Lung Tumors Recap Developmental Patterns

Principal component analysis of gene expression signatures may help determine prognosis

Principal components of gene expression data for mouse lung and normal human lung compared to that of various types of human lung cancer. The mouse lung development profile (blue dots) marches to the right over time. The most malignant forms of lung cancer (small cell lung cancer) more closely resemble early lung development in the mouse, while the least malignant forms (adenocarcinomas) more closely resemble later lung development in the mouse and normal human lung tissue. Carcinoids (purple triangles) are known to be quite different from the other types of cancers and have a pattern of gene expression that clusters perpendicular to and below the others. Carcinoids can look like small cell lung cancer under a microscope, but the two types of cancer require different treatments. This gene expression tool might help to distinguish them. Researchers have long speculated that many of the genetic programs responsible for rapid growth of tumors are also important for the growth that occurs during normal embryonic development.


Now, researchers at the Children’s Hospital Informatics Program at Harvard have found not only a relationship between tumors and lung development, but also a trend: The tumors with genetic profiles that resemble early lung development are deadlier than those with profiles that resemble later lung development. Separating out the least aggressive tumors from the more dangerous ones might help some lung cancer patients avoid unnecessary toxic chemotherapy. The work was published in PLoS Medicine in July 2006.


“Until now, lung cancers were classified through clustering of gene expression data, without seeing the trend from the point of view of development,” says Hongye Liu, PhD, research fellow in the Children’s Hospital Informatics Program at Harvard and MIT. “But we’ve found that the development trend can predict which cancer is worse.”


Earlier work by Liu’s co-authors, Alvin Kho, PhD, and Isaac Kohane, MD, PhD, showed that the gene expression profiles for each of several different types of brain tumors form distinct clusters when projected onto the gene expression profile of mouse genomic cerebellar development. The work by Liu and colleagues confirms these findings in the lung cancer context and takes them one step further by finding a connection between tumors, development and prognosis.


Charles Powell, MD, professor of clinical medicine at Columbia University College of Physicians and Surgeons, says Liu’s work is important in emphasizing the link between cancer and development, but prognostic indicators in this paper need to be tested prospectively. More interesting, he says, is the potential for insight into the origins of lung cancer. “The steps that transform a damaged cell into lung cancer of one type or another are likely to be similar to normal development in the lung,” he says. “If we can follow-up this paper to understand those steps then we should be able to discover novel insights into lung carcinogenesis.”


All submitted comments are reviewed, so it may be a few days before your comment appears on the site.

Post new comment

The content of this field is kept private and will not be shown publicly.
This question is for testing whether you are a human visitor and to prevent automated spam submissions.
Enter the characters shown in the image.