Sex steroids as mediators of phenotypic integration, genetic correlations, and evolutionary transitions

Highlights

• Regulation of gene expression by sex steroids has several important evolutionary implications.

• Sex steroids shape patterns of phenotypic variance and covariance that are available to selection.

• Sex steroids shape genetic correlations that determine evolutionary responses to selection.

• Sex steroids mediate evolutionary transitions via regulatory coupling and decoupling.

• Genomic and transcriptomic data from non-model species facilitate increasingly powerful studies.

Abstract

In recent decades, endocrinologists have increasingly adopted evolutionary methods and perspectives to characterize the evolution of the vertebrate endocrine system and leverage it as a model for developing and testing evolutionary theories. This review summarizes recent research on sex steroids (androgens and estrogens) to illustrate three ways in which a detailed understanding of the molecular and cellular architecture of hormonally mediated gene expression can enhance our understanding of general evolutionary principles. By virtue of their massively pleiotropic effects on the expression of genes and phenotypes, sex steroids and their receptors can (1) structure the patterns of phenotypic variance and covariance that are available to natural selection, (2) alter the underlying genetic correlations that determine a population's evolutionary response to selection, and (3) facilitate evolutionary transitions in fitness-related phenotypes via subtle regulatory shifts in underlying tissues and genes. These principles are illustrated by the author's research on testosterone and sexual dimorphism in lizards, and by recent examples drawn from other vertebrate systems. Mechanistically, these examples call attention to the importance of evolutionary changes in (1) androgen- and estrogen-mediated gene expression, (2) androgen and estrogen receptor expression, and (3) the distribution of androgen and estrogen response elements in target genes throughout the genome. A central theme to emerge from this review is that the rapidly increasing availability of genomic and transcriptomic data from non-model organisms places evolutionary endocrinologist in an excellent position to address the hormonal regulation of the key evolutionary interface between genes and phenotypes.

Introduction

Over the past several decades, the field of evolutionary endocrinology has emerged as a synthesis between evolutionary genetics and comparative endocrinology (Cox et al., 2016a; Nepomnaschy et al., 2009; Zera et al., 2007). Consequently, it is now routine for comparative studies of hormonal phenotypes to adopt formal phylogenetic methods (Goymann et al., 2018; Husak and Lovern, 2014; Johnson et al., 2018a; Vitousek et al., 2018). Likewise, comparative endocrinology has leveraged the availability of genetic and genomic data from non-model vertebrates to characterize the molecular evolution of hormone receptors (Eick and Thornton, 2011; Filowitz et al., 2018; Frankl-Vilches and Gahr, 2018; Sumiya et al., 2015), genes in their signaling pathways (Lorin et al., 2015), and recognition motifs in their target genes (Frankl-Vilches et al., 2015; Fuxjager and Schuppe, 2018). In wild populations, phenotypic selection analyses have been adapted to measure natural selection on both baseline hormone levels and those induced by experimental stress or hormone manipulation (Bonier and Cox, 2020, Bonier and Martin, 2016; Cox et al., 2016b; John-Alder et al., 2009; McGlothlin et al., 2010; Patterson et al., 2014). To assess the evolutionary consequences of such selection, quantitative genetic studies of heritability and genetic correlations are also being used to characterize the genetic architecture of endocrine phenotypes and thereby predict how they will respond to selection (Cox et al., 2016b; Iserbyt et al., 2015; Pavitt et al., 2014; Stedman et al., 2017).

The approaches described above represent the successful application of foundational evolutionary techniques to explore the evolution of the endocrine system itself. However, there are also many ways in which the study of endocrinology can inform broader theoretical issues in evolutionary biology. For example, the divergence of androgen, progestogen, corticosteroid, and mineralocorticoid receptors following gene duplication of an ancestral estrogen receptor has been used to explore evolutionary concepts ranging from exaptation and irreducible complexity (Bridgham et al., 2006; Thornton, 2001) to the importance of historical contingency and resultant predictability and reversibility of evolution (Bridgham et al., 2009; Harms and Thornton, 2014; Starr et al., 2017). Likewise, the recognition that most major axes of the vertebrate endocrine system are both evolutionarily conserved and responsible for pleiotropically regulating numerous genes and phenotypes has stimulated considerable discussion about the extent to which this type of regulatory architecture acts as a source of evolutionary potential versus constraint (Fuxjager and Schuppe, 2018; Hau, 2007; Ketterson and Nolan, 1999; Lema, 2014; McGlothlin and Ketterson, 2008). In essence, this is an endocrine-centric version of the much broader question of whether and how evolutionary trajectories are influenced by the functional mapping of genes to phenotypes (Pigliucci, 2010; Wagner and Zhang, 2011).

The goal of this paper is to call attention to several ways in which the study of hormones in general, and sex steroids in particular, can inform broader questions in evolutionary biology. For simplicity, this review focuses on androgens and estrogens, though many of the underlying principles are generalizable other hormones, particularly those that exert genomic effects by altering transcription of target genes. Although sex steroids can produce rapid (non-genomic) cellular responses without altering gene expression (Lösel and Wehling, 2003), their ability to pleiotropically regulate the transcription of hundreds to thousands of genes is key to their evolutionary significance, so this review begins with a brief discussion of the mechanisms by which sex steroids link genes to phenotypes via transcription. The coordinated regulation of multiple phenotypes by androgenic and estrogenic pathways is predicted to structure patterns of phenotypic integration (Fig. 1), thereby shaping the trait combinations that are available to selection. Although the underlying principle of hormonal pleiotropy (Fig. 1A) is widely recognized, it has yet to be rigorously incorporated into a quantitative genetic framework centered on phenotypic variance and covariance. Moreover, because sex steroids regulate gene expression, they also have the potential to shape the underlying patterns of genetic covariance (Fig. 1C) that determine how populations evolve in response to natural and sexual selection on correlated traits. For example, sex steroids can orchestrate the developmental breakdown of genetic correlations between males and females, thereby facilitating sex-specific phenotypic evolution despite the constraint of a shared genome.

Shifting from a quantitative genetics (microevolutionary) framework to a phylogenetic (macroevolutionary) perspective, several recent examples are used to illustrate how evolutionary transitions in sexually selected phenotypes have been linked to the molecular and cellular coupling (or decoupling) of phenotypes to (or from) androgenic and estrogenic signaling. These couplings can occur at the tissue level, through changes in receptor density, cofactor availability, and resultant sensitivity to sex steroids, or at the gene level, through the gain or loss of recognition motifs in regulatory regions. In extreme cases, evolutionary reversals in sexual dimorphism may be achieved by reversing the effects of a hormone on a given phenotype, as illustrated by the “bipotential” nature of testosterone as both a promoter and an inhibitor of growth across reptiles. The cellular and molecular mechanisms that facilitate such evolutionary reversals are not yet known, but represent a promising path for future research. Collectively, these topics illustrate how the increasing accessibility of genomic and transcriptomic data for non-model organisms can be leveraged to address emerging hypotheses about the roles of sex steroids in shaping phenotypic integration, genetic correlations, and evolutionary transitions.