Introduction

Domesticated guinea pigs (GP; Cavia porcellus) have been used for over 200 y in biomedical research and continue to be valuable models of human disease; this use can be attributed to their gentle temperament, commercial availability, and extensive historical uses in research.⁵⁹ GPs are used as models in many disciplines including cardiology, immunology, infectious disease, nutrition, otology, and pharmacology.⁵⁹ The outbred Dunkin Hartley GP is an established and common model of progressive, chronic, and adult-onset diseases, including osteoarthritis, Alzheimer disease, tuberculosis, and asthma.^1,64 Many of these conditions have strong sex predilections, with more pronounced phenotypes in adult males.^1,64 Models that require the use of adult male GPs pose difficulties in balancing scientific objectives and providing species-appropriate social housing.

GPs are social creatures that thrive in groups.⁴⁶ They tend to eat communally, lay together, and interact frequently with members of their group.^3,9 The Guide for the Care and Use of Laboratory Animals recommends cohousing social species like GPs unless a scientific exemption, medical reason, or individual incompatibility exists.^27,46 A social environment improves animal welfare by promoting species-specific behaviors and provides social support during stress that can be inherent to some studies. This benefit is well described across a variety of species, including GPs and other rodents.^{35,45,48,50,54,57} In the wild, Cavia aperea, one of the most widely studied Cavia species, typically live in communities consisting of one male and several females.^2,50 Once offspring reach sexual maturity, which occurs at about 2 mo of age in females and 3 mo of age in males, the males leave the group.^46,50 Some Cavia species show variations in this social structure that include the presence of a smaller sexually mature male in a restricted range of the primary social group and roaming males waiting at a group’s periphery until the existing dominant male dies or can be overthrown.² Although previously thought to descend from Cavia aperea, domesticated GPs are most closely related to Cavia tschudii; the natural social structure of this C. tschudii has not been described but may mirror one of the many systems described in other Cavia species.⁶¹ Both sexes of GP can be housed in iso- or heterosexual groups before sexual maturity.^{6,32,40,46,58} Existing literature suggests isosexual housing of adult male GPs is associated with significant aggression and may not be possible in a research setting, particularly if the male is reared with or cohoused in proximity to females.^36,51,52,59 The creation of de novo adult male social pairs in a biomedical research setting has not yet been systematically assessed. Furthermore, published studies that use adult male GPs generally do not report housing details. These factors, along with personal communications between the authors and the vendors who recommend individual housing of retired male breeders, may discourage attempts to pair-house sexually mature males.

The goal of this study was to assess the compatibility of 24 previously unfamiliar adult male Dunkin Hartley GPs who were placed in new social pairs after arrival at our facility. These GPs were concurrently enrolled in a long-term approved primary study. Given their social nature and enrollment in a long-term study, isosexual pair housing was attempted and evaluated to promote species-specific behaviors and animal welfare. To evaluate the stability of isosexual pairs, we performed serial wound scoring and observations to assess fighting, affiliative, and agonistic behaviors.

Materials and Methods

Literature review.

To assess the prevalence of social housing in adult male GPs, we conducted a systematic literature review. First, we performed a PubMed search using the following search criteria: (“guinea pigs”[Majr] OR “guinea pig*”[ti] OR guineapig*[ti] OR “cavia porcellus”[ti] OR “c porcellus”[ti] OR “cavia aperea”[ti] OR “domestic cavies”[ti]) NOT (“Systematic Review”[Publication Type] OR “Meta-Analysis”[Publication Type] OR “Guideline”[Publication Type] OR “Editorial”[Publication Type] OR “Comment”[Publication Type]). Results were filtered to include articles in English and those published in the last 5 y (date range 5/17/2018 to 5/17/2023), yielding 1,121 articles for further analysis. Results were exported in PubMed format and uploaded to Covidence for a 2-step review process: screening and final review. Covidence (Veritas Health Innovation, Melbourne, Australia, www.covidence.org) is a web-based collaboration software platform that streamlines the production of systematic and other literature reviews. For screening, articles were excluded that did not use live GPs (examples include in vitro and ex vivo work, computational studies, and published protocols or methodologies), were not primary research (i.e., corrections, review articles, and case reports), involved client-owned or wild populations of GPs, could not be obtained, or involved only short-term terminal procedures or tissue harvest. Two hundred and ninety-five articles were excluded during the screening process. For the final review, methodology sections were assessed for descriptions of GP sex, age, and housing density. Articles that described group housing of adult male GPs were included. Adult male GPs were defined as at least 3 mo of age or weighing at least 600 g at the study start.^46,50 The systematic literature review is summarized in Figure 1 according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and adapted from the autopopulated schematic from Covidence.

View Figure

• Download as Powerpoint
• Download Figure

Wounds scoring.

GPs were assessed individually without sedation for the presence of wounds. Initial assessment occurred immediately before pair housing (i.e., at receipt). Subsequent wound scoring occurred 24 h and at 1, 2, 3, and 4 wk after pair housing and monthly thereafter until they reached endpoints related to their primary study. Wounds were counted, and their individual location and severity were recorded at each time point. Wound locations were recorded as mouth, pinna, remaining face (all areas of the face other than mouth or pinna), dorsum, ventrum, genitalia, forelimbs, or hindlimbs. Wound severity was scored according to the following criteria: 0, no wounds; 1, ‘punctate’ superficial wounds not penetrating the dermis; 2, ‘punctate’ wounds penetrating the dermis; or 3, lacerations (Figure 2). Each individual wound on a given animal received a severity score of 0 to 3. Severity scores corresponding to each wound were then summed for a single animal, yielding the total wound score. A total of 9 wound examinations were performed on each guinea pig during the 4-mo study. All identified wounds healed by secondary intention without the need for systemic antibiotics or primary closure.

View Figure

• Download as Powerpoint
• Download Figure

Results

Literature review of social housing in adult male GPs.

The PubMed search using the terms mentioned above yielded 1,121 articles. Of these, 295 were excluded during the screening process because they met exclusion criteria and 12 were not accessible. Eight hundred and fourteen articles passed the screening phase and went on to the final review. Only 20 articles (2.5%) described group housing of adult male GPs (Table 2).^{4,7,13,15,16,18,19,22-24,31,34,39-42,56,60,65,66} Eight of these had obtained the GP from an onsite breeding program, 7 obtained GP from a commercial vendor, one received GP from another institution, and 4 did not report the source of the GP. In one study, male GPs were castrated after arrival from the commercial vendor and prior to group housing.²⁷ Twenty-one publications (3%) reported that male GPs were housed individually. The vast majority of articles (n = 568, 70%) did not report housing density; 39 of these also omitted age and sex. Sixty-seven (8%) publications involved breeding (either mating or receipt of time pregnant dams that were subsequently nursing offspring), and 90 (11%) reported group housing of females of any age and juvenile males. Finally, 48 (6%) publications described individual housing of juvenile GP and adult females or did not provide the sex (n = 5) or age (n = 1) of singly housed GP (Figure 3).

Table 2.Studies that documented group housing in adult male guinea pigs

No. of adult males	Age	Weight (g)	Source of animals	Housing density	Strain	Reference
25	N/R	500–700	Commercial vendor	2	N/R	4
4a	Approximately 2 y	N/R	N/R	2	Dunkin Hartley	7
31	N/R	151–900	Onsite breeding colony	N/R	13/N	13
48b	7–20 wk	350–1,000	Onsite breeding colony	Up to 8	Dunkin Hartley	15
30b	10–12 wk	435–849	Onsite breeding colony	2–4	Dunkin Hartley	16
24	4–12 mo	N/R	N/R	2–3	Rosetta, Angora	18
15	N/R	600–1,200	Commercial vendor	1–2	N/R	19
41	16 wk	N/R	Commercial vendor	5–12	Dunkin Hartley	22
11	2–9 mo	430–950	Onsite breeding colony	N/R	N/R	23
17	N/R	400–1,200	Onsite breeding colony	N/R	N/R	24
24	3 mo	N/R	N/R	2	Dunkin Hartley	31
20	3–12 mo	N/R	Commercial vendor	N/R	Dunkin Hartley	34
18	14–22 mo	N/R	Onsite breeding colony	9	N/R	39
30	N/R	867–922	Onsite breeding colony	10	N/R	40
30	6–24 mo	N/R	Onsite breeding colony	10	N/R	41
26	6 mo	896–1,071	Commercial vendor	2–3	Dunkin Hartley	42
14a,b,c	12–24 mo	832–1,132	Commercial vendor	N/R	Dunkin Hartley	56
3	7–11 wk	503–525	Commercial vendor	N/R	Dunkin Hartley	60
96	6 mo	550–650	Another institution	4	Dunkin Hartley	65
2	N/R	400–750	N/R	N/R	N/R	66

N/R, not reported.

^aRetired breeders.

^bIncludes unspecified number of males and females.

^cAll males were castrated.

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Observed affiliative behaviors.

We observed 159 affiliative behaviors among the paired GPs during the 11 observation sessions (Figure 4A). The most common affiliative behaviors observed were proximity (78 total events) and social contact (60 total events; Figure 4A). The pairs showed wide variation in the number of observed affiliative behaviors. However, despite this variation, the number of prosocial events observed increased significantly (t = 2.47, P = 0.02) over time (Figure 4B) based on a generalized linear mixed-effects model.

View Figure

• Download as Powerpoint
• Download Figure

Observed aggressive behaviors.

Agonistic or dominant behaviors, jointly referred to hereafter as aggressive behaviors, were rare, with 12 total events observed among the pairs across the 4-mo study (Figure 5A). Prevalence of these behaviors did not correlate with the duration of pair housing. The initiator of the aggressive behavior was not recorded. Of the behaviors observed, mounting (n = 4), chasing (n = 3), and chattering (n = 3) were the most common (Figure 5A). Among the 11 pairs, 6 were never observed to show any agonistic behaviors. Three pairs showed aggression at 0.5 h, with 2 of these pairs also showing aggression during follow-up observations. Two pairs showed no aggression at 0.5 h but did show at least one instance of aggression during follow-up. During each observation session, 70% or more of the pairs displayed no agonistic behaviors (agonistic score of 0; Figure 5B). At any given observation session, no more than 2 agonistic behaviors were observed within a pair, resulting in a maximum observed agonistic score of 2. Furthermore, an agonistic score of 2 was only observed within the first 2 h after pairing. Thereafter, pairs either displayed no or only one agonistic behavior during observation sessions, indicating that the frequency and severity of agnostic behaviors decreased over time (Figure 5B).

View Figure

• Download as Powerpoint
• Download Figure

Wound scores and severity.

All wounds were counted and scored for individual GPs. A total of 72 wounds were observed during 9 scoring sessions, with the greatest number of wounds being present at 24 h after pairing (n = 43) (Figure 6A). Fewer wounds were noted 1 wk after pairing and were rarely seen thereafter (Figure 6A). Wounds were most commonly present on the mouth (n = 23), dorsum (n = 25), and the remainder of the face (excluding the mouth and pinna) (n = 14; Figure 6A). No wounds were present on the genitalia at any time point, and wounds were rare in other locations. Eight of the 11 pairs (73%) presented with wounds at 24 h after pairing. The remaining 3 pairs (27%) showed no wounds at any time. Severity scores for all wounds seen on a given animal were summed at each time point, yielding a total wound score. The total wound scores were significantly higher at 24 h after pairing as compared with wound scores obtained immediately before pairing (t = 3.45, P = 0.002; Figure 6B). Wound scores then slowly tapered off over time, reaching significantly (t = 7.21,P < 0.0001) lower levels at all subsequent time points compared with the 24-h score (Figure 6B). Means differences from the 24-h score were 2.1, 2.4, 2.5, 2.6, 2.6, 2.6, and 2.6, respectively, for weeks 1, 2, 3, and 4 and for months 2, 3, and 4. Of the 72 wounds observed, 83% (n = 60) received a wound score of one, indicating that the wounds were typically minor and superficial despite conspecific fighting (Figure 6C). Twelve wounds rated a severity of 2 (17%; Figure 6C). One pair of GP was deemed to be socially incompatible approximately 1 h after pairing due to a cumulative score of 11 in one GP, including a single wound with a severity of 3 and frequent agonistic behaviors. None of the other GPs had a wound severity score of 3 throughout the study (Figure 6C). The incompatible pair was used only to determine the frequency of successful pairing within the cohort (11 of 12 pairs, 92% success rate). Due to the limited data collected from this pair (receipt and 0.5 h after pairing), their data were excluded from all other analyses.

View Figure

• Download as Powerpoint
• Download Figure

Discussion

This study demonstrates the successful establishment of new, compatible social pairs in 11 of 12 pairs (92%) of adult male Dunkin Hartley GPs. The need for a systematic evaluation of compatibility in previously unfamiliar adult male GPs is indicated by 1) the paucity of peer-reviewed studies reporting social housing in adult male GPs (2.5% of the studies assessed); 2) the large number of reviewed publications that omit details of GP housing density, age, and/or sex (69.8% of studies); 3) existing literature, including the Laboratory Animal Medicine textbook, indicating that social housing of adult male GPs may be unfeasible in a research setting; and 4) vendor recommendation to singly house adult males (personal communication).^36,51,52,59 The failure to report subject and housing parameters persists despite increasing pressure to improve rigor and reproducibility in animal research and the widespread endorsement of reporting standards by journals and funding agencies.⁴³ The ARRIVE Guidelines, which have been endorsed by over 1,000 journals, include a comprehensive checklist of experimental and methodological parameters that should be reported in publications.⁴³ Because these guidelines were first published in 2010, we reviewed publications from the last 5 y, assuming that older studies would be more likely to lack information on housing density, age, and/or sex. Although animal age and sex are included in the ‘essential 10’ reporting components and housing and husbandry details are recommended in the ARRIVE Guidelines, 568 of the studies we assessed did not report at least one of these parameters, and 39 of these failed to report all 3.⁴³ Only 20 of the studies we reviewed reported social housing in adult males.^{4,7,13,15,16,18,19,22-24,31,34,39-42,56,60,65,66} Eight of these had onsite breeding programs, which can facilitate the establishment of isosexual groups in juvenile males or littermates before their use as adults.^{13,15,16,23,24,39-41} In addition, one study that obtained male GPs from a commercial vendor castrated them before creating social groups;⁵⁶ hormonal alterations associated with castration may improve the creation of isosexual groups and be a viable option when scientific objectives do not require reproductively intact GP. Familiarity with cohoused conspecifics before puberty, including genetic relation or age at introduction, was not described in any of the publications, making the present study the first to report the de novo creation of stable isosexual pairs in adult male GP and to systematically evaluate conspecific interactions for prosocial and aggressive behaviors.

In the present study, we achieved social housing in stable pairs despite their history of individual or harem (1 male: multiple females) housing at the commercial vendor. Upon arrival at our facility, GPs were immediately paired with the conspecific that shared the same primary shipping container. The shipping containers had been partitioned to prevent physical or visual contact between GP but allowed the perception of auditory and olfactory cues. A shared stressful event can facilitate social bonding in other species, leading us to pair GP that arrived in the same primary shipping container.^21,29,63 When selecting which unfamiliar animals to pair, alternative approaches that can be considered and may be equally or more successful in predicting compatibility include, but are not limited to, temperament, body weight, and duration of prior individual housing. Future studies could also assess the compatibility of adult male GPs without shipping to a new facility. We assessed compatibility after pairing by serial scoring of wounds and observing for affiliative, agonistic, and dominant behaviors. Wound scoring data indicated that most of the newly paired GPs fought, resulting in wounding. Despite the prevalence of wounding, lesions were small and superficial, with 83% and 17% of wounds receiving severity scores of one and 2, respectively (Figure 6C). Facial (54%) and dorsum (34%) wounds were the most common; no genital wounds were observed (Figure 6A). This finding contrasts the predilection sites of wounding in male mice, with dorsum and genital wounding being common and facial wounds being rare.^12,33 The GP’s face and dorsum should be inspected thoroughly when examining them for fight wounds.

In addition to wound severity and distribution, our peak total wound score occurred at 24 h after pairing, with a steady decline in scores thereafter. When wounding occurred during the first 24 h after pairing, one GP of the pair had a higher total wound score. This suggests an emerging dominance hierarchy as alpha male GPs have previously been shown to bite more often than nonalpha males.⁵⁰ The pattern of the same GP having a lower wound score was generally true throughout the study but was not always the case. We did not identify the initiator of agonistic or affiliative behaviors in the present study; therefore, we inferred the presence of an alpha male based on the wound scores at 24 h after pairing. This inference may not fully describe the social dynamic within the pair. Furthermore, 3 pairs had no wounds throughout the study, preventing any assumption about social dominance. Future studies could evaluate other parameters to more fully describe the social hierarchy formed between adult male GPs. After acclimation and presumed establishment of social hierarchies, wounding rarely occurred; wound scores were significantly higher at 24 h after pairing than at all other time points assessed (Figure 6B). The subsequent decreased propensity for wounding indicates the stability of pair social dynamics. Stable social hierarchies result in predictable behaviors and social interactions.^50,51 The predictability of interactions is important to ensuring animal welfare because social incompatibility can lead to stress, injury, or death.²⁷ Furthermore, bonding between conspecifics, indicating compatibility, has been shown to attenuate the stress response in GPs, as indicated by diminished production of cortisol after exposure to unfamiliar conspecifics or environments.⁵¹ Together, these data suggest that transient stress likely occurs as social rank is established and stable bonds are formed. The benefits of long-term social housing, indicated by a reduction in wounding and an increase in affiliative behaviors, likely outweigh the initial stress. Nonetheless, increased monitoring of newly paired adult male GPs is warranted, particularly during the first 24 h after pairing.

Although the presence of wounds confirmed fighting between pairs, we rarely observed aggressive behaviors, particularly after 2 h of cohabitation. This was true despite the more frequent observations made during the first 48 h after introduction, which allowed the assessment of evolving pair dynamics and aligned with existing recommendations.²⁷ The agonistic or dominant behaviors included in our ethogram are well described in GPs, but our ethogram did not include all possible aggressive behaviors.^{9,36,49,53,54} Other displays of dominance or aggression include scent marking, vocalization other than teeth chattering, snorting, teeth display with an open mouth, standing on rear legs, and barbering.^36,49,53,54 Future studies could monitor additional aggressive behaviors to further characterize group dynamics. Many factors likely contributed to the paucity of observed aggressive behaviors. For example, ear nibbling has been described in crowded or stressful environments, but ear wounds were rare in our study.⁴⁶ We used best practices and previously published recommendations to mitigate aggression when pairing unfamiliar adult males, although we did not test the efficacy of such measures. These measures included providing enclosures with ample floor space (24-in. × 32-in. × 10-in.), a separate hut for each GP as a visual barrier and retreat, 2 water bottles and separate feeders for pelleted diet and hay to reduce resource guarding, and increased monitoring during the introductory phase.^3,59 Future studies could evaluate the value of these and other factors for promoting the successful pairing of unfamiliar adult male GPs. Our facility housed only male GPs during the time of this study. Studies of adult males reared in isolation or social settings have shown that males are extremely aggressive in the presence of females.^51-53 The presence of females in the same room as socially housed adult males could result in severe wounding and prolonged distress. Enclosure size, enclosure complexity, and proximity to females should be considered when socially housing adult male GPs.

Affiliative behaviors were frequently observed in this study, including social contact, hut sharing, and eating together, with the observed affiliative behaviors significantly increasing over time. These behaviors are consistent with previously described prosocial behaviors in GPs.^9,40,59 While existing literature suggests that boars may not be socially compatible, we achieved successful stable pairs as indicated by infrequent agonistic behaviors, frequent affiliative behaviors, and low total wound scores in 11 of the 12 pairs. Although fighting has been well described among adult males, a wound-scoring system has yet to be published for GPs. Our wound scoring system was adapted from a published system used in nonhuman primates.¹⁷ Our observed peak in wound scores at 24 h after pairing correlated with the more frequently observed agnostic and less frequently observed affiliative behaviors at early time points, suggesting this wound scoring system is a good proxy for fighting. Only one pair was deemed to be incompatible, requiring separation due to fighting and wound severity approximately 1 h after pairing. This pair had been introduced prematurely during transit due to a divider failure, allowing unsupervised, physical contact between the 2 GPs before arrival at our facility. At the initial examination, both GPs had minor superficial wounds (wounds score of 6 and 1), presumably due to fighting during transit. Approximately 1 h after pairing, one animal sustained additional wounds, yielding a score of 11, including the only wound with a severity score of 3. The incompatible pair was used to determine the frequency of successful pairing in the 12 pairs (92%), but the pair was excluded from all other data analyses. This approach to data analysis was used in at least one other study investigating the success of social housing in nonhuman primates.³⁰ In the present study, we used a wound score of greater than 10, the presence of lacerations (Figure 2A), or prolonged agonistic behaviors as objective measures for incompatibility. Establishing such parameters should involve input from the IACUC, veterinarians, and researchers, as the parameters we used may not be appropriate for all GPs, facilities, or studies. Staff education is also critical to establishing expected amounts of aggression, reporting procedures, and parameters for intervention. Facilities should also have a plan for managing GPs that are deemed to be incompatible after pairing. This plan could include a slower return for pairing the same 2 GPs, creating new pairs to assess compatibility with other conspecifics, or providing a veterinary justification for permanent individual housing after documented incompatibility. Because our GPs were involved in an ongoing concurrent study, a limited number of GPs were available, and we were concerned about the individual propensity to fight, we did not reintroduce the 2 GPs that were incompatible in the present study. Further studies are needed to determine if reintroduction or the creation of new pairs would yield stable social relationships if pairs are deemed incompatible at first introduction. The premature introduction of these 2 GPs during transit may have contributed to their unsuccessful pairing, although other unassessed factors, including temperament and housing density immediately before shipping, may also have contributed. Direct contact with unfamiliar adult male GPs should be avoided during transit. Inadvertent introduction during transport may hinder successful pairing at the receiving facility.

While this study indicates that boars can be housed in stable social pairs, facilities should consider the animal’s overall welfare and study objectives when socially housing previously unfamiliar males. We inferred social compatibility in the present study based on the frequency of affiliative behaviors, rarity of agonistic behaviors, and decreased wounding after an introductory period. Field studies assessing a variety of Cavia species indicate several possible social schemes, including harems with single males, multimale-multifemale groups, and roaming adult males existing at the periphery of harem territories.^2,50 Additional work is needed to determine the social or solitary housing preferences of domesticated adult male GPs in captivity. While we rarely saw agonistic behaviors in the present study, a live individual made the behavioral observations; this may have limited the ability to capture a wider range of aggressive and/or affiliative behaviors. Video recording was not feasible in the present study due to equipment design and enclosure complexity that impeded visualization of the entire enclosure. Observer presence has been shown to affect a variety of animal behaviors, including pain expression, prosocial or agonistic behaviors, self-directed behaviors, stereotypic behaviors, and activity levels.^{10,20,25,28,37,44,47,55} Despite the presence of an observer, compatibility of unfamiliar conspecifics has been successfully assessed in other species.^11,30 Video recordings could eliminate observer effects. Our finding of a high incidence of prosocial behaviors and comparatively rare occurrence of aggressive behaviors suggests pair compatibility; however, further analyses of behavior frequency, time spent performing each behavior, and various group sizes are an important future direction. Likewise, future studies could measure hormone levels, including cortisol, testosterone, norepinephrine, and oxytocin, to further characterize physiologic indicators of stress and social compatibility when creating new social groups.^{3,11,36,40,54,62} We did not have a complete housing and breeding history of the GPs included in this study, including duration of single housing as compared with breeding or number of litters sired. We also did not know the genetic relationship between males; therefore, some of the GPs may have been littermates, living together with their dam before weaning (approximately 4 mo before the study started). These and other aspects of their social history could affect pairing success. Future studies could compare the compatibility of adult males with known histories, including use as a breeder and housing densities. Future studies could also assess the social compatibility of different GPs to identify strain differences that could affect pairing success.

Our data demonstrate the compatibility of pair-housed adult male Dunkin Hartley GPs. Because wounding, and therefore agonistic behaviors, occur more frequently during the first 24 h after pairing, increased monitoring is warranted during this time. Investigators should consider study objectives and duration, pairing-related stress, and overall animal welfare because social housing may not be appropriate in all situations, particularly if female GPs are present. With sufficient space and resources, unfamiliar adult male GPs can form stable social pairs in a research setting.