Transcriptomic support for the Immunocompetence Handicap Hypothesis but not the Oxidation Handicap Hypothesis

Sexually selected traits are hypothesized to be honest signals of individual quality due to the costs associated with their development or expression. Testosterone, a sex steroid known to influence the production of sexually selected traits, has been proposed to underlie the costs associated with sexually selected traits via its immunosuppressive effects (i.e., the Immunocompetence Handicap Hypothesis) or by influencing an individual’s exposure/susceptibility to oxidative stress (i.e., the Oxidation Handicap Hypothesis). Previous work testing these hypotheses has primarily focused on physiological measurements of immunity or oxidative stress, but little is known about the molecular pathways by which testosterone could influence immunity and/or oxidative stress pathways. To measure the molecular consequences of experimentally elevated testosterone, we used previously published RNA-seq data from studies that measured the transcriptome of individuals treated with either a testosterone-filled or an empty (i.e., control) implant. Two studies encompassing two species of bird and three tissue types fit our selection criteria. We found strong support for the Immunocompetence Handicap Hypothesis, but no support for the Oxidation Handicap Hypothesis. More specifically, testosterone-treated individuals exhibited strong signatures of immunosuppression, encompassing both cell-mediated and humoral immunity. Our results suggest that testosterone enforces the honesty of sexually-selected traits by influencing an individual’s immunocompetence rather than their exposure or susceptibility to oxidative stress.


37
There is a long-standing interest in understanding why sexually selected traits have evolved and 38 one hypothesis suggests that mates have selected for traits that are costly to develop or bear (i.e., 39 the handicap hypothesis; Zahavi, 1975). An important assumption of the handicap hypothesis is 40 that an individual's investment in sexually selected traits correlates with their investment in other 41 traits that also influence their reproductive success or survival (Grafen, 1990;Andersson, 1994). 42 Individuals face tradeoffs when fitness-related traits exhibit negative correlations and, as a result, 43 individuals can incur survival costs from their reproductive investments (Stearns, 1992). These 44 costs arise because the development and/or expression of traits important for reproduction (e.g., 45 sexually selected traits) and traits important for survival (e.g., immune function) are dependent 46 2 on the same mechanism (Zera and Harshman, 2001). As such, our understanding of the evolution 47 of sexually selected traits is dependent upon our understanding of the pleiotropic nature of the 48 mechanisms that underlie their production (Kokko et al., 2003).

50
Testosterone is a sex steroid that is known to influence the development and/or 51 expression of sexually selected traits (Hau, 2007;Fusani, 2008;Ball and Balthazart, 2009). In   Here, we use published RNA-seq datasets to further examine the effects of testosterone 108 on the transcriptome. Specifically, we re-analyze studies that compared gene expression between     Table 1). The yellow module was primarily enriched for broad immune categories, e.g.,

193
"immune system process" and "immune response", whereas the dark green module was  Table 1). The green module (795 genes, r=0.61) was primarily enriched for broad 204 metabolic activity and protein modification processes.  Table 2).

225
In this study, we quantified transcriptional responses to experimentally increased circulating 226 testosterone in two species of bird. Our gene network analysis revealed that both manakin and 227 quail exhibit immunosuppression following testosterone treatment, supporting the 228 6 Immunocompetence Handicap Hypothesis. However, we did not find support for the Oxidation 229 Handicap Hypothesis, as there was no enrichment of genes expressed related to oxidative 230 damage, nor suppression of genes related to antioxidant defenses in either species. These results 231 suggest that high levels of circulating testosterone can be costly to maintain partly due to their 232 potential negative effects on an individual's immune response and not the individual's 233 susceptibility or exposure to oxidative stress. Importantly, oxidative stress could still be involved 234 in enforcing the costs of reproduction or sexually selected traits; however, our results suggest 235 that this cost is not borne out via molecular pathways that are sensitive to testosterone, at least in 236 the tissues and species examined here.

238
Our analyses revealed that transcriptomic immunosuppression was broad, encompassing 239 aspects of both innate immunity (e.g., leukocyte activation and cytokine signaling) as well as 240 adaptive immunity (e.g., antigen processing and presentation) across both species (Table 1) Table 1) and manakin (Table 1). In addition to suppression of T cell activity in 254 manakin, we also identified MHC class IA as a hub gene in the manakin dark turquoise module.   individual sampling approach as this allows for a more robust test of testosterone's effect on 290 gene expression (Williams, 2008). Overall, these integrative, mechanistic approaches will 291 ultimately provide novel insights into the evolution of sexually selected traits.

293
Acknowledgements 294 We would like to thank Ignacio Moore, Dana Hawley, and Christopher Balakrishnan for 295 providing helpful feedback on an earlier version of this manuscript.