Scientists report that they were able to distinguish between the patterns of gene expression for two common types of non-small cell lung cancer, squamous cell carcinoma and adenocarcinoma, in the current issue of the Proceedings of the National Academy of Sciences*. The study was a collaboration between government, academic, and industry scientists.

"We hope that the molecular signatures we identified will eventually allow us to predict the clinical course of lung cancer and provide potential targets for rational therapy of the disease," said Jin Jen, Ph.D., the principal investigator, currently at the National Cancer Institute's Center for Cancer Research. "This is a first step toward that goal."

Lung cancer is the leading cause of cancer death for both men and women and kills more patients than the next five most common cancers combined. The vast majority (85%) of patients who develop lung cancer die from it. Nearly 80% of lung cancers are non-small cell lung cancer (NSCLC).

Using a gene expression technique known as SAGE (serial analysis of gene expression), the scientists compared the patterns of gene expression of squamous and adenocarcinomas lung tumors to normal lung epithelial cells. They created a total of nine SAGE libraries -- collections of expressed genes or transcripts -- from various tissue sources: primary lung tissue removed from 2 patients with squamous cell carcinoma and 2 with adenocarcinoma; 2 primary epithelial cell cultures from the small and large airways for normal controls, and a lung adenocarcinoma cell line. After sequencing more than 300,000 gene transcripts from the SAGE libraries, the researchers used various statistical methods to determine the genes that were uniquely expressed in each tumor type compared to the normal controls. They discovered that 115 genes could differentiate between the two tumor types and normal tissue. This is the first report to look at gene expression patterns in different lung tumor types using SAGE analysis.

"We were surprised to get such a clear-cut distinction between normal and tumor cells," said Jen. "And it didn't seem to matter which patient the tumor came from. The genes expressed from the same tumor types clustered together."

The researchers found that genes associated with detoxification and antioxident properties are highly expressed in squamous cell tumors. These genes include glutathione peroxidase, glutathione S-transferase, carboxylesterase, and aldo-keto reductase. Their presence in squamous cell lung cancers which are usually centrally located in the lung and associated with tobacco smoking, probably reflects a response by the bronchial epithelium to carcinogenic insults from the environment.

In contrast, genes that were highly expressed in adenocarcinomas are related to small airway-associated proteins and immunologically-related proteins. The researchers hypothesized that this finding reflects the fact that these tumors are generally found in the peripheral portion of lungs and are thought to be derived from small airway epithelial cells. They also found that many of the genes underexpressed both in the primary adenocarcinoma and the adenocarcinoma cell line were associated with squamous cell differentiation.

Because the SAGE technique was used on only 4 tumor samples, the researchers corroborated their results with two more commonly used gene expression techniques, real-time PCR (polymerase chain reaction) and cDNA microarray analysis. They found that 5 highly-expressed genes they had identified by SAGE analysis in either adenocarcinomas or squamous cell carcinoma were also highly expressed using real-time PCR. Similarly, microarray analysis revealed similar patterns of underexpression for genes most highly differentially expressed by SAGE analysis.

Two previous papers appearing in the November 20th issue of PNAS also looked at global gene expression patterns in lung tumors, using microarrays, and found distinctive patterns for the different lung tumor subgroups. Mariana Nacht, Ph.D, first author of the paper and senior scientist at Genzyme Molecular Oncology in Framingham, Mass., was pleased at the agreement between the groups, especially because the earlier studies used a much larger sample size. "I think the degree of overlap is very comforting. Interestingly, we were able to draw many of the same conclusions with just a handful of samples because SAGE data is much more quantitative and reproducible than microarray data."

There have been several attempts to look at global gene expression patterns in other tissues. Previous reports identified subtypes of tumors for B-cell lymphomas, breast cancers, and leukemias. Besides tumor subtypes, there have been several efforts to use global gene expression patterns to search for candidates markers for early detection, to describe the molecular events in biological pathways, and to discover molecular profiles that predict disease prognosis or response to treatment.

Most lung cancers are classified as either non-small cell lung cancer (NSCLC) or small cell lung cancer (SCLC) based on the appearance of the tumors cells when viewed under a microscope. Patients with NSCLC are treated differently than from those with small cell tumors. SCLC spreads rapidly and is more likely than the other types to have spread to other parts of the body at the time of diagnosis. This type accounts for about 18% of all lung cancers.

Of the three types of NSCLC, most are squamous cell carcinoma and adenocarcinoma; about 10% are classified as large cell carcinoma. Subtyping NSCLC by appearance under the microscope is difficult and it is not reliable in predicting patient outcome. Generally, patients diagnosed with the same stage squamous or adenocarcinomas receive similar treatment. It is hoped that detailed molecular portraits of tumor types can facilitate tumor classification and will be able to predict clinical prognosis and response to therapy.

The authors are from Genzyme Molecular Oncology in Framingham, MA, the National Cancer Institute (NCI) and National Center for Human Genome Research in Bethesda, MD, Johns Hopkins Medical Center in Baltimore, MD, and BioChain Institute Inc., in Hayward, CA.

*The article is titled, "Molecular characteristics of non-small cell lung cancer". The authors are Mariana Nacht, Taniana Dracheva, Yuhong Gao, Takeshi Fujii, Yidong Chen, Audrey Player, Viatcheslav Akmaev, Brain Cook, Michael Dufault, Minday Zhang, Wen Zhang, MingZhou Guo, John Curran, Sean Han, David Sidransky, Kenneth Buetow, Stephen L. Madden, and Jin Jen. PNAS 2001;98 (26):15203-15208.

Back to Top