Posted: June 7, 2017

TCGA Data Used to Study Ties Between Tumor Genetics and Evolutionary History

Amy E. Blum, M.A.

Image: Epifluorescence microscopy of microtentacles forming on a breast tumor cell. Credit: Stuart S. Martin, National Cancer Institute / Univ. of Maryland Greenebaum Cancer Center

A new study used public genomic data to investigate whether tumors take after our single-celled ancestors. By analyzing seven cancer types in The Cancer Genome Atlas (TCGA), researchers at the University of Melbourne observed that a common feature of cancer is repression of genes that were acquired during the transition from unicellular to multi-cellular life and simultaneous activation of ancient genes and cell behaviors1. Published in May in Proceedings of the National Academy of Sciences, the study also suggests that targeting abnormal interactions between multicellular gene pathways and unicellular pathways may lead to clinical strategies that would apply to a wide variety of cancers.

The Atavistic Model

Hallmarks of cancer include boundless growth, avoidance of programmed cell death, and de-differentiation toward a less specialized cell type. These characteristics are common to single-celled organisms, which thrive by growing and dividing until halted by their environment, but differ somewhat from multicellular organisms, which develop specialized cells and tightly control cell growth and death.

The theory that cancer reverts to ancient cellular processes, called the atavistic model, has been previously proposed and is debated among scientists because there are many differences between single-celled organisms and tumors and because it is difficult to define unicellular and multicellular genes2. Ancient genes play foundational roles in multicellular organisms and may even participate in the complex interactions characteristic of multicellular organisms. Another critique of the atavistic theory is a lack of evidence. To address these issues, this study provided the first large-scale molecular analysis of the relationship between tumors and evolutionary history to explore the atavistic model and better understand the genetic basis of cancer.

How Old Are Cancer Genes?

The researchers began by dating, for each gene expressed in the cell, the time period in which the gene arose based on the most distant species that carries a version of that gene in their genome. These dates were consistent with gene functions; older genes corresponded with fundamental processes and newer genes with more complex processes. The researchers then applied these dates to seven cancer types in TCGA RNAseq data and observed that genes expressed in cancer tissue tend to be older than the genes expressed in matched normal tissue. They found activation of genes dating back to our unicellular ancestors and repression of more recent genes, with an inflection point at the transition between unicellular and multicellular life.

The investigators also observed that the same sets of genes were affected consistently across the seven cancers, indicating that cancers may undergo convergent evolution toward expression of particular ancient gene pathways. For example, stress responses conserved within unicellular species were often activated in cancer and stress responses exclusive to multicellular organisms were repressed. The authors suggest that this convergent evolution by tumors may be explained by tumors experiencing similar environmental stressors to unicellular organisms, such as hypoxia and DNA damage.

Lost Connections

Next, the researchers developed a network of correlations in gene expression to reflect the interconnectedness of gene pathways, or how expression of one gene set influences expression of another gene set. The results indicated that expression correlations between two unicellular-specific pathways were maintained in tumors, as were connections between sets of genes particular to multicellularity. However, the pairs of gene sets in which one gene set is ancient and the other more recent followed a different pattern. In tumors, these pathways were more likely to be negatively correlated, suggesting that connections between unicellular and multicellular gene sets were severed in cancer, resulting in reliance on one pathway or the other.

By looking deeper into an interaction disrupted by cancer, the authors hypothesized they might identify genes that mediate interactions between ancient and more recent genes across cancers. Focusing on cellular junction organization and chromosome organization, two process that were negatively correlated in TCGA cancers, they identified 12 genes that mediate the relationship between these processes. Many of these genes were previously associated with chromosomal instability and lower cancer survival, suggesting that genes that mediate between single-celled and multicellular process may be promising targets for therapy across many cancer types.

Though it is difficult to untangle the effects of a gene’s age from its indispensability to living organisms, these results highlight the important role of ancient genes in promoting diverse cancer types. This analysis using the publicly available TCGA dataset provides new information about the theory that tumors revert to unicellular gene expression pathways and suggests that further investigation could lead to the discovery of novel pan-cancer targets.


  1. Trigos, A.S, Pearson, R.B., Papenfuss, A.T., and Goode, D.L. (2017) Altered Interactions Between Unicellular and Multicellular Genes Drive Hallmarks of Transformation in a Diverse Range of Solid Tumors. Proceedings of the National Academy of Sciences doi: 10.1073/pnas.1617743114
  2. Davies, P.C.W, and Lineweaver, C.H. (2011) Cancer Tumors as Metazoa 1.0: Tapping Genes of Ancient Ancestors. Physical Biology, 8:1. Doi: 10.1088/1478-3975/8/1/015001