Genetic diversity within tumors suggests continuous evolution

Large-scale genomic analysis finds that tumors evolve to adapt and survive

By analyzing tumors from more than 2,600 patients and from 38 cancer types, researchers from The University of Texas MD Anderson Cancer Center and fellow member institutions of the international Pan-Cancer Analysis of Whole Genomes Consortium have characterized the extensive genetic diversity across cancer and within individual tumors.

The study, published today in Cell, found that 95% of the analyzed tumors had at least one subclone, or genetically distinct group of tumor cells, and these subclones were often very diverse — even in the same tumor. The findings suggest that tumors continue to evolve in ways that help the cancer survive.

Intra-tumor heterogeneity, or the genetic variation of cells within a tumor, has long been a challenge as a mechanism of therapeutic resistance. Distinct subclones within a tumor may respond differently to a given treatment, resulting in subsets of cancer cells that remain alive after therapy to regrow the tumor. To better understand this heterogeneity, Consortium researchers examined the whole genomes of 2,658 tumor samples.

“We found that every tumor tells its own story of evolution,” said senior author Wenyi Wang, Ph.D., professor of Bioinformatics and Computational Biology. “Using a very large number of samples, we observed that each tumor is different, but similar evolutionary patterns emerge that characterize cancer types. Each cancer type shows distinctive patterns of how its subclones develop and grow.”

Knowing that subclones can form at various points during disease progression, the study team aimed to understand what motivates subclones to evolve. The subclones may develop as a result of certain pressures during tumor development, such as cancer treatments or anti-tumor immune responses, Wang explained.

Supporting this hypothesis, the researchers discovered that subclones appear to follow an evolutionary pattern of accumulating changes that benefit the tumors, meaning that beneficial genetic alterations are more likely to persist.

“Cancers are constantly changing over time, so it’s important to recognize that a sample taken from a tumor reflects a single point in time and the cancer will continue to evolve after this,” said corresponding author Peter Van Loo, Ph.D., group leader of the Cancer Genomics Laboratory at the Francis Crick Institute. “They can grow into a patchwork with sections driven by different mutations and evolutionary pressures. Understanding more about the evolution of subclones, why they develop in one direction over another, as well as how common they are, could help doctors better predict the levels of and types of variation likely to be present in a specific cancer type.”

A greater understanding of intra-tumor heterogeneity will help physicians and researchers make more informed decisions about the best treatment plans for each patient, as they will be better able to predict a patient’s response and likelihood of relapse.

“The analytical challenge of understanding the development of cancer subclones over time and their consequences, using high-throughput sequencing data from tumor tissue specimens, is huge,” Wang said. “In this study, we were finally able to scratch the surface of characterizing these subclones through years of team work across statisticians, computer scientists and computational biologists. We hope researchers around the world can utilize this resource as they devise their own hypotheses and studies of cancer evolution.”

As a result of this research, the Consortium has created an open-access resource, including a curated summary of the genetic variations identified in the tumor samples, and shared its computational methods for any researcher who would like to conduct cancer genome analyses. In the future, similar analyses may be performed on metastatic tumors to understand the heterogeneity in those samples as well, Wang noted.

MD Anderson researchers now are using these findings to inform ongoing efforts and design new projects in collaboration with MD Anderson’s Moon Shots Program® and other investigators, focusing on colorectal, thyroid, bladder, prostate and ovarian cancer.

In addition to Wang and Van Loo, senior authors on this study include Quaid D. Morris, Ph.D., of the University of Toronto, Toronto, Canada, and Memorial Sloan Kettering Cancer Center, New York City; and David C. Wedge, Ph.D., of the University of Oxford, Oxford, United Kingdom, on behalf of the Pan-Cancer Analysis of Whole Genomes Consortium. The full list of authors and their disclosures can be found here.