Reproductive behaviour, pair-bonding, and biparental care

Females, although capable of reaching reproductive maturity and becoming sexually active at the time of weaning, require exposure to an unrelated, intact male or his soiled bedding to attain reproductive maturity (e.g., increased uterine weight and displays of behavioural oestrus in the form of lordosis; Carter et al., Reference Carter, Getz, Gavish, McDermott and Arnold1980, Reference Carter, Witt, Schneider, Harris and Volkening1987; Witt et al., Reference Witt, Carter, Carlstead and Read1988). Mating in prairie voles occurs in extended bouts often lasting more than 24 hours (Witt et al., Reference Witt, Carter, Carlstead and Read1988). Unlike most mammals, after mating, prairie voles form long-term, quasi-monogamous social pair bonds, as well as engage in biparental care.

Although not all individuals are strictly sexually monogamous (Solomon et al., Reference Solomon, Keane, Knoch and Hogan2004), evidence of pair bond formation has been recorded both in the field and laboratory (Getz et al., Reference Getz, Carter and Gavish1981; Getz and Hofmann, Reference Getz and Hofmann1986; Roberts et al., Reference Roberts, Williams, Wang and Carter1998; Barrett et al., Reference Barrett, Keebaugh, Ahern, Bass, Terwilliger and Young2013). Evidence of such relationships in the field comes from live-trap data of specific animals regularly captured near particular nests or repeatedly trapped with the same opposite-sex partner over time (Getz et al., Reference Getz, Carter and Gavish1981; Getz and Hofmann, Reference Getz and Hofmann1986). In laboratory settings, pair bonding is operationally measured as a preference for spending time with a familiar opposite-sex partner with which an individual has mated or cohabited over a novel opposite-sex stranger, as well as increased aggression towards strangers of both sexes (Getz et al., Reference Getz, Carter and Gavish1981; Winslow et al., Reference Winslow, Hastings, Carter, Harbaugh and Insel1993; Roberts et al., Reference Roberts, Williams, Wang and Carter1998; Barrett et al., Reference Barrett, Keebaugh, Ahern, Bass, Terwilliger and Young2013).

Paternal care has been inferred from field observations of males residing in nests with partner females and their offspring, as well as recorded in laboratory settings (Getz and Hofmann, Reference Getz and Hofmann1986; Roberts et al., Reference Roberts, Williams, Wang and Carter1998; Terleph et al., Reference Terleph, Jean-Baptiste and Bamshad2004; McGuire and Bemis, Reference McGuire and Bemis2007; Ophir et al., Reference Ophir, Phelps, Sorin and Wolff2007). Variability in paternal care has also been observed between populations. Male prairie voles from Kansas, for example, are less paternal than males from Illinois, potentially due to differences in responses to peripheral vasopressin, neural sensitivity to oestrogen, and their sometimes more xeric environments (Roberts et al., Reference Roberts, Williams, Wang and Carter1998; Cushing et al., Reference Cushing, Martin, Young and Carter2001, Reference Cushing, Razzoli, Murphy, Epperson, Le and Hoffman2004).

Oxytocin and vasopressin

Central oxytocin and vasopressin activity, often respectively in the female nucleus accumbens (NACC) and male ventral pallidum (VP), have long been associated with social behaviours in prairie voles (Winslow et al., Reference Winslow, Hastings, Carter, Harbaugh and Insel1993; Williams et al., Reference Williams, Insel, Harbaugh and Carter1994; Liu and Wang, Reference Liu and Wang2003; Lim et al., Reference Lim, Wang, Olazábal, Ren, Terwilliger and Young2004; Keebaugh and Young, Reference Keebaugh and Young2011; Barrett et al., Reference Barrett, Keebaugh, Ahern, Bass, Terwilliger and Young2013; Berendzen et al., Reference Berendzen, Sharma, Mandujano, Wei, Rogers, Simmons, Seelke, Bond, Larios, Goodwin, Sherman, Parthasarthy, Espineda, Knoedler, Beery, Bales, Shah and Manoli2023). Research has demonstrated that infusions of oxytocin into the lateral ventricle of ovariectomized females induce a preference for a partner male over a novel stranger, while simultaneously infusing an oxytocin antagonist blocks partner preference formation (Williams et al., Reference Williams, Insel, Harbaugh and Carter1994; Liu and Wang, Reference Liu and Wang2003). Furthermore, overexpression of oxytocin receptors (OTRs) in the female NACC also induces more affiliative behaviours towards partners by females and more alloparental behaviours towards unfamiliar pups (Keebaugh and Young, Reference Keebaugh and Young2011). Notably, however, one recent study utilizing prairie voles lacking functional OTRs has called into question the necessity of oxytocin for pair bonding in prairie voles, as both males and females lacking OTRs still exhibited pair bonding (Berendzen et al., Reference Berendzen, Sharma, Mandujano, Wei, Rogers, Simmons, Seelke, Bond, Larios, Goodwin, Sherman, Parthasarthy, Espineda, Knoedler, Beery, Bales, Shah and Manoli2023). Similar studies of the role of vasopressin in social monogamy in male prairie voles demonstrate that blocking V1a receptors (V1aRs) in the brain before mating can prevent selective aggression towards unfamiliar conspecifics and partner preference formation, while downregulating V1aRs in the VP impairs partner preference formation without impacting alloparental behaviours (Winslow et al., Reference Winslow, Hastings, Carter, Harbaugh and Insel1993; Barrett et al., Reference Barrett, Keebaugh, Ahern, Bass, Terwilliger and Young2013).

Dopamine, corticotropin releasing factor, corticosterone, and opioid receptors

Research into the roles of other potential neurochemical and endocrine mediators of social behaviour highlights several additional mechanisms, with many acting in the NACC and other parts of the striatum (e.g., caudate putamen [CP]). In the NACC, the binding of dopamine to D2-like receptors (D2Rs) facilitates pair bond formation while binding to D1-like receptors (D1Rs) prevents it (Liu and Wang, Reference Liu and Wang2003; Aragona et al., Reference Aragona, Liu, Yu, Curtis, Detwiler, Insel and Wang2006; Loth and Donaldson, Reference Loth and Donaldson2021). CRF activity in the NACC, likely mediated through CRF1 and CRF2 receptors, helps facilitate partner preference formation in males (Lim et al., Reference Lim, Liu, Ryabinin, Bai, Wang and Young2007). In females, μ-opioid receptors (MORs) in the rostral dorsomedial area of the NACC shell and CP help facilitate partner preference formation, while MORs in the dorsal striatum are important for mating behaviour (Burkett et al., Reference Burkett, Spiegel, Inoue, Murphy and Young2011; Resendez et al., Reference Resendez, Dome, Gormley, Franco, Nevárez, Hamid and Aragona2013). Additionally, κ-opioid receptors in the NACC shell play a role in selective aggression, although MORs in this region do not (Resendez et al., Reference Resendez, Kuhnmuench, Krzywosinski and Aragona2012). Peripherally, CORT, a stress hormone released downstream of the actions of CRF, blocks the formation of new pair bonds but does not effect established pair bonds when administered to female prairie voles (DeVries et al., Reference DeVries, DeVries, Taymans and Carter1995; Cacioppo et al., Reference Cacioppo, Cacioppo, Capitanio and Cole2015a).

Effects of social stress on social mammals

For social mammals, extended social isolation is a form of chronic stress that affects such organisms through the same systems as more standard stressors and has comparable effects on health and well-being (Hawkley et al., Reference Hawkley, Cole, Capitanio, Norman and Cacioppo2012; Cacioppo et al., Reference Cacioppo, Cacioppo, Capitanio and Cole2015a). In humans, perceived loneliness is associated with premature death and reduced mental, cardiovascular, cognitive, and metabolic health (Cacioppo et al., Reference Cacioppo, Grippo, London, Goossens and Cacioppo2015b; Valtorta et al., Reference Valtorta, Kanaan, Gilbody, Ronzi and Hanratty2016; Holt-Lunstad, Reference Holt-Lunstad2018; Sutin et al., Reference Sutin, Stephan, Luchetti and Terracciano2020; Henriksen et al., Reference Henriksen, Nilsen and Strandberg2023). Additionally, isolation following a stressful life event is associated with greater symptoms of clinical illness following rhinovirus exposure (Cohen et al., Reference Cohen, Janicki-Deverts, Doyle, Miller, Frank, Rabin and Turner2012).

Given the highly social nature of prairie voles and extensive research on the biological basis of their social behaviours, these animals have long been considered an excellent model organism for exploring questions about the impact of social stressors on social mammals in general and social isolation more specifically. In social isolation studies utilizing prairie voles, animals are generally housed either with a same-sex sibling or in isolation for 4–8 weeks, then behaviourally assessed before euthanasia and tissue sample collection (Grippo et al., Reference Grippo, Cushing and Carter2007a, Reference Grippo, Gerena, Huang, Kumar, Shah, Ughreja and Carter2007b, Reference Grippo, Lamb, Carter and Porges2007c, Reference Grippo, Wu, Hassan and Carter2008; Normann et al., Reference Normann, Cox, Akinbo, Watanasriyakul, Kovalev, Ciosek, Miller and Grippo2021).

Behavioural assessments typically entail placing animals in one or more behavioural assays to measure proxies of anxiety and depression. The elevated plus maze (EPM), forced swim test (FST), and sucrose preference test (SPT) are among those that have been utilized (Grippo et al., Reference Grippo, Cushing and Carter2007a, Reference Grippo, Gerena, Huang, Kumar, Shah, Ughreja and Carter2007b, Reference Grippo, Lamb, Carter and Porges2007c, Reference Grippo, Wu, Hassan and Carter2008; Normann et al., Reference Normann, Cox, Akinbo, Watanasriyakul, Kovalev, Ciosek, Miller and Grippo2021). The EPM is used to assess behavioural signs of anxiety (Walf and Frye, Reference Walf and Frye2007; Normann et al., Reference Normann, Cox, Akinbo, Watanasriyakul, Kovalev, Ciosek, Miller and Grippo2021). The FST and SPT are behavioural tests of depressive symptomology (Grippo et al., Reference Grippo, Moffitt and Johnson2002, Reference Grippo, Beltz and Johnson2003, Reference Grippo, Wu, Hassan and Carter2008; Strekalova et al., Reference Strekalova, Spanagel, Bartsch, Henn and Gass2004). Sometimes, additional physiological measures (e.g., heart rate and heart rate variability) are also collected (Grippo et al., Reference Grippo, Lamb, Carter and Porges2007c, Reference Grippo, Wu, Hassan and Carter2008).

Analyses generally show isolated prairie voles engage in more behaviours indicative of anxiety and depression than paired controls (Grippo et al., Reference Grippo, Cushing and Carter2007a, Reference Grippo, Gerena, Huang, Kumar, Shah, Ughreja and Carter2007b, Reference Grippo, Lamb, Carter and Porges2007c, Reference Grippo, Wu, Hassan and Carter2008; Normann et al., Reference Normann, Cox, Akinbo, Watanasriyakul, Kovalev, Ciosek, Miller and Grippo2021). Additionally, isolated prairie voles demonstrate more signs of cardiovascular dysfunction, including tachycardia and decreased heart rate variability, likely due to dysregulated autonomic nervous system (ANS) activity (Grippo et al., Reference Grippo, Lamb, Carter and Porges2007c, Reference Grippo, Wu, Hassan and Carter2008). Research on social isolation’s cardiovascular and autonomic effects indicates isolated prairie voles physiologically respond to non-threatening environments in the same manner they respond to threatening ones, as well as take longer to recover from additional social stressors like being placed in the cage of an unfamiliar conspecific as part of a resident intruder test (RIT; Grippo et al., Reference Grippo, Lamb, Carter and Porges2007c).

Studies examining the neurophysiological and hormonal effects of social isolation on prairie voles beyond the ANS, although sometimes inconsistent, generally reveal changes in parts of the brain and in central and peripheral neurohormones associated with both stress and social behaviour such as increased CRF immunoreactivity in the paraventricular nucleus of the hypothalamus (PVN) and vasopressin immunoreactivity in the supraoptic nucleus of the hypothalamus (Ruscio et al. Reference Ruscio, Sweeny, Hazelton, Suppatkul and Sue Carter2007).

Some sex-specific differences in responses to social isolation have also been found in prairie voles. Females, but not males, show an increase in circulating OT, an increase in OT immunoreactive cell density in the PVN, and a decrease in AVP immunoreactivity in the PVN (Grippo et al., Reference Grippo, Gerena, Huang, Kumar, Shah, Ughreja and Carter2007b; Ruscio et al. Reference Ruscio, Sweeny, Hazelton, Suppatkul and Sue Carter2007). Findings regarding the impact of social isolation on circulating CORT have been inconsistent (Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b; Grippo et al., Reference Grippo, Gerena, Huang, Kumar, Shah, Ughreja and Carter2007b; Ruscio et al., Reference Ruscio, Sweeny, Hazelton, Suppatkul and Sue Carter2007).

Research assessing whether social isolation alters how prairie voles respond to additional social stressors, such as those experienced in the RIT, shows that following the RIT, isolated females exhibit elevated circulating oxytocin, vasopressin, and CORT, as well as more oxytocin and CRF immunoreactive cells in the PVN (Grippo et al., Reference Grippo, Cushing and Carter2007a, Reference Grippo, Gerena, Huang, Kumar, Shah, Ughreja and Carter2007b). Circulating ACTH has sometimes been shown to be elevated in isolated females following the RIT (Grippo et al., Reference Grippo, Cushing and Carter2007a, Reference Grippo, Gerena, Huang, Kumar, Shah, Ughreja and Carter2007b). The impact of social isolation on responses to other social stressors for males appears less pronounced, with isolated males showing increases in oxytocin immunoreactive cells in the PVN and circulating oxytocin following the RIT but no changes in CRF immunoreactive cells in the PVN or circulating ACTH or CORT (Grippo et al., Reference Grippo, Gerena, Huang, Kumar, Shah, Ughreja and Carter2007b).

Compared to other rodent models of stress (e.g., chronic unpredictable mild stress [CUMS]), the behavioural and physiological effects of social isolation on prairie voles are broadly consistent (Grippo et al., Reference Grippo, Moffitt and Johnson2002, Reference Grippo, Beltz and Johnson2003; Walf and Frye, Reference Walf and Frye2007; Jianguo et al., Reference Jianguo, Xueyang, Cui, Changxin and Xuemei2019).

Prairie vole gastrointestinal anatomy and physiology

Early research examining the prairie vole gastrointestinal system largely focused on basic anatomy and physiology, often approaching questions from an ecological perspective. This work elucidated how food intake by prairie voles appears to be limited by the amount of time necessary for digestion, thus requiring animals to engage in regular feeding bouts every 2–4 hours, interspersed with periods of rest (Zynel and Wunder, Reference Zynel and Wunder2002).

Such studies also provided morphometric and functional details concerning different prairie vole gastrointestinal areas, revealing the caecum to have a greater nominal surface area compared to other hindgut regions and considerable acetate abundance in the caecum and proximal colon (Hume et al., Reference Hume, Karasov and Darken1993). This work also demonstrated proline uptake in the small intestine via an active transport system, greater acetate uptake in the distal colon than in the caecum or proximal colon through a passive mechanism when acetate is present in high concentrations, and equal levels of acetate uptake in the proximal and distal colon via a mediated system when acetate is present in low concentrations, as might occur during periods of fasting or starvation (Hume et al., Reference Hume, Karasov and Darken1993).

Prairie vole gastrointestinal development

More recently, a limited number of studies have examined histological, cellular, and molecular aspects of the prairie vole gastrointestinal system. One of the most comprehensive studies, Supeck et al. (Reference Supeck, Assefa, Meek, Curtis and Köhler2018), provided a detailed description of the postnatal maturation of the intestinal epithelial barrier in prairie vole pups. Using animals ranging in age from 2 to 8 weeks, Supeck et al. (Reference Supeck, Assefa, Meek, Curtis and Köhler2018) characterized the development of the prairie vole intestinal epithelial barrier using a variety of techniques. Dextran permeability assays revealed greater intestinal permeability in 2-week-old animals compared to 3-week-old animals. Histological comparisons of colon tissue from 1- and 3-week-old animals demonstrated that 3-week-old animals had a greater number of goblet cells, increased smooth muscle width, and greater gland length and distance. Examination of the expression of genes involved in tight junction formation and paracellular permeability showed several such genes were upregulated at 3 weeks compared to 2 weeks. Subsequently, Supeck et al. (Reference Supeck, Assefa, Meek, Curtis and Köhler2018) concluded the prairie vole intestine is mature by 3 weeks of age, suggesting this would be the minimum appropriate age for using prairie voles in MGBA research, given that a mature intestinal epithelium would decrease permeability to macromolecules that could act as confounding factors.

Later work by Kenkel et al. (Reference Kenkel, Kingsbury, Reinhart, Cetinbas, Sadreyev, Carter and Perkeybile2023), examining the impact of caesarean delivery on prairie vole behaviour and MGBA physiology, found that adult male prairie voles delivered in this manner, when compared to animals delivered vaginally, exhibited a lower expression of multiple proteins relevant to gut maturation. This study also found that such animals exhibited lower abundances of the bacterial family Prevotellaceae and higher abundances of the taxa Clostridiales and Ruminococcus. No procedures were performed, however, to determine if these reported differences were causally linked or accompanied by abnormalities in gastrointestinal histology or functioning.

Prairie vole gut microbiome characterization

Other investigations of the prairie vole microbiome include the assembly of five metagenome-assembled genomes from members of the prairie vole gut microbiome, a report on the general composition of the prairie vole gut microbiome, two studies examining the impacts of social isolation, and two studies assessing bacteria for potential probiotic properties or effects (Assefa et al., Reference Assefa, Ahles, Bigelow, Curtis and Köhler2015; Curtis et al., Reference Curtis, Assefa, Francis and Köhler2018; Donovan et al., Reference Donovan, Lynch, Mackey, Platt, Washburn, Vera, Trickey, Charles, Wang and Jones2020a; Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b; Donovan et al., Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023; Nuccio et al., Reference Nuccio, Normann, Zhou, Grippo and Singh2023). In addition to the study focused solely on the characterization of the prairie vole microbiome, such data are also reported in both social isolation studies and one of the probiotic studies (Curtis et al., Reference Curtis, Assefa, Francis and Köhler2018; Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b; Donovan et al., Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023; Nuccio et al., Reference Nuccio, Normann, Zhou, Grippo and Singh2023).

Reports on the general composition of the prairie vole gut microbiome indicate Bacteroidetes and Firmicutes are the most abundant phyla, while Muribaculaceae, Prevotellaceae, Ruminococcaceae, and Lachnospiraceae are the most abundant families (Curtis et al., Reference Curtis, Assefa, Francis and Köhler2018; Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b; Donovan et al., Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023; Nuccio et al., Reference Nuccio, Normann, Zhou, Grippo and Singh2023). These findings are consistent with descriptions of the gut communities of humans and other rodent models at the phylum level; they also bear some family-level similarities (Krych et al., Reference Krych, Hansen, Hansen, van den Berg and Nielsen2013; Nagpal et al. Reference Nagpal, Wang, Solberg Woods, Seshie, Chung, Shively, Register, Craft, McClain and Yadav2018).

Notably, though, there have been some differences in the exact proportional abundances of some taxa between studies, as well as within studies. For example, the first study to characterize the prairie vole gut microbiome reported Firmicutes to make up 58.4% of gut communities and Bacteroidetes to make up only 26.96% of gut communities (Curtis et al., Reference Curtis, Assefa, Francis and Köhler2018). However, other researchers have found a greater proportion of the gut microbiota comprising Bacteroidetes, with some reporting this phylum to make up at least 50% of the prairie vole gut microbiome and Firmicutes to account for 40% or less of its composition (Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b, Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023). Additionally, one study that used two different sequencing techniques to process the same samples found inconsistencies in community composition (Donovan et al., Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023).

The impact of social isolation on the prairie vole gut microbiome

The two studies that examined the impacts of isolation on the prairie vole gut microbiome both followed the basic experimental design of previous work in this area, but included faecal collection and analysis for microbiome comparison (Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b; Nuccio et al., Reference Nuccio, Normann, Zhou, Grippo and Singh2023).

Donovan et al. (Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b) entailed the placement of prairie voles in either paired or isolated housing conditions for 6 weeks, with faecal collections gathered both before and after this period. Microbiome comparisons revealed isolated voles to exhibit a decrease in the abundance of an uncultured Anaeroplasma bacterium and an increase in the abundance of an uncultured Candidatus Saccharimonas rumen bacterium, an uncultured Desulfovibrio bacterium, and Gastranaerophilales. Paired animals exhibited a decrease in uncultured C. Saccharimonas rumen bacterium and an increase in a Clostridiales vadinBB60 group species. Additionally, sex-specific changes were observed, including a decrease in an uncultured Ruminococcaceae member in isolated females. Notably, members of the Desulfovibrio genus are broadly associated with disease, although some members may provide benefits to their hosts, while members of the Ruminococcaceae family are known for breaking down plant material from a host’s diet and the production of SCFAs, an important source of energy for colonocytes and a potential regulator of metabolic health (Flint et al., Reference Flint, Duncan, Scott and Louis2007; Wu et al., Reference Wu, Wu, He, Wu, Wang and Liu2018; Zhou et al., Reference Zhou, Huang, Sun and Chen2024).

Other analyses indicated isolated animals showed increased behaviours associated with anxiety, a greater expression of CRF in the NACC, an increased expression of CRF and brain-derived neurotropic factor in the amygdala, and, in isolated males, increased circulating CORT.

The second, Nuccio et al. (Reference Nuccio, Normann, Zhou, Grippo and Singh2023), entailed the placement of female prairie voles in either paired or isolated housing for 4 weeks, with faecal collections gathered before the start of this period, followed by weekly collections for the next 4 weeks, as well as collection of faecal content directly from the colon. Emphasized findings included a greater abundance of Anaerostipes in paired animals and a greater abundance of Staphylococcaceae and Enterococcaceae in isolated animals when data from all faecal time points were pooled, as well as greater abundances of Lactobacillaceae in paired animals and Vampirovibrio in isolated animals when data from the final time point were considered alone. Notably, Lactobacillaceae and Anaerostipes, respectively, are associated with the regulation of pathogen growth and the greater production of SCFAs, while Enterococcaceaeus and Staphylococcaceae are families with prominent members associated with disease (Kant et al., Reference Kant, Rasinkangas, Satokari, Pietilä and Palva2015; Fiore et al., Reference Fiore, Van Tyne and Gilmore2019; Chia et al., Reference Chia, Mank, Blijenberg, Aalvink, Bongers, Stahl, Knol and Belzer2020; Bui et al., Reference Bui, Mannerås-Holm, Puschmann, Wu, Troise, Nijsse, Boeren, Bäckhed, Fiedler and deVos2021; Zawistowska-Rojek et al., Reference Zawistowska-Rojek, Kośmider, Stępień and Tyski2022).

This study also reported that isolated animals exhibited behavioural signs of both anxiety and depression. Furthermore, upon examining the faecal and serum metabolomes of their animals, the authors reported several metabolomic differences between paired and isolated animals that they argued were indicative of declining health in isolated animals. Specifically, they suggested that lower serum lactic acid and higher serum sorbitol and glyoxylic acid observed in isolated animals were consistent with the development of pre-diabetes or type-2 diabetes, while increased faecal succinate in isolated animals was consistent with an earlier characterization of colitis in a rodent model (Preston and Calle, Reference Preston and Calle2010; Nikiforova et al., Reference Nikiforova, Giesbertz, Wiemer, Bethan, Looser, Liebenberg, Ruiz Noppinger, Daniel and Rein2014; Osaka et al., Reference Osaka, Moriyama, Arai, Date, Yagi, Kikuchi and Tsuneda2017; Mora-Ortiz et al., Reference Mora-Ortiz, Nuñez Ramos, Oregioni and Claus2019).

Despite the differences in specific bacterial taxa found by the two studies when isolated and paired animals were compared, both studies seemed to reveal isolated animals to have greater abundances of taxa associated with poor health and lower abundances of putatively beneficial taxa, including those associated with SCFA production. The metabolite profiles reported by Nuccio et al. (Reference Nuccio, Normann, Zhou, Grippo and Singh2023) also indicate that isolated animals were less healthy. However, neither study directly examined the physiological mechanisms that mediated the differences reported. They also did not directly examine whether the microbiome or metabolite profiles observed resulted in gastrointestinal or metabolic disease.

Probiotics

Of the two published, peer-reviewed probiotic studies utilizing prairie voles, one focused on the identification of possible probiotic bacteria, while the other focused on the effects of administering what was thought to be a probiotic species (Assefa et al., Reference Assefa, Ahles, Bigelow, Curtis and Köhler2015; Donovan et al., Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023).

The former characterized 30 Lactobacillus strains found in the prairie vole gut, assessing them for their probiotic potential (Assefa et al., Reference Assefa, Ahles, Bigelow, Curtis and Köhler2015). Five were found to adhere well to intestinal epithelial cells, tolerate low pH environments and bile, and demonstrate anti-fungal and anti-bacterial properties, inhibiting the growth of Candida albicans, Escherichia coli, Pseudomonas aeruginosa, and Staphylococcus aureus.

The latter study examined how administering Limosilactobacillus reuteri to prairie voles impacted their behaviour, neurophysiology, and gut microbiota (Donovan et al., Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023). In this study, however, the researchers did not have a true control condition, instead comparing animals receiving live Lactobacillus reuteri to those receiving heat-killed L. reuteri. Interestingly, they found animals administered live bacteria exhibited less social behaviour, more CRF, and fewer vasopressin receptors in the PVN, and less CRF and fewer CRF2 receptors in the NACC, while males exhibited more CRF in the amygdala. Upon examining the possible impact of administering live or heat-killed L. reuteri on the gut microbiome, the researchers found numerous differences in taxa, including an increase in the potentially beneficial species, Bifidobacterium boum, in females administered heat-killed L. reuteri and decreased abundances of the potential pathogen, Capnocytophaga canis. Additionally, when performing correlational analyses on different neurophysiological measures and gut bacteria, they found in females positive correlations between Methanobrevibacter oralis and CRF in the PVN, Prevotella and CRF in the PVN, Desulfovibrio and vasopressin receptors in the PVN, and Desulfovibrio and both V1aRs in the PVN and CRF2 in the NACC.

The authors of this study attributed differences between animals receiving live and heat-killed L. reuteri to the administration of the bacterium, although, as with the social isolation studies, to incorporate a microbiome component, a precise mechanism was not identified through the work directly carried out in the study. However, the authors did note that the administration of heat-killed Lactobacillus species has been reported to yield physiological and behavioural effects on animals in previous studies, potentially through bacterial proteins, lipids, or metabolites that remain after bacteria are killed. They also suggested that the lower sociality exhibited by animals administered live L. reuteri was potentially linked to greater amounts of CRF in the PVN and deficits in CRF and CRF2 in the NACC in these animals, as the actions of CRF on CRF2 in the NACC previously have been linked to social behaviour in prairie voles. Additionally, it is worth noting that, given this study did not have a true control condition, it is unclear whether the administration of live L. reuteri led to some of the apparent deficits in these animals, had no effect, or perhaps improved animals compared to an unmeasured baseline but not as much as heat-killed L. reuteri.

Microbiological research in prairie voles beyond the gut

Microbiological research utilizing prairie voles for investigations outside the gut has been limited. One study from the social isolation literature that examined the effects of social isolation and aggregate stressors on immune functioning using a bacterial killing assay found that plasma from males housed with same-sex siblings showed greater antibiotic capabilities against E. coli than plasma from isolated counterparts (Scotti et al., Reference Scotti, Carlton, Demas and Grippo2015). It also demonstrated plasma from isolated females to have less antibiotic capability following placement in the RIT than that of paired females, suggesting social stressors, including isolation, can have an aggregate effect in attenuating immune functioning (Scotti et al., Reference Scotti, Carlton, Demas and Grippo2015). A second study which, to the best of our knowledge, is the only published study to examine any component of the prairie vole microbiome outside of the gut and the only prairie vole microbiome study to use animals housed outside of a standard lab environment, characterized the oral microbiome of prairie voles in a lab setting, as well as that of animals transferred to semi-natural outdoor enclosures (Sabol et al., Reference Sabol, Close, Petrullo, Lambert, Keane, Solomon, Schloss and Dantzer2023). This study found Firmicutes to be the dominant phylum in the prairie vole oral microbiome, followed by Proteobacteria, Bacteroidetes, Actinobacteria, and Fusobacteria, regardless of whether animals were housed in a lab or an outdoor enclosure. At the family level, they established what they described as a core oral microbiome comprising 14 families. Upon comparing the oral microbiomes of prairie voles housed in the lab and in the outdoor enclosures, they found prairie voles exhibited greater alpha-diversity when in the outdoor enclosures and speculated this was due to differences in diet or oral contact with soil when burrowing or foraging. Changes in oral communities after animals were placed in outdoor enclosures entailed a decrease in the abundance of core families and an increase in others. Notably, the authors also reported that sociality had no influence on oral microbiome communities and that the oral microbiomes of siblings became more dissimilar as animals spent more time in the outdoor enclosures, suggesting the immediate environment may have a greater influence on the prairie vole oral microbiome than genetic relatedness.

Procedural considerations in microbiome research

In microbiome research, there are numerous decisions researchers must make that can impact the results and interpretation of their work. These include how samples are collected and stored, the means by which bacterial cells are lysed, which hypervariable regions of the 16S rRNA gene are selected for amplification, which sequencing method is implemented, and how quality control measures are set before taxonomic assignment (Robinson et al., Reference Robinson, Brotman and Ravel2016; Gloor et al. Reference Gloor, Macklaim, Pawlowsky-Glahn and Egozcue2017; Nearing et al., Reference Nearing, Douglas, Hayes, MacDonald, Desai, Allward, Jones, Wright, Dhanani, Comeau and Langille2022).

Differences in such decisions regarding protocol may, in fact, account for some of the discrepancies in the composition of the prairie vole faecal microbiome seen across studies in which the general compositions reported were similar, but specific percentages for the proportional abundance of even major taxa were quite different (Curtis et al., Reference Curtis, Assefa, Francis and Köhler2018; Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b, Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023; Nuccio et al., Reference Nuccio, Normann, Zhou, Grippo and Singh2023). Such considerations were also highlighted in one prairie vole microbiome study in which multiple sequencing techniques were used on the same samples, but yielded different results (Donovan et al., Reference Donovan, Mackey, Lynch, Platt, Brown, Washburn, Trickey, Curtis, Liu, Charles, Wang and Jones2023).

Similarly, the tools and analyses used for statistical comparison can influence one’s results, as such tools and analyses may differ in their false discovery rates and their compatibility with the compositional nature of microbiome datasets (Gloor et al. Reference Gloor, Macklaim, Pawlowsky-Glahn and Egozcue2017; Nearing et al., Reference Nearing, Douglas, Hayes, MacDonald, Desai, Allward, Jones, Wright, Dhanani, Comeau and Langille2022). Potentially, differences such as tools and analyses may help explain the inconsistencies in findings from different microbiome studies, such as those of the two social isolation studies that examined the prairie vole microbiome (Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b; Nuccio et al., Reference Nuccio, Normann, Zhou, Grippo and Singh2023).

Animal models

Despite the noted value of animal models in biomedical research, there is growing concern over conflicting results across studies using animal models and the implications this can have for translating findings from lab animals to their wild counterparts or to humans in clinical settings (Bleich and Fox, Reference Bleich and Fox2015; Omary et al., Reference Omary, Cohen, El-Omar, Jalan, Low, Nathanson, Peek and Turner2016; Rosshart et al., Reference Rosshart, Herz, Vassallo, Hunter, Wall, Badger, McCulloch, Anastasakis, Sarshad, Leonardi, Collins, Blatter, Han, Tamoutounour, Potapova, Claire, Yuan, Sen, Dreier, Hild, Hafner, Wang, Iliev, Belkaid, Trinchieri and Rehermann2019). Suggested remedies include stricter and more standardized reporting requirements for variables like age, sex, genetic background, bedding composition, and diet (Omary et al., Reference Omary, Cohen, El-Omar, Jalan, Low, Nathanson, Peek and Turner2016).

Of the variables enumerated above, diet is of particular importance for MGBA research, as it can impact gut anatomy, physiology, and microbiome composition, as well as behaviour, cognition, and metabolic health (Pellizzon and Ricci, Reference Pellizzon and Ricci2020; Thigpen et al. Reference Thigpen, Setchell, Saunders, Haseman, Grant and Forsythe2004). However, despite its importance, the authors of relevant reviews highlight that information regarding diet composition, nutrient concentration, and the possible inclusions of non-nutritive components (e.g., phytoestrogens, heavy metals, mycotoxins, and endotoxins), to the extent such details can be known, are generally omitted from published research reports. Moreover, the authors of such reviews note how even commonly used laboratory diets purchased from manufacturers can vary considerably and affect animal phenotype (Pellizzon and Ricci, Reference Pellizzon and Ricci2020). Specifically, such authors note how grain-based diets can vary in the concentration of the ingredients, nutrients, and non-nutritive components present, while purified diets, although more consistent and less prone to contamination, may be deficient in the kinds of fibres used by gut microbes for fermentation (Pellizzon and Ricci, Reference Pellizzon and Ricci2020).

Additionally, for work examining the microbiome, researchers have highlighted that free-living animals, including humans, evolved in a “microbial world,” and that those living in “sanitized environments” are ecologically and microbially different (Rosshart et al., Reference Rosshart, Herz, Vassallo, Hunter, Wall, Badger, McCulloch, Anastasakis, Sarshad, Leonardi, Collins, Blatter, Han, Tamoutounour, Potapova, Claire, Yuan, Sen, Dreier, Hild, Hafner, Wang, Iliev, Belkaid, Trinchieri and Rehermann2019). This has been repeatedly demonstrated, including in studies examining the microbiomes of different rodents (Rosshart et al., Reference Rosshart, Herz, Vassallo, Hunter, Wall, Badger, McCulloch, Anastasakis, Sarshad, Leonardi, Collins, Blatter, Han, Tamoutounour, Potapova, Claire, Yuan, Sen, Dreier, Hild, Hafner, Wang, Iliev, Belkaid, Trinchieri and Rehermann2019; Schmidt et al., Reference Schmidt, Mykytczuk and Schulte-Hostedde2019; Sabol et al., Reference Sabol, Close, Petrullo, Lambert, Keane, Solomon, Schloss and Dantzer2023). It has also been suggested that failure to take this into account can have devastating results for human participants in clinical research, in part, due to differences in the immune functioning of humans and standard lab mice that are downstream of the latter having spent their entire lives in sanitized environments (Rosshart et al., Reference Rosshart, Herz, Vassallo, Hunter, Wall, Badger, McCulloch, Anastasakis, Sarshad, Leonardi, Collins, Blatter, Han, Tamoutounour, Potapova, Claire, Yuan, Sen, Dreier, Hild, Hafner, Wang, Iliev, Belkaid, Trinchieri and Rehermann2019).

Suitability of prairie voles for MGBA research and comparison to other models

The extent to which different concerns about the use of more standard lab animals, such as mice, apply to the use of prairie voles is largely unknown, given the dearth of research on this topic. Animal age and sex are routinely reported in prairie vole research, while possible sex differences are regularly discussed, if not examined directly. Genetically, animals are usually reported as being descended from wild prairie voles initially captured in central Illinois and regularly outbred to maintain genetic diversity within colonies.

Future opportunities

Moving forward, there are ample opportunities for further exploring both basic and applied questions pertaining to the prairie vole MGBA. At present, it is unknown if and how the prairie vole gut microbiome varies between subpopulations or changes in females based on reproductive factors. Additionally, further investigation is required to determine the extent to which the prairie vole microbiome differs between wild and lab-housed animals, as well as the effect of such potential differences on prairie vole immune functioning.

As discussed, prairie voles offer relatively unique opportunities for researchers looking to examine social components of the MGBA. The utilization of the standard design implemented in social isolation studies in conjunction with measures of gut permeability offers researchers the opportunity to examine social stress’s effects on gut integrity (Donovan et al., Reference Donovan, Mackey, Platt, Rounds, Brown, Trickey, Liu, Jones and Wang2020b; Nuccio et al., Reference Nuccio, Normann, Zhou, Grippo and Singh2023). By also measuring the presence of inflammatory markers such as LPS and proinflammatory molecules, researchers can examine whether social isolation is associated with systemic inflammation, as has been done in the CUMS literature (Zhang et al., Reference Zhang, Li, Yang, Li, Zheng, Wang, Sun, Huang, Zhang, Song and Liu2023). Additionally, given that elevated LPS in circulation and systemic inflammation are associated with poor metabolic health, one could test for indicators of declining metabolic health associated with LPS, such as increased fasted glycaemia and insulinaemia (Cani et al., Reference Cani, Amar, Iglesias, Poggi, Knauf, Bastelica, Neyrinck, Fava, Tuohy, Chabo, Waget, Delmée, Cousin, Sulpice, Chamontin, Ferrières, Tanti, Gibson, Casteilla, Delzenne, Alessi and Burcelin2007). One could also test whether the administration of specific probiotics can attenuate those or other negative health effects spurred by social isolation.