Female reproductive senescence across mammals: A high diversity of patterns modulated by life history and mating traits

Highlights

• Reproductive senescence is widespread across mammals.

• Onset and rate of reproduction are influenced by the phylogeny.

• Onset and rate of reproduction vary with the species pace of life.

• Promiscuous females display an earlier age at the onset of reproductive senescence.

• Females that display an induced ovulation mode show an earlier at the onset of reproductive senescence.

Abstract

Senescence patterns are highly variable across the animal kingdom. However, while empirical evidence of actuarial senescence in vertebrates is accumulating in the wild and life history correlates of actuarial senescence are increasingly identified, both the extent and variation of reproductive senescence across species remain poorly studied. Here, we performed the first large-scale analysis of female reproductive senescence across 101 mammalian species that encompassed a wide range of Orders. We found evidence of reproductive senescence in 68.31 % of the species, which demonstrates that reproductive senescence is pervasive in mammals. As expected from allometric rules, the onset of reproductive senescence occurs later and the rate of reproductive senescence decreases with increasing body mass and delayed age at first reproduction. Moreover, for a given pace of life, females displaying a high level of multiple mating and/or with induced ovulation senesce earlier than females displaying a low level of multiple mating and/or with spontaneous ovulation. These results suggest that both female mating behavior and reproductive physiology shape the diversity of reproductive senescence patterns across mammals. We propose future avenues of research regarding the role played by environmental conditions or reproductive features (e.g. type of placentation) on the evolution of reproductive senescence.

Introduction

From an evolutionary biology perspective, the process of senescence corresponds to a decrease in the age-specific contribution to fitness (Cohen et al., 2020; Gaillard and Lemaître, 2020; Monaghan et al., 2008). Therefore, senescence is mostly studied through a decline in demographic rates with increasing age (i.e. actuarial and reproductive senescence). Historically, senescence was thought to not show up in the wild because animals were expected to die from harsh environmental conditions or anthropogenic effects (e.g. hunting) before experiencing a decline in age-specific survival or reproductive rates (Nussey et al., 2013). Thanks to the increasing number of long-term field studies - where individuals are monitored from birth to death (Clutton-Brock & Sheldon 2010), and the subsequent demographic analyses of these populations, the view that demographic senescence does not occur in the wild has been totally revisited (Nussey et al., 2013). Yet, after pioneer descriptive studies suggesting that mortality increases in late life (e.g. Caughley, 1966), empirical evidence currently indicates that actuarial senescence is pervasive across vertebrates with determinate growth in the wild (Promislow, 1991; Gaillard et al., 1994; Brunet-Rossinni and Austad, 2006; Nussey et al., 2013), whereas reproductive senescence was considered to be inexistent in the wild for a much longer period (e.g. Caughley, 1976). On the contrary, the negative effects of age on reproductive performance have been documented in laboratory rodents (e.g. Leslie and Ranson, 1940) and human populations (e.g. Tietze, 1957) for a while. The first reports of reproductive senescence in the wild only emerged in the late 70 s (Coulson and Horobin, 1976; Ollason and Dunnet, 1978; Perrins and Moss, 1974; Sinclair, 1977), following the insightful work by Emlen (1970), who stated that age-specific reproductive rates should increase to a peak, then stay relatively constant during a prime-age stage (i.e. the period of adulthood between the beginning of the reproductive peak and the onset of reproductive senescence) before declining in late life.

Empirical evidence of reproductive senescence from wild populations of vertebrates is now compelling (Nussey et al., 2013). So far, these case studies have been almost entirely performed on females (but see Lemaître and Gaillard, 2017 for a review of evidence in males), simply because accurate paternity assignments in the wild require the use of molecular tools. Age-specific declines in a wide array of traits reflecting female reproductive success have been documented, such as breeding proportions (e.g. Photopoulou et al., 2017), birth rates (e.g. Lee et al., 2016), number of offspring produced at birth (e.g. Sparkman et al., 2017) or at the end of parental care (i.e. at weaning or fledgling, e.g. Thorley et al., 2020), and offspring survival (e.g. Karniski et al., 2018). Interestingly, it has been suggested that female reproductive senescence patterns can markedly differ across species (Baudisch and Stott, 2019; Jones et al., 2014), even between closely related species. For instance, a population-level comparison of reproductive senescence in terms of breeding success across three albatross species has revealed that the age at the onset of reproductive senescence was much earlier in the wandering albatross (Diomedea exulans) than in both the black-browed albatross (Thalassarche melanophris) and the grey-headed albatross (Thalassarche chrysostoma) (Froy et al., 2017). Likewise, the southern fulmar (Fulmarus glacialoides) showed clear evidence of senescence in breeding success, whereas the snow petrel (Pagodroma nivea) did not (Berman et al., 2009). Despite these reports of contrasted patterns of age-specific decline in reproductive performance with increasing age, there has been so far no attempt to quantify the occurrence and the variation in reproductive senescence across a wide range of species, or to identify the ecological and biological factors underlying the diversity of patterns. This contrasts with the increasing number of comparative analyses that focused on actuarial senescence (Bronikowski et al., 2011; Lemaître et al., 2020c; Péron et al., 2019b; Ricklefs, 1998, 2006; Ricklefs and Scheuerlein, 2001), which have notably revealed that life-history strategies (e.g. Garratt et al., 2013; Ricklefs, 2010) and environmental conditions (Colchero et al., 2019; Lemaître et al., 2013; Tidière et al., 2016) modulate both the onset and the rate of actuarial senescence. So far, most comparative analyses of age-specific reproductive data in mammals have focused on the evolution of post-reproductive lifespan (e.g. Alberts et al., 2013; Cohen, 2004; Ellis et al., 2018) and did not investigate among-species differences in the onset or rate of reproductive senescence. Only Jones and colleagues (2008) analyzed interspecific differences in female reproductive senescence patterns (i.e. measured as the age-specific decline in the number of recruited offspring) across 19 species of birds and mammals. They found that the rate of reproductive senescence decreased with increasing generation time, a reliable measure of the species position along the slow-fast continuum (Gaillard et al., 2005). Thus, slow-living species had a lower rate of reproductive senescence than fast-living species (Jones et al., 2008; see also Gaillard et al., 2016).

The aim of this work is twofold. First, we assess whether female reproductive senescence is the rule rather than the exception across mammals by accurately quantifying reproductive senescence using a large sample of mammalian species that display a high diversity of lifestyles and life history strategies. This first step allowed quantifying among-species variation in both the onset and the rate of reproductive senescence. Second, we investigate the role of several ecological, biological and life history traits in shaping the diversity of reproductive senescence patterns observed across mammalian females. Among these factors, we focus on the role played by the phylogeny, the species position along the slow-fast continuum, the mating behavior and the ovulation mode.

Following comparative analyses of many life history traits (e.g. Kamilar and Cooper, 2013; Wootton, 1987), notably the rate of actuarial senescence (e.g. Lemaître et al., 2020b), we expected closely related species to share more similar reproductive senescence patterns than distant species across the phylogeny. We also expected the onset and the rate of senescence to occur earlier and to increase, respectively, with increasingly fast-living life history as a direct consequence of the covariation among all biological times (i.e. life history traits expressed in time units, see Gaillard et al., 2016; Ronget and Gaillard, 2020). Moreover, we investigated whether the number of mating partners and the ovulation mode influenced the timing and intensity of reproductive senescence. In most mammals, females generally copulate with more than one male during a given reproductive event (Gomendio et al., 1998; Hayssen and Orr, 2017; Soulsbury, 2010). The propensity of females to mate repeatedly within a reproductive event (or estrus) increases from monogamous species to polyandrous or polygynandrous species, which is considered as a way to increase fitness through genetic benefits (e.g. increased genetic diversity among offspring, Jennions and Petrie, 2000; Stockley, 2003). However, multiple mating promotes sperm competition (i.e. when sperm from two or more males compete to fertilize a given set of ova, Parker, 1970), which in turn can lead to intense sexual conflicts between sexes (Stockley, 1997). Physiological and subsequent fertility costs associated to these sexual conflicts have been widely documented (see Arnqvist and Rowe, 2005; Stockley, 1997 for reviews) and might increase with female age. Thus, the risk of contracting infectious diseases increases steadily with age in species displaying high levels of multiple mating (Nunn et al., 2014) and the negative consequences of such infections in terms of female fertility are likely to be amplified at old ages due to the progressive (and taxonomically widespread) decline in immune performance throughout the lifetime (Peters et al., 2019). We thus tested whether the degree of multiple mating by females leads to an earlier and/or steeper reproductive senescence. We also investigated whether the ovulation mode influences mammalian reproductive senescence. Ovulation is a physiological process governed by complex interactions between hormonal levels and external cues (e.g. photoperiod, mating) (Hayssen and Orr, 2017). Mammalian females are typically divided between spontaneous and induced ovulators (Soulsbury and Iossa, 2010). When spontaneous, ovulation is mostly triggered by endogenous hormonal changes, whereas with induced ovulation, the mating event triggers a physiological cascade that leads to ovulation. Comparative analyses performed so far have revealed that the intensity of sperm competition is stronger in species with spontaneous ovulation (Iossa et al., 2008; Soulsbury and Iossa, 2010). Since sperm competition can be associated with long-term aging costs in females, including a decline in age-specific physiological performance and fertility in females (Lemaître et al., 2020a; Stockley, 1997), we expect females from species with spontaneous ovulation to suffer from an earlier and/or steeper reproductive senescence than females from species with induced ovulation.