Stress can be defined as the organism’s biological response to a threat or a disturbance to its homeostatic state. An organism’s stress response involves multiple important biological systems, particularly the autonomic nervous and neuroendocrine systems. The first drives alterations in the cardiovascular and gastrointestinal systems and delivers short duration responses (Cannon, 1929; Kemeny, 2003; Ziegler, 2012). In contrast, the neuroendocrine system response provides slower but potentially longer-lasting responses to challenges. In mammals, the main neuroendocrine response to stress involves the hypothalamic-pituitary-adrenal (HPA) axis, which affects many aspects of biology including metabolism, reproduction, and development of immunocompetence. The main output of the HPA axis are the glucocorticosteroids cortisol or corticosterone (depending on species) which regulate carbohydrate, fat, and protein metabolism, boost energy supplies, increase blood pressure and sugar as well as modulate the anti-inflammatory response. When animals are faced with challenges that quickly resolve, the role of the HPA axis and other stress response systems is adaptive. However, when challenges persist or are severe over the long-term, repeated or continual activation of the axis can be highly demanding on animal resources and lead to maladaptive outcomes such as reductions in immunocompetence, reproductive ability, growth development, productivity and welfare, as reflected by findings in many domesticated species including cattle (West, 2003), chickens (Zaboli et al., 2019), and pigs (Rutherford et al., 2006).

In pig production systems, sources of stress can be categorized into environmental, animal handling/management, and social causes (Martínez-Miró et al., 2016). Pigs are susceptible to ambient temperature, particularly high temperature (Mutua et al., 2020) since they lack functional sweat glands and have relatively small lungs to effectively dissipate excess heat (D’Allaire et al., 1996; Patience et al., 2005). Farming operations involving animal handling (such as vaccination, snaring, blood sampling, ear tagging or tattooing) or transport (uploading, unloading, travel duration and vehicle design) are also stressful for pigs (Prunier et al., 2005; Goumon and Faucitano, 2017). Additional sources of stress include for example, high stocking density, which leads to restricted space allowance, reduced feeder space per animal, lower access to the feeder, and increased competition between the animals, resulting in more frequent fighting, and higher levels of aggression (Randolph et al., 1981; Andersen et al., 2004), and animal regrouping, which can cause changes in the social dynamics of the population, leading to fights and subsequent harassment of the defeated animals (D’Eath, 2002; Desire et al., 2015a; Foister et al., 2018).

Stress and intestinal microbiome appear to be closely linked. Bailey et al. (2011) found that mice exposed to social disruption exhibited an increased relative abundance of Clostridium and decreased relative abundance of Bacteroidetes in comparison to their non-stressed counterparts. Within the pig industry, weaning represents a highly stressful challenge because piglets suddenly experience new nutritional, physiological, and psychological challenges. Weaning-induced stress is reported to result in a decline of some benefit bacterial genera such as Alloprevotella and Oscillibacter, while some opportunistic pathogens such as Campylobacter, Clostridium XlVa, and Clostridium XlVb increased in colon (Li et al., 2018). In addition, studies over the last decade have provided important information on a bidirectional microbiome-gut-brain axis, in which the gut microbiota and their metabolites interact with the host’s brain over metabolic, immunological, endocrine and neural pathways (Cryan and O’Mahony, 2011; Grenham et al., 2011; Wang and Kasper, 2014; Martin et al., 2018; Valles-Colomer et al., 2019; Barandouzi et al., 2020). As suggested by Bohórquez et al. (2014) and Bohórquez and Liddle (2015), the gut microbiota and their metabolites communicate through the gut connectome (i.e., the complex neural network in the gut involving the gut glial, intrinsic neurons and enteroendocrine cells), playing an important role in gut enteric nerves development and enhancement of the host’s ability to sense and use nutrients. Additionally, Messaoudi et al. (2011) found that a probiotic based on Lactobacillus helveticus R0052 and Bifidobacterium longum R0175 contributed to reducing anxiety-like behaviors in rats and alleviated psychological distress in humans. Furthermore, when transplanting the fecal microbiome of depressed human patients to antibiotics-treated rats, Kelly et al. (2016) found that the rats exhibited anxiety-like behaviors and tryptophan metabolism disruption, and proposed the vital role of the intestinal microbiome on development of depression in humans.

Most studies on the bidirectional microbiome-gut-brain axis have been performed in mice, for example by using germ-free or specific-pathogen-free (regulated by antibiotics) animals, or by performing fecal microbiome transplants (Martin et al., 2018). Previous authors have investigated the shift of the intestinal microbiome due to weaning stress in pigs (Guevarra et al., 2018; Li et al., 2018), but overall, much less literature is available on the microbiome-gut-brain axis on livestock, particularly on growing pigs. The present study is the first to use a social stress model to investigate the effect of stress on the intestinal and fecal microbiota in finishing pigs. A combination of high stocking density, low feed space and regular regrouping was employed to ensure the reliable establishment of stress in pigs. Each of the above stressors has been found to increase stress levels in pigs, with many biological consequences such as abnormal maternal behaviors (Jarvis et al., 2006; Ison et al., 2010), tail biting and aggression behaviors (Turner et al., 2001; Cornale et al., 2015), cortisol and testosterone elevation (Rutherford et al., 2006; Coutellier et al., 2007; Escribano et al., 2015), and negative cognitive bias (Scollo et al., 2014).

Pigs have previously been shown to be a highly relevant model species for studying the effect of environmental factors (e.g., nutrition, milk delivery method, antibiotic treatment) on early-life microbiota establishment in humans (Roura et al., 2016). Pigs have an omnivorous nature, with diet, nutritional requirements, and size similar to humans (Heinritz et al., 2016). Regarding physical and physiological traits, the digestive systems of pigs and humans are very similar, including the transit rate of liquids and food in the gastrointestinal tract, and the digestive and absorptive processes (Graham and Åman, 1987; Miller and Ullrey, 1987). Humans and pigs are essential colon fermenters (Graham and Åman, 1987) of plant/fibrous dietary components. Considering the gut microbiota composition, Firmicutes and Bacteroidetes are the dominant phyla in humans and pigs (Heinritz et al., 2013). Due to its size and temperament, the pig provides an easier approach for sample collection, such as saliva, blood samples and cannulas in the gastrointestinal tract, than other species, e.g., mice. Enhanced by innovative techniques, such as genetically modified, germ-free, gnotobiotic and human-microbial associated pigs, the pig is emerging as a powerful, translational model of gastrointestinal microbiota studies (Rose et al., 2022). In contrast to mice, pigs and humans have the common HPA axis’s output of cortisol with a similar circadian rhythm in responding to stress (Ruis et al., 1997), resulting in similar effects on the immune system, gastrointestinal and neuroendocrine alterations (Gimsa et al., 2018). Therefore, these authors concluded that social stress in pigs can reflect partly the intensity of psychosocial stress in the human society. These findings suggest that application of the social-stress model in pigs provides a great opportunity to research the modulation effect of stress on the gut microbiota, as well as the potential use of microbiota as therapeutic or preventive tools to cope with the stress challenge in humans. The potential microbial biomarkers of social stress are expected to have also implementation for dietary intervention (e.g., development of probiotics) and animal breeding for pathogen resistant pigs.

The objectives of this study were (i) to demonstrate the effectiveness of a porcine stress model, (ii) to assess the effects of social stress on porcine performances and intestinal (caecum and colon) and fecal microbiota, (iii) to identify social stress microbial biomarkers in the intestinal and fecal samples of pigs with potential for broader applications in other species, and (iv) to identify potential differences of animals in susceptibility to intestinal pathogen growth.

Our results suggest that social stress has a substantial negative impact on several aspects of pig welfare (increased lesions) and productivity, leading to reduced feed intake and growth rate as well as alterations in the caecal, colonic, and fecal microbiota profiles.

Our porcine social stress model successfully increased the release of cortisol in saliva cumulatively over time suggesting that the severity of the stress increased throughout the experiment. This increase suggested that alterations in the HPA activity, the main stress response axis of the body (Smith and Vale, 2006), were persistent throughout the experiment. This finding was in agreement with previous studies that showed increased cortisol concentration in pigs exposed to different stressors, such as shipping (McGlone et al., 1993), altered ambient temperature (Hicks et al., 1998), and social stress (Rutherford et al., 2006; Escribano et al., 2015; Casal et al., 2017). One outcome of the social stress model was the substantial increase in skin lesions, most likely due to the regrouping of animals, which caused vigorous physical aggression between unfamiliar individuals, potentially because a new hierarchy had to be repeatedly established (Desire et al., 2015a; Foister et al., 2018). Skin lesions are related to animal welfare and are also influenced by animal genetics (Desire et al., 2015b; Agha et al., 2022).

The stress condition led to reductions of 12.4% in the voluntary feed intake, 16.1% in the growth rate, and 5.9% in feed use efficiency in comparison to the controls, which may lead to a delay in reaching the slaughter weight, negatively impacting productivity. Hyun et al. (1998) reported decreased growth rates of 7.1 and 15.7% in pigs subject to regrouping and crowding, respectively, but did not find significant differences in voluntary feed intake between treatment groups.

The impact of stress on the intestinal and fecal microbiome adjusted Shannon index was not significant. However, animals in the stress group had microbiota profiles at the genera level with significantly higher richness than the control animals. The beta-diversity differences observed between the stress and control groups were not consistent throughout the sample collection sites. A negative control sample (containing no DNA) and a positive mock control sample with a known bacterial composition (in duplicate) alongside the experimental samples were not available and therefore could not be used to get further insight into the accuracy of the sequencing results. These findings are consistent with the review of Kuo and Chung (2019), in which the authors concluded that the impact of neuro-psychiatric conditions and depression on the stability of human microbiome diversity is inconsistent throughout the literature. Likewise, in a study on physically healthy women, the fecal microbiome diversity of participants with anxiety and depression was not significantly different from that of the psychologically healthy woman (Kleiman et al., 2017). Michels et al. (2019) claimed that even though higher diversity is often recognized as a good health indicator, its association with the brain’s function and health remains unclear.

When interpreting the results, it has to be considered that stressed pigs may have encountered a higher bacterial load due to weekly mixing, potentially contributing to larger microbiota richness in the intestine compared to the control group. However, the control pigs were kept in the same room so that due to ventilation the different exposure may not be substantially different. In addition, pigs from each litter were distributed across experimental groups, ensuring a balanced early-life microbial exposure. Furthermore, the experiment was based on grown up pigs in the final 4 weeks of their life so that different microbial exposure due to mixing may be of negligible impact on the intestinal microbiota. Generally, it is reported that a stress impaired immune system is closely linked to intestinal microbiota profiles (Gimsa et al., 2018). Hence, the observed larger microbiota richness in the stressed group is most likely attributed to this stress-induced immune modulation.

We identified 32 genera whose abundances differed significantly between the stress and control groups across all sampling sites, and 10 of them were identified in all locations (Anaerovibrio, Faecalibacterium, Dialister, Prevotella, Catenibacterium, Oribacterium, Fusicatenibacter, Agathobacter, Clostridium, and Parabacteroides). Five out of seven stress-enriched biomarkers that were identified in the caecum and colon were also identified in the feces, suggesting that the feces represent the most suitable sample type for microbiota-based studies on stress in pigs, which is advantageous given the ease of access to this sample type.

We have identified the genus Clostridium to be significantly enriched in stressed animals consistently over all sample sites. Some Clostridium species are part of the commensal porcine core microbiota, and they are involved in saccharolytic and proteolytic metabolism (Holman et al., 2017). However, Clostridium perfringens types A and C and Clostridium difficile are opportunistic pathogens that can cause significant enteric porcine infections (Songer and Uzal, 2005) and have been found to severely infect hospitalized humans (Kelly and LaMont, 1998).

Our results agree with previous observations of increased relative abundance of Terrisporobacter, Marvinbryantia and Romboutsia (all belonging to the Clostridiales order, such as Clostridium) in stressed animals. Whereas Terrisporobacter and Marvinbryantia belong to the Peptostreptococcaceae family, which has previously been observed to be enriched in patients with Clostridium difficile infection (Stewart et al., 2019), Romboutsia was increased in mice with chronic unpredicted mild stress (Sun et al., 2019). The authors found that Romboutsia’s abundance was positively correlated with mice anxiety- and depression-like behaviors.

Streptococcus, Parabacteroides and Desulfovibrio were also enriched in pigs in the stress group. In humans, Streptococcus and Desulfovibrio have been reported as increased in major depressive disorder patients (Simpson et al., 2021), and may be involved in the modulation of inflammatory response in depression patients (Evrensel and Ceylan, 2015; Barandouzi et al., 2020). Desulfovibrio and its product hydrogen sulfide were found to contribute to inflammation modulating and to the development of different inflammatory bowel diseases (Loubinoux et al., 2002; Bisson-Boutelliez et al., 2010; Kushkevych et al., 2018, 2021). Kverka et al. (2011) found that orally fed membranous fraction of Parabacteroides distasonis lysate to dextran sulfate sodium-induced colitis mice reduced the production of both proinflammatory and anti-inflammatory cytokines in the colon of treated mice. Considering that the relative abundances of these two genera were positively correlated through the three sampling sites, the inflammatory modulation effect can be the common function that led to the strong associations between them.

Treponema (a known opportunistic pathogen) was enriched in the stress group (colon and feces samples) and is therefore proposed as a potential biomarker for stress in pigs. Treponema is associated with porcine colonic spirochetosis, a diarrheal disease resulting in hindered performance. Treponema hyodysenteriae, recently reclassified as Brachyspira hyodysenteriae, can cause swine dysentery (Burrough, 2017). Another species, Treponema pallidum, is linked to diarrheal illness in pigs, as well as in humans (Tsinganou and Gebbers, 2010).

The enrichment of Turicibacter in the stress group is interesting because this may be due to the pig’s ability to adapt to stress. The relative abundances of Turicibacter in the pigs’ intestinal samples were highly consistent with the correlation of 0.88 between caecal and colonic datasets whereas they were not significantly associated with salivary cortisol (p > 0.05). These findings suggested an involvement of another mechanism led to the enrichment. In a recent study in mice, Fung et al. (2019) discovered that Turicibacter sanguinis can co-evolve with the host to induce serotonin production. Serotonin is a neurotransmitter involved in many aspects of neural activities, such as stabilizing mood, social behaviors and increasing happiness. The authors found that T. sanguinis had a special receptor named CVW-0748 that was structurally and functionally similar to the serotonin transporter of the host. As a result, T. sanguinis could uptake the host’s serotonin and then increase in abundance in the intestine. Fung et al. (2019) also demonstrated the bidirectional microbiome-gut-brain axis and recommended further investigations into the spore-forming bacteria and gut-derived serotonin (Hoffman and Margolis, 2020), especially in stress conditions as suggested by the results in growing pigs in this study. Moreover, T. sanguinis is involved in the regulation of the host’s steroid and lipid metabolism and is reported to be associated with Parkinson’s disease (Jin et al., 2019) and depression (Jackson et al., 2018).

Even though Sphaerochaeta belongs to the same family as Treponema (Spirochaetacea), it does not have the two key features of pathogenic spirochaetes, i.e., helical or spiral morphology and motility (Ritalahti et al., 2011; Miyazaki et al., 2014; Dong et al., 2018). Caro-Quintero et al. (2012) found that Sphaerochaeta genomes have relatively higher numbers of fermentation and carbohydrate metabolism genes than other spirochetes, promoting a fermentative lifestyle. Their increase is expected to have no negative effects on the host, suggesting that even commensal bacteria in the porcine gastrointestinal tract were influenced by stress.

The bacterium Candidatus Saccharimonas, and the archaea Methanosphaera and Methanobrevibacter were also enriched due to social stress. These bacteria may have a role in degrading fiber and gut microbiota stabilization (Chaucheyras-Durand et al., 2010; Albertsen et al., 2013; Kindaichi et al., 2016; Mi et al., 2019). Interestingly, the two methane-production genera Methanosphaera and Methanobrevibacter were plotted relatively close to each other, reflecting their positive correlation (Figure 8).

One highlight of our research is that many bacteria reported to be positively related to health were depleted in the stress group, including Prevotella, Faecalibacterium, Butyricicoccus, Dialister, Megasphaera, and Mitsuokella. Prevotella is a dominant population in the porcine intestine and metabolizes hemicelluloses and pectin to acetate, an energy supply for the host and other bacteria. Faecalibacterium and Butyricicoccus produce butyrate (Duncan et al., 2004) which is a key energy source for colonocytes and has potent anti-inflammatory properties, thus benefiting host health (Hamer et al., 2007; van der Beek et al., 2017).

Recently, Valles-Colomer et al. (2019) found that Faecalibacterium were positively associated with a majority of the quality of life scores in the RAND-36 Health Survey (Hays and Morales, 2001), including both mental and physical aspects, such as social functioning, emotional well-being, vitality, and physical functioning. In addition, Dialister was significantly positively associated with the physical RAND scores and significantly depleted in depression patients (Valles-Colomer et al., 2019). In our study, Faecalibacterium and Dialister were in all sampling sites found to be depleted in the stress group and strongly associated with each other. Coprococcus, Agathobacter, Fusicatenibacter and Oribacterium are from the Lachnospiraceae family, and the latter three genera are closely associated with each other. Lachnospiraceae was previously reported to be depleted in patients with major depressive disorder (Jiang et al., 2015). Similarly, Coprococcus (Liu et al., 2016; Huang et al., 2018; Valles-Colomer et al., 2019), Fusicatenibacter (Chen et al., 2021) and Agathobacter (Hua et al., 2020; Chen et al., 2021) showed negative associations with depression symptoms. These results suggest that the imposed social stress in pigs affects the same bacteria associated with depression and quality of life in humans. Generally, these findings provide evidence of many similarities in the microbiota-gut-brain axis in humans and pigs.

Dialister and the Anaerovibrio, Megasphaera, Mitsuokella and Allisonella are all belonging to the family of Veillonellaceae and were all depleted in the stress group. Megasphaera elsdenii is known to reduce acidosis in ruminants by metabolizing lactate to SCFA. In a caecal in-vitro porcine model Megasphaera elsdenii was shown to substantially increase the production of butyrate supporting its beneficial health effects and therefore was suggested as a probiotic agent (Tsukahara et al., 2006). Two novel Megasphaera species isolated from the human gut had diverse and unique sets of Carbohydrate-Active enzymes (CAZymes), that produced SCFA, vitamins and essential amino acids which further promote the potential health effects on the host (Shetty et al., 2013). Mitsuokella jalaludinii isolated from the porcine gut was found to inhibit Salmonella growth and invasion due to the production of SCFA and reduction in pH (Levine et al., 2012).

SCFA have anti-inflammatory effects (Parada Venegas et al., 2019), therefore their decrease may directly contribute to the potential microbiota-mediated inflammation in the anxiety/depression groups in humans (Simpson et al., 2021). In this study, SCFA-producing genera were depleted in the stress group, such as Catenibacterium, Fusicatenibacter (Takada et al., 2019), Agathobacter (Rosero et al., 2016), Alloprevotella (Downes et al., 2013), Butyricicoccus (Eeckhaut et al., 2008), Subdoligranulum (Holmstrøm et al., 2004). Our results indicate a correlation between the abundance of these genera (clustered together in the PLSDA biplots, Figure 8), which might be explained by their common role in SCFA production.

The impact of commensal genera Sutterella, Libanicoccus, and Holdemanella on porcine health and stress are unclear. Sutterella is reported to be commensal with high prevalence in the human gut. It has minimal impact on epithelial homeostasis and is possibly involved in regulating the gut immune system (Hiippala et al., 2016). Libanicoccus and Holdemanella are identified in the human gut and feces, but research on their impact on human and porcine physical and mental health is still lacking (Bilen et al., 2018). The depletion of these three genera in the stress group was probably due to changes in host immune response, feed intake and efficiency.

Our results support the hypothesis that stress altered the intestinal and fecal microbiota profiles. However, the stress group also showed decreased voluntary feed intake. This suggests that stress may have directly influenced the microbiota profiles, or that this influence occurred in an indirect manner, due to changes in feed intake that led to alterations in the microbiota, or even both simultaneously. Diet is one of the most important factors influencing the gastrointestinal microbiota, since changes in nutrient availability will favor or impair/inhibit the growth of microbial communities. However, the composition of the diet did not change, so only less nutrients of the same composition were available in the stressed animals. Therefore, we hypothesis that the change in microbiota profiles as a results of stress are mainly due to the microbiota-gut-brain axis, supported by the substantial higher salivary cortisol concentration in the stress in comparison to the control group (Misiak et al., 2020).

Based on the low abundance of the opportunistic pathogenic bacteria Clostridium, Treponema and Streptococcus in the gut of most of the experimental animals even under stress and relatively few animals with extremely large abundances, we hypothesize that most of the animals in this experiment were resistant to the growth of these pathogens. Resistance against these pathogens, which might be explained by host-genetic factors, is an important attribute in livestock production (Bishop and Woolliams, 2014; Bai et al., 2020) to help animals maintain their performance with minimal treatment. This resistance might be also beneficial for pigs while adapting to challenging social stressors. Breeding for improved adaptation to stressors has been previously discussed, such as selection for higher heat tolerance to adapt to heat stress in tropical climate (Gourdine et al., 2019) or higher resistance to reduce disease risk and prevalence (Tsairidou et al., 2019). The identification of biomarkers indicating the extent to which an animal is protected against pathogen growth in the gut, in particular under stress conditions, therefore is of great interest for supporting breeding for the resistance of pigs.

Campylobacter is one of the leading food-borne zoonoses worldwide (Havelaar et al., 2015), and Campylobacter coli is the main species in pigs that cause gastrointestinal infection in humans (Fosse et al., 2009). Campylobacter infection in humans is the most frequently reported gastrointestinal disease in Europe with an estimated annual cost of 2.4 billion Euros (Elliott et al., 2012; EFSA and ECDC, 2018). Whereas this bacterium causes gastroenteritis in humans, pigs do not show any symptoms. In this study, there was a tendency of the enrichment of Campylobacter in the stress group; however, the probabilities of being enriched were relatively low at 59, 94 and 65% in caecum, colon and feces, respectively.

Our porcine stress model with weekly regrouping of animals and reduced space allowance during the last 4 weeks of the finishing period was shown to be effective, based on the sharp increase in stress-induced salivary cortisol concentration, which gradually increased throughout the experimental period. These changes in salivary cortisol levels reflect HPA activity as the major stress response axis. Although there are differences in stress-induced bacterial biomarkers between sampling sites, most of the important stress biomarkers were identified in feces. Opportunistic pathogenic genera such as Clostridium and Treponema were significantly enriched due to social stress, demonstrating that stress impairs the ability of animals to defend against pathogenic bacteria, and confirming the influence of the HPA axis on the composition of the gut microbiota. In particular, the distribution of abundances of the opportunistic pathogenic bacteria Treponema and Campylobacter indicates the differential ability of animals to cope with stress and, assuming a genetic component, should be investigated as a selection criterion for breeding for resistance to stress-induced pathogenic bacteria growth in the gut. Bacteria that are depleted in the stress group are very informative, as they have previously been shown to be associated with health benefits in various animal species due to their metabolic products, especially SCFA, which might be involved in the inhibition of the growth of pathogenic bacteria. Of particular interest are the bacteria Dialister and Faecalibacterium, which were reduced by social stress in our porcine study and depleted in humans with depression and low quality of life. Therefore, bacteria depleted by social stress in our study provide biomarkers to identify stress and may potentially be used as probiotic agents to reduce the effects of stress on host health. In addition, Turicibacter enriched due to stress is a highly interesting candidate biomarker to explore antidepressant probiotics. In general, our results confirm the microbiota-gut-brain axis communication as demonstrated by increased cortisol levels regulated by the HPA axis and the alteration of microbiota composition, particularly bacteria known to be associated with poor mental health.

The data presented in the study are deposited in the European Nucleotide Archive repository, at: https://www.ebi.ac.uk/ena, accession number PRJEB60501.

The animal study was approved by SRUC’s Animal Welfare and Ethical Approval Body. The study was conducted in accordance with the local legislation and institutional requirements.

RR, KR, and MA: conceptualisation. TN: statistical analysis and writing original draft. TN, MM-Á, JL, and RR: formal analysis, methodology, and software. TN, MM-Á, JL, MA, KR, GS, RD, and RR: investigation and validation. TN, MM-Á, JL, MA, KR, GS, RD, EB, and RR: review and editing. All authors contributed to the article and approved the submitted version.