew the most about, and we
4
have had to say you know what, we missed a lot, and
5
we have had to go back and take the safety and
6
functional effects that we have seen and reevaluate
7
them in all of these animal models again, and I
8
think that should be a lesson for going forward.
9 Is it time for randomized trials, and, if
10
so, who? The think the questions we have to ask is,
11
is it safe, who are the right patients, what are
12 the
appropriate endpoints.
13 I show this every time. If we do this
14
wrong, we are really going to doom what I think is
15
an exciting field. I don't want to be standing up
16
here talking to you about the gene therapy lessons
17
that we could have learned and didn't.
18 I think what we can learn from gene
19
therapy is in 6 open-label trials, they were all
20
positive, and 4 out of 5 randomized, double-blind,
21
control trials, they weren't. Is
that because
22
patients got better care, or because there was a
23
better placebo effect? Well, we
won't really know
24
until we do the randomized trials, but we need to
25
under-promise and over-deliver, not conversely.
202
1 These are the people who did all the work,
2
and I thank you.
3 [Applause.]
4 DR. RAO: Thank
you, Doris.
5 We are open for questions.
6 Q&A
7 DR. ROSE: Dr.
Taylor, I enjoyed your
8
talk, but I must say, as a clinical cardiologist, I
9
just have difficulty in envisioning how these
10
myotubules, that presumably result from the
11
injection of the myoblasts into scar, increase
12
systolic wall thickening, which is what I would
13
accept as evidence of improved regional function.
14 Unfortunately, many of your images weren't
15
showable, but was the increase in thickening that
16
you showed on that on that one graphic slide, was
17
that from your larger animal models as opposed to
18
your smaller animal models?
19 DR. TAYLOR:
No, that was from rabbit.
20 DR. ROSE: You
mentioned early in your
21
talk about a disconnect between the numbers of
22
cells and the functional improvement.
Is that true
23
for the myoblast injection? In
other words, there
24
was probably a range of responses, I would imagine,
25
some had perhaps better systolic thickening
203
1
improvement than others, and by histologic
2
examination, were there more myotubules or were the
3
myotubules oriented in a different direction than
4
the animals that perhaps had less improvement in
5
systolic thickening?
6 DR. TAYLOR:
Have we quantified that
7
unequivocally? No, because it is
very difficult to
8
quantify the number of cells in an infarct and know
9
that they are actually myoblasts in the cells you
10
injected because we are giving autologous cells.
11 One of the rate limiting steps in this
12
field is having good markers for the cells, and
13
that is one of the reasons we went to iron, so that
14
we could actually stain our sections later for
15
Prussian blue and look for iron and say, okay,
16
these are the cells we injected.
We shouldn't see
17
that in normal cells.
18 So, we are starting to now look and try to
19
answer that question. What we do
know is that with
20 a
number of studies that have been done by a number
21
of groups, the number of cells that you retain in
22
the scar after you inject them is about 15 to 25
23
percent of the cells you initially put in there
24
probably. On a good day, maybe
30, 40 percent of
25
the cells you put in there, but that is rare.
204
1 What we also know is that numbers of cells
2
can die over time, but we also know that these
3
cells can proliferate over time, and when we look
4
for proliferation, we in fact can find that in
5
these scarred regions.
6 I don't actually think that we are seeing
7
evidence of long myofibers necessarily.
We see
8
cells that line up with each other and connect with
9
each other, and we believe that these cells are
10
being mechanically induced to contract.
11 We know when you stretch muscle, it
12
contracts. Do I think they are
electrically
13
coupled with the rest of the heart?
No. Do I
14
think they are mechanically coupled?
Yes. What we
15 see is we measure left ventricular pressure
going
16
up, we see pressure plateau, and right at the time
17
that pressure within a few milliseconds after
18
pressure plateaus, we see our crystals move in that
19
region of the heart, and then pressure decreases,
20
and then we see that again.
21 We can track that beat after beat after
22
beat in the scarred region. We
can go from
23
negative work loops where the rest of the heart is
24
being pushed by the remainder of the heart to
25
positive work loops when we put the
205
1
cells--actually, I have got it backwards--negative
2
in this direction versus positive work loops, that
3 actually
correspond to systole.
4 So, I believe that there is wall
5
thickening and actual contraction going on in that
6
region, and I think our sonomicrometry data are the
7
most convincing data.
8 Is it possible to get the image to show,
9
so I could try to show one of the functional wall
10
thickening images, so you can actually see that?
11
If you just give me the slide thing back, I will
12
try.
13 I am going to try this first.
14 So, this is, in theory, it is beating
15
here. No, apparently I can't,
maybe you can, but
16
we are able to actually show thickening and I will
17
pull my computer out to show anybody later who
18
wants to see it.
19 You can actually see in the
sham-treated
20
animal, here is the scarred region and only in the
21
very center of the scar is it not thickening
22
anymore. The scar was from about
here to here, by
23
contrast.
24 The scar of the sham-injected animals over
25
here is not thickening at all.
Here, we only don't
206
1
see thickening right in the very center of the
2
scar, and we used contrast enhancement, so we took
3
10 slices through the rabbit heart, in a long axis
4
view, we used contrast, and only where we saw
5
contrast gadolinium did we call that infarct, and
6
then measured thickening in that region, and we did
7
that in a blinded way.
8 We are pretty convinced that we see wall
9
thickening in that region.
10 DR. BLAZAR:
You made a good point that
11
the larger animals allow you to assess function
12
much more directly and extrapolate to humans. With
13
the smaller animal models, you have a higher
14
throughput. So, the question is,
what is the data
15
that says that the smaller animals extrapolate, the
16
large animals extrapolate to the human as you go
17
through all of these different examples, because
18
one would hear your presentation and think that it
19
really just should be restricted to large animals
20
minus a few, more esoteric.
21 DR. TAYLOR:
What I can say is we have
22
made measurements in rabbit and pig for many, many
23
years, and I swore we would never use mice or rats,
24
and then we decided that we wanted to start making
25
comparisons with bone marrow derived cells, and we
207
1
didn't have the markers to isolate those cells in
2
rabbit or pig, so we went back and did the same
3
experiments in mice, and we got the same results.
4 So, now I have to bite my tongue and show
5
mice data even though I swore we never would. I am
6
convinced at least with MRI, that we can make
7
meaningful measurements that show us the same sort
8
of thing.
9 DR. BLAZAR:
Although you also made a
10
point as to the dose, location, et cetera, it would
11
seem that that is going to be extremely difficult.
12 DR. TAYLOR: It
is very difficult. That
13
being said, what we do is we inject the cells in
14
the center of the scar in a mouse.
Do I believe we
15
get the same percentage that we get in larger
16
animals? No. We really don't know what number we
17
actually get in.
18 We have started doing some biodistribution
19
studies to try to answer that, and we are mostly
20
doing those in rabbit and pig, because I have no
21
confidence for the numbers we get in mice. In
22
terms of doing stem cell studies, though, I think
23
it is critical that you use something where you can
24
clearly define the cells.
25 DR. ALLAN: A
little bit of a follow-up,
208
1
which is use of nonhuman primates.
Somebody
2
mentioned that there were some primate studies.
3
This morning, I think somebody just referred to
4
them.
5 DR. TAYLOR:
Right, with bone marrow
6
cells.
7 DR. ALLAN: Is
it with bone marrow cells?
8
Because you can use those markers on many of the
9
nonhuman primates, you can look
at stem cells.
10 DR. TAYLOR:
Right.
11 DR. ALLAN:
Maybe to use that model as a
12
step between "large" animals and humans, because
13
anything that you have derived that looks promising
14
could through that nonhuman primate.
15 DR. TAYLOR: I
think that is good point,
16
and we have actually started collaborating with
17
some people in California who do nonhuman primate
18
studies. I can tell you that
they are God-awful
19
expensive, they are hard to do, and I personally
20
find them hard to do.
21 Sometimes we just have to get over it, but
22 I
think if the pig data are good enough, I would
23
rather stick with pigs, and the fact that we can
24
use some of the same stem cell markers has made me
25
focus more in that area.
209
1 DR. BORER:
That was a very interesting
2
presentation. I have a question
that sort of falls
3
into are we looking at the data the right way or
4
oversimplifying box. You can't
argue with success,
5
and I believe that some contractility is probably
6
better than done and perfect is the enemy of good,
7
and all that stuff, but, you know, you have put in
8
skeletal muscle cells, and they contract, and yet
9
to get useful work from the heart, forced
10
generation is just part of the equation, you have
11
to have forced transmission, as well, and I haven't
12
heard anything yet about remodeling at the
13
extracellular matrix and regeneration of dystrophin
14
ECM hookups, and whatever, that might demonstrate
15
that we are actually developing a forced
16
generation/forced transmission system.
17 I don't know how important that is
18
ultimately, but it makes me wonder, the fact that
19
we haven't heard about that, and I am not sure that
20 I
could expect it would happen, that we are
21
actually looking at the data the right way, so
22
could you talk a little bit about that?
23 DR. TAYLOR:
Sure. The one piece of data
24
that we do have, that I think addresses that, is
25
the compliance data where we looked at changes in
210
1
strain with the different cell types, and that
2
begins to address matrix and what is going on in
3
that matrix - does it look at signaling, does it
4
look at MMPs, does it look at any of that? No.
5 I think the bottom line is when we started
6
this 15 years ago, our goal was to show it worked,
7
and then once we showed it worked, we thought we
8
would go back and figure out how it worked.
9 As soon as we showed it worked, God and
10
everybody wanted to do it clinically, and so we are
11
having this discussion rather than understanding
12
how it worked, which we have had to develop in
13
parallel. I think we are still
catching up in
14
terms of trying to develop, trying to understand
15
the mechanisms by which it works.
16 Five years ago, at American Heart, Michael
17
organized a session, and there were probably five
18
or six people talking about this.
If you look at
19
American Heart now, there are two days of people
20
talking about this. People
didn't believe it five
21
years ago.
22 Now we have a critical mass in the field
23
and we can start asking those questions, and I
24
think the data will emerge over the next couple of
25
years.
211
1 DR. RAO: Dr.
Ruskin.
2 DR. RUSKIN:
Doris, you described a very
3
significant prolongation of action potential
4
duration in the myoblasts from about 20 to 120
5
milliseconds. That suggests a
change in iron
6
channel expression. Do you have
any information as
7
to how that came about?
8 DR. TAYLOR:
Gus Grant told me it did?
9
No.
10 DR. RUSKIN: I
will buy that.
11 DR. TAYLOR:
Short answer is no except
12
that we know that we started seeing a plateau phase
13
which wasn't there before, and we didn't change the
14
rate of rise of the action potential.
Do I know
15
any more than that? No.
16 Have we looked for channel markers? You
17
know, here I am saying, well, the only thing I will
18
believe that if you tell me it's a cardiocyte, is a
19
channel marker, have we looked for channel markers?
20
No, because we did all of that in rabbits, and the
21
darn markers don't exist.
22
Are we now trying to figure
that out in
23
some of these other animal models?
Yes, and I
24
think that is where some of the mice genetic models
25
of changes in electrical activity in the heart may
212
1
actually be really useful in terms of trying to
2
dissect what myoblasts can do.
3 DR. HARLAN:
You had a great quote from
4
Einstein. I will have to give
you a quote I have
5
from Osler, where he talks about stern iconoclastic
6
spirit that we need, and that you reflected.
7 My question is I think I misunderstood,
8
you implied that there was not the cellular
9
specificity that people assume, that you have seen
10
some contractility with myoblasts, but also with
11
bone marrow. I wonder how
extensively you have
12
studied that with other cell types.
13 DR. TAYLOR: We
have looked with
14
myoblasts, we have looked with bone marrow stroma,
15
and we have looked with bone marrow mononuclear
16
cells, and we see an improvement with all of those.
17
We haven't gone back and dissected the bone marrow
18
mononuclear cell populations, we are starting to do
19
that now.
20 Other people have seen the same thing with
21
MAPC cells. So, have we
dissected that in detail?
22
No, but by the criteria that we have used, which is
23
sonomicrometry and MRI, we see the effect, and yet
24
we don't, when we look at histology, we don't see
25
the same degree of muscle formation with all of
213
1
those different cell types, which begins to argue
2
that mechanism is more complicated than we thought.
3 DR. HARLAN:
Let me ask it this way. Are
4
there cells that you have looked at, that you
5
inject, where they don't work?
6 DR. TAYLOR:
Fibroblasts.
7 DR. HARLAN:
Fibroblasts don't work.
8 DR. TAYLOR:
Fibroblasts actually improve
9
compliance, but make systolic function worse in our
10
hands.
11 DR. RAO: Dr.
High.
12 DR. HIGH: I
want to ask one question to
13
try to reconcile some of your data preclinical
14
studies with some earlier clinical work that we
15
heard about.
16 Was that in rabbits or pigs that you said
17
that you needed at least 10
8 cells to see an
18
effect, was that rabbits?
19 DR. TAYLOR:
That was rabbits--no, 3 x 10
7
20
in rabbits, 108 was in pigs, I am sorry.
21 DR. HIGH:
Okay. Then, the 10
9
cells that
22
are being injected in the clinical study would be
23
roughly appropriately correlating in terms of--
24 DR. TAYLOR:
Right, we actually see a
25
better effect with 3 x 10
8 cells in pig. We
214
1
haven't gone up to 10
9 cells yet although we have
2
plans to do that. Philippe can
tell you, you start
3
dealing with massive numbers of cells, and when you
4
are doing this in an autologous way, and you are
5
dealing with cells that you have to keep at low
6
confluence, it gets out of hand pretty quickly.
7 DR. HIGH: You
said that in, is it the pig
8
studies, about 15 to 20 percent of the injected
9
cells are retained?
10 DR. TAYLOR:
That is actually not data
11
from my lab, that is data from other groups, and
12
that has been in some pig studies, and I believe in
13
some--I know of pig, and I can't remember what
14
else.
15 In our hands, in rabbit, we see a little
16
bit higher than that, on the order of 20 to 25
17
percent, but that has only been in a few studies
18
with indium-labeled cells, so I don't trust those
19
numbers yet.
20 DR. HIGH: Is
that known to be a function
21
of time after injury?
22 DR. TAYLOR:
Actually, what we found is
23
that the longer we wait after injury, the easier it
24
is to get cells to hang around in the heart, that
25
if we inject cells at 2 weeks, we get fewer cells
215
1
retained than if we inject cells at a month.
2
What we also know is if
we inject cells in
3
normal heart, they all go in the cardiac vein and
4
get carried elsewhere, that the junctions in the
5
myocardium are so tight that it is really hard to
6
get those cells into the normal heart.
7 Again, it gets back to injection. We
8
have come up with a way where we inject the cells,
9
we see a bleb, we wait for the bleb to go away, we
10
inject more cells, but that is just empiric,
11
because it works for us. Do we
know how it is
12
being done by other groups? No
clue.
13 DR. SCHNEIDER:
Doris, you summarized
14
nicely both the cellular complexities and the
15
technical complexities that are involved here, and
16 I
wanted to comment about one in each of those
17
categories.
18 You talked about cryoinjury versus
19
coronary artery ligation, and I wanted to agree
20
with your comment that relatively little of the
21
work in the field is being done with ischemia
22
reperfusion injury, which more closely resembles
23
the clinical situation particularly in an era of
24
stenting and reperfusion therapies.
25 A further complication there, though, is
216
1
that much as was learned over a period of years in
2
investigations of stunning, it may be necessary to
3
distinguish between open-chested ischemia
4
reperfusion injury and the chronically instrument
5
close-chested animal that undergoes ischemia
6
reperfusion injury, which is something that a few
7
labs have been able to develop as a means to
8
minimize potential artifactual effects of the
9
surgical procedures.
10 DR. TAYLOR:
Right.
11 DR. SCHNEIDER:
With respect to cellular
12
heterogeneity, you talked about the possibility
13
that SP cells in the skeletal muscle population
14
might be important. Michael Rednicke has identified
15
in skeletal muscle, and investigators at Indiana
16
have, as well, a scar-positive, LIN-negative,
17
CD34-negative, CD45-negative population that lacks
18
any ability to undergo hematopoietic
19
differentiation and very closely resembles the
20
scar-positive cells we found in adult heart.
21 DR. TAYLOR: I
have seen data from the
22
University of Minnesota like that, as well, and I
23
also should say that Johnny Heward at Pittsburgh
24
has found that as you increase the passage number
25
of cells, and these are actually old data, from '98
217
1 I
believe, that as you increase the passage number
2
of cells in vitro that you see differences, I think
3
passages 3 and passage 5, or something like that,
4
give you much better functional results in
5
engraftment than passages 1, 2, and 4, and the
6
desmin staining of those changes.
7 So, I think we probably are selecting for
8
different cells.
9 DR. SCHNEIDER:
Along those lines, do you
10
know if the scar-1-positive fraction goes up or
11
goes down in your skeletal muscle cells over time?
12 DR. TAYLOR:
Scar-1, we haven't look at
13
scar-1 in our population of cells.
We have looked
14
primarily for SP cells. What I
will say is that we
15
grow our cells a little differently than most
16
people. As you know, we make an
explant and
17
actually allow our cells to grow out from the
18
explant, and we are getting a
much higher
19
percentage of more immature cells than other people
20
as a result, and I think that is because we are not
21
throwing them all away when we do the filtration
22
and enzyme digestion, and those kind of things, and
23
that may impact some of our functional data.
24 DR. RAO: We
will take on last question.
25
Dr. Simons.
218
1 DR. SIMONS:
Doris, you mentioned that up
2
to 90 percent of cells that they injected in the
3
heart die soon thereafter. So,
the cells that are
4
still there, do they need to be concentrated at a
5 certain
per sort of square area of the cell, or can
6
you spread them as much as you like, so is it
7
really a mechanical effect or is it something that
8
the cells make, because you are making a point that
9
it matters where you actually inject them?
10 DR. TAYLOR:
What we have found is that we
11
can do three parallel injections, and we get the
12
same effect as if we do one injection, as if we do
13 a
star-shaped injection, so we spread them out in
14
different--I mean we have given them under
15
different geometries, and we see the same effect,
16
the same dose of cells.
17 What we don't know is how many die under
18
each of those conditions.
19 DR. SIMONS:
So, if you can spread them
20
around, that would imply that that this is probably
21
not a purely mechanical effect?
22 DR. TAYLOR: I
think there is absolutely
23
some truth to that. I don't
think it's purely
24
mechanical. I also don't
think--that is what I
25
said when we use these muscle cells, there are
219
1
multiple populations of cells in there, and I think
2
different ones have different effects.
3 Absolutely, I think mechanism is an open
4
question, we really don't know.
5 DR. RAO: In
the interests of time, I
6
guess we move on. Thank you,
Doris.
7 I apologize for not recognizing members of
8
the public, but this is the part of the meeting
9
where we have to give priority to the committee
10
members in terms of questions.
People in the
11
audience can address the committee in the open
12
session.
13 Dr. Itescu.
14 Preclinical Models - Hematopoietic and
15
Mesenchymal Cell Therapies for Cardiac Diseases
16 DR. ITESCU:
Thank you very much for
17
inviting me here today. As if
the talks haven't
18
been complex enough, I have got the difficult task
19
of speaking for 30 minutes and covering a variety
20
of animal models, as well as a variety of cell
21
types, so I hope it will be cohesive enough.
22 The issues to consider in this field, I
23
have tried to address some of them here in this
24
slide in terms of small animal models or large
25
animal models, and it is the precise
220
1
characterization of cell type and population to be
2
used to define the cell source and process for
3
isolation to determine if there is a need for ex
4
vivo culture and expansion to identify the
5
mechanisms of action for inducing cardiac repair,
6 to
identify appropriate animal models, in other
7
words, small versus large, species-specific versus
8
those that use human products and those that
9
involve acute versus chronic ischemia.
These are
10
all very, very important
questions.
11 Finally, experiments that address
12
dose-ranging studies for functional correlation and
13
toxicity, and the last question which has not been
14
touched on yet today, which I will at the end of my
15
talk, is that between autologous versus allogeneic
16
products.
17 The adult mammalian bone marrow contains
18
two stem cell populations at least.
The one that
19
we are most familiar with are hematopoietic stem
20
cells defined as being CD34-positive, and more
21
importantly, CD45-positive.
22 These cells form blood-forming elements,
23
such as monocyte and macrophage lineage cells,
24
these account for about 10 to 20 percent of the
25
CD45-positive fraction, and more recently, cells
221
1
that are endothelial progenitor cells or
2
angioblasts that express these markers plus several
3
others such as AC133 and c-kit.
4 The second population, the
5
non-hematopoietic stem cell fraction is typically
6
CD45-negative and CD34-negative, and at least three
7
cell types have been defined within this fraction -
8
the mesenchymal stem cells, which really is poorly
9
termed as stem cells since these cells are defined
10
based on their in-vitro culture characteristics.
11
The way they are isolated is very crudely, very
12
grossly based on density properties and plastic
13
adherence. In fact, the
population cells that is
14
pulled out initially is very heterogeneous.
15 A second stem cell type is the MAPC that
16
you have heard about. These
cells are potentially
17
more homogeneous in nature, however, the fact that
18 they are dependent on negative immunoselection
19
means that again we don't really know what the cell
20
type that is ultimately derived is, and the cell
21
culture conditions are very laborious and require
22
low density for outgrowth.
23 Finally, there is a third population of
24
mesenchymal precursor cell or progenitor cell which
25
can be defined on the basis of several surface
222
1
markers and can be selected by immunoselection,
2
positive immunoselection freshly from bone marrow,
3
and we are using these cells currently in my
4
laboratory, and I will touch on them a little bit
5
towards the end of the talk.
6
In respect to the hematopoietic
stem cell
7
fraction, the CD34/CD45 fraction, there is high
8
frequency within the bone marrow compartment, as I
9
have mentioned, 10 to 20 percent of macrophages, 1
10
to 2 percent of the CD34 progenitor cells that can
11
be freshly isolated without any requirement for
12
in-vitro culture and expansion, and are well
13
characterized for many years with established
14
isolation protocols.
15 In this slide, I tried to summarize the
16
interaction and cooperation between the
17
CD45-positive and 45-negative subsets in terms of
18
vascular network formation, so that in the bone
19
marrow, the primordial circle stem cell may give
20
rise to the CD45-positive/34-positive endothelial
21
progenitor cells, which egress to the embryonic
22
organs where you get the initial vasculogenic
23
small, thin-walled capillary development.
24 At the same time, the CD45-negative
25
parasite fraction derived from a mesenchymal
223
1
progenitor cell produces a variety of factors
2
including VEGF, FGF, angiopoietin, and SDF-1, which
3
can interact with these vasculogenic capillaries to
4
give rise to the more mature vascular network
5
through a process of angiogenesis.
6 In addition, the parasites may also
7
migrate and give rise to the smooth muscle outer
8
wall, so that you really have this cooperation
9
between both cells and factors to give rise to the
10
permanent new vessel formation.
11 With this as a background, I will like to
12
address several studies that have looked at how
13
angiogenesis per se can improve cardiac function.
14
This study from Kobashi [ph] and Collins in 2000, I
15
think was the first to demonstrate or one of the
16
first to demonstrate that whole bone marrow in a
17
small rodent model could induce transient
18 angiogenesis
and improve cardiac function.
19 You can see here the implantation of whole
20
bone marrow into the ischemic myocardium of a rat
21
gives an increase in local production of VEGF
22
protein within 24 hours, but by 7 days this protein
23
production is gone, and in parallel, the
24
angiogenesis that is occurring in the heart is
25
again transient, maximal at 1 week
224
1
post-implantation, gone by 4 weeks.
2 So, I think this particular model gives us
3
some pause and suggests that the whole bone marrow
4
approach may not be a way to give rise to permanent
5
vasculature. Others have used
this sort of
6
approach now in larger models, in pigs.
Work from
7
Kemiharder [ph] and colleagues demonstrated that in
8 a
LAD-ligated porcine model, again whole bone
9
marrow implantation gives rise to angiogenesis.
10
This is within 6 weeks post-implantation, you can
11
see improvement in collateral flow associated with
12
some degree of improvement in function.
13 In parallel studies they published about
14
two years ago now, the same group demonstrated that
15
in a more chronic ischemia model in pigs using the
16
ameroid constrictors, they were able to again
17
induce angiogenesis using mononuclear cells from
18
bone marrow, as well as mononuclear cells from
19
peripheral blood to a lesser extent.
You can see
20
improvement in regional blood flow and improvement
21
in global parameters of cardiac function including
22
ejection fraction DPDT, and a reduction in the
23
ischemic area.
24 But again, all of these studies were done
25
in the setting of an acute ischemic or perhaps more
225
1
subacute ischemic model in this scenario, and the
2
animals were followed up for no more than 6 weeks,
3
so if we take that into consideration that in the
4
smaller rodent, the neovascularization and the
5
cardiac function improvement was transient, but I
6
think the 6-week follow-up in these larger models
7
is really not adequate especially in a pig where
8
the physiology much more closely resembles that in
9
man.
10 The next approach is to look at perhaps
11
subsets of some of these cells, and I would like to
12
show you some work on endothelial progenitor cells
13
defined by surface markers.
14 Isner and colleagues originally described
15
these cells demonstrating that endothelial
16
progenitor cells were present in the bone marrow,
17
were released to the circulation under certain
18
signaling and present from growth factors under
19
ischemic conditions, and that these cells can
20
incorporate into regions of ischemia.
21 So, we asked the question if one could
22
identify these cells in the adult marrow of humans,
23
could they potentially improve function through a
24
process of neovascularization.
25 The model that we use in my lab is one of
226
1
using the incompetent nude rat, which is an athymic
2
rat model that lacks T cells and B cells, continues
3
to have some degree of natural killer cell
4
function, in other words, it is a linking model,
5
but it is able to tolerate certain types of human
6
cell injections, particularly human cells that have
7
not been in vitro cultured and expanded.
8 Our objective was to cause a permanent
9
occlusion of the anterior descending left coronary
10
artery, in contrast again to other models that
11
perhaps are using the reperfusion type of model.
12
Our objective was to induce an infarct and see
13
whether we could then protect the subsequent
14
territories of myocardium still at risk in the
15
periphery.
16 We mobilized in the progenitor cells from
17
healthy human donors using GCSF, think that, in
18
fact, in the future this would
have been a way to
19
move towards clinical trials, in other words, being
20
able to harvest large numbers of progenitor cells.
21 In our studies, we then immunoselected
22
these on the basis of surface markers, CD34 and
23
CD117, injected these into the tail vein of acute
24
ischemic rats to see whether they homed to the
25
myocardium.
227
1 I would just like to pause for a minute to
2
say that given the recent paper in the Lancet, I
3
think a week or two weeks ago, I think that
4 certainly raises a
question about how such cells
5
are going to be accessed in large numbers if they
6
are to be used at all at the time of an acute
7
myocardial infarct because I think under a local
8
anesthetic, a bone marrow aspirate gives rise to
9
extremely few numbers of these types of cells, and
10
you really do require either large harvesting or
11
large numbers to be mobilized, and I think at this
12
point in time, GCSF is not the agent to be used
13
post-infarct.
14 In any case, the point of our studies were
15
to see whether intravenous administration at the
16
time of acute infarct would lead to selective
17
homing to the myocardium, and we have met these
18
trafficking pathways based on the type of chemokine
19
expression that occurs post-infarct, and we found
20
that B cells selectively migrate to the
21
peri-infarct region, where they within two weeks
22
induce both incorporation into vessels, and this is
23
staining with antihuman CD31, and you can see cells
24
that are previously labeled with dye/Dil,
25
incorporate into vessels of the peri-infarct
228
1
region, but in addition, a very dense induction of
2
angiogenesis at the more distal areas as defined by
3
rat-specific anti-CD31 antibodies.
4 So, again, akin to the type of data shown
5
by Dr. Epstein, it appears that these cells are
6 able
to not only incorporate and induce
7
vasculogenesis, but presumably secrete a variety of
8
antigenic factors.
9 From a pathophysiologic perspective, we
10
see that when we induce a vascular network at the
11
peri-infarct region, you can see these large
12
capillaries. There is nice
viability of the
13
cardiomyocytes at the peri-infarct region as
14
opposed to these apoptotic cardiomyocytes. You see
15
reduction in scar formation and clearly a viability
16
and survival of the myocytes.
17 But in addition to that, we were quite
18
surprised to see that as early as five days
19
post-neovascularization, there seems to be
20
induction of cell cycling by endogenous
21
myocyte-like cells, and you can see these by
22
confocal microscopy small cells of the peri-infarct
23
region that express alpha-sarcomeric actin in blue,
24
but more interestingly, in yellow, the staining
25
with the rat-specific anti-KR67 antibody, which
229
1
recognizes cycling cells only of rat origin. You
2
can see lots of these type of small cells in the
3
peri-infarct region only in the situation where we
4 also have neovascularization being induced by the
5
human cells.
6 Within two weeks, in these same areas,
7
what you see is that the cells that express
8
sarcomeric actin now express troponin, and you can
9
see an example of that large mature differentiated
10
cardiomyocyte where the nucleus continues to be in
11
cell cycle, and we see about 8-fold increase in
12
numbers of these type of cells at the peri-infarct
13
region that has received the human progenitor
14
cells.
15 We think that the cells or the new
16
capillaries are secreting factors that are
17
presumably inducing cycling of these
18
cardiomyocytes, and we are looking at a variety of
19
anti-apoptotic genes that are increased in
20
expression in these cells including redox-related
21
anti-apoptotic genes.
22 The end result is a very significant
23
salvage of the anterior wall of myocardium. You
24
can see that here, and you can see in this scenario
25
obviously a very dramatic left ventricular scar
230
1
formation and aneurysmal formation here.
2 We can show a dose-dependent reduction in
3 scar formation at the left ventricle as we
increase
4
the number of progenitor cells that we inject.
5 Now, others have published recently that
6
perhaps the CD34 cells may have the ability to
7
become multipotential, perhaps transdifferentiate,
8
however, the data are fairly scant, and I would say
9
that given that only 1 in 7 healthy animals
10
demonstrated HLA human troponin co-staining, I
11
think at this point, it is open to debate whether,
12
in fact, these cells do transdifferentiate or
13
whether these cells have the capability to simply
14
fuse with the donor cells.
15 Nevertheless, whatever the precise
16
mechanism, it appears that hematopoietic stem cell
17
injection, either intravenously or
18
intramyocardially, does result in significant
19
global improvement in cardiac function, as we
20
showed here, at least 30 percent improvement in
21
ejection fraction, which is permanent.
This was 15
22
weeks of follow-up post-LAD ligation, and we can
23
look at fractional shortening or DPDT and
24
demonstrate similar sort of things.
25 In this scenario, you can see that there
231
1
are other cell types that we use as controls,
2
CD34-negative cells, saphenous vein endothelial
3
cells, and what this sort of study shows is that
4
you must always use different cell types when you
5
are trying to evaluate a particular therapy. It
6
cannot just be compared to saline or existing
7
control. You have got to
demonstrate specificity
8
in the product that is being tested.
9 Moreover, the question that we asked was
10
what about if we combine this type of an approach
11
with existing therapies, because really that is the
12
question you want to address, and the example of
13
restenosis with stenting is just one type of
14
combination therapy between a new product and
15
existing therapies.
16 What you see here is if we combined CD34
17
cells with both ACE inhibitors and beta blockers in
18
the same rodent model, we had very significant
19
synergy in outcome, and histologically, the reason
20
for that was really not because the two products
21
worked in a similar way, but because they had very
22
separate mechanisms.
23 The ACE inhibitors and beta blockers had
24
no impact on neovascularization, but, in fact,
25
prevented fibrosis in the posterior wall as they
232
1
are known to do, so we have two very different
2
mechanisms of action working in synergy, but it
3
could have just as easily worked the other way
4
around, and that is why we did the experiment.
5 It could have been that by reducing wall
6
strain or reducing fibrous replacement, we may have
7
reduced the drive for endothelial cells to induce
8
neovascularization and improve cardiac function.
9 So, what can be concluded from preclinical
10
data using non-cultured
cells? The objectives
11
that one should look for when you do these kind of
12
studies include identifying the predominant
13
mechanism by which a given cell type induces
14
cardiac recovery.
15 A comparison of efficacy of one given cell
16
type with others, identification of the tissue
17
distribution of that cell type following the
18
preferred mode of delivery, unique short- or
19
long-term risks associated with the preferred cell
20
type, unique risks associated with the method of
21
cell acquisition or isolation. I
have given you
22
the example of GCSF administration, and perhaps
23
alterations in efficacy or safety following
24
coadministration with standard therapies.
25 Now, I would like to shift attention a
233
1
little bit to the second population I want to talk
2
about, the non-hematopoietic stem cells, the
3
CD34-negative, CD45-negative fraction.
4 These cells are rare.
They account for
5
less than 0.01 percent of bone marrow cells, and
6
they do require, for this reason, in-vitro culture
7
and expansion.
8 I apologize about the complexity of this
9
slide. I will just take you
through it a little
10
bit. I mentioned these three
types of mesenchymal
11
lineage cells that people are working with. The
12
so-called mesenchymal stem cells are more likely
13
committed progenitors, but the point is that they
14
are tolemerase-negative.
15 The multipotent adult progenitor cells are
16
MAPCs, tolemerase-positive, and the stromal or
17
mesenchymal precursor cells are also
18
tolemerase-positive. These cells
may be related to
19
each other, but all three types of cells give rise
20
to mesoderm lineage cells. The
MAPCs have also
21
been shown to give rise to endodermal and
22
ectodermal cells. It appears that the mesenchymal
23
precursor cells can also give rise to endoderm and
24
ectoderm.
25 With respect to mesoderm, which is really
234
1
what we are interested in primarily here, cardiac
2
muscle and smooth muscle lineage cells can both be
3
differentiated from the mesoderm, but also these
4
cells can give rise to other cell types with
5
mesodermal lineage.
6 The so-called mesenchymal stem cells are
7
cells that are derived from whole bone marrow and
8
then isolated by simple density centrifugation to a
9
particular layer. These cells
are then taken from
10
the interface, put down on plastic, adhered for 24
11
to 48 hours, and then the cells that continue to
12
adhere are then cultured again fairly crudely with
13
generally fetal bovine serum for weeks at a time.
14 So, if you understand how this process is
15
initiated, you understand that really, the
16
isolation of these cells is so crude that you are
17
starting out with a very heterogeneous population
18
simply based on density characteristics, and that
19
the true multipotential cell within this fraction
20
probably is not more than 1 in 1,000 to 1 in 10,000
21
of the cells you are starting out with.
22 In any case, after, say, 10 to 14 days of
23
culture and passage, you see the sort of
24
monomorphous type of fibroblastoid phenotype, and
25
you see the cells that have survived this period of
235
1
culture are fairly homogeneous in terms of the type
2
of markers that are being used.
3 There are specific antibodies that can
4
define surface characteristics of these cells, SH2,
5
SH3, endoglin, and I am not sure what SH3 actually
6
defines. But in any case, these
cells, after two
7
weeks in culture, become fairly homogeneous.
8 I borrowed this slide from Dr. Epstein,
9
his recent paper in Circulation Research, which
10
shows that these cells, after several passages in
11
culture, demonstrate production following induction
12
of ischemia of a variety of factors that are
13
associated with both angiogenesis and
14
arteriogenesis, and in a nice model or rat hind leg
15
ischemia, you can see--I think you showed this
16
slide already--demonstration of improvement in
17
perfusion following MSC infusion in these cells in
18
ligation of the hind leg artery.
19 But in addition to secretion of
20
pro-arteriogenic factors, the interest in these
21
cells lies in the fact that they seem to have
22
multipotential capability in that under appropriate
23
culture differentiation conditions, they can be
24
differentiated to adipocytes, chondrocytes, and
25
osteocytes.
236
1 Work now since at least 1999 has
2
demonstrated that if you use appropriate inductive
3
signals in these cells, in this case, 5-azacytidine
4
to demethylate the cells, but there have been
5
reports of other agents that can do similar sort of
6
things, you can push the cells toward a
7
cardiomyocyte-like lineage with appropriate marker
8
expression and electromicroscopic criteria
9
consistent with BT myotubes.
10 Probably the best study to date to
11
demonstrate that these cells can actually do
12
something in vivo is work from Victor Dzaus' lab
13
published last year in Nature Medicine.
Here, we
14
are talking about again rat mesenchymal lineage
15
cells cultured in the way I have just defined with
16
density separation and long-term culture with
17
bovine serum, fetal calf serum, and what you see is
18
after injection of these cells into the
19
peri-infarct region of rats following LAD ligation,
20
these cells appear at least phenotypically to
21
acquire markers of cardiomyocytes, so that they
22
express myosin heavy chain, cardiac troponin,
23
sarcomeric actin, and connexin 43.
This is within
24
two weeks of implantation.
25 However, the majority of these cells just
237
1
don't survive when you put them in.
They die after
2
early transplantation. Causes
appear to be
3
ischemia again, competition for oxygen nutrients,
4
inflammatory and oxidative stress in the
5
post-infarct myocardium, loss of survival signals
6 from cell to cell or cell matrix
interactions, and
7
probably lack of tolemerase activity and
8
self-renewal capability by these cells, and I said
9
to you earlier that these cells just don't express
10
tolemerase, certainly when they are in culture.
11 The work by Mongi, et al., in fact, showed
12
that if you genetically modify these cells with an
13
AKT transgene, you could significantly prevent
14
these cells from dying following implantation,
15
significantly reduce the apoptotic index, increase
16
survivability, and overall improve function almost
17
to the level of the non-infarcted animals, quite
18
impressive, but it required essentially making them
19
resistant to death or apoptosis by AKT
20
overexpression.
21 I will just move along to the second
22
population of cells, the MAPCs.
These cells, as I
23
mentioned, are defined based on negative expression
24
of all known markers, CD45 and many other known
25
markers of lineage-specific differentiation, and
238
1
following long-term culture, these cells have been
2
shown to be capable of differentiating to a variety
3
of tissue types requiring cocktails of various
4
cytokines and differentiation factors.
5 In particular, following activation with
6
VEGF and implantation into appropriate medium, in
7
this case, tumor model in the scid mouse, these
8
cells appear to be able to incorporate into first
9
neovascularization, and you can see they can induce
10
networks of capillaries, neocapillaries really, in
11
wound-healing tissues, so they may contribute to
12
vascularity.
13 With that as a backdrop, Gallegos
and
14
colleagues at the University of Minnesota decided
15
to look at whether again fresh non-differentiated
16
MAPCs from dogs could be implanted into an acutely
17
LAD-ligated model and could perhaps improve
18
function in some form.
19 The model was to induce again a complete
20
LAD ligation in the dog, autologous cells were
21
taken from the marrow, they were expanded for about
22
four weeks, and then four weeks later, injections
23
were put directly into the myocardium.
24 What they found was that a month
25
post-injection of the cells,
there appeared to be
239
1
increase in myocardial perfusion reserve defined as
2
perfusion under stress relative to perfusion at
3
rest compared to the controls, and this is in four
4
animals.
5 But significantly, although there was some
6
improvement in regional myocardial contractility
7
within the infarct area and the peri-infarct zone,
8
defined as circumferential shortening and radial
9
thickening, there was actually no improvement in
10
global function of cardiac measurements as defined
11
through systolic improvement by fractional
12
shortening or ejection fraction.
13 Now, just to the last area that I would
14
like to touch on, and that is the possible use of
15
cultured allogeneic stem cells for cardiac disease.
16
The issues to consider here are whether or not
17
cells constituitively or under inductive conditions
18
express surface molecules, surface HLA molecules,
19
and co-stimulatory molecules, secondly, whether
20
stimulation of recipient T cell responses occur in
21
vitro, and, thirdly, whether these cells might
22
induce inflammatory responses after in vivo
23
injection.
24 This work may or may not be familiar to
25
many people here, but many labs now have completely
240
1
demonstrated that mesenchymal lineage cells from
2
humans, from primates, and from smaller animals
3
routinely actually do not express Class II HLA
4
molecules until they are induced by
5
gamma-interferon. They certainly
do all express
6
Class I, but in addition to that, do not express a
7
variety of co-stimulatory molecules, such as CD40,
8
CD80, and CD86.
9
All of these molecules are
absolutely
10
critical in inducing T cell immune responses when
11
you transplant cell or an organ between
12
individuals.
13 What I want to show you here is that
14
mesenchymal stem cells clearly do not induce a T
15
cell response following standard mixed leukocyte
16
reactions in vitro. This, you
can see in
17
comparison to allogeneic mononuclear cells, which
18
induce a vigorous T cell response.
19 In addition, even activation with
20
interferon gamma to upregulate Class II HLA has no
21
effect on mesenchymal stem cell induction of T cell
22
responses, where you can see the potency of
23
gamma-interferon at inducing allogeneic responses
24
if you do it to dendritic cells or to allogeneic
25
mononuclear cells.
241
1 So, there is something very special about
2
these cells, that they do not seem to express
3 co-stimulatory
molecules, they do not seem to
4
induce an allogeneic T cell response at least in
5
vitro, and moreover, they seem to have the ability
6
to suppress ongoing immune responses.
7 What you can see here is that this is now
8 a
third party mixed leukocyte reaction where you
9
have T cells from one donor proliferating. This is
10
to allogeneic mononuclear cells, to allogeneic
11
dendritic cells, or to PHA, and you can see that if
12
you add third party mesenchymal stem cells, they
13
will suppress this proliferative response, and this
14
suppressive effect is in a dose-dependent manner.
15 Additional studies have suggested that, in
16
fact, part of this suppressive effect requires
17
cell-cell contact inhibition, and part of it
18
requires secretion of a variety of
19
anti-inflammatory cytokines, such as TGF-beta, but
20
the precise mechanism at this point has not been
21
defined.
22 The only human trial that I am aware of
23
using allogeneic mesenchymal stem cells to date is
24
one where allogeneic stem cell transplants were
25
performed with third party allogeneic mesenchymal
242
1
stem cells, and I understand the results of those
2
were better engraftment and reduction in
3
graft-versus-host disease, suggesting again that
4
the third party mesenchymal stem cell, not only
5
does not induce an immune response, and is
6
presumably not itself rejected, but is also
7
beneficial in outcome.
8 But just to go back to the cardiac studies
9
now, using that again as the backdrop, Brett Martin
10
and colleagues from Ceros Therapeutics--and these
11
are two slices that he has given me--presented at
12
the American Heart Association last year, their
13
findings using allogeneic mesenchymal stem cells
14
from pigs in acute myocardial injury, and this is
15
now in a reperfusion model to persecute myocardial
16
infarction, what they was generate a panel of
17
allogeneic porcine mesenchymal stem cells that are
18
matched at blood group antigen, they used very
19
large numbers of these cells distributed over the
20
infarct zone. We are talking
about a reperfusion
21
model without any immunosuppression.
22 What you see here is that within two weeks
23
of implantation, the cells were still present, so
24
they hadn't been rejected. You
can see the
25
presence of cells two weeks later.
However, these
243
1
cells did not differentiate to cardiomyocytes, so
2
they did not induce a rejection episode, but they
3
also did not appear to really do much in terms of
4
integration or differentiation.
5 With respect to function, what you see is
6
actually improvement again in diastolic parameters,
7
as we have heard earlier, with increasing diastolic
8
wall thickness at various time points including
9
well beyond six weeks of study, and a reduction in
10
N-diastolic pressures, however, again no
11
improvement in systolic function, which is again
12
consistent with perhaps alterations in matrix,
13
alterations in ground substance, but no real effect
14
on contractility or myocytes.
15 This table, what you see is really a
16
summary of the type of the cells that are being
17
used of mesenchymal lineage. The
MAPCs, so far we
18
have seen only modest efficacy in cardiac models in
19
dogs in the absence of predifferentiation in vitro.
20 Mesenchymal stem cells have shown good
21
function in rats, but modest only in pigs, and
22
mesenchymal precursor cells are currently being
23
investigated. We think they may
be far more potent
24
than either of these two because you are able to
25
isolate them by surface characteristics when they
244
1
are fresh.
2 The final slide here is that I think many
3
remaining hurdles is the message.
The appropriate
4
clinical indication needs to be defined. The more
5
proof of principle animal studies, we have got to
6
determine optimal doses for efficacy, careful
7
registry of adverse events in our animals, let
8
alone humans.
9 We have got to optimize the ex vivo
10 culture
process, which I haven't even talked about.
11
The autologous versus allogeneic question is
12
critical because it is going to impact both on the
13
process and on the cost of this whole therapy,
14
determine the best route of administration and
15
really how do you improve engraftments, we have
16
heard all about that before.
17 Thank you very much.
18 [Applause.]
19 DR. RAO: Thank
you, Dr. Itescu, for a
20
really nice summary and trying to help keep us on
21
time, as well.
22 In the interests of trying to get us back
23
on track, I am going to ask people to limit their
24
questions to things which are directly relevant to
25
the talk and if there are specific burning
245
1
questions on this issue.
2 Q&A
3 DR. CUNNINGHAM:
I just have one question.
4 I
was curious, in all your rat studies you
5
published in Nature, did the rats die, and what did
6
they die of, you know, the ones that didn't get to
7
the endpoint of the experiment, were there any
8
adverse events that you noticed?
9 DR. ITESCU: You
mean unrelated to the
10
initial surgery? We have
something like 30, 40
11
percent death rate from the initial surgery, but
12
pre-cellular implantation, but after cellular
13
implantation is what you are asking me, we carried
14
the animals out. There was no
real adverse events
15
that I am aware of in these studies.
246
1 DR. KURTZBERG:
You made a passing comment
2
as you were speaking that you thought GCSF wouldn't
3
be safe after an MI. Can you
expand on that?
4 DR. ITESCU: I
think in the recently
5
published study, just to summarize my take on that
6
study, a randomized study where one group received
7
GCSF subcutaneously to mobilize the endogenous
8
population of marrow, another group received GCSF
9
subcutaneously to mobilize and then had the
10
mobilized cells harvested, not immunoselected, and
11 I
think the cells were then delivered by I believe
12
intracoronary routes, whole unfractionated cells.
13 The conclusion of the study was that the
14
patients that received the cell therapy had a
15
significant improvement in cardiac function,
16
whereas, the group that received GCSF alone did
17
not.
18 However, in both populations, the study
247
1
was cut short because of this complication, but I
2
think if you pooled both populations, 7 out of 10
3
patients developed significant restenosis at the
4
site of the stent implantation.
5 Now, these are bare stents. This is the
6
pre-repamycin days. I think the
anticipated
7
restenosis rate would have been maybe 25 percent at
8
most in that population. I think even with
9
repamycin, you would expect to reduce the rate--if
10
70 percent is right--diabetic patients actually
11
have a restenosis rate with repamycin stents still
12
of about 20 to 25 percent, 50 percent without
13
repamycin, so these patients are more severe than
14
diabetics are.
15 I wouldn't anticipate that you would be
16
able to lower that to below 35, 40 percent, and
17
that is obviously totally unacceptable.
18 Now, the question is why is it that GCSF
19
was associated with this high restenosis rate, and
20 I
think we can all make speculations, but I think
21
at least one possibility is that the cells that are
22
mobilized, CXER-4 positive cells from the marrow,
23
many cell types express CXER-4 including smooth
24
muscle progenitor cells, and I think that is
25
probably the simplest explanation, that there are
248
1
smooth muscle progenitors that are mobilized, that
2
migrate to the site of the stent, and I am not sure
3
how one can get around that actually, but that is
4
just a guess really.
5 DR. TAYLOR: That was great. I have one
6
quick question. You didn't
mention this except
7
that in vitro, a lot of these cells can become a
8
number of mesodermal cell types, and I think there
9
are some data in rats that showed early on that if
10
you inject these cells in the center of an infarct,
11
they, in fact, at times become adipocytes or
12
chondrocytes or osteoblasts, or something like
13
that.
14 I wondered if you would comment on what
15
cells you think are likely to do that and whether
16
all of them are.
17 DR. ITESCU:
Obviously, that is the worst
18
case theoretical question in this whole field.
19
What we don't want is to develop bone in the middle
20
of their hearts.
21 I am not aware really of a lot of studies
22
that have demonstrated that sort of abnormal
23
differentiation. Certainly, adipocytic
24
differentiation has occurred, I am not sure that I
25
have seen bone differentiation.
249
1 It really depends on how well defined the
2
cell population is, I think, and to what extent the
3
cells have been predifferentiated or to what extent
4
they may be still very multipotential.
It is not
5
clear to me whether you need to start with a cell
6
that is very multipotential or that is fully
7
differentiated, and perhaps the culturing process
8
where you are pushing the cells, and again using
9
fetal calf serum I think is an unfortunate--it is
10
the only thing that as been done to date--but it is
11 I
think probably the worst way to be culturing
12
cells, because you don't know what is in your
13
culture medium, you are pushing these cells to many
14
different lineages.
15 When we look at these cells
16
following--very few studies have been published
17
that have looked at this--but we have looked
18
ourselves in this way, and you can see that after
19
three or four weeks of culture, you can see cells
20
that express markers of mature smooth muscle cells,
21
of mature, some bone differentiation, some
22
differentiation to cartilage.
23
Whether or not that is
relevant when you
24
put the cells back in, whether there is going to be
25
differential potential for outgrowth of one cell
250
1
type over another, I think is really a totally open
2
question, but I would push towards putting cells
3
that are less differentiated in rather than more
4
differentiated, because I think the less
5
differentiated cells have got the ability to still
6
proliferate, to be pushed towards the appropriate
7
lineage under appropriate inductive signals that
8
may still be present in the heart.
9 There may be more differentiation towards
10
maybe a cardiomyocyte lineage, but at least towards
11 a
smooth muscle lineage, so you might get some
12
degree of arteriogenesis, as well as I think it is
13
the more undifferentiated cells that are the ones
14
that produce the very rich supply of arteriogenic
15
factors.
16 Just getting back to what I was saying,
17
the very undifferentiated cells express markers
18
really of parasites, so if you are thinking of
19
parasite implantation, it is one way of maintaining
20
viability of endothelial cells and integrity of the
21
endothelium, and maybe it is a way of building the
22
vascular network.
23 DR. RAO: This
is the last question.
24 DR. BLAZAR:
The issue of
25
transdifferentiation in vivo is striking that you
251
1
can get cells there and they just sit there, and I
2
think this has been seen with a number of
3
non-hematopoietic cell sources, MSCs being one.
4
I guess the question is whether
there are
5
inhibitory signals that are present that prevent
6
differentiation in vivo under certain conditions,
7
and has anyone ever taken these cells back out of
8
the heart and show that they can, in fact, be
9
induced to differentiate in vitro?
Do we know
10
anything about the inhibitory factors present at
11
the site that are precluding differentiation?
12 DR. ITESCU:
Those are obviously great
13
questions, very important questions, and again I
14
think it speaks to having a good understanding of
15
the surface markers of these cells because if you
16
know what they express, you can do the sort of
17
experiments that you are suggesting.
You know what
18
you are putting in, and then you can actually take
19
them out again based on maybe immunoselection.
20 We are actually trying to do those type of
21
experiments to ask exactly those questions, but I
22
think in many ways you are limited then to human
23
cells, because many of the well-defined surface
24
markers, there just aren't enough reagents.
25
Perhaps mouse cells is the only other.
The mouse
252
1
system is the one that is well enough developed and
2
the human system, and other than that, we are
3
really missing reagents where you can do those type
4
of experiments. DR.
RAO: Thank you,
5
Doctor.
6 DR. HARLAN:
Can I still ask a quick
7
question? You were careful to
say so in your talk
8
when you were talking about allogeneic cells, and
9
you said at least in vitro, these allogeneic cells
10
don't activate T cells, but I just wish to
11
emphasize the point that in vivo, it is so much
12
more complicated than that, and the presence or
13
absence of B7, I thought Bruce maybe was going to
14
speak to this, you didn't make the point, but I
15
think it is important to emphasize that whether or
16
not a cell in vitro stimulates a T cell response is
17
really a very poor predictor of whether that will
18
be rejected in vivo.
19 DR. ITESCU:
Well, I am not sure I agree
20
with that at all actually.
21 DR. HARLAN:
Then, we will have more
22
discussion. I saw the pig data,
that is
23
interesting.
24 DR. ITESCU: I
am actually a transplant
25
immunologist. This is what I do,
I take care of
253
1
cardiac transplant patients, and I am going to tell
2
you that 12 months ago, when I first heard about
3
these sort of data, I was extremely skeptical,
4
along similar lines to what you are saying, you
5
know, is it an in vitro artifact.
6 However, you go back and you see how many
7
labs are reproducing these data, which is really
8
what I am surprised about, and I can tell you that
9
the mixed leukocyte reaction is about the best
10
single assay that we have in transplant medicine to
11
predict allogeneic rejection and allogeneic
12
sensitization.
13 In the old days, it used to be used
14
routinely for kidney transplant selection,
15
donor-recipient selection, and it is still the best
16
assay. I am not sure whether it
reflects whether
17
these cells are going to be accepted long term in
18
vivo, but it is certainly a very good marker for
19
biology. I don't know what it
means, but it is
20
routinely being reproduced using these type of
21
cells.
22 DR. HARLAN: To
say it is the best, and I
23
won't disagree with it, is fine, but it is still a
24
very poor predictor of in vivo function, and it
25
goes back 30 years to the two-signal model,
254
1
thinking if you could get rid of the antigen
2
presenting cells within a graft, that you would
3
take, and it seems to work in rodents, but it just
4
doesn't work in higher animal
models.
5 Bruce, do you want to step in?
6 DR. BLAZAR: I
think there is the point of
7
somewhere in between because clearly, there are
8
cell-to-cell contact phenomenon that happen in
9
vitro with regulatory T cells.
You can show a
10
TT-dependent inhibition of responses throughout
11
TGF-beta, and depending on the models in vivo,
12
those either are true or they fall apart.
13 We know in matched sibling donor
14
transplants, mixed leukocyte reaction culture does
15
not predict graft versus host disease.
So, I think
16
that if you do see suppression, it is encouraging
17
to try to go forward for in vivo, and that is
18
probably as good as you are going to do, but it
19
doesn't necessarily mean that those mechanisms take
20
place in vivo particularly as those cells
21
themselves change in their own ability to elaborate
22
cytokines or express other molecules.
23 DR. ITESCU: I
agree entirely and I am
24
saying the exact same thing, that I think it is a
25
phenomenon that is reproducible in vitro, it's an
255
1
unusual and unique scenario. Whether or not it has
2
implications in vivo remains, but I think at least
3
it ought to be tested, and I think it may have
4
implications on dendritic tolerance-inducing
5
mechanism, which truly is a different way of
6
thinking about this all together.
7 It may have nothing to do with the ability
8
of these cells themselves to escape surveillance in
9
the periphery, but perhaps they end up in the
10
thymus, and they may actually be able to reeducate
11
the immune response, but I think all of that
12
remains open to test.
13 DR. RAO: Last
comment.
14 DR. NEYLAN: A
very quick
15 transplant-related
question, again, maybe bring
16
some of these last discussions to a very practical
17
safety concern.
18 That is, given the ability of these
19
allogeneic cells to abrogate or reduce the immune
20
response of the host, is it possible that the
21
migration and homing of these cells may differ to
22
autologous cells in a way, maybe akin to the
23
micro-chimers and observations of solid organ
24
transplantation, that potentially pose a safety
25
risk to the use of allogeneic cells, homing to or
256
1
disseminating and finding a welcome home in other
2
tissues.
3 DR. ITESCU: I
think that is a fair point,
4
in other words, if they are immunoregulatory and
5
you are injecting them and they find their way into
6
the thymus, for example, will they induce a state
7
of tolerance to an exogenous antigen that the
8
patient sees at exactly the same time, which is the
9
concern whenever we use an immunoregulatory new
10
drug.
11 Now, there are a couple of studies
12
actually that have just recently been published
13
that suggest that, in fact, while they induce
14
tolerance to themselves and maybe tolerance to an
15
alloimmune reaction, they don't seem to have
16
induced tolerance to an exogenous pathogen, for
17
example.
18 Now, again, that is two studies, more work
19
needs to be done, but I think the question is fair,
20
it's an absolutely valid point, and obviously, you
21
have to worry about that when you do your studies
22
and follow up the patients very closely.
23 DR. RAO: We
will visit it tomorrow. We
24
should move on.
25 Dr. Taylor.
257
1 From Mouse to Man:
Is it a Logical Step
2 for Cardiac Repair?
3 DR. TAYLOR: It
is a logical step because
4 I
think it raises actually the issues that we are
5
just talking about, whether or not data we get from
6
rodents actually can be translated to humans or to
7
larger animals.
8 I also want to say actually, in terms of
9
an apology, I realized when I sat down that one of
10
the reasons my last talk was so disjointed was that
11
the version that was up here was not the version
12
that I had on my computer, so it was a kind of
13
foreign talk to me, so I apologize and hope we will
14
do better this time.
15 From mouse to man, is it a logical step?
16 I
am going to start with another comment from
17
Ghandi, which I have had on my office door for the
18
last 10 years about this field, which is, "First,
19
they ignore you, then, they laugh at you, then,
20
they fight you, then you win."
21 I think it raises the point that we are
22
somewhere in the continuum in this field, and it is
23
time for us to start asking the hard questions, so
24
that we can have the fight and then win.
25 As we are talking about moving from mouse
258
1 to man, I think we have
to talk about cell type,
2
and I think I would be remiss if I didn't say that
3 I
think cell type depends on what we are trying to
4
do here, and that if we are trying to look at
5
chronic ischemia or hibernation, we are probably
6
looking at cells that are more likely to induce
7
angiogenesis, such as the cells you just heard
8
about - bone marrow mononuclear cells, angioblasts,
9
some subpopulation of stromal cells, growth
10
factors, or maybe even myoblasts plus growth
11
factors, but if you want functional repair and
12
contractile cells, you are either want cells that
13
are contractile, such as skeletal myoblasts or
14
cardiocytes, maybe cardiac stem cells, or you want
15
bone marrow cells that can become contractile
16
cells.
17 But the most important issue probably in
18
this whole field, which is why I think we are going
19
to have to talk about moving from mouse to man is
20
probably this arrow, and the fact that we need both
21
angiogenesis and myogenesis if we are going to have
22
an appropriate outcome, and that, in fact, it may
23
be not just one cell type, but multiple cell types
24
that we end up needing for cardiac repair.
25 Unfortunately, as we are looking at these
259
1
cells, we don't have the opportunity to do those
2
anywhere except in rodent at the present or in
3
humans, so we are going to have to move from mouse
4
to man at least with many of these cells unless
5
industry provides us with the tools that we need to
6
do the studies in between, because right now we
7
really don't have the capability of moving to a
8
larger animal model.
9 I guess I want to start by asking the
10
question what the appropriate preclinical animal
11
models are and how quickly can we move forward by
12
saying, you know, I presented this slide a minute
13
ago, we had 15 years of preclinical data in rabbit
14
and dog and pig and rat, mouse, and sheep, and we
15
thought myoblast transplantation was safe,
16
effective, and feasible.
17 But we missed a lot of things, and we
18
missed--I apologize, I thought there was another
19
part down here, we are off to a great AV start, but
20
that's okay--so what did we miss?
We missed the
21
fact that these cells might be electrically
22
incompatible with the remainder of the myocardium.
23 We missed questions about location of
24
injection, we missed questions about some of the
25
dosing phenomena, we missed a number of things in
260
1
our early preclinical models despite the fact that
2
we used both large and small animal models.
3 So, the question then really is what do
4
you really need to do and when do you move to
5
clinical studies, and I am going to give my jaded
6
perspective for a minute and say that I think
7
sometimes you move to clinical studies because your
8
institution wants you to and kind of forces you to
9
either because there is a financial incentive or
10
that there is no such things as bad PR, but I would
11
like to say that the appropriate time to move to
12
clinical trials is when the data warrant it and
13
that again we have to underpromise and overdeliver.
14
So, if we believe that every
cell injected
15
seems to work, and that thus the heart is easier to
16
repair than we thought, let's take, for example,
17
the possibility that that is really the case, and
18
if that is the case, when do we move to the clinic.
19 I guess I could start by saying we have
20
already moved to the clinic, but that being said,
21
if we look at the clinical data, does it support
22
the fact that the myocardium is easier to repair
23 than
we thought.
24 Well, yes, everything works, but none of
25
the clinical cell studies are placebo controlled,
261
1
and 8 patients or even 53 patients can show you
2
anything, especially when follow-up is short and we
3
aren't considering age, gender, or heart failure
4
status.
5 I want to get back to these, and I will in
6
this talk, that we haven't talked at all about
7
factors like age and gender and how they may be
8
really relevant, and there is a reason that drug
9
trials involve thousands of patients with at leave
10
five-year follow-up.
11 Also, most of the clinical cell studies
12
out there were designed as Phase I safety studies
13
as they should have been, yet, many of these claim
14
efficacy despite the fact that they were either
15
revascularization studies or had other
16
co-treatments involved, and I think it really
17
raises questions about what we need to do.
18 So, if we believe that the myocardium is
19
easier to repair than we thought, what does that
20
tell us about moving to clinical studies? I am
21
sorry, those two slides are actually backwards. If
22
every clinical cell works, what does that mean?
23 I think it means that we have no clue how
24
they work, whether they create angiogenesis or
25
myogenesis, unloading of the heart, recruitment of
262
1
stem cells or whatever.
2 What questions does that raise? It raises
3
questions about patients, which raises the same
4
questions about injury models, preclinical injury
5
models. It raises questions
about dose and timing
6
of cells, which then have to interact with the
7
injury models. It raises
questions about route of
8
administration or location of the cells, and how we
9
measure the outcome, and those all affect which
10
animal model you can choose.
11 Well, the genie is out of the bottle, as
12
the people at Mayo have said.
Clinical trials have
13
started, so what do we do? I
think we educate
14
people about what the appropriate situation is.
15 There are a number of my clinical
16
colleagues that I have talked with, and it scares
17
me a little bit, who don't even realize that if you
18
are going to use bone marrow derived cells, that
19
the FDA needs to be involved if you are going to
20
put them in the heart, and I think that is an issue
21
that we really have to address.
22 I think we need to require enough
23
preclinical data, and then I think we need to quit
24
rewarding people for doing it wrong.
What do I
25
mean by that? Well, what I mean
by that is we have
263
1
to quit doing science by the Washington Post, as
2
you said earlier, and we have to quit focusing on
3
the fact that this is a multibillion dollar market
4
every year, and focus on the patients instead.
5 So, what do myoblasts tell us about moving
6
forward? Well, as I said, what
we knew and what we
7
missed is that there are electrical events, the
8
route of administration wasn't clear, location and
9
timing wasn't clear, culture medium, we thought we
10
knew, but it didn't turn out to always be the case.
11 Autologous serum has been reported to be
12
safer than non-autologous serum.
Different
13
vehicles have been used to deliver the cells and
14
been associated with different outcomes, and we
15
don't know anything about biodistribution, and we
16
really didn't look.
17 So, what issues are there? The issues are
18
safety and efficacy obviously, and as we move up
19
this continuum, we can address
these issues
20
differently. Safety obviously
involves cells,
21
delivery, dose, and we have to do that in relevant
22
models.
23 Efficacy involves the right model, acute
24
MI potentially, looking at various cell types in a
25
side-by-side way. We have to
begin to
264
1
differentiate between diastolic heart failure. I
2
think the preclinical data that exists so far
3
suggests that everything works to begin to improve
4
remodeling and diastolic effects including cells
5
that don't work in systolic heart failure.
6 Systolic heart failure in our lab at
7
least, we know fibroblasts don't work, stromal
8
works less well than some other cell types, and I
9 think the issues are
open questions that have to be
10
addressed.
11 If you then believe that the issues effect
12
are impact cells, delivery, and effect, how does
13
that translate to the animal models?
Well, in
14
small animals, I think we can ask questions about
15
the cells. We can ask questions
about deriving the
16
cells, we can ask questions about markers for the
17
cells, we can ask in vitro questions.
18 I think as we begin to move towards
19
delivery and effect, we have to move up the animal
20
continuum. Delivery, we really
can't do in small
21
animal models, relevant delivery, we can't do.
22
Effect, again, I don't think we can do in small
23
animal models, we have to do it in larger animal
24
models.
25 In addition, when we start looking at the
265
1
mechanism of effect, to some degree, if we are
2
looking at angiogenesis, we can measure capillary
3
density in small animals, but it is very difficult
4
to get adequate measures of relevant vessels or
5
perfused vessels in small animals, and I think we
6
have to move up the continuum.
7 In terms of myogenesis, we can get some
8
data about ex vivo in isolated heart preparations
9
and whether or not there are gross improvements in
10
contractility. We can begin to
measure wall
11
thickness, ejection fraction, and you can certainly
12
do exercise studies in small animals and move up
13
the continuum, but in terms of electrical
14
compatibility and mechanical compatibility, you are
15
never going to get it done in a rodent model, you
16
have to do it in large animals.
17 So, I think it is only relevant to work
18
with small animals when you have no choice about
19
the cells, but as you start to measure the
20
important parameters of physiology, you have to
21
move up the model.
22 So, what are the possible effects of these
23
cells? I think they can cause
unloading of the
24
heart or reverse remodeling simply by altering the
25
mechanical properties of the scar.
They could
266
1
possibly engraft and become muscle and contribute
2
to contraction.
3 They can obviously potentially form
4
vessels or secrete factors that recruit cells that
5
improve blood flow. You know, if you think about a
6
lot of the data we have seen so far, it is possible
7
that the cells we are putting in are doing nothing
8
but recruiting bone marrow derived cells and
9
actually ramping up endogenous repair.
10
It is possible that the
cells have a
11
paracrine effect and either change the cytokine
12
supply to the scar or the remainder of the heart,
13
or recruit other stem or progenitor cells.
14 Another possibility that is not really
15
talked about, and again you can do these studies in
16
small animal models better than large animal
17
models, is fusion with existing cardiocytes. It is
18
possible that the cells we inject actually fuse
19
with cardiocytes that are hibernating and save
20
those cells, and thereby contribute to contraction.
21 If we are going to talk about the
22
different possibilities, cell delivery and effect,
23
and which animal models to use, I want to briefly
24
say that we have to define our populations of
25
cells, and that is easier in some animal models
267
1
than others.
2 Again, when you start looking at
3
myoblasts, fibroblasts, and other cells, being able
4
to do that in rodents is much easier than being
5
able to do that in larger animals.
I think I have
6
beat that horse, so I won't keep saying it, but
7
cells, I think are studies we can do in rodents.
8 You have seen this slide before where we
9
begin to talk about how the cells might work and
10
whether or not we have an effect on reverse
11
remodeling or growing new cells, and whether or not
12
we think we can do those studies in large animals
13
or small animals. I would submit
that we are going
14
to have to do those in large animals because we
15
can't really measure remodeling and reverse
16
remodeling in all of the new cells in the small
17
animal models.
18 Now, we took the approach that if we are
19
going to start moving back up this cascade of
20
events, that might, in fact, take a combination of
21
angiogenesis and myogenesis, and it might take
22
cells plus genes or multiple combinations of cells.
23 Initially, we looked at cells plus genes
24
by virally infecting myoblasts with VEGF
25
adenovirus, and we measured the effect of those
268
1
cells on capillary density in peripheral skeletal
2
muscle and a hind limb ischemia model, and these
3
are old data, so I am not going to spend much time
4
on them other than to say that we found that
5
myoblasts had a relatively significant effect in
6
terms of increasing capillary density greater than
7
VEGF virus alone or MT virus, but when we
8
overexpress VEGF in the myoblasts, we got back up
9
to about 75 percent of the control without the side
10
effects that we saw with VEGF virus alone,
11
angiomas.
12 So, we decided to move into the
13
myocardium, and these are data done with cells that
14
overexpress another angiogenic factor where we
15 looked
at MRI, and I apologize, this is percent
16
change and ejection fraction, and our historical
17
controls, and these really are historical, they are
18
not done at the same time, the active study was
19
myoblasts versus angiogenic myoblasts versus
20
historical shams.
21 We found that we increased capillary
22
density significantly in these animals.
These were
23
studies done in mice. We
couldn't do these studies
24
actually in rabbit because rabbit cells don't have
25
the cell surface receptors to actually take up or
269
1
be transfected with some of the viruses that we
2
were trying to use, so that becomes a problem.
3 But we believe that angiogenesis was
more
4
important than to promote cell survival or
5
proliferation, and we also looked at Victor Dzaus'
6
data, I wanted to show briefly.
You saw it when
7
Silviu presented it, that if you overexpress a
8
survival factor AKT in bone marrow stromal cells
9
and transplanted them into the heart, survival
10
increased.
11 So, we know angiogenesis increases
12
function, increasing survival increases function,
13 so if angiogenesis helps and increasing survival
14
helps, why use a gene, why not use a mixture of
15
cells.
16 We based that on data that we have gotten
17
now in our hands where we were able to show that if
18
we gave cells from young apoE animals, gave bone
19
marrow cells from young apoE animals to apoE
20
animals that were fed on a high fat diet, normally
21
develop pretty bad atherosclerosis, that we could
22
actually prevent this atherosclerosis.
23 We took, we actually have now begun to
24
take these cells and deliver them in combination
25
with myoblasts to look at effects on function. I
270
1
don't have a slide, but I can tell you that we are
2
beginning to see better effects on function with a
3
combination of cells and with individual cells
4
alone. We couldn't do these
studies except in
5
mice, so there are times when mouse cells are
6
relevant.
7 We did another study where we began to
8
look at bone marrow mononuclear cells.
These are
9
data that were just presented at the ACC last week,
10
where we actually infused mononuclear cells into
11
the circulation of animals where we created a
12
vascular injury.
13 We are able to show in our sham-treated
14
animals or animals treated with other cells that
15
you had neointimal proliferation, that we were able
16
to prevent with bone marrow mononuclear cells.
17 So, again, we have now started the
18
approach of delivering these cells in combination
19
with myoblasts to see if we can have a more
20
dramatic effect on not only myocardial repair, but
21
on vascular repair, as well.
22 So, what are the other factors that are
23
likely involved that are going to affect the model
24
we use? The timing after injury,
whether or not we
25
can really grow old animals that replicate the six
271
1
to seven years that are needed, the type of injury,
2
it is not probably going to be feasible to do
3
dilated cardiomyopathy studies in mice and inject
4 significant
numbers of animals, plus when we start
5
trying to treat ischemic and chronic and acute
6
animals, mice, the issue we have to consider is we
7
have got a 1 mm infarct, we got a 1 mm infarct, and
8
the cells are microns in diameter.
9 Those cells are the same size essentially
10
in rat, rabbit, pig, human, they are not much
11
different in size, so the whole geometry of putting
12
those cells in and getting an improvement is going
13
to be much different than you are going to see in
14
larger animal models.
15 There are two issues that really don't
16
affect what animal you choose, but they are not
17
being discussed at all, and those are gender and
18
age.
19
Most of the preclinical
data that we have
20
published are in female rabbits, so we went back
21
and started doing studies in male rabbits, and what
22
we found is that male myoblasts die under
23
conditions where female cells survive, and that is
24
true both in vitro and in vivo, and that really
25
surprised us, and we had to go back and begin to
272
1
reevaluate what we think is going on here, and we
2
are just beginning to follow up on that, but I
3
think it is an interesting point that we are going
4
to have to consider going forward.
5 In terms of age, all of the studies we
6
have done have been in old animals, but that is
7
rarely the case. It is rarely
the case that old
8
mice are used in these studies, that old pigs are
9
used in these studies. In fact,
typically, people
10
use young pigs because they want to keep them
11
small, and don't want them to grow significantly
12
over the duration of the studies.
13 I think the numbers and kinds of cells
14
that you can obtain are going to be very different.
15
Other than cells, we have to consider the culture
16
conditions, and we can't ignore the fact that
17
autologous serum has been touted as one reason that
18
there is a Spanish study where there aren't any
19
abnormal electrical events even though there have
20
been in all the other human studies.
21 So, the injury models currently that we
22
are using don't match the patients, and I think we
23
are going to have to really think about that going
24
forward. We don't have heart
failure models, we
25
just don't. Nobody is using
heart failure models,
273
1
but every patient is a heart failure patient.
2 As I said before, I apologize, in terms of
3
whether or not we can use small animals, I think
4
for an isolated heart prep, mouse and rat are fine,
5
but as you are going to start doing physiologically
6
relevant studies, you have to move up the animal
7
continuum, but there are limitations there, as
8
well, so we have actually chosen, and I think more
9
and more people are choosing, to use sonomicrometry
10
or cine MRI. Fortunately, you
can use that for all
11
of the animal models that have been proposed so
12
far.
13 I am not going to show those data.
14 I actually want to end with two slides
15
that show something that I have tried to gather
16
from the clinical data that exists, but I think
17
they make a point.
18 We are talking about different animal
19
models, but we are also talking about different
20
cell types, and people are constantly saying how do
21
myoblasts and bone marrow derived stem or
22
progenitor cells compare.
23 The point I want to make is they don't.
24
If you look at the studies that exist so far--and
25
these are clinical studies, not preclinical
274
1
studies--if you look at myoblasts, the dose varies
2
widely.
3 If you look at bone marrow derived stem or
4
progenitor cells, the dose is significantly less
5
and the physiologically relevant cells are a very
6
small subset of those, as you just heard.
7 Moreover, the patients differ greatly.
8
With myoblasts, the patients are from greater than
9
one month to end-stage heart failure, but with the
10
progenitor cells, the patients are 3 to 9 days
11
post-MI or have refractory angina.
12
The delivery methods differ
significantly.
13
They are intracardiac for myoblasts, surgically or
14
percutaneously. They are
intracoronary for the
15
bone marrow derived cells. Yet,
people are trying
16
to compare the outcomes from these, and I think
17
that is true, not just clinically, but
18
preclinically, as well, as people are trying to
19
make the argument for their cell type.
20 So, until we are doing side-by-side
21
studies with the same cell type and the same animal
22
model, I really think we can't draw conclusions
23
about what is going on.
24 So, the questions that I think are really
25
out there are: Is there a best
cell? I don't
275
1
think we know that there is yet.
2 Should we use just growth factors? I
3
think the GCSF data would suggest not yet, we don't
4
understand it well enough.
5 Is there a future for biologic devices?
6
Probably. Are we going to be putting cells on
7
stents? I don't know, but we
probably are going to
8
be putting them on patches and other devices in the
9
very near future.
10 There is a real question, dose, timing,
11
and patients, probably, but our models need to
12
mimic that.
13 I will stop there.
14 [Applause.]
15 Q&A
16 DR. RAO: Thank
you, Doris. I wanted to
17 start off by asking you one question first. That
18
was that implicit in all of these talks throughout
19
has been the fact that it seems to be important
20
when choosing a model that you have to have the
21
right markers. That really is
because you are
22
doing syngeneic transplants. You
want to put the
23
same animal species cells back into the animal to
24
do the follow-up.
25 Nobody, at least in this field, seems to
276
1
think that you can immunesuppressed mouse and use
2
human cells directly, or do some equivalent
3
otherwise, and that that is reasonable.
4 Now, is that true?
That is important
5
because this is going to be really important in the
6
future. I just wanted to get your feel for that.
7 DR. TAYLOR: I
think that is fairly
8
accurate, and I think the problem has been that we
9
haven't had the opportunity to really measure the
10
effects of these cells in mice very well until the
11
last couple of years.
12 Until the last probably year and a half to
13
two years, you couldn't do MRI in a mouse reliably.
14
You certainly couldn't get a good enough image to
15
measure regional versus global function.
16 Sonomicrometry was hard to do in a mouse.
17
The pressure volume catheters weren't quite up to
18
snuff, and so the ability to make those
19
measurements weren't true. Moreover, most of us
20
believed early on at least that autologous cells
21
were more likely to be clinically accepted, and
22
patients certainly liked the concept of getting
23
their own cells better than the concept of getting
24 somebody else's.
25 So, I think we chose those cells because
277
1
they made sense to us clinically, and we were
2
technically limited by our ability to make the
3 measurements with other
cells.
4 In terms of using human cells in
5
immunocompromised rodents, I think we can do that
6
now, but whether or not--I think it is an open
7
question about whether or not we are going to get
8 the best functional
outcomes.
9 DR. ITESCU:
Can I just maybe add a little
10
bit to that? You know, we use
the
11
immunocompromised rodents pretty well, but I think
12
what we are learning as we move forward is that
13 even the so-called immunocompromised rodents
are
14
not fully immunocompromised, and you have really
15
got to start understanding which kind of lineages
16
in their immune system remain active, and what
17
impact does that have on the cells you are putting
18
in.
19 We are now at the point where we are
20
adding cocktails to try to remove even the residual
21
immune function in these kind of animals.
22 On the other hand, I think that if you are
23
trying to use a cocktail of immunosuppressive
24
agents in a normal animal, I think then you are
25
going to run into the problems of what effects all
278
1
these drugs have on the cells that we are putting
2
in, are they inhibiting differentiation, are they
3
inhibiting function. There is a
whole range of
4
issues that I don't think we want to get into.
5 DR. RAO:
Absolutely, that is important.
6 DR. HARLAN: I
will make a comment on
7
that, and then I had a question.
I agree, in the
8
islet transplant field that I know best, for
9
instance, in order to correct a mouse with human
10
islets, you need about 1,500 islets, whereas, you
11
need about 400 rodent islets, and presumably it is
12
because there are species differences in the
13
factors that support the growth of those cells.
14 Then, I agree and appreciate your talk,
15
but I would extend it in two ways.
I think a theme
16
of your talk was that the large animal models in
17
general tend to be better than rodents in
18
predicting things, and I think that is true, but I
19
wish to point out that all models are models.
20 In our transplant studies, we did things
21
in primates, testing various immunotherapies in
22
different systems, and it worked beautifully and
23
failed miserably in the clinic, so even large
24
animal models, even using nonhuman primates, are
25
models, and they have variables that are hard to
279
1
predict.
2 DR. TAYLOR: I
would add one comment to
3
that, which is all of us are looking at progenitor
4
cells and the effects of these progenitor cells on
5
cardiovascular function, and we know that the
6
animals we are treating are relatively acutely ill
7
even if we have given them heart failure, but
8
patients are very ill.
9 They have had things impacting their
10
progenitor cells for years, that we know nothing
11
about, and it is not just aging, it's drugs, it's
12
other things, so the cells we get out and the cells
13
we put back in are going to be very different than
14
cells we get out and put back in, in animals that
15
we have only made sick for a month or a year.
16 DR. HARLAN:
And because we don't know all
17
the factors that made people sick, generating that
18
perfect animal model may be an impossible ideal,
19
that's the only point, I think we are agreeing.
20 The second thing is, though, that you
21
didn't mention is in any model, and when you go to
22 the clinic, I think it is important and we
started
23
the day that way, identifying the patients for whom
24
existing therapies just have failed, so that is one
25
way where you can go. When you
have got nothing
280
1
else to offer, and you have the patients in a bad
2
position, then, I think it passes the threshold.
3 DR. TAYLOR:
But those aren't the models
4
that people are using.
5
DR. HARLAN: Well, we will talk about that
6
tomorrow, I agree.
7 Then, the third point, I say this in jest,
8
and I don't buy it, but I like your Ghandi quote,
9
and I will cite another philosopher, W.C. Fields,
10
who said, "If at first you don't succeed, try
11
again; if it fails again, you might as well give
12
up, there is no sense being a damn fool about it."
13 I don't agree with that, it's just the
14
other take.
15 DR. RAO: Dr.
Epstein.
16 DR. EPSTEIN: I
just wanted to make a
17
point, which I think is important because I sense
18
that one could very easily come to the conclusion
19
without thinking it through further that a large
20
animal model is the really only valid preclinical
21
model.
22 I think Doris sort of made this point, but
23
it might have been lost. It
depends on what you
24
are looking at. For example, if
are interested in
25
myogenesis, I would agree, ultimately, you have to
281
1
go to the large animal model because the small
2
rodents, mice, become difficult to draw conclusions
3
about, but if you are looking at angiogenesis, the
4
mouse model is much better, I think, than a large
5
animal model.
6 We have excellent ways of measuring
7
perfusion now, superb ways of measuring perfusion,
8
and what comes up in our laboratory a lot is if we
9
give an intervention that has an effect, and then
10
we want to see whether we could further enhance
11
that, like stromal cells and then genetic
12
engineering of stromal cells, in a pig model of
13
myocardial ischemia, it starts out with 85 percent
14
of normal.
15 If you get it up to 95 percent of normal
16
with your first intervention, you have no room to
17
look at the next step, whereas, with mice, that
18
could be modulated much more easily, and we are
19
able to demonstrate a primary effect and then an
20
additional effect on it.
21 So, I think, you know, I wouldn't like the
22
FDA to go away with the conclusion that you have to
23
do an efficacy model in a pig or a dog to go to the
24
clinic. Again, it depends what
you are looking at.
25 I will make a point about the
282
1
immunosuppressed animals.
Angiogenesis, an
2 intrinsic
component of the angiogenic process is
3
inflammation, so if you put in cells in a model
4
that is immunocompromised, it is a laboratory model
5
where you have taken away one of the normal
6
modulating influences, so you have to be very
7
careful using an immunocompromised model at least
8
to look at a process in which inflammation is a
9
very critical component.
10 DR. RAO: A
point well taken.
11 Bruce.
12 DR. BLAZAR: Doris,
I wanted to ask you,
13
since you have reviewed all of the preclinical
14
data, and you had the tenet that the data should
15
drive the studies, what of the preclinical data is
16
sufficiently compelling that this would have driven
17
the studies to go forward, are the models done so
18
far incomplete to be able to decide on appropriate
19
studies?
20 I know we are going to talk about that
21
tomorrow, but since you have reviewed in two talks
22
this issue--
23 DR. TAYLOR: I
think there is certainly a
24
lot more preclinical data from myoblasts than there
25
are for some of these bone marrow derived
283
1
progenitor cells. I think that
is in part because
2
we have known about myoblasts since 1961 and we
3
have know about these bone marrow derived cells for
4
the last five or six years, so it is not
5
surprising.
6 I think myoblasts have been used in
7
peripheral models, they have been used in cardiac
8
models of injury, they have been used in large
9
animals, they have been used in small animals, and
10
there is a confluence of data, all of which say if
11
you give these cells, the animals get better.
12 I think most of the cardiac models were
13
relatively acutely after injury, within a month
14
after injury, but nonetheless, they said if you use
15
these cells, the animals get better.
16 I think what is starting to happen in the
17
bone marrow mononuclear cell field and in the bone
18
marrow stromal cell field, and in even the MAPC
19
field, is that we are seeing isolated studies with
20
cells that are called a given thing, but aren't
21
necessarily defined the same way.
22 The way one group defines an EPC, and the
23
way another group defines an EPC may be very
24
different, so there is not necessarily a confluence
25 of
data yet, and you can't necessarily even compare
284
1
some of the preclinical studies because the cells
2
are very different, or we don't know what the
3
criteria are for those cells.
4 So, I think that what has to happen is we
5
have to know what the definition of the cells are,
6
and that is only now becoming true for endothelial
7
progenitor cells.
8 DR. BLAZAR:
So, a corollary of this
9
question is, is it pretty well established by those
10
of you in the field, what the bar is that you have
11
to get over.
12 It seems the way you have described this,
13
it's a systolic function bar that is--no.
14 DR. TAYLOR: I
think it depends on what
15
you believe the mechanism of action is for the
16
cells, and I think again we don't know that, but it
17
looks like what is coming about generally is that
18
for bone marrow derived cells, the goal may be
19
angiogenesis, and for muscle-derived cells, the
20
goal may be myogenesis, and those have different
21
criteria and different preclinical studies that I
22
think you ought to do.
23 DR. BLAZAR:
But that presumes for
24
clinical applications that you know exactly the
25
pathophysiology of the lesion you are trying to
285
1
treat.
2 DR. TAYLOR: I
think it begins to argue
3
for timing after injury.
4 DR. BLAZAR:
Okay.
5 DR. RAO: If
anybody has a really brief
6
comment, otherwise, we will take Dr. Mul and then
7
Dr. Schneider.
8 Do you have a quick comment?
9 DR. MENASCHE:
Go ahead. I just had a
10
comment, but I can wait.
11 DR. MULE: I
would just like to follow up
12
on Bruce's questions. This has
been an incredibly
13
frustrating afternoon for me from the standpoint
14
that I think the presentations have nicely pointed
15
out the strengths and weaknesses of small animal
16
models versus large animal models.
17 Overlaid on that is the fact that none of
18
these models really are good models for the actual
19
disease state in humans. What I
was hoping to hear
20
this afternoon was taking the strengths and
21
weaknesses of each of the models and maybe laying
22
out, hopefully tomorrow, what an ideal, if we could
23
go that way, what the ideal recommendations would
24
be for the field to help the clinicians, such as
25
Philippe, in conducting the next generation trials.
286
1
That is the real concern to me,
is that at
2
the end of the day, the horse is out of the gate,
3
and Philippe and others are going to be conducting
4
these trials rather rapidly, and I think we need to
5
help them to establish some guidelines as to
6
whether or not we should abandon animal models,
7
move on to the clinic, and design the clinical
8
trials in such a way that we get the best
9
scientific data available and the best clinical
10
situations that can be defined, and define the
11
endpoints of the clinical trial with the
12
appropriate placebo.
13 So, I just wanted to lay out that my
14
deepest concern is that years from now, we will be
15
using the same models that you have very nicely
16
summarized, small animals, large animals, fully
17
aware that these limitations continue to exist, and
18
whether or not those data that are generated over
19
the next several years will help Philippe and
20
others to characterize how we should go forward in
21
conducting these clinical trials.
It is just some
22
comments I had.
23 DR. RAO: I
think you are echoing what the
24
FDA is feeling, I guess, right now.
25 Dr. Schneider.
287
1 DR. SCHNEIDER:
Doris, you know that I
2
share your general cautionary note and share many
3
of the specifics, but let me disagree with two
4
specific points that you made about the lack of
5
utility or impediments to the use of the mouse.
6 One of them that you singled out is the
7
impediment to using the mouse as a model for
8
studying the electrical connectivity of donor cells
9
to the host environment, which has become an issue
10
of prominence because of the issue of ventricular
11
tachycardia.
12 I think Loren Field has shown very
13
convincingly you can use 2-photon microscopy to
14
study the propagation of action potentials on the
15
epicardial surface, you can study the propagation
16
of calcium transients in the mouse heart with
17
grafted cells, so that is a non-issue as of this
18
year.
19 DR. TAYLOR: I completely
disagree, but we
20
can have that discussion--
21 DR. SCHNEIDER:
Perhaps so, but
22
Circulation doesn't, and there is adequate
23
peer-reviewed data out there that shows that it is
24
technically feasible with some esoteric
25
instrumentation.
288
1 The same thing I think is true in terms of
2
the issue of mechanics. As you
say yourself, cine
3
MRI levels the playing field across all of these
4
species, and at a few centers, cine MRI is even
5
being combined with spam, so that people can do
6
finite element analysis in the small mammals.
7 So, to me, and I share Dr. Mul's
8
frustration, I tend to focus myself on a different
9
issue, and that is not whether the small mammal is
10
adequately achievable for the endpoints that we
11
want to study, but is the small mammal adequately
12
predictive of the pathophysiology we want to study.
13 There, the issues include the fact that
14
neither the small nor the large mammal is done in
15
an aged animal against a background of diffuse
16
atherosclerotic disease.
17 DR. TAYLOR: I
want to just comment on
18
that looking at electrical connectivity is not the
19
same thing as looking at ventricular tachycardia or
20
arrhythmias, and that it is going to be difficult
21
to do those studies when the heart is beating at
22
600 beats per minute, and that is I think one of
23
the main issues, as well as the fact that most
24
electrophysiologists tell you that the larger
25
geometry is much more conducive, of the human
289
1
heart, is much more conducive to electrical
2
abnormalities than the smaller geometry of a rodent
3
heart, so you are not going to see the same
4
properties even if you have an injury there.
5 DR. SCHNEIDER:
Again, it becomes an issue
6
of whether it is a predictive biology, not an issue
7
of whether it is technically achievable.
8 DR. TAYLOR:
Sure.
9 DR. MENASCHE:
To some extent, Dr. Mul
10
has anticipated my comment. What
I wanted to
11
emphasize from a clinical standpoint is that
12
regardless of the animal model we are going to
13
use--and I have advocated large animal models, as
14
well--there is not a single animal preparation
15
which can realistically model the very complex
16
situation of the patients we are dealing with.
17 There is not a single model which can
18
reproduce the situation of a 70-year-old person
19
with Class III heart failure, two previous bypass,
20
seven angiopathies, collaterals, and so on, so at
21
what point we really have to be prepared to move
22
forward and to go across the gap.
23 Now, considering skeletal myoblasts, we
24
have clear evidence regardless of the limitations
25
of the models that the technique can be implemented
290
1
easily in patients.
2 Number two, we have reasonable evidence
3
that it is safe provided some precautions are
4
taken. We don't fully understand
the mechanisms,
5
but many interventions are currently using patients
6
without an extensive understanding of the
7
mechanisms. I think it took
years before people
8
understood how aspirin was working.
9 So, as we are continuing to try to
10
understand the mechanisms, I think that given the
11
huge population we have to deal with, and the
12
number of patients without any option, it is
13
critically important and timely appropriate now to
14
move to the efficacy studies, and we are certainly
15
looking for advice and help for designing the study
16
in such a way that they can draw meaningful
17
conclusions and answers to the two fundamental
18
questions, does it improve function in the areas
19
where cells are put, does this improvement have an
20
impact on the clinical outcomes.
21 DR. TAYLOR:
Philippe, the one comment I
22
would make to that is I think you are absolutely
23
right, we don't have good animal models, I don't
24
anyone in the room will argue that.
25 I think the thing I would add to what you
291
1
said is you mentioned no-option patients, but those
2
aren't the patients that are involved in some of
3
the studies that are going forward, and I think
4
those are the patients who should be involved in
5
the studies going forward, and I hope that is one
6
of the things that will emerge from this, that it
7
may be responsible to go forward in those groups of
8
patients, it may not be responsible to go forward
9
in some other groups without more preclinical data.
10 DR. RAO: I
would like to just remind
11
everyone that part of the discussion is not
12
specific to one particular cell type, but is in
13
general, and maybe the conclusion may be that it is
14
just simply we have to have differing criteria
15
depending on the cell type or the model, or
16
whatever, when we discuss it, at least its options,
17
and that may be as far as one can go.
18 So, keep that thought in mind, that
19
nothing that you hear
necessarily means this is
20
absolute or anything in that fashion.
21 DR. ITESCU:
Along those lines, I think
22
the point you are making is
exactly right. Some
23
cell products are based on characterization of the
24 surface phenotype, based on many years of
25
immunoselection with markers only present in humans
292
1
and not present in any other species.
2 Some cell types have been characterized by
3
very different biologists, who have used functional
4
outcomes and functional criteria.
So, you are
5
talking about apples and oranges really, so you
6
can't have the same clinical models or preclinical
7 models
for those, so you have got to define
8
appropriate models for each of those cell types,
9
because they have been characterized from totally
10
different perspectives.
11 DR. RAO: I
thought we might use that to
12
start the conversation tomorrow, yes.
13 Everybody looks like they need a break, so
14
we will take a short 10-minute break.
15 [Break.]
16 DR. RAO:
Welcome back, everybody.
17 We are going to change topics a little bit
18
and talk about devices now. It
is finally going to
19
be the turn of Dr. Jensen to talk on cardiac
20
catheters.
21
Cardiac Catheters for Delivery of Cell Suspensions
22 DR. JENSEN: I
am going to go ahead and
23
get started because we are running a little bit
24
slow here, and I am also going to try and see if I
25
can catch up a little bit in terms of time.
293
1 My name is Nick Jensen. I work in the
2
Division of Cardiovascular Devices in FDA's Center
3
for Devices and Radiological Health, and I have
4
been asked to briefly introduce the cardiac
5
catheters that have been used for delivery of cell
6
therapies intended to treat cardiac disease.
7 My presentation today will be limited to
8
potential questions that relate to the interaction
9
between cell therapy suspensions and the cardiac
10
catheters used to deliver these therapies.
11 Further, the questions that I list today
12
are among the standard questions that we currently
13
suggest to all sponsors either of cell therapy INDs
14
or for these investigational catheters.
15 As examples, we will discuss two types of
16
cardiac catheters that have been used to deliver
17
cell therapies to the heart.
18 First, is infusion of cells into a
19
coronary artery during a balloon occlusion of the
20
artery. Second, is needle-tipped
injection
21
catheters designed to permit percutaneous
22
transendocardial injection into the myocardium.
23 Again, these simply represent the methods
24
and catheter types that have been most commonly
25
used to date, and they are again
meant to provide
294
1
useful examples for discussion.
I think it is hard
2
to predict what types of devices, what delivery
3
methods may be used in the future.
4 The first method, infusion of cells into a
5
coronary artery offers potential advantages that
6
include simplicity and ease of use.
As has been
7
mentioned earlier today, this method may not be
8
suitable for all types of cell suspensions.
9
Potential limitations include the potential
10
requirement that infused cells be able to migrate
11
from the vasculature into the myocardium, and a
12
potentially increased risk of embolization or
13
microembolization if some types of cells are
14
infused using this method.
15 Published case series using this method
16
have demonstrated preliminary clinical feasibility
17
both when used within hours to days following acute
18
myocardial infarction, and this is often following
19
emergency stent placement at the site of thrombotic
20
occlusion.
21 It has also been used in patients who
22
suffer from chronic myocardial infarction and
23
ischemia.
24 In the studies reported to date, balloon
25
catheters have typically been used to temporarily
295
1
occlude the coronary artery proximal to the
2
treatment region. The desired
cell suspension is
3
then infused into the artery distal to the inflated
4
balloon either using a lumen within the balloon
5
catheter, typically, a guidewire lumen, or using a
6
separate infusion catheter placed lateral to the
7
balloon catheter, such that it lies between the
8
inflated balloon and the artery wall.
9 Use of balloon occlusion permits infusion
10
of cells at pressures that exceed coronary artery
11
pressure, and it has been hypothesized that
12
increased infusion pressures may provide benefits
13
that include increased dispersion of cells within
14
the vasculature, increased adhesion of cells to the
15
vascular endothelium, and increased migration of
16
cells across the vascular endothelium and into the
17
myocardium.
18 This familiar illustration from a
19
publication by Strauer, et al., illustrates the use
20
of a coronary artery balloon catheter to infuse
21
cells into a region of acute myocardial infarction.
22 In this illustration, the balloon catheter
23
has been inserted into a large artery, directed
24
retrograde through the aorta, then, into a coronary
25
artery, and then directed distal within that artery
296
1
to the site of acute thrombosis.
2 Following inflation of the balloon to
3
obstruct the artery, a syringe is used to infuse
4
cell suspension through the guidewire lumen of this
5
catheter and into the coronary artery distal to the
6
balloon.
7 Finally, although it is not obvious in
8
this illustration, in this clinical study, the
9
balloon was inflated at the site of acute
10
thrombosis, and more specifically, at a site where
11 a
coronary artery stent was placed as an emergency
12
treatment for the myocardial infarction, and that
13
is potentially important because balloon deployment
14
within a coronary artery stent can largely protect
15
the artery from one potential concern that we will
16
discuss briefly today, and that is the potential
17
for damage to the artery caused by balloon
18 inflation
and subsequent stretching of the artery.
19 Studies of this cell delivery method
20
reported to date have commonly used balloon
21
angioplasty catheters to occlude the artery. The
22
catheters are originally designed to stretch the
23
lumen of an occluded fibrotic atherosclerotic
24
coronary artery to a specific diameter that has
25
been selected in advance by the treating physician.
297
1 They are also used to expand
coronary
2
artery stents and again to prespecify diameter.
3
Although these catheters were not originally
4
designed for occlusion of an artery, they can be
5
used for that purpose, followed by cell suspension
6
and, additionally, as noted, if these catheters
7
have a central guidewire lumen, that lumen can then
8
potentially be used to infuse the cell suspension.
9 Potential considerations when you are
10
using an angioplasty catheter for this purpose
11
include the following. First, contact between the
12
catheter lumen materials and the cell suspension
13
can potentially adversely affect either the
14
viability or the functionality of the infused
15
cells.
16 Additionally, those cells may contact
17
various lubricants that are commonly applied to the
18
guidewire lumens during manufacturing for the
19
purpose of facilitating guidewire passage.
20 As a note, we are not aware yet of
21
published reports that have examined whether
22
catheter lumen materials may adversely affect
23
viability or functionality of cells, however, this
24
published animal study cited in the slide evaluated
25 cardiac
delivery of a transgene by an adenovirus
298
1
vector.
2 The investigators found that some lumen
3
materials tested for use in a prototype needle
4
injection catheter, the second type of catheter we
5
will discuss today, adversely affected both the
6
viral activity and viral transduction.
7 They also found that a change in lumen
8
materials was sufficient to completely eliminate
9
these adverse side effects.
10 Additional considerations. Number two. A
11
second consideration when balloon angioplasty
12
catheters are used for this purpose, infusion of
13
cells, is that the balloon was originally designed
14
to stretch a coronary artery, in other words, to
15
controllably damage the artery, and for this use,
16
it must instead be used to occlude that artery,
17
hopefully, without damaging it.
This is
18
potentially important.
19 The degree of artery wall stretch that is
20
typically created during balloon angioplasty will
21
also subsequently induce arterial stenosis due to
22
multiple mechanisms. We think it
is therefore
23
essential that safe methods for balloon inflation
24
be developed and demonstrated if you want to use an
25
angioplasty catheter for this purpose.
299
1 Importantly, development of safe methods
2
for this new use could be complicated because
3
balloon angioplasty catheters have widely varying
4
pressure diameter relationships, in other words,
5
compliance can vary greatly between different
6
catheter models, so therefore the methods that are
7
developed for one catheter may not be applicable to
8
another one.
9 A third potential concern related to this
10
is the potential that concentrated cell suspensions
11
may clog the catheter lumen.
12 A fourth concern is that the lumens and
13
connectors of angioplasty catheters are primarily
14
designed for passage of a guidewire.
They may not
15
have been tested for the ability to sustain high
16
pressures that can occur during infusion of
17
concentrated cell suspensions.
18 The second type of catheter we will
19
discuss today are needle-tipped injection
20
catheters. This method of cell
delivery also
21
offers potential advantages. Notably,
first, the
22
ability to directly inject cells into desired
23
myocardial locations. Second,
the potential for
24
use with all types of cells.
25 The investigational cell delivery systems
300
1
developed for this therapy consists either of a
2
catheter or of a system comprised of a catheter
3
plus delivery sheaths, that include a retractable
4
distal injection needle. None
are approved
5
currently for sale in the U.S., however when we
6
consider these new devices, it may be useful to
7
note that some design requirements, and thus,
8
potentially catheter characteristics, may be
9
similar to requirements for other currently
10
marketed cardiac catheters, potentially including
11
both cardiac electrophysiology ablation catheters
12
and endocardial biopsy catheters.
13 More specifically, all three types of
14
catheters or catheters plus sheaths would generally
15
require a steerable or deflectable tip in order to
16
facilitate direction of the catheter tip to various
17
endocardial locations, and all three types of
18
catheters must be sufficiently stiff to permit the
19 user to maintain stable contact between the
20
catheter tip and the moving endocardial surface of
21
the ventricle.
22 These illustrations are from a publication
23
by Dr. Perin's group, and they illustrate one,
24
investigational needle-tipped injecting catheter.
25
The photo on the left illustrates a complete
301
1
catheter including controls on the catheter handles
2
that deflect the tip and that extend and retract
3
the needle.
4 The needle would normally be retracted
5
into the catheter except when briefly extended for
6
each injection that is made during a therapy
7
session, and the syringe, potentially loaded with
8
therapy suspension, is attached to an infusion
9
port, also on the catheter handle.
10 The drawing on the right illustrates
11
delivery of the catheter retrograde through the
12
aortic arch, then, through the mitral valve, and
13
into the left ventricular cavity.
In this drawing,
14
as you see, the catheter tip is deflected and the
15
catheter is being used to make multiple injections
16
from the endocardial surface of the ventricle.
17
There are also potential
concerns that
18
attach to the use of needle injection catheters for
19
delivery of cell therapies for cardiac disease.
20 First, we believe this type of catheter
21
may potentially be particularly prone to clogging
22
by cell suspensions. Factors
that might contribute
23
to this include the following:
24 The potential desirability of using very
25
small injection volumes plus highly concentrated
302
1
cell suspensions, the potential desirability of a
2
small diameter injection needle that will also,
3
thus, have a small diameter lumen, and the fact
4
that injections may be made at more than 20
5
locations during a treatment session, thus,
6
increasing the potential for clogging due to the
7
repeat injections.
8 A second concern.
As noted previously, it
9
may be important to ask whether cell viability or
10 functionality could be adversely affected by
11
contact with catheter lumen materials.
In this
12
type of catheter particularly, we think it may also
13
be important to ask whether shear forces produced
14
during infusion of cells through a long,
15
potentially a very small diameter injection lumen
16
might also adversely affect the cells.
17 Third, it may be important to consider the
18
safety, and this was brought up in one earlier
19
session, of whether the safety of the cell
20
suspension is accidentally delivered in the
21
systemic circulation.
22 Of note, with this type of catheter, it
23
may be very difficult to maintain continuous
24
contact between the catheter tip and the moving
25
endocardial surface of the beating heart, and when
303
1
contact is broken, therefore, you could have
2
injection into the ventricular cavity and into the
3
circulation.
4 There is a note, for folks in the
5
audience, I don't know if you can see the
6
reference, but if I understand correctly, we will
7
have cells on the web site, is that correct? Okay.
8
Otherwise, I was going to read the reference to
9
you.
10 One note, if people want to look into
11
this, this has been studied for cardiac ablation
12
catheterization electrophysiology.
I have noted
13
one good study up here where they used intracardiac
14
ultrasound imaging catheters to evaluate the
15
difficulty of maintaining continuous contact when
16
they thought contact was perfect.
17 A fourth consideration. Should we assume
18
that it is important to control or
limit the
19
maximum needle extension?
Factors to consider
20
might include the following: Is it critical to
21
avoid injection or laceration of the organs that
22
surround the heart?
23 It may also be important to consider
24
whether there may be safety concerns if the cell
25
suspension is inadvertently injected into the
304
1
pericardial or thoracic spaces, or if it is drained
2
from these spaces by the lymphatics and then
3
delivered into the systemic circulation. Relating
4
to needle injection, curves or bends in many
5
catheter designs, including the 180-degree bend
6
around the aortic arch that will normally be
7
present, could affect the needle extension length
8
of the catheter.
9 Finally, particularly in hearts that have
10
minimal epicardial fat surrounding the left
11
ventricle, it may be difficult to avoid occasional
12
injection completely through the wall of the
13
ventricle and into the pericardial space.
14 Factors that might contribute to this
15
could include the following:
16 First, is locally thin regions in the
17
ventricle, possibly related to myocardial
18
infarction, possibly related to the normal
19
indentations that separate the muscular trabeculae
20
of the ventricle on the endocardial surface of the
21
heart.
22 Second, is compression or stretching of
23
the ventricular wall where the catheter tip is
24
again pressed into contact with the wall, and,
25
finally, the possibility that a forceful injection
305
1
could simply potentially separate both myocardial
2
cells and epicardial cells, allowing cell
3
suspension to flush completely through the
4
ventricular wall.
5 A fifth and final question regarding
6
needle injection catheters is the following: Are
7
injection depth and the spread of injection of the
8
injected cells potentially important therapy
9
parameters?
10 For example, will injection of cells near
11
the more ischemic endocardial surface of the heart
12
provide therapy that is identical or equivalent to
13
injection near the less ischemic epicardial
14
surface, or is a minimally dispersed bolus of cells
15
at each injection site equivalent to wider
16
dispersion of cells at each injection site?
17 We currently suspect that catheter design,
18
cell suspension characteristics, and injection
19
speed can all affect injection depth and spread.
20 If a clinical study is performed using a
21
specific injection catheter and a specific cell
22
suspension, will the same therapy then be delivered
23
if a different injection catheter is used to
24
deliver that same cell therapy?
25 Of note, you can use animal studies, and
306
1
this is where the large animals become important,
2
to characterize the depth and spread of the cell
3
suspensions produced using a specific catheter.
4
Finally, and this is a
question I think
5
that unites all the questions listed above on this
6
slide, when an investigational therapy is studied,
7
how important is it that the therapy delivered be
8
characterized? When an investigational
therapy is
9
poorly understood, is the characterization of
10
therapy more or less important?
11 Today's meeting is focused on scientific
12
discussion of cell therapies for cardiac disease,
13
and because of both the focus of this meeting and
14
the time constraints, this is not a good forum for
15
discussion of regulatory concerns related to
16
cardiac catheters, however, if you would like to
17
discuss cardiac catheters intended for delivery of
18
cell therapies, you may contact either myself or my
19
branch chief, Mr. Elias Mallis.
20 I have listed our contact information on
21
this slide. Again, it will be
posted on the web
22
site.
23 Finally, because this is one of the final
24
presentations today, I have one final slide. My
25
manager has repeatedly asked the following
307
1
question: Whether there are
earlier or predicate
2
devices that have been used to provide catheter
3
delivery of biologicals.
4 He has also asked what we can learn from
5
any earlier devices. Initially,
I was unable to
6
define a useful predicate for catheter delivery of
7
biologicals. Then, I did find a
useful example,
8
and it was among photos from the Minnesota State
9
Fair.
10 The obvious lesson from this photo, I
11
would say is that intense concentration may be
12
required during catheter manipulation.
13 [Laughter.]
14 DR. JENSEN:
Thank you.
15 [Applause.]
16 DR. RAO: Thank
you, Dr. Jensen.
17 I am going to suggest that we wait and
18
hold off questions until we hear from Dr. Lederman,
19
as well, since he may be perhaps answering some of
20
those questions, and then direct questions to both
21
people at the end of that talk.
22
Transcatheter Myocardial Cell Delivery: Questions
23 and Considerations from the Trenches
24 DR. LEDERMAN:
I am going to be quick. I
25
am grateful for the opportunity to speak before
308
1
this audience and this committee.
Thank you for
2
your service.
3 I am speaking to you as a clinical
4
cardiologist, so I like to think of myself as Joe
5
six pack of clinicians. I will be talking about the
6
considerations, the frustration that many
7
investigators feel when we would like to talk with
8
the agency to get some guidance about how to start
9
bringing these interesting therapeutics to clinic
10
assuming we have determined that the timing is
11 right
to bring therapeutics to clinic.
12 I am sorry there will be a bit of
13
repetition. I will try to go
quickly through
14
repetitive slides.
15 We are dealing with integrated therapies
16
and unfortunately, we are also dealing with a
17
morass of regulatory purviews that don't
18
necessarily intersect. You have
seen already that
19
we have considerations of delivery devices, as well
20
as cellular agents, to combine in therapy, and we
21
haven't discussed much in this room combinations of
22
novel mobilization agents should we choose to use
23
that route to drive our cells, and it becomes
24
difficult when we have proof of concept in some
25
animal models to find an appropriate proof of
309
1
concept to support our clinical trial and to
2
support our safety considerations when some of our
3
colleagues outside the U.S. have kind of moved
4
ahead.
5 So, let's see if I can generate some
6
interesting questions for the committee, and that
7
is really what I hope to end on.
8 You have seen this slide from Strauer. I
9
will review again just the different approaches to
10
cell delivery, into coronary cell delivery is
11
attractive because it is very easy, there is a wide
12
dispersion into the target territory, and there are
13 a
lot of available devices to be used although they
14
must be used off label.
15 The disadvantages have been mentioned that
16
there is a potential coronary artery injury. One
17
of the real clever innovations by Strauer's group
18
and by Sawyer's group and the Hanover group is that
19
they have chosen to deliver cells through an
20
occlusive balloon deployed at the site of a
21
recently deployed stent, so there is really almost
22
no possibility of coronary injury from the delivery
23
device, however, there is a possibility of a
24
coronary microembolism, and when you test this is
25
the setting of a recent acute myocardial
310
1
infarction, that coronary embolism may be difficult
2 to
detect, so clinicians have gotten away with it,
3
or at least I should say their patients have gotten
4
away with it.
5 Certainly, there is a great potential for
6
direct washout of injected cells.
Very few people
7
have actually measured this or reported this, but
8
there is some evidence that there is a low
9
fractional retention of delivered cells, and really
10
this intracoronary cell delivery is yet another way
11
to expose the entire patient to the therapeutic
12
agent.
13 This kind of approach is probably
14
unsuitable to certain patient populations,
15
especially those with chronic myocardial ischemia
16
when the inflow arteries are occluded. Most
17
investigators have taken advantage of transient
18
coronary flow interruption ostensibly to improve
19
local retention, but it is just not clear the value
20
or importance of this transient coronary flow
21
interruption. Certainly, a lot
of patients can't
22
tolerate prolonged coronary flow interruption
23
without incremental myocardial injury.
24 There has been discussion about both
25
surgical and transcatheter cell injection.
311
1
Certainly, most of us will recognize that
2
catheter-based injection is less morbid than
3
surgical epicardial injection.
Primary surgery has
4
been unattractive in investigational studies when
5
the surgery is offered only for the sake of cell
6
delivery.
7 The problem is in small studies, combining
8
cell delivery with an effective therapy, I think
9
has been mentioned by several people before,
10
doesn't really generate meaningful safety or
11
efficacy data, because the assessment of toxicity
12
events is confounded by the concomitant surgical
13
procedure, and the assessment of efficacy events is
14
very easy to ascribe to the concomitant effect of
15 therapy.
16 So, I think this kind of approach should
17
probably be discouraged in small, single-center
18
studies. I hope some
investigators in the room
19
have already overcome this in moving to larger
20
studies.
21 Direct catheter injection of the
22
myocardium is attractive because we can achieve a
23
high local cell density and probably a high total
24
dose. It is certainly very easy
also, with the
25
variety of catheters that I will describe, and the
312
1
entire myocardium is for the most part accessible
2
irrespective of the patient's individual coronary
3
anatomy.
4 But these devices are disadvantageous in
5
the U.S. because there are no approved devices,
6
although a few are available through
7
investigational device exemption.
8 I will show you some data that there also
9
is low retention of injected cells and that direct
10
myocardial injection is yet another means of
11
systemic exposure of the patient to the cellular
12
agent. We are left even in the
best situation with
13
multifocal cell accumulation, meaning a
14
heterogeneous dispersion of the cellular agent.
15 Whether or not that is important isn't
16
clear, and there is the potential for damage to the
17
myocardium or to the chordal structures or the
18
valve structures, however, this potential has not
19
been supported by experience with comparably
20
aggressive or more aggressive intramyocardial
21
catheters especially in the fairly large laser
22
myocardial "revascularization" experience or
23
angiogenic gene transfer.
24
There are a bunch of
variants of
25
myocardial injection catheters.
There are
313
1
techniques to access the pericardium and bathe the
2
epicardial surfaces of the heart with the cell
3
preparation of interest.
4 Patients with chronic myocardial ischemia
5
who have undergone coronary bypass surgery are, for
6
the most part, not eligible, so this is a difficult
7
approach in early clinical trials.
8 There are investigators who have
9
demonstrated satisfactory delivery of genes and
10
cells by retrograde coronary venous approach. I
11
will show you a picture to show what that means. A
12
company has commercialized tangential transvenous
13
intramyocardial injection.
14 Then, there are a bunch of endocavitary
15
catheters that go across the aortic valve
16
retrograde and are pretty successful in delivering
17
cellular agents. There are an
array of what I call
18
"dumb" catheters that we use just under x-ray
19
guidance that are very attractive because they are
20
quick and easy. Two examples are
Boston Scientific
21
Stilletto and Biocardia Device.
22
There a couple of smarter
devices, it is
23
not clear that they are better, but they employ a
24
Static Roadmap like the Cordis Biosense
25
electromagnetic guidance system with some
314
1
non-fluoroscopic guidance. There
are also devices
2
that have an integrated ultrasound, and the
3
smartest devices, I will show you an example, in
4
research mode only, of instantaneous imaging both
5
of the tissue and the device.
6 We should all open our hearts to the
7
possibility that surgical videothoracoscopic, a
8
minimally invasive surgical procedure may
9
accomplish cell delivery with very little morbidity
10
even in a primary surgical procedure.
11 This is just a demonstration of one of the
12
so-called "dumb" catheters.
This is a Boston
13
Scientific catheter going from a femoral artery of
14 a
pig, across the aortic valve, and can
15
successfully guide whatever agent you want with
16
centimeter, not millimeter, precision to any aspect
17
of the endocardial surface.
18 Medtronic has recently bought
19
Transvascular, which is an interesting device that
20
has an integrated ultrasound to guide the
21
deployment of a needle through a coronary vein and
22
can access target myocardium through any of the
23
coronary veins in a tangential fashion.
This has
24
been tested in clinic in a very small number of
25
patients, and so far there haven't been any safety
315
1
disasters.
2 A retrograde transvenous approach has been
3
demonstrated I believe only in animals.
The
4
Stanford group, Keith Marsh's group, have been
5
interested in this for delivery of dyes or gene
6
agents or even some cellular agents.
I am sorry
7
the pictures aren't very attractive.
It is not
8
clear to me how this can possibly work, but the
9
proof of principle has been shown.
10 It is also attractive in that the inflow
11
coronary artery anatomy is not a problem since
12
coronary vein patency is maintained in virtually
13
all patients. Dr. Perin is an
expert in the use of
14
the Cordis electromagnetic guidance system, the
15
Biosense system that has its advantages in that it
16
has been widely used and tested for a variety of
17
investigational approaches.
18 It is disadvantageous in that it is a
19
prior roadmap of the heart that may vary over time,
20
and so it is not clear to me that you accomplish
21
millimeter scale precision of your injections, but
22
it is also not clear that that is very important.
23 I think Dr. Epstein's group also has great
24
expertise in the use of the device for cell
25
therapy.
316
1 Just to brag about some work in my lab for
2 a
moment, we have used real-time MRI to guide cell
3
injection to very small targets with great
4
precision and great ease. This is, as you see, a
5
multi-slice real-time movie of the heart. You can
6
see it in long axis and short axis, of a pig in
7
which we have caused a tiny, little
8
microinfarction, shown in white.
9 The catheter is shown in red and green,
10
and we can see with great 3-dimensional sense where
11
we are steering our catheter, and if we like, we
12
can label cells, say, mesenchymal stromal cells,
13
very easily.
14 Here, the mesenchymal stromal cells are
15
showing up in black, so you can actually see
16
interactively, as you deliver the cells, that at
17
least some are attained in the target myocardium,
18
and this is almost ready for clinical application.
19 I want to show a little bit of data. This
20
is from my youth when I was just out of fellowship,
21 a study that was funded by Boston Scientific
while
22 I
was at still at University of Michigan.
23 We injected neutron-activated microspheres
24
into the heart with an endocavitary catheter,
25
direct surgical approach or postmortem, and what
317
1
was interesting is that we inject and then kill the
2
animal within minutes, and most of what we inject,
3
both surgically or transcatheter approach, is lost
4
immediately.
5 There is some effect of volume. Smaller
6
volumes were associated with slightly greater
7
retention in tissue. This is 10 microliters. That
8
is a tiny, tiny injection compared with 20 or 100.
9 A better study was published by
Smits from
10
the Thoraxcenter group in Rotterdam using
11
scintigraphy and radiolabeled albumin, either plain
12
radiolabeled albumin or a colloid, and they showed
13
also loss of the majority of injectate after just a
14
minute by scintigraphy.
15 Their colloidal preparation had great
16
retention, which is interesting,
and that there
17
might, of course, be some interaction between
18
biological agents and the myocardial interstitium,
19
so conceivably, cells won't necessarily be lost.
20
These are studies that can be done in the lab in
21
healthy animals.
22 But I think it is easy to say that local
23
myocardial injection is at best an exaggeration,
24
that most injectate is lost rapidly and exits
25
either by backflow directly into the myocardial
318
1
cavity, which we can see directly, or with
2
intracardiac ultrasound or even with
3
high-resolution x-ray, that there is also a clear,
4
what I call "intravasation" or return to the
5
coronary circulation or coronary lymphatics.
6 When we inject too deeply and directly
7 into
the pericardium, that is another mode of exit,
8
but clearly, the interstitial myocardial target
9
retains only a fraction of what we intend to
10
deliver there.
11 So, where does this material go and is
12
that really important? I think
it is important at
13
least that we assume that what we think we are
14
injecting by any route goes everywhere, and I think
15
that means that conventional toxicology or
16
biodistribution experiments can be conducted in
17
uninfarcted animals without the needle of interest
18
just by modeling it as a left atrial or left
19
ventricular cavitary injection that is not device
20
specific.
21 I think also it is interesting to talk
22
about open- label autologous unfractionated bone
23
marrow data, what is the incremental value of
24
animal safety or tox experiments in light of the
25
fact that provisional safety has been shown in
319
1
open-label studies if we are convinced that the
2
safety reporting has been complete, and I am pretty
3
comfortable that it has been.
4 For autologous leukapheresis products, say
5
we apherese CD34 cells for a direct myocardial
6
injection, those cells are circulating already. It
7
is not clear to me what is the value of incremental
8
biodistribution experiments regarding systemic
9
exposure.
10 I will qualify that by saying that
11
allogeneic material perhaps should be treated
12
differently, but for autologous material, it is not
13
clear to me that these animal data are valuable,
14
and to require it of investigators before going to
15
clinic sounds dubious in my opinion.
16 Also, it is worth noting that the
17
importance of targeting is just not established.
18
While I am very interested in precise anatomic
19
targeting, it is not clear why we need it.
20 Delivery targets certainly vary by
21
application. We may want to
target infarct
22
borders. Doris Taylor had some
data today that
23
infarct borders may be unattractive for certain
24
therapies.
25 Do we want to target ischemic zones, or do
320
1
we want to avoid ischemic zones, do we want to
2
avoid thin myocardium? Is
roadmap data worse than
3
blind data or worse than instantaneous real-time
4 MRI data? It is just not clear.
5 Certainly, good targeting is attractive in
6
that it may reduce overlapping injections and waste
7
of injections, and overlapping injections may
8
increase systemic loss. It is
hard to imagine that
9
we might be exceeding some therapeutic index.
10 So, in other words, the value of targeting
11
is just not clear to me, and if we are able to show
12
some efficacy, it is not clear how much we must ask
13
of investigators to establish these catheter-based
14
information before going to clinic.
15 A point that has not been mentioned, but
16
that I have encountered in animal studies is that
17
operators need feedback regarding delivery of their
18
therapeutic agent, and I would like to encourage
19
the committee and the regulatory agency to consider
20
contrast labeling at the time of cell delivery. It
21
is certainly clearly more important than needle
22
stability measures.
23 There are lots of ways to label injection
24
mixtures. You can admix contrast
into your
25
injection cocktail. Iodinated
radiocontrast is
321
1
clearly tolerated in myocardium.
We inject
2
high-dose, full-strength intracoronary
3
radiocontrast, replacing blood inflow for many
4
seconds in patients in all settings, acute MI,
5
acute and chronic ischemia, and healthy myocardium.
6 It is certainly very well
tolerated. It
7
is hard to believe that iodinated radiocontrast
8
injected into the myocardial interstitium is not
9
tolerable. It is certainly well
tolerated in
10
animal experiments, but this kind of feedback under
11
x-ray guidance, for example, is critical in knowing
12
that we are delivering the cells into the target
13
that we think we are.
14 If we are doing injections under MRI,
15
then, certainly we can admix gadolinium-based MRI
16
contrast agents in very dilute form just like the
17
agent that reaches the myocardium after systemic
18
exposure, and these are very easy to be tested
19
biocompatible in vitro.
20 Certainly, we can do test injections of
21
contrast to test the purchase of our needle in the
22
myocardium before injecting the cell of interest,
23
but the problem with these test injections of
24
radiocontrast is that the catheter and hub dead
25
space often exceeds the volume of the desired cell
322
1
injection, so that is a problem that is difficult
2
to overcome.
3 Alternatively, we can label our cells.
4
That is certainly easy if we want to deliver our
5
cells under MR or under echo, but it is not clear
6
what options we have under x-ray, so to the members
7
of this committee, please facilitate solutions to
8
this clinical need.
9
Some engineering concerns
have been
10
mentioned in the very excellent guidance materials
11
supplied to members of the committee.
I just want
12
to speak to some of them.
13 The issues of biocompatibility of lumens
14 and potential clogging of lumens, this is
easy to
15
test on benchtop and doesn't require animal
16
experiments.
17 The issues of balloon injury of target
18
coronary arteries is an important one.
You have
19
seen creative solutions by investigators in Europe
20
by protecting the target coronary artery, inflating
21
their occlusion balloon inside a recently deployed
22
stent.
23 Certainly, there are noninjurious
24
compliant occlusion balloons that are clinically
25
approved for a variety of peripheral artery
323
1
applications, that are used widely in the cerebral
2
circulation, that are also used for coronary
3 protection
and substantial equivalence data are
4
already widely available.
5 We have also heard discussion of
6
considerations of the pressure capacity of balloon
7
wire lumens. The European
investigators, for
8
example, are administering their cells via the wire
9
lumen of an inflated coronary balloon.
10 As a practicing interventional
11
cardiologist, I don't consider that an important
12
concern because every balloon, every over-the-wire
13
design balloon that I put into a patient, I expect
14
to use for intracoronary angiography, and I use for
15
intracoronary angiography with a fairly high
16
pressure system.
17 There are many times I need to know that
18 my
balloon is in the right place, so I pull out the
19
wire and I inject contrast at a fairly high
20
pressure directly through the balloon lumen.
21 That is not, of course, an indicated use,
22
but it is a wide use by all operators of coronary
23
artery balloons, and I think the test of time has
24
already been past, but if you like simple benchtop
25
pressure data, they are easy to acquire.
324
1
Regarding endomyocardial
injection
2
catheter engineering concerns, the same
3
biocompatibility and clogging concerns are easy to
4
answer on benchtop tests. This
issue of variable
5
needle extension is probably an important one if
6
injection depth proves to be important, and it is
7
not clear that it is, but this can be addressed in
8
benchtop testing.
9 It has been mentioned that purchase
10
stability is important to assure injectate reaching
11
the target tissue. My assertion
about marking or
12
labeling injection cocktails with contrast might
13
address that concern.
14 The report from UC/SF from Jonathan
15
Coleman using an old intracardiac echo device
16 reporting
instability is actually a spurious
17
observation because of through-plane motion of the
18
target that the UC/SF group was inspecting with a
19
fixed vena caval or a right atrial intracardiac
20
echo device. In other words, I
don't think it has
21
been shown that the contact of EP catheters or
22
myocardial injection catheters cannot be
23
maintained, in fact, just the opposite, especially
24
from the Biosense device which has a local cardiac
25
electrogram capability to assure contact stability.
325
1 So, I think this is not really an
2
important problem for us to worry about.
3 The issue of potential myocardial
4
perforation is often raised when we discuss the
5
possibility of delivering cells directly into
6
patients after a recent large myocardial
7
infarction, and I think that is an important one.
8
It is interesting that you refer to myocardial
9
biopsy devices as predicate devices, because as I
10
view the biotomes as some of the most dangerous
11
devices we ever laid hands on, they are so
12
incredibly stiff and indeed perforations do
13
sometimes occur.
14 Fortunately, there is an animal
experience
15
from my lab. I guess we should
probably get it out
16
there, of a large number of injections directly
17
into freshly infarcted myocardium, and I think this
18
kind of data is easy to obtain.
19 But the bigger issue is not that freshly
20
infarcted myocardium can be safety injected, it is
21
that the device companies can't really control the
22
operators. I have seen this so
many times. An
23
engineer walks into a lab and cringes as the
24
interventional cardiologist effectively abuses the
25
device. How do you model
operator misbehavior? It
326
1
is kind of difficult.
2 In reality, proof of principle has
been
3
established.
4 The issue of inadvertent pericardial
5
injection probably has little or no clinical
6
importance especially when compared with the loss
7
of injectate via other routes, and its only value
8
is that you are not delivering what you think
9
directly into target tissue again, the value of
10
instantaneous visualization of injections.
11 The issue of distribution of injected
12 material
within the target myocardium, I think it
13
may be reasonable to assume that this distribution
14
is different in normal myocardium versus fresh
15
infarct versus chronic scar, but the value of these
16
data are just not clear compared with the efficacy
17
data in support of preclinical or early clinical
18
experiments, so having this information of how many
19
cubic centimeters of myocardium are exposed to
20
target cell based on a given volume or dose of
21
cells, it is just not clear why we need that
22
information. This kind of
information ultimately
23
can only be valuable in patients.
24 Are endomyocardial injection catheters
25
generic? In my opinion, assuming
benchtop
327
1
biocompatibility has been determined, and assuming
2
that mechanical performances are satisfactory
3
compared with predicate devices, I think that a
4
myocardial injection catheter is pretty much the
5
same from one to another.
6 One needle device should be translatable
7
to another, and the scientific and regulatory value
8
of additional data from large mammals, healthy
9
ischemic infarcted, is really pretty small and hard
10
to justify, in my opinion. So, I
keep giving this
11
message, nihilistic message about the large animal
12
models investigators have been asked to provide.
13 To summarize my opinions, I think
14
engineering and biocompatibility concerns can be
15
addressed with benchtop data. I
think that animal
16
model safety experiments matching a given catheter
17
device with a given putative therapeutic agent
18
don't meaningfully contribute to patient safety and
19
are, in fact, potentially misleading.
20 I wish that there were a way to get a
21
screening IDE or IND capability to support testing
22
new cell preparations without repeating unnecessary
23
preclinical experiments as we switch from device to
24
device, and ultimately, careful human
25
experimentation is what is most important.
328
1 Let me just make a few more points before
2 I
turn to some questions and step off the podium,
3
so people can go home.
4 I want to reiterate some points made by
5
other speakers today, and I want to reiterate it
6
especially to the regulatory officers here. I
7
think that blinded placebo groups are mandatory
8
even in first experiments.
9 Why would you conduct an experiment
10
without a suitable matched control in the name of
11
safety? That is just bad
science. Interestingly,
12
there has not been a single open-label or "Phase I"
13
safety trial that fails to make an efficacy claim
14
without a suitable matched control.
15 Unfortunately, the agency is inadvertently
16
discouraging blinded controls, for example, when
17
they ask for a delay in between exposing a given
18
subject within a given group of patients, asking
19
for seven days or four-week delay between patients
20
to look for safety of individual patients. This
21
often frightens investigators away and makes them
22
drop placebo groups.
23 I think that in cardiology, we rarely
24
conduct classic Phase I studies in end-stage
25
subjects in spite of the conversation in Doris
329
1
Taylor's speech.
2 So, I would like to encourage people in
3
the committee and encourage regulatory agencies to
4
facilitate inclusion of blinded placebos in early,
5
first in man even, experiments.
6 I was also asked to talk a little bit
7
about the safety of direct myocardial injection,
8
and unfortunately, there are no large series of
9
direct myocardial injection of cells or any other
10
agents, however, a related catheter has been tested
11
in a few hundred patients.
12 Cordis had a myocardial--they called it a
13
DMR, direct myocardial revascularization procedure,
14
but it was a way to burn the myocardium from a
15
transcatheter approach. In this
Cordis-sponsored
16
study presented by Martin Leon and Ron Kornowski a
17
few years ago, a Cordis Biosense derivative, a
18
device much like the myocardial injection catheter
19 shown
today, was used to steer into the myocardium
20
of 300 patients with refractory ischemia, mild or
21
moderate left ventricular dysfunction, and preserve
22
wall thickness.
23 One hundred patients underwent sham
24
procedures, placebo burns of the myocardium, and
25
200 more received laser in two different doses.
330
1
This is just the clinical complications. I will
2
point you to left ventricular perforation. There
3
were none in the 100 placebo patients, there was 1
4
out of 100 in the highest laser dose, which is
5
comparable to some other laser trials.
6 I think this establishes a relative safety
7
base from a perforation perspective.
These other
8
events unfortunately weren't very well described,
9
and this study unfortunately has not been
10
published, and I am not sure it ever will be
11
submitted for publication, but the acute safety of
12
this device, I think is relatively self-evident.
13 So, is placebo and the myocardial
14
injection safe in principle? I
think yes, and it
15
is not a reason to discourage these placebo groups
16
in early first clinical studies.
17 So, again, we are trying to bring
18
therapies to clinical testing, and we are trapped
19
between delivery devices and cellular agents that
20
we would like to use together.
21 I want to just ask a few hypothetical case
22
questions to the committee before I step down. I
23
certainly don't have answers, but I hope you find
24
it provocative.
25 Let us say that an investigator identifies
331
1 a
novel marker, an HL321, let's call it, that
2
identifies some kind of progenitor cell for which
3
there is no clear animal homolog.
4 There are some limited preclinical
5
efficacy that when an enriched human HL321
6
population is injected to intrinsically
7
immunocompromised rat infarct, it causes functional
8
recovery compared with a population of known cells
9
that are relatively depleted, so this is not the
10
most robust type of experiment, but based on
11
experiments like this, and based on patients who
12
are clamoring for therapy of their massive
13
myocardial infarction, investigators may want to
14
bring some test of this therapy to clinic even now.
15 How could we test local autologous HL321
16
cells assuming we had a feasible way to mobilize
17
and recover these cells? How
could we test that
18
for safety and efficacy?
19 In the example I want to give you, there
20
already is a commercial cell system available,
21
marked with a CE in Europe, and in Europe, hundreds
22
of patients have successfully undergone bone marrow
23
transplantation with a population positive
24
selection for this marker, and also in Europe,
25
dozens of patient underwent local cardiac delivery
332
1
trials in a variety of applications, and Phase II
2
trials have been offered.
3 Is it unreasonable to permit U.S.
4
investigators to conduct similar experiments now
5
that provisional safety has been tested in Europe,
6
if incompletely reported in Europe?
7 For this kind of clinical experiment, are
8
additional animal data really necessary when human
9
studies have already been conducted, and when there
10
is no animal homolog of that positive selection
11
marker? What animal model is
really adequate?
12 Can we use the experience of investigators
13
like Dr. Epstein, investigators like Dr. Perin,
14
using undifferentiated bone marrow to support the
15
local delivery of other autologous cells, and are
16
individual cell preps, autologous cell preps
17
substantially equivalent in this case when they
18
have just been derived directly from the patient
19
irrespective of sources? Is bone
marrow that much
20
different from apheresis product, from a
21
mobilization product?
22
So, how can we apply
non-U.S.A. human
23
safety data to support U.S. clinical trial
24
proposals, or should we continue the way we are now
25
and just sit and wait for others to do their
333
1
experiments without us?
2 Thank you very much for your attention,
3
and I hope you found these questions provocative.
4 [Applause.]
5 DR. RAO: We
have time for a few
6
questions.
7 Q&A
8 DR. NOGUCHI: I
think that was a terrific
9
representation of the tension between belief and
10
what is published or not published, and what we
11
might call the paradigm for FDA, which is absence
12
of evidence is not evidence of absence.
13 I will just challenge you a little bit.
14
When we see safety data whether
it is from
15
another country or here, one of the critical
16
questions for an adverse event is, well, is that
17
really showing that if you give a product that you
18
have a lack of an adverse event, or could it be
19
that, in fact, you didn't give a product, which by
20
your own arguments you would say most of the time,
21 you lose most of the product all over the place.
22 So, some of it falls into the category of
23
our experience is that adverse events actually
24
other than from the actual injections of others for
25
biological products don't occur unless you actually
334
1
also have bioactivity, and sometimes if you don't
2
have any cells, you may not have any bioactivity.
3 So, I think that you have a number of very
4
genuine points very well worth arguing, but I would
5
just caution that it is very simple to say
6
something is safe. We rarely say
look at all the
7
published data and it's safe, because we don't
8
really know if a product was actually being used
9
there, and perhaps that is one of the points you
10
might want to just think about.
11 DR. LEDERMAN:
I, of course, can't even
12
answer your question, and I want to point out I
13
want to thank the regulatory officers and the FDA
14
for trying to protect our patients, and to try to
15
protect the American public. You
are in a very
16
difficult position that is often a thankless
17
position, but I just hope you are open to this kind
18
of conversation.
19 DR. NOGUCHI:
Absolutely. That is why we
20
would like to have you end it here, because I am
21
sure it will provide the focus of discussion for
22
tomorrow.
23 DR. RAO: Any
other questions?
24 DR. EPSTEIN:
Bob, I really enjoyed that
25
presentation, it was really great.
I would just
335
1
like to raise one point. For
example, to my
2 knowledge,
no one has really tested in depth the
3
safety of the transvascular administration of an
4
angiogenic agent or its cells.
5 The reason I feel that might be different,
6
for example, than a transendocardial or a
7
transepicardial injection is because you are
8
injecting it right around a large artery. It is
9
conceivable that there can be pro-atherosclerotic
10
or pro-restenotic effects.
11 DR. LEDERMAN:
I am sorry. Let me
12
interrupt your question. Do you
mean intracoronary
13
approach or the Medtronic transvascular approach?
14 DR. EPSTEIN:
The Medtronic approach, so
15
where you are injecting it, not downstream at the
16
small vessel level, but at the large vessel level,
17
so if you are injecting it, for example, through
18
the venous system, it is contiguous with the
19
arteries that have atherosclerosis in it.
20 So, I would think that for that special
21
case, the FDA would require that you have to show
22
that there is no deleterious effect in terms of a
23
pro-atherosclerotic effect. I would be interested
24
in your thoughts about that.
25 DR. LEDERMAN:
I think that point is well
336
1
taken and every such safety request or demand is
2
interesting and valuable, but how do we answer that
3
kind of question satisfactorily.
4 Let's take the question you just asked, a
5
tangential myocardial needle or perhaps a
6
retrograde venous administration of agent X, and
7
the problem of an unrecognized atherogenic effect,
8
how on earth do we test that?
Are apoE knockout
9
mouse experiments satisfactory?
10 DR. EPSTEIN:
For this particular, I would
11
injure a vessel in a pig, and then inject whatever
12
agent you are interested in transvenously in the
13
area of that injured vessel and just see whether
14
there is an increase in the neointimal response.
15 I don't know how you carry--I mean your
16
question would be so what, whatever you see, and
17
that would be a good question.
18 DR. LEDERMAN:
But that is exactly right,
19
that might reassure us, but that is also not
20
atherosclerosis.
21 DR. EPSTEIN:
That's right, but the AMI
22
studies, you know, you are doing angioplasty, so if
23
you were to increase the incidence of restenosis in
24
the pig, you know, quite predictively, it would
25
certain add a major cautionary note to approval of
337
1
such a protocol.
2 DR. TAYLOR: I
would ask two questions.
3
One, very short, but one is you said that you would
4
argue that exogenous delivery of any given cell
5
population is equivalent essentially, but I would
6
think that the GCSF, or if you have given one bone
7
marrow cell population, is it right to go ahead
8
with all the others without necessarily more safety
9
data?
10 I would argue that the GCSF data that just
11
came out would actually argue the converse, that
12
the only difference there was mobilization of cells
13
that would otherwise be endogenous to that same
14
patient, and yet an increased number of those cells
15
clearly caused some negative effect.
16 DR. LEDERMAN:
I wasn't actually making
17
that assertion. I was making the
assertion that
18
needle injection catheters are ultimately very
19
similar.
20 DR. KURTZBERG:
But the difference, there
21
have been, I don't know how many tens of thousands
22
of patients have had GCSF, we have gotten bone
23
marrow in their right atrium, so I mean the
24
dissemination of those cells is not the issue, it
25
is combining that with a local technical injection
338
1
and trauma to that site that is different.
2 I mean there is experience with these
3
cells disseminated through the human body for two
4
decades, so that is not the issue.
The issue is
5
what do the cells do in the setting of a local
6
technical injection into an artery or other part of
7
the heart that is sick.
8 DR. TAYLOR: I
think that is sort of the
9
point I am trying to make, that exogenous delivery
10
is not necessarily the same thing as mobilization
11
of cells, and that having more cells there, that we
12
don't understand, we can't just interpolate from
13
other data.
14 I want to ask a very short question. What
15
do you think about clinicians moving forward who
16
don't have experience with preclinical studies? I
17
mean one of the things that probably enabled
18
Philippe to do the studies he did is he had that
19
six years of preclinical experience making mistakes
20 or whatever.
21 What do you think about any clinician
22
moving forward in a trial without having previous
23
preclinical experience?
24 DR. LEDERMAN:
My short answer is who
25
cares what I think, and we all operate as parts of
339
1
teams with expertise in our respective areas. In
2
cell therapy, for example, it would be outrageous
3
for me to do an early clinical study without the
4 close collaboration of
cell therapy experts like
5
some that are fortunately in the room.
6 And would we need to have our local
7
on-site preclinical experiments?
It is not clear
8
to me how important that is. It
is more important
9
to me that our agents be well characterized, that
10
the studies be well conducted, and that they be
11
designed in a way that the data can be interpreted
12
rather than open-label, early clinical experiments
13
that are very difficult to interpret.
14 DR. RAO: I
really agree with on the
15
emphasis you made about the fact that you should
16
have a placebo-controlled trial, but then on the
17
same token, you know, you also said that one
18
catheter is much like the other, but I don't think
19
that we can extrapolate that from saying one is
20
much like the other, you know.
21 I mean we worry about drugs when we say
22
whether it is a generic formulation or whether it
23
is a formulation which contains the same active
24
ingredient, and to me, when you are looking at the
25
device, you are making it with cells, you have to
340
1
worry about how clearly or how similar the device
2
is to any other device.
3 We can't simply say, well, you know, the
4
benchtop pressure was the same.
You know, I can
5
take syringes and I can show you that the benchtop
6
pressure on that cell agent injection is exactly
7
the same because of how I do it, but, you know, I
8
can put cells through it, and I can guarantee you
9
that there would be a difference.
10 I mean Dr. Menasch showed in his data you
11
use a 27-gauge needle, and it is very different
12
from using a 29-gauge needle. It
doesn't matter
13
whether you have got the same pressure or not.
14 So, I think that it would be hard put at
15
least for me to be convinced that most catheters,
16
even if they are giving delivery externally in much
17
the same way, that one can logically extend it and
18
say that it will probably be the same.
19 DR. LEDERMAN:
So you are telling me--and
20 I
don't mean to belabor the point--but if you have
21
two catheters by two different vendors, that have
22
satisfactory benchtop testing for biocompatibility,
23
have satisfactory hydraulic characteristics, that
24
what you inject at one end comes out the other end,
25
and one such catheter has satisfactory efficacy
341
1
data in some kind of preclinical model, that you
2
would require a repetition of that preclinical
3
model for another catheter that is virtually the
4
same?
5 DR. RAO: I am
saying that right now we
6
can't make the assumption that it will be the same,
7
and the reason I say that is that we know that when
8 we
make minor manipulations to cells which we are
9
delivering, for example, if we
take CD34 cells,
10
which have been kept in culture for 48 hours as
11
opposed to 12 hours, we have a very different
12
endpoint result. We know that.
13 We don't know what the interaction will be
14
with the catheter, and we can't make the assumption
15
that because we know five parameters, that those
16
will be adequate in making a reasonable prediction,
17
so until I have a lot more data, I will be very
18
surprised that one could make that statement or
19
anybody would agree that that is okey.
20 DR. LEDERMAN:
The end result is that we
21
have an unmanageable number of permutations, an
22
unmanageable number of permutations that makes it
23
hard to make progress.
24 DR. RAO: But
again I think this is to
25
reemphasize what Phil said, it is not that we can't
342
1
do it, so that means it shouldn't be done. It is
2
to try and identify what is critical, so that you
3
make sure that the critical points are done, so
4
that is the critical issue to me.
5 DR. RIEVES:
Dr. Lederman's presentation
6
was excellent, and I think it raised some excellent
7
points. I think the important
part will be to
8
discuss them tomorrow, and can give one example,
9
because for every point that was raised in that
10
discussion, there is always the other hand. It is
11
like the two-armed economist. I will give you one
12
example right now.
13 It is true that one study with what is a
14
laser TMR system, there were very few perforations.
15
There was another blinded, randomized study,
16
completed in the U.S., published in JACC, 140
17
patients. Only the treated
patients, 70 treated
18
patients actually were catheterized.
19 Now, in that study, 5 of them had
20
perforations, so it is often difficult to--just as
21
one example, there is always another side to this,
22
and the important thing is I think you have raised
23
some excellent points. We have
left the tough
24
questions for our committee members to address.
25 DR. SIMONS: To
come back to the point
343
1
that catheters are different, not only are they
2
different from each other, they are different from
3 the cells depending on the cell type used. I
4
absolutely do not think there could be universal
5
device.
6 DR. LEDERMAN:
And these questions were
7
not answered in benchtop testing?
8 DR. SIMONS:
No.
9 DR. RAO: I
think your point is well made,
10
though, that I think the way devices need to be
11
regulated is somewhat different from cells, because
12
of the number of variables one might have to
13
consider are somewhat different, and I think that
14
is a very valid point.
15 If there is no more questions, we will go
16
to the open part of the question and answer
17
session.
18 Open Public Hearing
19 Before we can have the open public
20
hearing, by law, I am required to read a statement.
21 I
will do that right now.
22 Both the Food and Drug Administration and
23
the public believe in a transparent process for
24
information gathering and decisionmaking. To
25
ensure such transparency at the open public hearing
344
1
session of the advisory committee meeting, FDA
2
believes that it is important to understand the
3
context of an individual's presentation.
4 For this reason, FDA encourages you, the
5
open public hearing speaker, at the beginning of
6
your written or oral statement, to advise the
7
committee of any financial relationship that you
8
may have with any company or any group that is
9
likely to be impacted by the topic of this meeting.
10 For example, the financial information may
11
include the company's or a group's payment of your
12
travel, lodging, or other expenses in connection
13
with your attendance at the meeting.
Likewise, FDA
14
encourages you at the beginning of your statement
15
to advise the committee if you do not have any such
16
financial relationship.
17 If you choose not to address this issue of
18
financial relationships at the beginning of your
19
statement, it will not preclude you from speaking.
20 There were two people who had asked to be
21
recognized before the meeting started.
The first
22
person is Dr. Vulliet.
23 DR. VULLIET:
Thank you for the
24
opportunity to come and present some data to you.
25
Am I supposed to make a statement I have no
345
1
financial interest in this?
2 DR. RAO: Yes.
3 DR. VULLIET:
Okay. I have no financial
4
interest in this.
5 This is an example of some studies that
6
were done very recently where we have been
7
investigating, using my research team, which is
8
myself, a cell biologist/pharmacologist, Dr.
9
Greeley is a pathologist, Mitch Halloran is a cell
10
biologist, Kristin McDonald and Mark Kittelson are
11
both board-certified cardiologists, so we have a
12
very interdisciplinary team.
13 We are at a vet school, which is probably
14
novel for this group, and we are very specifically
15
interested in animal models. I was
very pleased to
16
hear quite the discussion of animal models. I
17
disagree with almost everything every one of the
18
speakers complained about, not being suitable
19
animal models.
20 I guarantee we see animal models that
21
definitely are real patients, that have real
22
disease. It is not induced, it
is a real disease.
23
It is there, it needs to be treated.
For that
24
reason, we have decided to investigate the
25
possibility of using cytotherapeutics to see if we
346
1
can produce a beneficial effect in animals.
2 A good example of our animal model--and
3
this slide is probably Peter's example of what they
4
think we do with animals--but this is Oscar, and I
5
guarantee you he will grow up to have somewhere
6
later in life, lumbar disease, lumbar disk disease,
7
either at the L2/3 or the L3/4, and he is a great
8
model if you are into disk disease, but that is a
9
different committee we talk to about that.
10 Steps for successful cytotherapeutics.
11
This is my perception, not the committee's
12
guideline, the first step is safety studies. This
13
is about where we are at. In
fact, most of the
14
stuff we are squabbling about right now is whether
15
these things can be done safely or not, if you
16
think about it.
17 Very little good data on dose response,
18
nothing on time course that I am aware of. Nothing
19
or very little on clinical endpoint.
What do I
20
mean by clinical endpoint? When
I am giving an
21
antibiotic, I can tell you I need to hit serum
22
concentrations of 1 microgram per ml.
23
Okay. I can design a pharmacokinetic
24
regimen, I can hit 1 microgram per ml.
If I am
25
treating dilated cardiomyopathy, what is my
347
1
endpoint? If I see something
four to six weeks
2
later, I will be lucky.
3 So, when I am administering cells, I
4
really don't have a defined clinical endpoint at
5
this point other than the lack of adverse reaction.
6
Think about it. It is a very
interesting point of
7
view.
8 Anyway, because we started off with safety
9
studies, that is what we did. We
asked a very
10
simple hypothesis, and we started with can a half a
11
million cells--we are using mesenchymal stromal
12
cells, we call them stromal cells rather than stem
13
cells because we are not convinced that primordial
14
germ layers have been demonstrated coming from MSC,
15
so we refer to them as stromal cells, you can call
16
them stem cells if you like.
17 We also use a terminology I don't think
18
anybody else has used in this room yet, is we are
19
using MICs [ph]. That is a million cells per
20
kilogram. We are very interested
in a
21
dose-response relationship.
Doses are a key in any
22
therapeutics.
23 So, we are giving half a million MSCs per
24
kilogram of body weight, and can they be safely
25
injected into the coronary arteries of the
348
1
anesthetized dog?
2 Simple experimental design, they are
3
autologous. We collect bone
marrow somewhere about
4 a
month later, we inject 10 million cells.
These
5
are 20 kilogram dogs, half-million cells per
6
kilogram.
7 Seven days later, the dogs were
8
anesthetized, physical exam, CV exam, necropsy and
9
histo and immunocytochemistry.
10 I should also point out everywhere in this
11 study,
all of these dogs, after recovering from
12
anesthesia, passed the cold nose test.
They were
13
perfectly normal, you would not be able to tell.
14
They jumped, their tails wagged, they licked your
15
hand. They were nice dogs.
16 So, in that regard, on physical exam, they
17
looked good.
18 Abbreviated methods.
Autologous, four
19
sites, collect them. Freshly
dispersed, and this
20
will be a key point at the end.
These are
21
autologous cells, freshly dispersed MSCs into the
22
circumflex artery. Catheter
placement verified
23
before injection, after injection, with
24
fluoroscopy, physical exam. All
dogs basically
25
appeared to be normal once the effects of
349
1
anesthesia had worn away.
2 This is the first dog we did. This is the
3
highest dose we did, and we did 1 million cells per
4
kilogram, injected in the coronary.
At 2 hours,
5
took a section, you can see the catheter placement
6
there, took a section of ventricular myocardium.
7
It's a lightly stained hematoxylin section. These
8 2
cells, basophilic cells here, translate to the
9
CMFDA-labeled cells or fluorescent cells there.
10 This is one of the things we are looking
11
for. Our research team, probably
different than
12
many of the people in this room, feel that the only
13
way we will get effective therapeutics is to have
14
intimate contact between these cells and the dead
15
and dying cardiomyocytes that they are going to
16
replace. We don't believe in
direct injection, and
17
we can talk about that later.
18 So, that was our first dog, 1 million
19
cells per kilogram. The reason
he went--it was at
20
two hours--was he went into V fib and died. Two
21
hours, so it is fairly easy to do your necropsy in
22
that. So, we then bacted those
off to a
23
half-million cells per kilogram and injected them,
24
and the injections are 5 injections of 2 ml each
25
and 1 million cells per ml, and this was the
350
1
control, the anesthetized dog, this is the ECG.
2 Dogs normally have inverted T waves, I
3
don't know how many people know that, and it has to
4
do with chest dimensions and chest geometry more
5
than anything else. This is a
normal ECG for a
6
dog.
7 After the fifth injection, the T wave has
8
converted to a normal position, but more
9
importantly, what you see here is ST elevation,
10
very profound ST elevation. This increase with each
11
dose and at the fifth injection it was the most
12
severe.
13 Twenty-four later, post-injection, you can
14
see you have got back to an inverted T wave, but
15
you have got bizarre complexes here.
One-week
16
post-injection you have got normal ECG, and this
17
was published recently in Lancet, so I am not going
18
to go through a lot of this published thing two
19
weeks ago.
20 ST elevation made us think of troponin.
21
We measured troponin at various times after
22
injection. You can see it goes
up, increases to
23
about 45 nanograms per ml. If
you ligate an LAD in
24 a
dog, you normally see ranges in the order of 150
25
to 200, so we had subnormal, if you will, levels of
351
1
troponin, but the time course it would be
2
consistent with some sort of myocardial ischemia.
3 One of the things you also don't see in
4
many of these preparations that you are seeing,
5
which is my personal--I will give you guys my
6
personal things that sort of irk me a little bit
7
about science--is we are using H&E histology.
8 This is the gold standard of pathology. I
9
don't care how many immunocytochemistry studies you
10
see, H&E is what medical pathologists use to
11
evaluate an outcome of a case.
It is essential, it
12
should be included, I don't know why people don't
13
use it, but it is a great technique, I like it
14
although I didn't do very well in pathology.
15 Here is a good example of a section of the
16
ventricle of one of the injected dogs.
What you
17
see is three areas here of hypercellularity. This
18
is the normal myocardium here as you zoom in. This
19
is at 4X or 40X, this is 100X and 400X, and we are
20
zooming in on this area right in here.
21 You see it is more hypercellular. What
22
you see in here is you see mononuclear cells,
23
rounded nucleus. You see some
elongated nucleus.
24
You see fibrosis. You see some
lytic lesions. I
25
looked at that as a cell biologist, and I said,
352
1
great, that's where my stem cells are, right?
2
Wrong. My pathologist looked at it and said, "Rick,
3
you have got a problem. Those
are macrophages.
4
You have just produced a heart attack in this dog."
5
And I said, "Oh."
6 So, to verify that, these were CM dye/I
7
labeled cells zooming in. You
can see that the CM
8
dye/I label is in the vicinity of this
9
hypercellularity, so both the MSCs and these
10
macrophages are here, but how do we know they are
11
macrophages?
12
Again, using
canine-specific antibodies
13
that we have available in the vet teaching
14
hospital, this is an H&E section, this is a CD18
15
monoclonal antibody specifically raised against
16
canine macrophages. You can see
you have punctate
17
lesions, it is very characteristic, and, indeed, to
18
confirm this, 7 days later, what we also see
19
characteristic of myocardial infarctions is
20
increased fibrosis and collagen deposition.
21 As you can see here, this is normal
22
myocardium, it stains very red in Masson's, and
23
what you see here is you see the blue here is
24
collagen fiber deposition. This
is very
25
characteristic of myocardial infarction.
353
1 Five cardinal signs of MI are ECG changes,
2
we showed that; proteins released from damaged
3
myocardium, we showed that; decreased wall motion,
4
we did not see that. We did
ultrasound, but I
5
don't know that we could. We
have the
6
sophistication to measure wall motion.
Myocardial
7
infarction is just not a common veterinary disease
8
that we see. Characteristic
cellular infiltrates,
9
we saw that. Collagen
deposition.
10 Basically, our conclusion is that at 0.5
11
million cells per MSC, will produce myocardial
12
ischemia, microinfarctions in these dogs.
13 Our original interpretation of this was
14
that this was a dose/rate of delivery problem. So,
15
our feeling was, because there really is not much,
16
is you critically look at the clinical studies out
17
there in humans, they have got doses all over the
18
place. It is very hard to
extrapolate what the
19
dose is. That is why use
milligrams per kilogram.
20 I
would encourage anybody doing these studies to
21
use some sort of a normalization like that.
22 I would encourage the committee to require
23
it, so that you can start comparing, but more
24
importantly, what we did, just to give you an in
25
idea, is possibly post-injection cell clumping. We
354
1
didn't consider this as a possibility.
We have
2
taken the holding media the cells were in, and the
3
injection media, to inject them into the cells.
4 At the end of 2 hours of holding them in
5
the holding media, I personally inspected most of
6
the dogs. We inspected the
cells. None of the
7
cells are clumped, so we assumed clumping was not a
8
problem.
9 Just to give you guys a little bit more
10
gray hair in terms of your job as far as making
11
decisions, what we didn't do is we didn't do the
12
right control, and the right control was to put
13
these cells in 100 percent serum, because that is
14
what you are doing, you are injecting them into the
15
artery. From the time they leave
that artery, they
16
are going into 100 percent serum.
17 We did that and the cells started
18
clumping. I don't know why. They didn't clump
19
when we pulled them out of the bone marrow because
20
they would have been clumped when we put them in
21
tissue culture dish.
22 They didn't clump then, but during the
23
process to 2 weeks of preparing
them, sticking to
24
plastic, their adhesive properties had changed.
25
One dog, and I have heard several people talk about
355
1
fetal bovine serum, I don't know if it is known to
2
this group or not, Darlen Procoff [ph] last year
3
had an abstract, I don't think he has published it
4
yet, had an abstract where he looked at fetal
5
bovine serum in cells. It
carries over, I believe
6
it was about a milligram per million cells if you
7
grow cells in FBS, at least with the MSCs, will
8
pinocytose and hold about a milligram per million
9
cells of fetal bovine serum.
10 It is not released by washing. These
11
cells were rinsed three times in Hanks' balanced
12
salt before they were injected.
If you want to
13
remove the fetal bovine serum, you have to grow
14
them for at least 48 hours to get rid of it.
15 This may explain some of the early dieoff
16
of the myoblasts that you are seeing, if you have
17
got FBS in there, because you will get a reaction
18
from it.
19 Basically, this research team does not
20
feel that we are in a position to perform these
21
studies on client-owned patients at this time,
22
although that is our long-term goal.
23 What do we think is happening? What we
24
think is happening is, we think the cells are
25
coming in, we think they are clogging, producing
356
1
microinfarctions by clogging either a second or
2
third order arterial, and causing areas of ischemia
3
and microinfarctions.
4 Are there clinical techniques to detect
5
this? That is one of those
things. Would you see
6
that, or does all the old imaging, does the modern
7 world ultrasound pick up something like that? I
8
don't think so. So, that is why
I am saying
9
histology is pretty important.
10 How do our cell preparations compare?
11 DR. RAO:
Dr. Vulliet, we would like try
12
and make sure that we stay on time.
13 DR. VULLIET:
Sure. Let me finish this
14
because that is actually what I think we are more
15
interested in.
16 This is bone marrow from canine bone
17
marrow. As you can see, you have
got metas and
18
milas and bands, and that kind of stuff. This is
19
canine MSCs. Our cell size is
about 19 microns, 20
20
microns. Mean cell size on this
population,
21
because they are much smaller, is 10 to 12 microns.
22 More importantly, look at this. The range
23
on these, again, these are characteristic things.
24
These are done in dog, they don't publish them in
25
humans. Range is 7 to 50
microns, okay, and it
357
1
could go even go higher, go as high as 80.
2 Plasticity in this population of cells,
3
this is probably being generous, saying it is about
4
0.01 percent. If you take the
canine population of
5
cells, and you go through a CFU selection process,
6
you can get plasticity on the order of 40 to 60
7
percent.
8 If you are starting to compare potency,
9
potential potency between these
two preparations
10
of cells, this preparation of cells, after it went
11
through a CFU selection process, would be about
12
4,000 to 40,000 times more potent than just crude
13
bone marrow in terms of plasticity potential, if
14
you will.
15
Intrinsic properties of
these cells, I
16
don't know. These cells have a tendency to lay down
17
collagen if you just leave them sitting in a tissue
18
culture dish.
19 Successes for safety studies,
20
therapeutics, efficacy studies need to be done in
21
well characterized model diseases and patient
22
diseases.
23 What I would like to do is leave you with
24 a
couple of philosophies. We believe that
bone
25
marrow stem cells have potential to treat many cell
358
1
loss diseases, especially the myocardium, and we
2
will continue using these even in spite of the
3
negative report. We believe this
is a technical
4 problem. We believe we will solve it by adjusting
5
the dose and a few other things.
6 However, strict scientific disciplines are
7
necessary to avoid a train wreck.
As Doris Taylor
8
said, we do not want to repeat the gene therapy
9
trial.
10 The other thing I can give you is those of
11
us who used to have gray in our hair, and now I
12
don't have hair, is the comment, when clinicians
13
read their own press clippings, patients are going
14
to suffer.
15 Thank you for your attention.
16 [Applause.]
17 DR. RAO: Thank
you, Dr. Vulliet.
18 Our next speaker is from Genzyme. Let's
19
try and restrict the time to about 10 minutes.
20 MR. DU MOULIN:
It is very tough being the
21
last speaker of the day, but it is absolutely worse
22
being the last speaker of the day and following a
23
vet who talks about puppies.
24 [Slide.]
25
Good afternoon. My name is Gary C. du
359
1
Moulin. I am vice president of
Quality Systems for
2
the cell therapy operations at Genzyme Corporation
3
in Cambridge, Massachusetts.
4 Genzyme Corporation is collaborating with
5
Professor Philippe Menasch of Paris, France, in a
6
multicenter, Phase II/III clinical study autologous
7
skeletal myoblast implantation in Europe.
8 [Slide.]
9 Ensuring the therapeutic success of
10
cardiac cell therapy is predicated on a rigorous
11
scalable autologous cell culture program based upon
12
the principles and practices of good manufacturing.
13 Our long experience with the scale-up and
14
delivery of cartilage and keratinocyte-based cell
15
therapy products and services to thousands of
16
patients has confirmed that each element of good
17
manufacturing practices contributes an essential
18
part of an overall program that optimizes chances
19
of providing cell therapy products expressing the
20
attributes of safety and consistent quality for
21
patients.
22 Controls required for the manufacturing
23
process begin at the collection site of the muscle
24
biopsy and ends approximately three weeks later as
25
the suspension of cells exhibiting quality and
360
1
safety characteristics that once implanted can
2
consistent initiate a robust repair process.
3 Maintaining the sterility of the cell
4
culture system and ensuring lot segregation are
5
critical attributes of success.
6 [Slide.]
7 All the elements of the GMP-based
8
manufacturing program are essential in order to
9
control the inherent variability representative of
10
autologous cell culture. Briefly, these controls
11
must be established based upon the following
12
aspects of GMP.
13 These include, and they are listed here, a
14
process that is validated, personnel who are
15
trained and certified to manipulate cell safely, an
16
appropriate facility expressing stringent
17
environmental controls, records and documentation
18
of all processes conducted, equipment that is
19
calibrated and validated, raw materials that have
20
been tested for their quality, accepted formally,
21
and released into the manufacturing stream. Unique
22
to an autologous cell process is to maintain
23
stringent patient lot segregation.
24 [Slide.]
25 This is a photograph of our sole
361
1
manufacturing facility. It is
approximately 10,000
2
square feet, contains about 70 biosafety cabinets
3
in which individual patient's tissues are
4
manipulated, but other organizational requirements
5
necessary beyond the GMPs include these elements
6
here, beyond the manufacturing, a purchasing
7
element, materials handling element, logistics for
8
the shipment of cells, customer care for
9
communicating with the surgeon and the patient,
10
engineering and facilities to maintain your
11
facility, manufacturing technical services
12
responsible for the training of personnel, process
13
development, and clinical manufacturing, a
14
formalized quality assurance, quality control, and
15
validation services program.
16 [Slide.]
17 In order to scale up manufacturing
18
activities for a clinical development paradigm, an
19
organization must be created to effectively manage
20 a
myriad of direct and ancillary responsibilities,
21
but here, quality controls including environmental
22
monitoring for the manufacturing facility are
23
critical components of the operational elements of
24
cell therapy productions.
25 Robust testing programs ensures--and I
362
1
have listed them here--that cell therapy products
2
meet the highest standards of safety,
3
effectiveness, and reliability as a therapeutic
4
modality, that transmission of communicable
5
diseases is prevented, that one ensure that all
6
manufacturing and processing controls are in place
7
and consistently followed, that there is compliance
8
with existing and anticipated regulatory
9
requirements, that validated assays are performed
10
which support lot release and performance
11
monitoring of materials and components, and,
12
finally, encouraging the development of new assays
13
which enhance product safety, and, finally,
14
generating and analyzing that data, the
15
quantitative data to support continuous
16
improvements to your process.
17 [Slide.]
18 Putting these concepts together, a
19
manufacturing process whose key manufacturing
20
events from biopsy receipt through final product
21
fill finish are well understood, can optimize the
22
ex vivo cell culture process, and supported by
23
validated quality controls can ensure safety and
24
product consistency.
25 [Slide.]
363
1 We believe that call product
2
characterization is possible based upon validatable
3
measures of viability, purity, identity, and yield,
4
with safety indicators of sterility, endotoxin in
5
the absence of mycoplasma.
6 Here are shown the percent viability and
7
percent CD56 expression of three cell therapy
8
products prepared during process validation studies
9
from cadaveric skeletal muscle tissues.
10 Included in these data is evidence of cell
11
product stability over a 72-hour time frame, one
12
reasonably to be expected if transportation of the
13
cells over a long distance to the patient is
14
required. Note that the lot
release parameters
15
remain stable over 72 hours.
16 Viability and CD56 expression, both flow
17
cytometric and validatable measures of cell product
18
identity can be consistently maintained above the
19
90 percent range.
20 [Slide.]
21 Measures of sterility and potency are
22
shown in this slide. It is
crucial to ensure that
23
sterility is maintained throughout the
24
manufacturing process. In the
case of short shelf
25
life cell therapy products utilizing automated
364
1
microbial detection systems can provide benefits to
2
improve time to detection should microbial
3
contaminants be present.
4 Finally, in order to demonstrate the
5
potential for a therapeutic effect, the presence of
6
myotubule formation can be used as a measure of
7
identification and perhaps potency.
Myoblasts are
8
undifferentiated muscle precursor cells which, when
9
fused, become the differentiated myotubules. The
10
presence of multinucleated muscle cells is a strong
11
visual indicator that muscle differentiation has
12
occurred and may be an important predictor of cell
13
function.
14 In these photographs, one can readily see
15
multinucleated myotubules indicated by the yellow
16
arrows from freshly prepared samples and in cells
17
after a simulated 72-hour shipment period.
18 [Slide.]
19 In conclusion, quality, safety, and
20
effectiveness must be designed into cell therapy
21
products. Quality cannot be
inspected in or tested
22
into cell therapy products.
Despite the fact that
23
we may, at this time, not know every aspect of cell
24
product characterization, institution and
25
maintenance of stringent manufacturing controls
365
1
through rigorous observance of GMPs can contribute
2
to the safety and consistency necessary for the
3
production of cell therapy products intended for
4
cardiac cell therapy.
5
Developers of cell therapy
products must
6
consider, understand, and incorporate quality
7
requirements at the earliest possible stages of the
8
clinical development program, and in so doing, can
9
optimize the therapeutic potential of these
10
promising technologies.
11 Thank you.
12 [Applause.]
13 DR. RAO: Thank
you.
14 If there are no questions, I would like to
15
ask if anybody else from the audience wishes to
16
make any comment at this time. I
am going to ask
17
that they limit their comments and be brief.
18 [No response.]
19 DR. RAO: If
there are no more comments,
20
then, I can declare the meeting adjourned.
21
[Whereupon, at 5:46 p.m.,
the proceedings
22
were recessed, to reconvene at 8:00 a.m., Friday,
23
March 19, 2004.]
24 - - -