Cancer cell evolution through the ages

The essential ingredients for evolution are variation, inheritance, selection, and time. Like all of life, cells within human bodies can evolve, because dividing cells accumulate many somatic mutations with age, some of which benefit their survival or reproduction. However, cellular evolution is constrained by the death of the host. But what if somatic cells could outlive their hosts by escaping to another host? On page 464 of this issue, Baez-Ortega et al. (1) document the remarkable odyssey of the canine transmissible venereal tumor (CTVT) using DNA exome sequencing of 546 CTVT samples from dogs throughout the world. This ancient tumor has been sexually transmitted between dogs for ~6000 years, indicating that there is no inherent limit on the number of mammalian cell divisions.

CTVT probably started from a macrophage (2), which evolved into a sexually transmitted parasite that can evade the canine immune system long enough to be transmitted to a new host. It is usually cleared by the host’s immune system before it becomes lethal. This rather benign course is common for sexually transmitted infections, the agents of which must rely on relatively healthy hosts to capitalize on infrequent mating to persist (3). Baez-Ortega et al. found that CTVT originated in Asia ~6000 years ago and started dispersing worldwide ~2000 years ago. They showed that in the past 500 years, CTVT has crisscrossed the globe, with the help of human travel. The same researchers previously sequenced CTVTs from an Australian and a Brazilian dog and found a pronounced stability in the CTVT genomes, despite considerable divergence from the original dog genome (4).

The generation of a cancer cell requires mutations in genes that confer growth advantages and immortality to the cell, called driver mutations. An open question in cancer biology is whether cancers ever reach a peak of fitness: a locally optimal strategy for surviving and reproducing in their environment. Approximately two-thirds of human cancers show evidence of ongoing natural selection, even after they accumulate sufficient mutations to make them cancerous (5). Current CTVT cells have had thousands of years to optimize their fitness, and their exomes (DNA sequences that encode proteins) are riddled with somatic mutations: ~38,000 per sample compared to ~100 per human cancer sample. Most genes in the CTVT genome had at least one protein-altering mutation among the 546 tumors. There are more than 200 known driver genes in humans that, when mutated, can increase cellular fitness. However, only five such driver genes were found to be mutated in the CTVT cells, which is approximately the same number of driver mutations estimated to occur in many human cancers (6). These driver mutations were found in all the CTVT samples, and therefore arose very early in CTVT evolution. Perhaps they were even present in the founder cancer cells (see the figure). The lack of further driver mutations suggests that there has been little ongoing selection since the CTVT line developed.

The evolution of transmissible tumors in dogs.

Canine transmissible venereal tumor (CTVT) arose in a normal dog ~6000 years ago in Asia. The five driver mutations identified in CTVTs today were likely present in the first CTVT cell. CTVT evolution has been neutral for most of its history, accumulating large numbers of passenger mutations.

The absence of ongoing selection was also assessed by Baez-Ortega et al. by measuring the ratio of mutations that change proteins to those that do not. The data suggest that CTVT has been evolving neutrally, accumulating mutations that do not change cell fitness. This lack of selection is in marked contrast to long-term cultures of the bacteria Escherichia coli, which continue to adapt over thousands of generations (7). CTVT appears to be on a fitness peak, or more accurately, a fitness plateau.

Only ~2000 genes in CTVT have been conserved. Mutations that changed those genes were probably detrimental to the CTVT cells and were quickly weeded out by natural selection. This implies that most genes in the mammalian genome are not needed by cancer cells, and only a handful of genes need to be tweaked to reach a fitness plateau. Once reached, there is little advantage in having a high mutation rate, and indeed CTVTs show no signs of high rates of DNA mutation or chromosomal instability that are common in human cancers. It is unclear if CTVTs were always genetically stable, or like human cancers, went through a period of genetic instability. The key to longevity may be the avoidance of declining fitness caused by the accumulation of deleterious mutations in essential genes, a problem known as Muller’s ratchet (8).

The lack of ongoing natural selection suggests that CTVT has not had much of an effect on dog survival and reproduction, which is consistent with the typical benign course of the disease. If CTVT had a strong selective effect on dogs, there would be evidence of a coevolutionary arms race, with dogs evolving defenses to CTVT and CTVT evolving adaptations to overcome those defenses.

Genome sequences also reveal mechanisms of mutagenesis, which produce characteristic base changes (9). CTVT mutations appear to be caused by many of the same mechanisms that cause mutations in human cancers, including the signature of aging or cell division. Another common signature was caused by ultraviolet (UV) light. The closer the dogs were to the equator, the more UV-induced mutations accumulated in their tumors. This suggests that the mutations were likely caused by sunlight striking CTVT cells on exposed genitalia.

The remarkable evolution documented by CTVT genome sequencing provides evidence that neoplasia is not inherently progressive. This gives hope that some relatively indolent human cancers, including many prostate cancers, could also be controlled for long periods of time when cure is not possible. The idea of such evolutionary or adaptive therapy (10) is to limit tumor growth rather than inevitably selecting for more aggressive or lethal subclones with attempts at curative therapy.

Mapping CTVT somatic cell evolution and its spread throughout the world has many parallels to mapping how mutant cells evolve and spread within a human body. Although the scales are different, the genomes that are increasingly being sequenced from human cells can also record what those cells do, how they spread, and what they were exposed to. This is valuable because it is difficult to observe the life and death of most human cells over time. Understanding how mammalian cancers can evolve over long periods of time will likely be important in future attempts to manage that evolution to prevent mortality and morbidity due to cancer.