Whales have 1000-fold more cells than humans and mice have 1000-fold fewer; however, cancer risk across species does not increase with the number of somatic cells and the lifespan of the organism. This observation is known as Peto's paradox. How much would evolution have to change the parameters of somatic evolution in order to equalize the cancer risk between species that differ by orders of magnitude in size? Analysis of previously published models of colorectal cancer suggests that a two- to three-fold decrease in the mutation rate or stem cell division rate is enough to reduce a whale's cancer risk to that of a human. Similarly, the addition of one to two required tumour-suppressor gene mutations would also be sufficient. We surveyed mammalian genomes and did not find a positive correlation of tumour-suppressor genes with increasing body mass and longevity. However, we found evidence of the amplification of TP53 in elephants, MAL in horses and FBXO31 in microbats, which might explain Peto's paradox in those species. Exploring parameters that evolution may have fine-tuned in large, long-lived organisms will help guide future experiments to reveal the underlying biology responsible for Peto's paradox and guide cancer prevention in humans. Similarly, the addition of one to two required tumour-suppressor gene mutations would also be sufficient. We surveyed mammalian genomes and did not find a positive correlation of tumour-suppressor genes with increasing body mass and longevity. However, we found evidence of the amplification of TP53 in elephants, MAL in horses and FBXO31 in microbats, which might explain Peto's paradox in those species. Exploring parameters that evolution may have fine-tuned in large, long-lived organisms will help guide future experiments to reveal the underlying biology responsible for Peto's paradox and guide cancer prevention in humans. This results in a lifetime risk of 0.5% and a rate of one division every 9 days lowers this below the human estimate to 0.2%. (c) Evolution of cancer gene families in mammalian genomes It should be relatively easy for a species to evolve redundant checks on neoplastic progression by duplicating tumour-suppressor genes, which would present as expanded gene families in those species. Alternatively, a species could decrease the risk of progression by deleting proto-oncogenes [ 17 ]. We developed a genome-wide BLAST search intended to find all genes within a gene family based on one representative. These functions evolved billions of years before multi-cellularity and are essential to all forms of life [ 19 ]. GK control cell proliferation and signalling by enforcing checkpoints to ensure that cells at risk for neoplastic transformation do not continue to propagate. We did not find a positive correlation between body mass and the number of genome hits for any of the cancer gene categories (proto-oncogenes, GK and CT; tumour-suppressor gene results are shown in figure 3 ). Figure 3. Cancer gene copy numbers across mammalian genomes. The number of tumour-suppressor genes does not increase with body mass ( a ). Based on our BLAST search, we find no positive correlation between tumour-suppressor genes as a whole, or GK and CT together with body mass. This was tested with a linear regression and is true on both the linear and log scale. The log (base 10) of the mass in grams is shown here to ease visualization of the range of masses. There is a strong linear correlation between the number of proto-oncogenes and GK ( b ). Based on our BLAST search for cancer gene families, the number of proto-oncogenes and GK found in a genome are highly correlated ( r 2 = 0.85, p -value < 0.001). We focused on increased copies of tumour-suppressor genes as it is difficult to confirm a gene deletion in draft genomes due to possible incompleteness and mis-assemblies. We used a comprehensive list of 830 human tumour-suppressor genes [ 20 ] and obtained the orthologous genes in 36 non-human mammals from E nsembl B io M art v. 72 (see electronic supplementary material, table S3). Genes that were found to have a ‘one : many’ relationship to the human tumour-suppressor gene in at least one mammal were considered for further analysis. This gene encodes an F-box protein that mediates the DNA damage response by promoting the degradation of cyclin D1 through polyubiquitination to induce cell cycle arrest in G1 [ 22 ]. Though the microbat is only 10 g, it can for live up to 34 years [ 14 ] so one hypothesis is that these additional tumour-suppressors may decrease the cancer risk of the bat, which would otherwise be heightened by their increased longevity [ 23 ]. The second highest gene copy number we came across was 12 which included TP53 , IL6 and LCN2. TP53 is mutated in the majority of human cancers and plays a crucial role in multiple tumour suppressive pathways including apoptosis, senescence and DNA repair [ 24 ]. Therefore, we might anticipate finding functionally redundant pathways or additional tumour-suppressor genes that act as a ‘back up’ in case of failure to existing pathways in animals that have evolved this tumour suppression mechanism better to combat cancer, which has been previously proposed as a solution to Peto's paradox [ 2 , 10 , 46 ]. The implications of this are not There may also be undetected cancer genes in non-human species with little homology to the human gene sequences. However, we added the time since the most recent common ancestor with human to our linear model, to account for the difficulty of detecting genes with low levels of homology due to evolutionary distance, and this did not change the results, suggesting that the number of genes we find in each species is not simply a function of how closely they are related to humans. Human tumour-suppressor genes were used for this analysis, but in doing so we made the assumption that they perform the same function in the other species. This has not been experimentally verified.

<h4>Background</h4>Cancer is a common diagnosis in many mammalian species, yet they vary in their vulnerability to cancer. The factors driving this variation are unknown, but life history theory offers potential explanations to why cancer defense mechanisms are not equal across species.<h4>Methodology</h4>Here we report the prevalence of neoplasia and malignancy in 37 mammalian species, representing 11 mammalian orders, using 42 years of well curated necropsy data from the San Diego Zoo and San Diego Zoo Safari Park. We collected data on life history components of these species and tested for associations between life history traits and both neoplasia and malignancy, while controlling for phylogenetic history.<h4>Results</h4>These results support Peto's paradox, in that we find no association between lifespan and/or body mass and the prevalence of neoplasia or malignancy. However, a positive relationship exists between litter size and prevalence of malignancy (P = 0.005, Adj. R2 = 0.212), suggesting that a species' life history strategy may influence cancer vulnerabilities. Lastly, we tested for the relationship between placental invasiveness and malignancy. We find no evidence for an association between placental depth and malignancy prevalence (P = 0.618, Adj. R2 = 0.068).<h4>Conclusions</h4>Life history theory offers a powerful framework to understand variation in cancer defenses across the tree of life. These findings provide insight into the relationship between life history traits and cancer vulnerabilities, which suggest a trade-off between reproduction and cancer defenses.<h4>Lay summary</h4>Why are some mammals more vulnerable to cancer than others? We test whether life history trade-offs may explain this variation in cancer risk. Bigger, longer-lived animals do not develop more cancer compared to smaller, shorter-lived animals. However, we find a positive association between litter size and cancer prevalence in mammals. [ 2 ] reported the first empirical evidence for Peto’s Paradox by analyzing cancer prevalence in 37 mammals. This study suggests that larger, longer-lived animals have enhanced cancer defense mechanisms. Additionally, extra copies of TP53 , a critical tumor suppressor gene, were reported in elephants, and functional studies identified this gene expansion as a potential mechanism of cancer defense in the largest extant land mammal. Along with body size and lifespan as predictors of cancer mortality, placental mammals may have higher rates of malignancy due to selection for invasive placental genes [ 20 , 21 ]. We predicted that species with the most invasive placenta type (hemochorial) would have higher rates of malignancy compared with animals with less invasive placentas (endotheliochorial and epitheliochorial). In this study, we retested Peto’s Paradox to answer the question: Do larger, long-lived mammals get more cancer? We then analyzed the association between cancer risk and life history traits in a phylogenetic context, including the degree of placental invasiveness. We report mortality records with attention to important interpretation details that were not available in the original report by Abegglen et al. [ 2 ] ( Supplementary Table S1 ). Data were filtered to exclude stillbirths, perinatal mortalities and animals less than 1 year of age with a low risk for developing cancer. Exclusion of these individuals decreases potential bias that would result in lower estimated cancer prevalence rates. We then refined the case definition of neoplasia to distinguish benign vs malignant neoplasms. A neoplasm is a general term for an abnormal growth that includes both benign and malignant tumors. Due to small post-mortem samples sizes, Asian and African elephant neoplasia and malignancy data were combined. Confidence intervals (95%) on lifetime neoplasia prevalence were estimated in PropCI package in R. Life history regression models We tested for a relationship between life history variables and cancer prevalence. Life history variables were collected from Pantheria [ 23 ] and AnAge [ 24 ]. Information on placental types was compiled from published sources [ 25 , 26 ]. While we used an estimated lifespan in elephants to be 65 years, we note that maximum lifespan in Asian elephant may be as high as 80 years [ 27 ]. We observed 45% neoplasia prevalence in our dataset (14/31 animals). The difference in prevalence suggests our current estimate for the black-footed prairie dogs and other animals may be noisy, due to our smaller sample sizes. Additionally, we find no reports of neoplasia or malignancy in moose and white-tailed gnus, large-bodied species of the Artiodactyla order. We also observed no cancer in armadillos. These results warrant further investigation to determine if armadillos and other large-bodied Artiodactyla are better at suppressing cancer compared with other species. Lastly, this report highlights the utility of well curated cancer across species data and provides exciting new opportunities for cancer comparative genomics and biology research. The elephant in the room We report higher cancer prevalence in elephants than previously reported [ 2 ]. Previous estimates were derived from the Elephant Encyclopedia Database ( n = 644 elephants) [ 2 ]. While this database is Why may the degree of placentation have no relationship with malignancy? The placenta is the site of intense evolutionary conflict between maternal–paternal genes [ 20 , 43 , 44 ]. As a consequence of this evolutionary conflict, the placenta is one of the most diverse mammalian organs. Across various mammalian species, placenta evolved different degrees of invasion, including multiple independent reductions in invasiveness over evolutionary time [ 45 , 46 ]. Indeed, malignancy risk may not be generalizable and highly invasive placental mammals (hemochorial) may have co-evolved heighten strategies to defend against inappropriate invasion mechanisms often co-opted by cancer cells.

AbstractEvolving to become bigger and/or longer lived should increase cancer susceptibility, but this predicted increase is not observed, a contradiction named Peto's paradox. A solution is that cancer suppression evolves to minimize cancer susceptibility, and the discovery of 19 retrogene (RTG) copies of the tumor suppressor gene TP53 in the African elephant (Loxodonta africana) is increasingly cited as a classic example of such adaptive suppression. However, classic examples need rigorous evaluation and an alternative hypothesis is that the RTGs spread by genetic drift. This study shows that before its duplication, the ancestral elephant RTG was already truncated from 390 amino acids to 157 by a frameshift mutation, and that 14 of the 19 copies are now truncated to ≤88 amino acids. There was no compelling evidence of either positive or negative selection acting on these 88 codons, and the pattern of RTG accumulation fits a neutral model with a duplication rate of ~10−6 per generation. It is concluded that there is no evidence supporting the hypothesis that the 19 elephant RTGs spread to fixation by selection; instead, the evidence indicates that these RTGs accumulated primarily by segmental duplication and drift. It is shown that the evolutionary multistage model of carcinogenesis (EMMC) predicts the recruitment of 1–2 independently acting tumor suppressor genes to suppress the increased cancer risk in elephants, so it is possible that one or a few RTGs may have been favored by selection resulting in the known enhanced sensitivity of elephant cells to DNA damage. However, the analysis does not provide any support for either a direct (via conserved TP53 activity) or indirect (via supporting canonical TP53 function) role of the RTGs sequences, so that the presence of multiple copies of TP53 retrogenes in elephants needs to be further justified before being used as a classic example of tumor suppression in large‐bodied animals. A solution is that cancer suppression evolves to minimize cancer susceptibility, and the discovery of 19 retrogene (RTG) copies of the tumor suppressor gene TP53 in the African elephant ( Loxodonta africana ) is increasingly cited as a classic example of such adaptive suppression. However, classic examples need rigorous evaluation and an alternative hypothesis is that the RTGs spread by genetic drift. This study shows that before its duplication, the ancestral elephant RTG was already truncated from 390 amino acids to 157 by a frameshift mutation, and that 14 of the 19 copies are now truncated to ≤88 amino acids. It is shown that the evolutionary multistage model of carcinogenesis (EMMC) predicts the recruitment of 1–2 independently acting tumor suppressor genes to suppress the increased cancer risk in elephants, so it is possible that one or a few RTGs may have been favored by selection resulting in the known enhanced sensitivity of elephant cells to DNA damage. Given the increasing availability of genomic data, the possible involvement of gene duplication can be tested by searching for duplicated tumor suppressor and other cancer‐related genes in large and/or long‐lived animals (Caulin et al., 2015 ) such as elephants (Vazquez & Lynch, 2021 ) and whales (Tejada‐Martinez et al., 2021 ; Tollis et al., 2019 ). A search of the genome of the African elephant ( Loxodonta africana ) resulted in the discovery of 19 retrogene (RTG) copies of TP53, with multiple copies also present in the genome of the Asian elephant ( Elephas maximus ; Abegglen et al., 2015 ). First, how much added protection against cancer is the elephant expected to evolve? Second, did any of the RTGs ever code for a fully functional p53? Third, is there evidence that the increase in the number of segmental duplicates containing the RTG was favored by selection acting to increase the number of TP53 RTGs? Fourth, is there any evidence that selection has acted within the coding sequence of the RTGs? 2. METHODS The adaptive response predicted by the EMMC, in terms of either the number of tumor suppressor genes recruited or the reduction in the somatic mutation rate, was estimated for the African elephant ( L . The predicted adaptive response to the risk of cancer The EMMC was used to estimate the magnitude of the adaptive change expected in the evolution from a hyrax‐like ancestor to modern‐day elephant or to a manatee (Table 1 ). This change was estimated either in terms of tumor suppressor recruitment (i.e., as in increase in M ) or as a reduction in the somatic mutation rate ( u ). The nature of the evolutionary response was different for different cancers, because some cancers Thus, while these sequence data are not definitive, they are consistent with a null hypothesis that the accumulation of the TP53 RTGs has been through a process of duplication followed by random genetic drift. 4. DISCUSSION TP53 is a crucially important tumor suppressor gene that has been named the “guardian of the genome” (Lane, 1992 ), and the finding of multiple retrogene copies of the gene in the genome of the African and Asian elephants prompted the hypothesis that these copies had been sequentially selected in concert with the increasing body size within the elephant clade (Abegglen et al., 2015 ; Sulak et al., 2016 ). As a result, any potential transcript is terminated before the p53 region involved in DNA binding, and the remaining 5, although they have part of this region, they all lack the dimerization site and the specific DNA‐binding site (Table S1 ). RTG#7 has retained the longest region of 5′ homology with the p53 protein (157 codons), but it lacks a start codon. The manatee and hyrax RTGs also show an extreme loss of potentially functional protein (Table S1 ). Estimates of the degree of adaptive change needed to keep the incidence of cancer in elephants at a low level (Table 1 ) are consistent with the selection in the elephant clade favoring a gain of the equivalent of 1–2 tumor suppressor genes.