Intergroup variation in oestrogenic plant consumption by black-and-white colobus monkeys

1 INTRODUCTION

Diet strongly influences the morphology, physiology and behaviour of primates (Porter, Gilbert, & Fleagle, 2014; Turner & Thompson, 2013), and recent studies have linked nutrition to epigenetic changes in the human genome (Haggarty, 2013). Although nutritional ecology often focuses on the influence of nutrients on genetic, physiological and evolutionary change (English & Uller, 2016), it also attempts to elucidate the importance of non-nutritive phytochemicals (Huffman, 2017; Villalba, Costes-Thiré, & Ginane, 2017; Wasserman, Milton, & Chapman, 2013). One type of non-nutritive compound found in some plants is endocrine-active phytochemicals (EAPs) (Wasserman et al., 2013). EAPs directly interact with vertebrate endocrine systems, potentially influencing animal physiology and behaviour (Lambert & Edwards, 2017). Common examples of EAPs include phytoestrogens and phytoandrogens.

Phytoestrogens are functionally similar to estradiol, the primary female sex hormone (Sirotkin & Harrath, 2014; Wadekar, Shah, Bagul, Bagul, & Patil, 2011). Human consumption of phytoestrogens is associated with an array of health benefits, including protection from certain cancers and heart disease, improved neurological health and postmenopausal therapy (Panay, 2011; Soni et al., 2014; Wadekar et al., 2011). However, these benefits are not conclusive due to a general lack of data (Rietjens, Louisse, & Beekmann, 2017). In contrast, phytoestrogens are also associated with infertility in humans and agricultural animals (Adams, 1995; Xia et al., 2013), although these relationships are variable among individuals (Qin et al., 2014).

Research that elucidates the role of EAPs in wild primate diets is limited. Simon, Kaplan, Hu, Register, and Adams (2004) found that aggressive encounters increased, and affiliative behaviours decreased when adult male cynomolgus macaques (Macaca fascicularis) were fed a diet high in oestrogenic isoflavones. However, this type of manipulated diet in a laboratory setting does not explain how wild primates interact with EAPs. Phytoprogesterone-containing Vitex has been suggested to alter progesterone levels and mating behaviour in leaf monkeys, chimpanzees and baboons (Emery Thompson, Wilson, Gobbo, Muller, & Pusey, 2008; Lu et al., 2011). For phytoestrogens, Whitten (1983) first explored their importance to non-human primates, while Wasserman, Chapman, et al. (2012) built on this research to find that red colobus phytoestrogen consumption in males was related to increased faecal estradiol and cortisol levels, along with decreased grooming and increased aggression and copulation rates. Whether these changes in physiology and behaviour that were correlated with phytoestrogen consumption influence fitness remains to be tested.

Based on effects of phytoestrogens on humans, captive animals and livestock, it is clear that small amounts of dietary phytoestrogens can have significant biological effects on individuals (Welshons et al., 2003). However, if phytoestrogen consumption acts as a selective pressure through modification of physiology and behaviour, then variation across individuals and groups must exist in either exposure or susceptibility. Most primates live in cohesive groups with individuals in the same group having similar diets, while different groups tend to show significant dietary variation (Chapman, Chapman, & Gillespie, 2002). Nonetheless, studies of EAP consumption are limited to contrasts within one group. Studying multiple groups of the same species can provide useful insights into differential exposure to, and biological effects of, EAPs. Thus, the goal of this study was to quantify variation in phytoestrogen consumption across different groups of black-and-white colobus monkeys (Colobus guereza) living in Kibale National Park, Uganda, and examine the relationship between consumption and grooming.

2 MATERIALS AND METHODS

Kibale National Park (KNP; 795 km²), located in western Uganda (0.13′–0.41′N and 30.19′–30.32′E), is mid-altitude forest in the foothills of the Rwenzori Mountains. Life-history traits, group size and home ranges of our study groups are described elsewhere (Harris & Chapman, 2007; Harris, Chapman, & Monfort, 2009; Harris & Monfort, 2006). Two researchers collected behavioural data from eight habituated groups near the Kanyawara research station in KNP from June 2006 to May 2007 for three to five days per week. Observers used instantaneous sampling, scanning the group every 15 min, typically from dawn until dusk. Groups in this study ranged from about 5–10 individuals, so sampling an entire group was feasible. Activities recorded included feeding, sleeping, playing, resting, moving, grooming and self-grooming. Here, we focus only on feeding and grooming behaviour given our interests in phytoestrogen consumption and its relationship to an affiliative social behaviour. All activities were averaged into monthly frequencies for statistical analysis. Staple food items that constituted >1% of a group's diet were compared across months for each group and across groups.

We identified oestrogenic plant foods using methods previously described by Wasserman, Taylor-Gutt, et al. (2012). In brief, plant food items were collected from KNP from 2007 to 2008 and dried at the field station. The dried plants were processed and screened for oestrogenic activity using transient transfection assays at the University of California, Berkeley. Samples were only screened for activity at oestrogen receptor beta (ERβ). Indeed, none of our previous primate plant food items had activity at ERα; those items with oestrogenic activity were only active at ERβ (Wasserman, Taylor-Gutt, et al., 2012). The transient transfection assay determines oestrogenic activity by quantifying light production from a human cell line transfected with ERE-tk-Luc, which consists of oestrogen response element linked to luciferase gene. The amount of light produced is measured with a luminometer after adding a plant extract to the cells. These samples are compared to a negative control of cells with nothing added and a positive control of cells with 10 nM of oestradiol added. If adding a plant extract to the transfected U2OS cells results in at least a twofold increase in light production as compared to the negative control, then the sample is classified as oestrogenic. All samples and controls were run in triplicate, and if oestrogenic activity was detected, that sample was assayed again to confirm the positive result. This transfection assay gives neither indication of the specific chemical responsible for this activity nor quantification of concentration.

Based on the transfection assays, we composed a list of all oestrogenic plants in the diet of each group and determined the percentage of oestrogenic plant items consumed each month. Single factor ANOVAs were used to identify significant differences in oestrogenic plant consumption between groups (n = 8 groups) over the year with month being considered independent. We also used Pearson's product–moment correlations to test for relationships between oestrogenic plant consumption and grooming activities.

All research herein complied with all regulations regarding the study of free ranging wildlife and with Ugandan and U.S. laws.

3 RESULTS

Thirteen plant food items were screened for oestrogenic activity and combined with 19 items previously screened by Wasserman, Taylor-Gutt, et al. (2012) to cover 82.3% of the overall black-and-white colobus diet. On average, 78.8% of the black-and-white colobus diet came from 14 dietary items (Table 1). Eight of the 32 items screened had oestrogenic activity (Table 2) and were from five species of four plant families: Fabaceae, Moraceae, Myrtaceae and Oleaceae. None of these oestrogenic items were from the 14 staple food items of the average diet. Summing consumption of all oestrogenic items (including all rarely fed on items, <1% of diet), at least 3.4% of the average black-and-white colobus diet consisted of oestrogenic plant parts.

All groups studied here relied heavily on Celtis durandii as their primary food source (Appendix A). This is consistent with previous studies on C. guereza diet in Kibale (Harris & Chapman, 2007). Other foods showed variation in consumption across the groups. Mean percentage of oestrogenic plant consumption across groups ranged from 1.24 ± 1.65% [Birungi] to 5.85 ± 4.51% [Zikuru] (Table 3), but there were no significant differences across groups (F = 1.64, p = 0.14, df = 7). The number of oestrogenic staple foods within groups ranged from zero to two and averaged 1.92% ± SD 1.27 of the overall diet.

There were significant differences between groups in per cent of time spent grooming (F = 3.84, p < 0.005, df = 7) and self-grooming (F = 3.24, p < 0.005, df = 7). However, only one group demonstrated a significant negative relationship between oestrogenic plant consumption and grooming (t = 2.61, p = 0.026, ρ = −0.64, df = 10), with the other seven showing no relationship.

4 DISCUSSION

Approximately 3.4% of the average C. guereza diet came from oestrogenic plant foods, and there was no significant difference across groups. This amount is less than half that previously observed for Uganda red colobus monkeys (Procolobus rufomitratus) of KNP and mountain gorillas (Gorilla beringei) of Bwindi NP (10.6% and 8.8%, respectively) (Wasserman, Taylor-Gutt, et al. (2012). The most consumed oestrogenic items across the eight groups were Ficus sansibarica (UF, YL, ML) (1.1%), Eucalyptus grandis (BA) (0.7%), Millettia dura (YL) (0.6%), Olea capensis(ML) (0.6%) and Erythrina abyssinica (YL, FL) (0.5%). These finds support those of Wasserman, Taylor-Gutt, et al. (2012) on the importance of Fabaceae, Moraceae and Myrtaceae as sources of phytoestrogens in primate diets.

After identifying the plant species acting as sources of phytoestrogens, determination of biological effects is needed. However, this is particularly difficult in a wild setting since the phytochemicals with oestrogenic activity vary from species to species and also vary in their range of physiological effects despite a similar hormone-mimicking mechanism of action. For example, oestrogenic coumestrol is 20% as potent as mammalian oestradiol, while phytoestrogens classified as formononetins are less than 0.1% as potent (Anderson & Garner, 1997). To further complicate the biological effects, if these less potent phytoestrogens are competing with endogenous estradiol for oestrogen receptors, then they could have an anti-oestrogen effect. In addition, some plant species even have their own variety of different oestrogenic compounds with different concentrations and potencies (Knight & Eden, 1996). Further, oestrogenic plants have spatiotemporal variation in phytoestrogen concentrations related to rainfall, temperature, UV radiation and other abiotic and biotic factors. For these reasons, future research should classify the specific phytoestrogens in primate diets and quantify their concentrations over spatial and temporal scales.

Nonetheless, Wasserman, Chapman, et al. (2012) found that as red colobus monkeys consumed more oestrogenic plants, the higher their faecal estradiol and cortisol levels, the more they fought and mated, and the less they groomed. Simon et al. (2004) found that monkeys fed higher levels of oestrogenic isoflavones displayed less body contact, more solitary behaviour and had more intense aggression and submission. Although we lacked data on hormone levels, mating or aggression for the black-and-white colobus, we did find that only one group showed a significant negative relationship between oestrogenic plant consumption and grooming, similar to the other previous studies. The lack of an effect in the other seven groups could be due to the lower levels of phytoestrogens in their diets and the presence of a threshold which only produces physiological and behavioural effects when crossed. To determine whether this threshold exists, future research should examine the hormone profile and behavioural repertoire of specific individuals over time, rather than exploring these trends at the group level. It will be particularly useful to explore differences before and after large amounts of oestrogenic plant consumption (Wasserman, Chapman, et al., 2012; Wasserman, Taylor-Gutt, et al., 2012).

Such studies will also be useful in testing the three hypotheses postulated to explain the ecological relationship between primates and oestrogenic plants: (a) the plant defence hypothesis suggests EAPs cause detrimental physiological effects for the primate, (b) the self-medication hypothesis suggests that these plants provide primates with a physiological benefit and (c) the biochemical coincidence hypothesis suggests there is neither benefits to plants nor primates (Wasserman et al., 2013). Testing these relationships requires documentation of the physiological effects of consumption but may also benefit from measuring exposure due to dietary niche differences in the chemistry of ingested plant parts based on plant–animal interactions such as herbivory and seed dispersal. Teasing apart functional trait diversity and phylogenetic trait diversity is increasingly common for explaining current ecological phenomena (Molnár et al., 2018). Additionally, exploring phylogenetic relationships may elucidate coevolutionary plant–herbivore relationships (Endara et al., 2017). Evolutionary relationships could therefore explain the distribution of EAPs across plant families, especially if they deter herbivory or promote mutualistic relationships.

CONFLICT OF INTEREST

The auths have no conflict of interest to declare.