1. Top of page
  2. Abstract
  3. References

In a classic paper, George Williams (1957) argued that alleles promoting reproductive success early in life may be favoured by selection, even if they reduce the lifespan of individuals that bear the allele. A variety of evidence supports the theory that such ‘antagonistic pleiotropy’ is a major factor contributing to the evolution of senescence (Ljubuncic & Reznick 2009), but examples of specific alleles known to fulfil Williams’ criteria remain rare, in both humans and other animals (e.g. Alexander et al. 2007; Kulminski et al. 2010). An intriguing example in this issue of Molecular Ecology (Fernandez & Bowser 2010) demonstrates that both natural and sexual selection may favour melanoma-promoting oncogene alleles in the fish genus Xiphophorus.

Antagonistic pleiotropy is expected under natural selection but may become even more pronounced under sexual selection. Trivers (1972) emphasized that sexual selection will often favour different life history tradeoffs (and hence mortality rates) for males and females. The potentially high advantage of success in competition for mates for males means that traits conferring a significant reproductive advantage early in life but imposing increased mortality later will frequently be favoured in the balance between sexual selection and natural selection, leading to shorter lifespan in males than females. Such sexually-selected traits are less likely to be favoured in females, where the reproductive rewards of successful competition for mates are typically lower. Nevertheless, females may evolve to prefer traits in males that will contribute to an early demise for their sons, if the advantage of their early reproductive success is sufficiently strong (Kokko et al. 2002).

Enter the swordtails and platyfishes (genus Xiphophorus). These fish are well-known to evolutionary biologists as premier subjects of research on sexual selection, including mate choice, aggression, competition for mates and alternative strategies (e.g. Ryan et al. 1990). Less well known is the role of these fish as models for analysing carcinogenesis. Experiments in the 1920s demonstrated that hybrids between platyfishes and swordtails showed a striking propensity to develop malignant pigment-cell tumours – melanomas—among the most deadly cancers of humans. Research since has revealed that melanomas develop within species as well as in hybrids, which has led to Xiphophorus becoming a model system for studies on the molecular biology and genetics of melanoma (MeierJohann & Schartl 2006).

The major gene underlying melanoma formation, Xiphophorus melanoma receptor kinase (Xmrk) was identified by Wittbrodt et al. (1989). This sort of gene is of special interest to cancer biologists, given a central role for high activity of such tyrosine kinases in progression of diverse forms of human cancer. In Xiphophorus, runaway tyrosine kinase activity, due to high expression of Xmrk, leads to development of a large black spot on the fish’s side—a cancer that destroys muscle, reduces swimming speed and causes early death (Fig. 1). Despite such deleterious effects, the oncogenic Xmrk allele persists in some natural populations of Xiphophorus, suggesting the presence of benefits early in life that outweigh late costs—the hallmarks of antagonistic pleiotropy.


Figure 1. Xiphophorus cortezi with the exaggerated Sc phenotype caused by the Xmrk oncogene. Photo by A. Fernandez.

Download figure to PowerPoint

In this issue of Molecular Ecology, Fernandez & Bowser (2010) investigate the expression of Xmrk, the spotted caudal (Sc) phenotype (the pigment pattern that expresses Xmrk) and melanoma in the context of body size in Xiphophorus cortezi. The Sc phenotype shows incomplete penetrance and hence fish carrying Xmrk may not express the spotted caudal pattern. Body size is known from previous research to be strongly predictive of reproductive success in a number of species of Xiphophorus, so a positive association of Xmrk with body size would fit the pattern predicted by Williams: a gene promoting reproductive success early in life but with a cost later on. Since larger body size increases reproductive success in Xiphophorus through mate choice and competition for mates, sexual selection is likely involved as well.

Fernandez & Bowser (2010) set out to investigate several key properties of this system. First, is the Xmrk gene common and widely distributed among populations? Second, does melanoma occur even when there is no opportunity for hybridization? Third, is there a sex-bias in the occurrence of Xmrk and of melanoma? And finally, are the Xmrk gene and melanoma correlated with a key life history feature, large body size, in these fish?

The authors collected fish from four localities in Mexico, spanning the range of X. cortezi, one of which, Arroyo Conchita, contained only this species, precluding the possibility that ongoing hybridization sustains the oncogenic Xmrk allele. They measured standard length for all individuals sampled and collected tissue for genetic analyses. Cloned sequences of the Xmrk gene from a closely-related species of Xiphophorus were used to design primers that allowed them to amplify both the Xmrk oncogene and a closely related proto-oncogene (a gene that when mutated, can become oncogenic). They also carried out detailed histological sampling of nine individuals (eight male, one female) from Arroyo Conchita, that appeared to have abnormal melanophore phenotypes on visual inspection. They analysed the data on males and females separately to investigate associations between the presence of the Xmrk genotype, the Sc phenotype, melanoma and body size.

Their results, in combination with those of previous studies, demonstrate that the Xmrk polymorphism is widespread throughout the range of X. cortezi. There was a pronounced sex bias in favour of males (replicating results from other populations and species), but both sexes showed high frequencies of the Xmrk genotype in some populations, with a maximum of 82% for males and 57% for females in the Arroyo Conchita population. The study definitively confirms that non-hybrid Xiphophorus develop melanomas in natural populations, as the Arroyo Conchita population shows a high frequency of Xmrk and melanoma. For both males and females, individuals carrying the Xmrk allele were significantly longer. For males, this difference was not associated with the Sc phenotype, whereas there was a marginally significant association for females. Finally, males with melanoma (and Xmrk) were significantly longer than both males with Xmrk but without melanoma and males without Xmrk (melanoma was not found in males without Xmrk).

The authors argue convincingly that both natural and sexual selection likely contribute to the high frequencies of Xmrk in natural populations. Large size thus provides a variety of advantages to both males and females in the context of natural selection, including increased access to resources and protection from predation. For females, substantial gains in fecundity would also be associated with increases in size.

Previous research has demonstrated that for many species of fish, including a variety of species of Xiphophorus, large size confers advantages to males in terms of attractiveness to females and success in competition for mates. Further evidence for the efficacy of sexual selection on the frequency of Xmrk comes from earlier studies by the first author. In a recent paper (Fernandez & Morris 2008), mate choice experiments were used to demonstrate that female X. cortezi prefer males with the Sc phenotype that is associated with Xmrk. Another study (Fernandez 2010) demonstrated a significant positive association between the Xmrk genotype and aggression in males, which engenders success in competition for mates. Hence the high frequency of Xmrk in males, relative to females, makes sense given the strength of sexual selection.

This research provides an outstanding example of antagonistic pleiotropy, in which the allele exhibiting the pleiotropic effects has been identified. It should now be possible to test the hypothesis that the benefits of the oncogenic Xmrk genotype outweigh its costs and how the benefit to cost ratio varies with the frequency of Xmrk. Intriguingly, Fernandez & Morris (2008) showed that in one population with a particularly high frequency of Xmrk, females preferred males without the Sc phenotype. Hence when Xmrk reaches high frequencies, negative effects such as the production of offspring with two copies of the Xmrk gene (which have low viability) may override the advantages associated with increased aggressiveness, attractiveness and large size (Fernandez & Morris 2008).

The authors offer several tantalizing insights into possible mechanistic connections between Xmrk and cancer. The Xmrk oncogene is derived via duplication from an epidermal growth factor receptor and hence may mediate both normal and abnormal growth through interactions with growth factors such as IGF-1. This type of interaction, involving constitutive over-expression of growth-related genes, is central to the development of numerous cancers in both humans and non-human animals (e.g. Hankinson et al. 1998). Further, Xmrk is near (and likely linked to) pigment production genes like Mc1r that are overexpressed in melanoma. Finally, previous work (Fernandez 2010) demonstrating a link between Xmrk and aggression suggests a role for androgen expression—and high androgen levels are associated with tumour formation in both humans and other animals (Summers & Crespi 2008).

Previous studies have connected greater height with increased human cancer risk, an observation that motivated the interests of Fernandez & Bowser (2010) in this phenotype for Xiphophorus. Recent studies have also identified candidate genes in humans that may mediate antagonistic pleiotropy, associated, as in Xiphophorus, with selection for enhanced growth due to sexually-selected or other benefits early in life, at the expense of later increased risk of cancer. Thus, polymorphisms in the High Mobility Group At-hook 2 (HMGA2) gene influence height in humans (Yang et al. 2010), HMGA2 variants with high levels of expression are associated with increased risk of cancer and poor prognosis (Young & Narita 2007) and a series of papers has demonstrated higher reproductive success of taller men (Sear 2006). Replicated associations between higher birth weight and increased risk of breast and prostate cancer in humans (Cnattingius et al. 2009; Oberg et al. 2009) hint at a broad swath of antagonistically pleiotropic effects in human cancer risk and high relevance of Xiphophorus as a model system for the genetic basis of human polygenic disease.

This paper validates a tractable system for investigating antagonistic pleiotropy in natural populations, which allows intensive investigation, from gene to natural population, of the mechanisms that connect reproductively advantageous traits, such as large size, to traits that increase mortality in adults, such as cancer. More generally, a new empirical field has been born, the molecular evolutionary ecology of cancer, which should stimulate similarly-novel studies that dissect the forces of selection that hold us and other animals, delicately poised at some mark between sex and death.


  1. Top of page
  2. Abstract
  3. References
  • Alexander DM, Williams LM, Gatt JM et al. (2007) The contribution of apolipoprotein E alleles on cognitive performance and dynamic neural activity over six decades. Biological Psychology, 75, 229238.
  • Cnattingius S, Lundberg F, Sandin S, Grönberg H, Iliadou A (2009) Birth characteristics and risk of prostate cancer: the contribution of genetic factors. Cancer, Epidemiology, Biomarkers and Prevention, 18, 24222426.
  • Fernandez A (2010) A cancer-causing gene is positively correlated with male aggression in Xiphophorus cortezi. Journal of Evolutionary Biology, 23, 386396.
  • Fernandez A, Bowser P (2010) Selection for a dominant oncogene and large male size as a risk factor for melanoma in the Xiphophorus animal model. Molecular Ecology, 19, 31143123.
  • Fernandez A, Morris M (2008) Mate choice for more melanin as a mechanism to maintain a functional oncogene. Proceedings of the National Academy of Sciences, USA, 105, 1350313507.
  • Hankinson S, Willett W, Colditz G et al. (1998) Circulating concentrations of insulin-like growth factor-1 and risk of breast cancer. Lancet, 351, 13931396.
  • Kokko H, Brooks R, McNamara JM, Houston AI (2002) The sexual selection continuum. Proceedings of the Royal Society of London Series B: Biological Sciences, 296, 13311340.
  • Kulminski AM, Culminskaya IV, Ukraintseva SV, Arbeev KG, Land KC, Yashin AI (2010) Beta2-adrenergic receptor gene polymorphisms as systemic determinants of healthy aging in an evolutionary context. Mechanisms of Ageing and Development, Apr 23 [Epub ahead of print].
  • Ljubuncic P, Reznick AZ (2009) The evolutionary theories of aging revisited – a mini-review. Gerontology, 55, 205216.
  • MeierJohann S, Schartl M (2006) From mendelian to molecular genetics: the Xiphophorus melanoma model. Trends in Genetics, 22, 654661.
  • Oberg S, Cnattingius S, Sandin S, Lichtenstein P, Iliadou A (2009) Birth weight-breast cancer revisited: is the association confounded by familial factors? Cancer, Epidemiology, Biomarkers, and Prevention, 18, 24472452.
  • Ryan MJ, Hews D, Wagner WE Jr (1990) Sexual selection on alleles that determine body size in the swordtail, Xiphophorus nigrensis. Behavioral Ecology and Sociobiology, 26, 231237.
  • Sear R (2006) Height and reproductive success: how a Gambian population compares to the West. Human Nature, 17, 405418.
  • Summers K, Crespi B (2008) The androgen receptor and prostate cancer: a role for sexual selection and sexual conflict? Medical Hypotheses, 70, 435443.
  • Trivers RL (1972) Sexual selection and parental investment. In: Sexual Selection and the Descent of Man 1871–1971 (ed. CampbellB), pp. 136179. Aldine Press, New York, NY.
  • Williams GC (1957) Pleiotropy, natural selection, and the evolution of senescence. Evolution, 11, 398411.
  • Wittbrodt J, Adam D, Malitschek B et al. (1989) Novel putative receptor tyrosine kinase encoded by the melanoma-inducing Tu locus in Xiphophorus. Nature, 341, 415421.
  • Yang TL, Guo Y, Zhang LS et al. (2010) HMGA2 is confirmed to be associated with human adult height. Annals of Human Genetics, 74, 1116.
  • Young AR, Narita M (2007) Oncogenic HMGA2: short or small? Genes and Development, 21, 10051009.