Corresponding author: W. Z. Amaral, MS, Harlow Center for Biological Psychology, University of Wisconsin-Madison, 22 N Charter Street, Madison, WI 53715, USA. E-mail: firstname.lastname@example.org
Individual variation in serotonergic function is associated with reactivity, risk for affective disorders, as well as an altered response to disease. Our study used a nonhuman primate model to further investigate whether a functional polymorphism in the promoter region for the serotonin transporter gene helps to explain differences in proinflammatory responses. Homology between the human and rhesus monkey polymorphisms provided the opportunity to determine how this genetic variation influences the relationship between a psychosocial stressor and immune responsiveness. Leukocyte numbers in blood and interleukin-6 (IL-6) responses are sensitive to stressful challenges and are indicative of immune status. The neutrophil-to-lymphocyte ratio and cellular IL-6 responses to in vitro lipopolysaccharide stimulation were assessed in 27 juvenile male rhesus monkeys while housed in stable social groups (NLL = 16, NS = 11) and also in 18 animals after relocation to novel housing (NLL = 13, NS = 5). Short allele monkeys had significantly higher neutrophil-to-lymphocyte ratios than homozygous Long allele carriers at baseline [t(25) = 2.18, P = 0.02], indicative of an aroused state even in the absence of disturbance. In addition, following the housing manipulation, IL-6 responses were more inhibited in short allele carriers (F1,16 = 8.59, P = 0.01). The findings confirm that the serotonin transporter gene-linked polymorphism is a distinctive marker of reactivity and inflammatory bias, perhaps in a more consistent manner in monkeys than found in many human studies.
Advances in genetic techniques allow researchers to identify specific genes and polymorphic variants affecting behavioral and physiological propensities associated with environmental sensitivity, which confers resilience and vulnerability. In particular, serotonergic activity and serotonin transporter (5-HTT) efficiency have been linked with reactions to social stressors, the likelihood of affective illness (Hariri & Holmes 2006; Meyer et al. 2004; Virkkunen et al. 1995), as well as immune functioning (Khan & Ghia 2010; Meredith et al. 2005; Mossner & Lesch 1998; Mossner et al. 2009; Paiardini et al. 2009; Yang et al. 2007). 5-HTT transcription is modulated by length polymorphism in the promoter region of the SLC4A6 gene (Heils et al. 1996). In vitro studies have shown that carriers of the shorter variant of 5-HTT-linked polymorphic region (5-HTTLPR) have reduced gene expression (Heils et al. 1996; Lesch & Mossner 1998; Praschak-Rieder et al. 2007) and slower 5-HT reuptake (Balija et al. 2011; Greenberg et al. 1999; Singh et al. 2011). Short (S) allele carriers also have increased amygdala activity (Canli & Lesch 2007; Hariri et al. 2002; Kalin et al. 2008), larger cortisol responses (Way & Taylor 2010a; Gotlib et al. 2008; Mueller et al. 2010, 2011), elevated norepinephrine (Otte et al. 2007) and differential drug responses (Rausch 2005). Similarly, the S allele has been linked to trait anxiety (Lesch et al. 1996), susceptibility to depression (Caspi et al. 2003), and aggressiveness (Kinnally et al. 2010). However, a number of studies in humans have also failed to replicate these findings (Sen et al. 2004). As a consequence, some have suggested more recently that this polymorphism should be characterized instead as an indicator of behavioral plasticity (Belsky et al. 2009; Chiao & Blizinsky 2010; Homberg & Lesch 2011; Kuepper et al. 2012).
Rhesus macaques have analogous variation of the 5-HTTLPR (Lesch et al. 1997). The S allele, both in homozygous and heterozygous animals, has been associated with differences in 5-HTT functioning (Bennett et al. 2002), as well as behavioral and hormonal reactivity (Champoux et al. 2002; Kalin et al. 2008). These propensities are most evident in monkeys raised by humans in a nursery setting without normal maternal stimulation and care (Bennett et al. 2002; Suomi 2006; Bennett & Pierre 2010). Thus, one goal was to investigate the strength of the association between the genotype and reactivity in normal mother-reared monkeys, evaluated both in an undisturbed setting and under moderate challenge induced by relocation. Prior primate studies have shown the potency of relocation to a novel cage or social group as a probe of emotional reactivity (Capitanio & Lerche 1998; Sloan et al. 2007; Willette et al. 2007).
Individuals also vary markedly in their proinflammatory physiology, especially following stressful challenges (Bartolomucci et al. 2005; Capitanio et al. 1998; Sgoifo et al. 2005). Our research took advantage of this strong relationship between neurobehavioral reactivity and inflammatory physiology (Avitsur et al. 2006; Capitanio 2010; Coe & Lubach 2003). For example, it has been known for decades that alterations in cell numbers and trafficking in the blood stream can be employed as a sensitive index of arousal and stress both in humans and animals (Cannon 1929; Coe 1993; Selye 1936). Psychological stressors increase the number and proportion of neutrophils in circulation at the same time that lymphocytes are redistributed out of the blood stream (Capitanio et al. 2011, Cole 2008; Cole et al. 2009, Costanzo et al. 2011, Dhabhar et al. 1996). In addition, cellular responses to in vitro lipopolysaccharide (LPS) stimulation were used to further characterize the monkeys' immune phenotype. Hormonal and autonomic responses to stress have been shown to differentially inhibit Interleukin-6 (IL-6) release and signaling (Ahmed & Ivashkiv 2000; Borovikova et al. 2000; Elenkov & Webster 2006; O'Connor et al. 2000) The a priori prediction was that the S allele would confer a heightened stress responsiveness, evident in both the neutrophil-to-lymphocyte ratios and cellular responses.
Materials and methods
Twenty-seven juvenile male rhesus monkeys (Macaca mulatta), mean age 1.9 years [SD = 0.5], were assessed in this research. Only male subjects were used in order to exclude the contribution of sex differences in behavior, physiology and social ranking. All were mother-reared, and similarly housed in stable social groups of 5–6 peers in standard pen cages (0.9 × 1.8 × 1.8 m) at the Harlow Primate Laboratory. Environmental conditions were standardized: room temperature was maintained at 21°C and light/dark cycles were 14:10 with lights on at 0600 h. Animals were fed commercial chow (PMI Nutrition International, St. Louis, MO, USA) daily at 0700 h, supplemented with fresh fruit several times a week, and water was available ad libitum. In addition to social housing, all animals received extensive environmental enrichment, including the provision of plastic toys, foraging devices, puzzles and music.
All 27 monkeys were initially evaluated under undisturbed baseline conditions in stable peer social groups and then a subset of 18 were examined again after a period of moderate disturbance. The social stress condition involved rehousing each monkey with a single unfamiliar partner in a smaller novel cage (0.9 × 1.8 × 0.9 m) for 1–3 weeks. Unfamiliar partners were of similar age and size, but mixed genders. Initially, a wire mesh panel separated the pair to ensure their compatibility and safety, and then it was removed, permitting the pair to interact. All procedures and housing arrangements were approved by the Institutional Animal Use and Care Committee.
Specimen collection and assays
Blood samples were collected at two time points: at baseline and approximately 3 weeks after rehousing. Specimens were always collected in the morning between 0830 and 1000 h. Blood (3–5.0 ml) was rapidly collected from nonanesthetized awake subjects by femoral venipuncture into ethylenediaminetetraacetic acid-treated vacutainers. To minimize the inadvertent influence of acute handling and/or a delay in collection, specimens were obtained from a maximum of three subjects on a given sampling day. The blood was used to determine a complete blood count (by General Medical Laboratories, Madison, WI, USA) and to set up whole blood cultures, which were stimulated with LPS (Escherichia coli derived). Blood was diluted 1:1 in culture medium (RPMI 1640; 1 mm sodium pyruvate; 1 nm nonessential amino acids; 25 µg/ml gentamicin; 1 U/ml penicillin G sodium; 1 µg/ml streptomycin sulfate; 2.5 ng/ml amphotericin B; 50 µm 2-mercaptoethanol; 2 mm l-glutamine; 0.075% NaHCO3). The blood was incubated in duplicate in 12-well plates, total volume of 500 µl per well, with or without 10 ng/ml LPS, for 24 h at 37°C and 5% CO2. Supernatants were then collected and frozen at −60°C until thawed for cytokine assays. Supernatant IL-6 concentrations were quantified with enzyme-linked immunosorbent assay kits using antibody targeted at human IL-6 (ELISA; RnD Quantakine, Minneapolis, MN, USA), but known to cross-react with the IL-6 protein of macaques.
On a different occasion, blood (3–5 ml) was obtained from each monkey to determine its 5HTTLPR genotype. DNA was isolated from fresh leukocyte preparations using a Puregene DNA Purification System (Qiagen, Valencia, CA, USA). Only DNA isolates with an A260/A280 absorbance ratio of at least 1.5 were used for the amplification. The polymerase chain reaction (PCR) amplification was undertaken using the Roche GC-Rich kit (Indianapolis, IN, USA) according to the manufacturer's directions. The PCR amplifications were carried out using the primer set 5HTTLPR-F (5′-CGT TGC CGC TCT GAA TGC CAG C-3′) and 5HTTLPR-R (5′-GGT GCC ACC TAG ACG CCA GGG C-3′) in a volume of 20 µl containing 200 µm each of dATP, dTTP, dCTP, dGTP, 0.375 µm forward and reverse primers, 50 ng DNA, 1 m Roche GC-rich resolution solution, 1 U enzyme, 1.5 mm MgCl2 and 1× enzyme buffer in a Perkin Elmer 9700 Thermocycler (Boston, MA, USA). The PCR conditions were as follows: 95°C × 3 min initial denaturation, followed by 32 cycles of 95°C × 60 seconds, 67°C × 30 seconds, 72°C × 60 seconds, followed by a final extension step of 7 min at 72°C. The PCR products were analyzed using electrophoresis on a 6% TBE, 6% urea, denaturing gel (Invitrogen, Carlsbad, CA, USA). The gels were visualized on a FMBIO II (Hitachi, Tokyo, Japan) using FMBIO II ReadImage 1.1 program. (Bennett et al. 2002). Genotyping revealed 16 LL and 11 LS/SS carriers. The baseline and challenge conditions comprised 16 LL and 11 S and 13 LL and 5 S subjects, respectively.
Neutrophil-to-lymphocyte ratios and cytokine levels were natural log-transformed to achieve normal distributions. Neutrophil-to-lymphocyte ratio data were analyzed with paired-sample t-tests to verify the effects of relocation, and with independent-sample t-tests for the effect of genotype during each housing condition. This use of t-tests allowed us to assess all subjects available, including the nine not available for a second sample in the rehousing condition. However, we did apply mixed model analyses of variance to the IL-6 data from the 18 monkeys with two samples. The influence of relocation and genotype were also considered separately because IL-6 levels were affected by leukocyte numbers in circulation; thus, IL-6 data were divided by the number of mononuclear cells (MNC) and reanalyzed. On the basis of finding an inhibitory effect of relocation on cytokine responses, the moderating influence of genotype was further tested by one-tailed, paired-sample t-tests.
In keeping with predictions, there was an overall effect of relocation stress on the neutrophil-to-lymphocyte ratio in the blood [t(17) = −3.80, P = 0.001] (Fig. 1a). The rehousing resulted in a marked cellular demargination, causing a sustained increase in the number and percentage of neutrophils in circulation, and an egress of lymphocytes from blood to tissue (Table 1). This redistribution of the two cell types was complementary and thus the total white blood cell (WBC) number was not affected. Carriers of the S allele had a higher neutrophil-to-lymphocyte ratio in the baseline condition [t(25) = 2.18, P = 0.02], a difference that became nonsignificant following rehousing [t(16) = −0.09, P = 0.85].
Table 1. Mean (SE) results from the complete blood counts for the L homozygous and S carrier monkeys during the baseline condition (N = 27) and after the relocation challenge (N = 18)
LL v. S carrier
Total white blood cells per ml
Following relocation, there was a significant decrease in the cellular IL-6 response to LPS stimulation (F1,16 = 14.29, P = 0.002) (Fig. 1b), as well as a stress-by-genotype interaction (F1,16 = 4.37, P = 0.05). The relocation challenge markedly inhibited IL-6 responses in S-carriers [t(4) = 3.00, P = 0.02], but had a more marginal effect on LL carriers [t(12) = −1.70, P = 0.06]. Because IL-6 release in vitro can be affected by MNC numbers in whole blood cultures, and both genotype and housing condition affected cell numbers, the IL-6 values were also examined after correcting for MNC number. After adjusting the IL-6 values, the effect of the rehousing manipulation retained statistical significance, but with decreased effect size (F1,16 = 8.59, P = 0.01), suggesting that the stress-induced shift in cell number had only partially accounted for the decrease in IL-6. Genotype continued to have a modulatory effect on the stress-induced inhibition after the cell count corrections (F1,16 = 5.52, P = 0.03). After this adjustment for MNC, the inhibitory effect of stress remained evident in S-carriers [t(4) = 2.52, P = 0.03], while attenuating the effect on LL carriers (t(12) = 0.61, P = 0.28), highlighting the interactive effect of genotype and reactivity to challenge.
Our assessment of young rhesus monkeys has confirmed that the arousal associated with rehousing has sustained effects on their immune responses, a finding that replicates and extends the reports of many other immune alterations associated with social stress in nonhuman primates (Coe 1993; Coe & Laudenslager 2007; Coe & Lubach, 2003). In keeping with previous articles indicating that S-carriers are more likely to exhibit an anxious temperament (Champoux et al. 2002; Hariri et al. 2006), we hypothesized these monkeys would have a higher neutrophil-to-lymphocyte ratio than their LL counterparts while in undisturbed housing conditions. We also predicted that S-carriers would undergo greater inhibition of cellular IL-6 responses to LPS stimulation after rehousing to a new cage and the arousal associated with forming a new social relationship with an unfamiliar monkey.
The presence of more neutrophils in circulation has long been associated with the aroused state, reflecting the detachment from vascular walls and translocation from tissue (Cole et al. 2009; Coe 1993; Dhabhar et al. 1996). While this cellular demargination can also be a marker of bacterial infection, the latter is typically manifested by a large increase in total leukocyte count (Busch et al. 1998; Nathan 2006). Apart from the effects on cell numbers, the release of glucocorticoids in response to acute stressors has also been associated with an inhibition of IL-6 responses to endotoxins (Ahmed & Ivashkiv 2000; Akira & Kishimot 1992; Tobler et al. 1992; Waage et al. 1990; Zitnik et al. 1994). This inhibition of cellular reactions is distinct from the in vivo increase in IL-6 secretion from many nonlymphoid tissues, which has described as the proinflammatory bias seen after sustained psychological stress and sometimes in glucocorticoid resistance states, such as posttraumatic stress disorder (Kiecolt-Glaser et al. 2003; Lutgendorf et al. 1999; Maes et al. 1998; Miller et al. 2002).
In our study, the cellular shift reflected the relative proportion of neutrophils-to-lymphocytes, without a change in the overall WBC number, which is more indicative of an aroused state. Monkeys from both genotypes showed the upward shift in the neutrophil-to-lymphocyte ratios after relocation, but the S-carriers had already exhibited higher neutrophil-to-lymphocyte ratios than LL monkeys in the undisturbed condition. Because of the similar response to rehousing, the influence of genotype was no longer statistically significant. Similarly, the relocation induced a marked inhibition of the IL-6 response to LPS in vitro. But the decrease was only partially because of the number of cells in the culture. After adjusting the IL-6 values by MNC counts, the main effect of rehousing retained its significance. The IL-6 value adjusted by cell number enabled us to detect the influence of genotype, with the extent of IL-6 inhibition being greater in the S carriers. Our findings reinforce the view that social and environmental context, especially with respect to arousing aspects of challenge and adversity, are important to consider when profiling the immune status of the individual (Avitsur et al. 2006; Capitanio et al. 2008). Rehousing these monkeys with unfamiliar mates induced sustained changes cell numbers in circulation and how the cells responded to stimulation with bacterial proteins.
Cell differentials were employed as some of the earliest indicators of the stressed state, used by Walter Cannon (1929) and later by Hans Selye in his classic research describing the ‘General Adaptation Syndrome’ (1932). It is important to consider that cell trafficking is a critical component of the host's adaptive reaction to infections, another type of threat and challenge to the body, when neutrophils are mobilized into circulation and lymphocytes egress to lymph nodes and spleen (Engler et al. 2004). While the shift in cell populations observed following rehousing was not of the same magnitude as seen during infection, the location of cells would influence the competence of the response to infectious pathogens and ability to survey for nonself antigens (Tseng et al. 2005). Accordingly, studies in humans have associated high neutrophil-to-lymphocyte ratios with poorer control of influenza virus infections as well as higher all-cause mortality rates (Dhabhar et al. 1996; Leng et al. 2005a,b; Poludasu et al. 2009; Smith & Wang 2012). In rodent models, detailed analyses have shown how the stress of rehousing or the introduction of a social intruder affects the MNC and macrophage populations in the spleen, impacting the response to pathogen exposure (Kinsey et al. 2008). Many of these cellular changes are induced by neuroendocrine activation, especially the potent immunomodulatory actions of cortisol (Engler et al. 2005). But the secretory release of cytokines, such as IL-6 and other chemokines, also provide soluble signals for cell migration and proliferation (Gordon et al. 1992; Hurst et al. 2001). Moreover, the release of serotonin by activated platelets in aroused individuals is also known to influence cell trafficking by affecting vascular permeability (Hutchison et al. 1959; Lekakis et al. 2010, Mercado & Kilic 2010). Although we did not measure serotonin levels in this study, previous studies in monkeys have shown that serotonergic functioning in S carriers differs from LL monkeys, and thus may affect the immune processes we assessed in this research. Many components of the serotonergic system, including the serotonin transporter, are expressed and functionally active in immune cells (Mossner & Lesch 1998). Lymphocytes, in particular, express the transporter and have been shown to vary in 5-HT uptake rates by 5-HTTLPR genotype (Singh et al. 2011).
Numerous articles on humans and monkeys have reported that carriers of the S allele are more reactive, emotionally and physiologically (Gotlib et al. 2008; Muller et al. 2009; Murphy et al. 2008; Way & Taylor 2010b). Our study took advantage of the known arousal associated with relocation to a novel setting and the formation of new social relationship in order to assess the differential reactions of the L and S carrier. Conclusions about the impact of this genotype in humans have been modified by several inconsistent findings, and the polymorphism now tends to be viewed as a marker of environmental sensitivity, rather than as a simple indicator of vulnerability to pathology, deviant behavior or mental illness (Beevers et al. 2011; Kuepper et al. 2012; Mileva-Seitz et al. 2011; Way & Lieberman 2010). The latter interpretation would concur with the research in monkeys demonstrating that the influence of the S allele becomes more pronounced following adverse rearing conditions, such as an early separation from the mother and nursery rearing (Bennett et al. 2002; Suomi 2006). But our study indicates that some physiological effects can also be found even in mother-reared monkeys when they are housed in undisturbed conditions.
Notwithstanding these straightforward findings in primates, several limitations of the current study should be acknowledged. Although it was possible to detect the influence of genotype, the number of S carriers was small for a thorough investigation of gene-by-behavior interactions. The prevalence of the S allele among the monkeys in this experiment was in keeping with its typical distribution in most rhesus monkey colonies, approximately 1:4, and the homozygous SS animal is more rare. Like other laboratory studies with relatively small number of animals, however, the chance of delineating significant genotype associations can be enhanced by the level of experimental control that lessens extraneous variation between subjects. For example, all animals in this study lived in standardized conditions, were raised in the same manner from birth and had the same diet, housing, light schedules and ambient environmental conditions. It should also be emphasized that the assessment of IL-6 responses was performed under controlled conditions in vitro, using an assay that evaluates the cellular capacity to respond to a bacterial stimulant. Other studies have evaluated levels of IL-6 in circulation (Fredericks et al. 2010; Murphy et al. 2008; Paiardini et al. 2009), where it provides a more general measure of synthesis and release by many different cells, including nonlymphoid tissues, such as adipocytes (Bastard et al. 1999; Kishimoto 1989). Then, it is interpreted as an index of the proinflammatory state of the host. On the basis of the current findings, it would be important to conduct an in vivo assessment of monkeys that differ in their rh5-HTTLPR genotype. In the light of recent research, we would predict that S carriers might present a more inflammatory phenotype. The differences in neutrophil-to-lymphocyte ratios and IL-6 responses to LPS between the LL and S carriers also speak to the need for a more detailed analysis of cellular receptors for signaling proteins, such as the serotonin receptors and transporters on blood cells. These have been well characterized in both humans and murine models, but less is known about these serotonergic features on different cell populations in monkeys, especially among ones differing in this allele polymorphism (Abdouh et al. 2004; Idova et al. 2008; Kubera et al. 2005; Maes et al. 2005; Muller et al. 2009; O'Connell et al. 2006).
Although further research is needed, the rh5-HTTLPR polymorphism appears to be a sensitive marker of physiological arousal and immune reactivity in the rhesus monkey. Our data suggest that the immune measures observed reflect genotype-related differences in temperament, which are downstream effects of limbic activation and neuroendocrine processes (Avitsur et al. 2006; Capitanio 2010), and the direct effect of genotype on immune cells via differences in serotonin transporter function (Singh et al. 2011). Larger surveys of monkeys from this colony have shown that temperament and reactivity to threat are associated not only with the rh5-HTTLPR polymorphism, but also with other genetic variation, including with genes linked to corticotropin releasing hormone and amygdala activity (Bakshi & Kalin 2000; Rogers et al. 2008), which may affect immune functioning both directly and indirectly. Collectively, this research shows the value of primate models for investigating genotypic influences, enabling us to explain how neurobehavioral and immune processes are regulated and modulated by individual genes in concert with other allele polymorphisms and myriad influences of the environment.
This research was supported in part by a supplement issued under the American Recovery and Reinvestment Act to a NIH grant (AI067518-05S1, C.L.C.) and by the NIMH grant (5T32MH018931) for the Training Program in Emotion Research. Some staff support was enabled by a Grand Challenges Explorations award from the Bill and Melinda Gates Foundation (C.L.C.). The invaluable assistance of D. Brar and A. Slukvina should also be acknowledged. The authors declare that there is no conflict of interest. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute Of Mental Health or the National Institutes of Health.