*E. L. Kinnally, Department of Psychiatry, Division of Neuroscience, Columbia University, College of Physicians and Surgeons, 1051 Riverside Drive, Room 2917, New York, NY 10032, USA. E-mail: email@example.com
The moderating effect of early experience on gene-behavior associations has been well characterized. The molecular events that allow for such moderation are not well understood, however. We assessed the impact of early experience and serotonin transporter linked promoter polymorphism (rh5-HTTLPR) genotype on peripheral serotonin transporter (5-HTT) regulation in response to a maternal/social separation and relocation stressor in infant rhesus macaques. We further tested the hypothesis that modulation of 5-HTT regulation by rearing and/or genotype is mediated by glucocorticoid (GC) availability. Fifty-three infant (3–4 months of age) rhesus macaques that were either nursery reared (NR) or mother reared (MR) were genotyped for rh5-HTTLPR. Infants were blood sampled within 2.5 h of maternal or social separation/relocation and again 5 h later. Infants were then administered dexamethasone, a synthetic GC and blood sampled 16.5 h later. 5-HTT RNA was quantified from peripheral blood mononuclear cells. Plasma cortisol was measured at all time points. The MR individuals upregulated 5-HTT significantly during maternal/social separation, while NR individuals did not. Concomitant increases in cortisol were not observed, but dexamethasone treatment stimulated 5-HTT expression regardless of genotype/rearing group, and 5-HTT expression in the post-stressor sample was correlated with plasma cortisol levels at all time points. Our data indicate that early experience exerted a strong influence on 5-HTT regulation during a stressor in infant rhesus macaques independent of rh5-HTTLPR genotype. We also showed that GCs may stimulate 5-HTT expression but that there likely exist faster-acting transcriptional regulators of 5-HTT that are in place as a function of experience.
Development of the monoamine neurotransmitter serotonin (5-HT) system is sensitive to various environmental and genetic contributors (Anderson et al. 2004; Higley et al. 1993; Rogers et al. 2004). The specific nature of these contributions to 5-HT function is of great interest because of the well-characterized association of 5-HT function with psychological processes and behavior in human and nonhuman primates (Arango et al. 2002; Asberg et al. 1976; Fairbanks et al. 2001; Mehlman et al. 1994). Early stress or adversity, for example, appears to reduce serotonergic functioning. In humans, self-reported early neglect predicts low serotonin function as indexed by cerebrospinal fluid (CSF) 5-HT metabolite 5-hydroxyindoleacetic acid (5-HIAA; Roy 2002). In rhesus macaques, early adversity in the form of nursery rearing (NR, maternal deprivation), as opposed to mother rearing (MR), results in reduced serotonin transporter (5-HTT) binding in multiple brain regions in adolescence (Ichise et al. 2006). Early environmental effects on 5-HT function may be exacerbated by the inheritance of 5-HT pathway polymorphisms that also reduce 5-HT function. Bennett et al. (2002) showed that NR infant rhesus macaques with a low activity, ‘short’ version of the rhesus macaque serotonin transporter linked promoter polymorphism (rh5-HTTLPR), exhibited the lowest 5-HT function (as indexed by low levels of CSF 5-HIAA) of any genotype/rearing group (Bennett et al. 2002). It is thought that 5-HTTLPR disrupts 5-HTT availability, limiting 5-HT reuptake from neurons, platelets or lymphocytes (Langer et al. 1986; Paul et al. 1981), as it is a noncoding insertion/deletion polymorphism that confers reduced/enhanced transcriptional efficiency of 5-HTT (Bradley et al. 2005; Heils et al. 1996; Hranilovic et al. 2004).
The mechanism by which early environment might moderate allelic influences on 5-HTT expression is not known. It is possible that transcription factors that are sensitive to experience regulate 5-HTT, potentially at rh5-HTTLPR. The hypothalamic–pituitary–adrenal (HPA) system, particularly the availability of glucocorticoids (GCs) such as cortisol, is well known to be sensitive to early experience (Capitanio et al. 2005; Levine 1994; Shannon et al. 1998) and has been shown to co-operate with the 5-HT system (Barr et al. 2004; Tafet et al. 2001). Furthermore, GC-receptor complexes interact with a GC-response element in the 5-HTT promoter region, stimulating 5-HTT expression, indicating that they act as transcription factors (Glatz et al. 2003; Slotkin et al. 2006). It is possible that individual differences in GC and/or GC-receptor availability may mediate sustained differences in 5-HTT expression levels between individuals with differing life experiences and genotypes.
The goals of this study were first to assess the effect of early experience and rh5-HTTLPR genotype on peripheral 5-HTT expression in infant rhesus macaques. We then tested the hypothesis that availability of GCs mediate rearing and/or genotype differences in 5-HTT expression. We predicted that MR infants would differ from NR infants in 5-HTT expression, but that genotype would further differentiate these groups. Moreover, if 5-HTT expression differences are mediated by GC availability, we expected that cortisol levels would predict 5-HTT expression, that 5-HTT expression would be stimulated differentially between rearing/genotype groups following synthetic GC dexamethasone treatment and that 5-HTT regulation would track cortisol changes during testing between genotype/rearing groups.
Materials and methods
Subjects were 53 (21 male and 32 female) infant rhesus macaques aged 90–120 days. Infants were selected from two rearing conditions: MR animals were raised with their mothers in half-acre outdoor enclosures at the California National Primate Research Center. Each field cage contained one large social group comprising at least six genetically distinct matrilines with extended kin networks. Nursery-reared animals were also born into the field cages but were separated from their mothers on the day of birth and were subsequently reared in an incubator for the first month of life. Following this month, infants were housed in indoor individual cages (0.46 × 0.61 × 0.69 m) with intermittent or continuous access to a sex-matched, same-aged pair-mate. Within rearing groups, all genotypes were represented [MR long/long (l/l), n = 22; MR long/short (l/s) or short/short (s/s), n = 10; NR l/l, n = 12; NR l/s or s/s, n = 5]. Four subjects’ genotypes were not available and were removed from analyses including genotype. Mean pairwise relatedness between subjects was <0.005%. There were 14 subject pairs (out of a possible 1378 pairs) that were 25% related, and two pairs of subjects were more than 25% related (pairwise kinship coefficients 0.37–0.39) in our dataset. These related individuals were distributed evenly across genotype/rearing groups and removal of these related pairs from the analysis did not change the significance of the results and were therefore included.
Blood sampling and drug treatment
Procedures for blood sampling and drug treatment have been described in detail previously (Capitanio et al. 2005). Briefly, at 90–120 days of age, animals were separated from mothers or pair-mates on the morning of testing at approximately 0830 h, relocated to the testing room and housed in individual indoor cages (0.58 × 0.66 × 0.81 m). Blood was sampled through femoral venipuncture four times over a 24-h period (only the first three blood samples are considered here), and each sample was decanted into ethylenediaminetetraacetic acid-treated collection vials. The first sample was collected at 1100 h (AM sample), approximately 2.5 h following social separation/relocation. The second blood sample was collected approximately 5.0 h after the first sample at 1600 h (PM sample). Subjects were then immediately injected intramuscularly with dexamethasone (500 μg/kg; American Regent Laboratories, Shirley, NY, USA). The next blood sample was taken through femoral venipuncture 16.5 h later, at 0830 h (DEX sample). Throughout this 24-h period, subjects were also tested on a variety of standardized behavioral challenges that will not be described here, as they are incidental to this analysis. Standardized procedures were designed to ensure that each subject had comparable experiences with all other subjects who underwent assessment. A detailed description of these behavioral tests can be found in a previously published report (Capitanio et al. 2005). During testing, subjects were allowed access to water and chow ad libitum.
Plasma and peripheral blood mononuclear cell extraction
Whole blood samples were centrifuged at 1500 g for 10 min at 4°C. Plasma was removed and decanted into 1.5 ml sarstedt (VWR, South Plainfield, NJ, USA) tubes for storage at −80°C. Peripheral blood mononuclear cells (PBMCs) were isolated from the remaining sample within 1 h of sampling. White blood cells were aliquotted to RPMI media (Invitrogen, Inc., Carlsbad, CA, USA), supplemented with 10% fetal bovine serum and applied to lymphocyte separation media (MP Biomedicals, Solon, OH, USA). Samples were centrifuged at 650 g for 30 min at 23°C. The purified phase was washed three times with media, centrifuged and then resuspended in Trizol RNA stabilizing reagent (Invitrogen). Samples were stored for no longer than 1 year at −80°C.
Genomic DNA was extracted from the organic phase of the Trizol-separated PBMC preparation according to standard phenol–chloroform extraction procedures. Polymerase chain reaction (PCR) analysis targeted a 21 base pair (bp) insertion/deletion region in the rhesus macaque serotonin transporter promoter region previously identified (Lesch et al. 1997). The insertion variants of rh5-HTTLPR were named ‘long’ and deletion variant named ‘short’ alleles by that group (Lesch et al. 1997). Genotyping was conducted in 25 μl reactions consisting of 1.5 mM MgCl (Promega, Madison, WI, USA), 3× enhancer buffer with betaine, 20 μM primer (F1 and R1), 10× amplification buffer (Promega), 2.0 μM deoxyribonucleotide triphosphate and 0.5 U Taq polymerase (Promega). Primers sequences were as following: STR F1 5′GGCGTTGCCGCTCT GAATGC 3′ and STR R1 5′ GAGGGACTGAGCTGGACAACCAC 3′. Polymerase chain reaction was performed for 35 cycles of initial 5 min denaturation at 95°C, annealing at 52°C, extension at 74°C, denaturation at 95°C and each step for 30 seconds. A final extension step lasted for 5 min. Following amplification, rh5-HTTLPR products were cleaved using restriction enzyme PstI for at least 1.5 h at 37°C. Samples were then run on a 3% agarose gel with ethidium bromide for DNA fluorescence. Animals with two long rh5-HTTLPR alleles yielded bands at 41, 45, 50, 87, 151 and 302 bp. Animals with two short alleles yielded bands at 45, 50, 87, 172 and 302 bp. Finally, animals with one of each long and short allele yielded bands at 41, 45, 50, 87, 151, 172 and 302 bp. Gels were visualized using ultraviolet light. Three samples for each genotype were confirmed through direct sequencing. Genotypes were grouped according to putative activity levels: high (l/l homozygotes) and low (l/s heterozygotes or s/s homozygotes). DNA for four subjects was lost, thus genotype data include only 49 subjects.
5-HTT expression analysis
Total RNA was isolated from purified PBMCs stabilized with Trizol reagent (Invitrogen). Lysed cells were subjected to phenol extraction and washed with ethanol. RNA was quantified using a spectrophotomer reading at 260/280 nm. One microgram of RNA was then treated with deoxyribonuclease (Ambion Inc., Austin, TX, USA) and incubated at 37°C for 60 min. Samples were then subjected to reverse transcriptase–PCR to synthesize complementary DNA (cDNA). Complementary DNA preparation entailed extension with random hexamers (GE-Amersham Biosciences, Piscataway, NJ, USA), and reverse transcriptase using Moloney Murine Leukemia Virus-Reverse Transcriptase (MMLV-RT; Invitrogen). Complementary DNA samples were stored at −20°C for no more than 6 months. Real-time PCR was conducted using ABI PRISM 7700 Sequence Detection System. A human Taqman quantitative gene expression assay (Applied Biosystems Inc., Foster City, CA, USA) targeting a region of the 5-HTT gene that we determined to be 100% homologous with rhesus macaques using the published sequence (GenBank accession number AF285761) was used for quantitative PCR. B-actin was selected as an endogenous control as it was determined to amplify at comparable efficiency with 5-HTT (the slope of the log input amount versus ΔCt < 0.1). Additionally, we confirmed that B-actin did not respond to the stressor or to dexamethasone treatment (all P > 0.05). B-actin probe (Applied Biosystems)/primer (Integrated DNA Technologies, Inc., Coralville, IA, USA) sequences are as follows: probe 5′ ACC ACC ACG GCC GAG CGG 3′; forward primer 5′ TGA GCG CGG CTA CAG CTT 3′ and reverse primer 5′ CCT TAA TGT CAC ACA CGA TT 3′. Complementary DNA (83.5 ng) was applied to the primer/probe cocktail and Universal Master Mix (Applied Biosystems) and run in duplicate. All samples were amplified as following: 2 min 50°C; 10 min denaturation at 95°C, 40 cycles each: 15 seconds at 95°C, 1 min at 60°C. Human RNA (Invitrogen) was included on each plate as a control to establish interplate variability. Interplate assay coefficients of variance (calculated as the SD of a human control sample/mean of the human control samples) were less than 3%. 5-HTT values were calculated using the 2−ΔΔCT method (5-HTT Ct–B-Actin Ct).
Plasma cortisol was assayed in duplicate using commercially available kits (Diagnostics Products Corporation, Los Angeles, CA, USA). Interassay and intraassay coefficients of variance (calculated as the SD of relevant samples/mean of relevant samples) were 5.8% and 7.9%, respectively.
5-HTT and cortisol were analyzed using a repeated measures analysis of variance (anova) with time of sample collection (AM, PM and DEX) as within-subjects variables and with rearing condition, genotype and sex entered as between-subjects variables using SPSS software (SPSS Inc., Chicago, IL, USA). Sex did not contribute significantly to any model and was therefore removed for subsequent analyses. Significant main effects and interactions were investigated using independent samples t-tests. Pearson’s correlations were conducted between 5-HTT and cortisol measures in AM, PM and DEX samples. Interplate variability for both 5-HTT and cortisol was calculated by dividing the SD of control samples from each plate by the average of control samples.
The anova model was composed of a 2 × 2 × 3 (genotype × rearing × time) design. The three-way interaction was nonsignificant (P > 0.05). There was, however, a significant main effect of time (F2,94 = 5.02, P < 0.01) such that 5-HTT generally increased across time points. There was also a time × rearing interaction (F2,94 = 3.51, P < 0.05). Post hoc analysis showed that the time × rearing interaction was because of a significant increase in 5-HTT expression between AM and PM in MR infants (F1,32 = 13.76, P < 0.001), while NR infants did not exhibit this increase (Fig. 1). The time × rearing interaction was also because of higher 5-HTT expression post-dexamethasone treatment in MR infants than in NR infants (t1,51 = 2.65, P < 0.05), although groups did not differ in regulation between PM and DEX samples (P > 0.05).
The anova model was composed of a 2 × 2 × 3 (genotype × rearing × time) design. The three-way interaction was nonsignificant (P > 0.05) indicating that cortisol regulation between time points was not significantly impacted by genotype, rearing or by an interaction between them. However, there was a significant between-subjects effect of rearing such that MR infants had significantly higher overall cortisol than NR infants (F1,47 = 7.77. P < 0.01; See Fig. 2).
5-HTT and cortisol
PM 5-HTT expression was significantly and positively correlated with cortisol measures at all time points: AM (r = 0.30, p < 0.05), PM (r = 0.28, p < 0.05) and DEX (r = 0.43, P < 0.001). 5-HTT expression from the DEX sample was correlated with post-dexamethasone cortisol sample (r = 0.36, p < 0.01), and nearly significantly correlated with the AM cortisol sample (r = 0.27, p = 0.055).
Expression of 5-HTT has been predominately attributed to variation in the regulatory region of this gene (Heils et al. 1996) and has therefore been presumed to be relatively trait-like within individuals. We present evidence that early experience exerted a profound impact on peripheral 5-HTT regulation in response to a stressor, while, surprisingly, rh5-HTTLPR genotype did not. These data point to an important role of experience in the developmental regulation of 5-HTT.
Previous studies have successfully showed allelic influences on 5-HTT expression in lymphocytes (Bradley et al. 2005; Hranilovic et al. 2004; Heils et al. 1996), while the impact of experience on gene expression has not been shown. We did not find a significant effect of rh5-HTTLPR genotype alone, a result consistent with one study in humans (Hranilovic et al. 2004) but inconsistent with others (Bradley et al. 2005; Heils et al. 1996). Because the frequency of the s/s genotype is low and short allele believed to be dominant, however, we combined the l/s and s/s rh5-HTTLPR individuals into one group. We cannot preclude the possibility that the s/s genotype may have been associated with lower gene expression from l/l or l/s individuals. Furthermore, we cannot preclude the possibility that other, yet uncharacterized polymorphisms in this region may play a role in 5-HTT expression (Hranilovic et al. 2004). Intriguingly, however, early experience greatly influenced 5-HTT regulation during a stressor, consistent with studies which have shown that rearing history significantly impacts 5-HT function, although in the central nervous system, not in PBMCs (Ichise et al. 2006; Suomi 2005). Previous studies have shown that rh5-HTTLPR–environment interactions guide neurobehavioral development; in the present study, genotype did not significantly differentiate 5-HTT regulation within rearing groups. The lack of a significant effect may be attributable to low sample size: our observed power for the three-way interaction was inadequate to detect a significant effect if present (10%). Future studies with larger sample sizes should be conducted to validate these findings. Even in the event that allelic influences are demonstrated, it should be noted that experience plays a substantial role in 5-HTT regulation.
We hypothesized that the effects of experience and/or genotype on 5-HTT expression may be mediated by individual differences in one aspect of HPA function: GC availability. It should be noted that our study design has assumed that GC-receptor availability is comparable between groups, which is not necessarily the case. If GC complexes stimulate 5-HTT expression, we predicted that (1) plasma cortisol would predict 5-HTT gene expression, (2) that dexamethasone, a synthetic GC, would stimulate gene expression and do so differentially among rearing/genotype groups and (3) that differences in cortisol regulation between genotype/rearing groups would mirror differences in 5-HTT expression between groups. One out of three of these predictions were fully supported. First, AM cortisol levels predicted PM gene expression. We expected that AM cortisol levels would be the best predictor of PM 5-HTT because these cortisol levels are the first reflection of the substantial increase in adrenal output following separation and relocation. The increased cortisol would result in binding with GC receptors, and the resulting GC complexes would act as transcriptional enhancers, a process which may take from minutes to hours. While it is true that PM and DEX cortisol levels also predicted PM gene expression, these results could reflect the trait-like nature of HPA response in primates, especially under conditions of challenge (Capitanio et al. 1998; Suomi 1983). It is also possible that 5-HTT gene expression and HPA measures are correlated merely because of their association with a third factor, reflecting some other physiological or psychological process. Clearly, more information is needed. Our second prediction that dexamethasone would stimulate 5-HTT gene expression differentially among genotype/rearing groups was partially supported. The result of this manipulation was that 5-HTT expression increased significantly, regardless of rearing, and, inconsistent with a previous report in humans, regardless of genotype as well (Glatz et al. 2003). Mother-reared infants had significantly higher 5-HTT expression in the post-DEX sample than NR infants, but as 5-HTT regulation between PM and DEX samples was nonsignificant, we cannot conclude that rearing differences in 5-HTT were the result of DEX treatment. Indeed, we acknowledge that DEX treatment effects were confounded by time, as all subjects experienced 16 more hours of a social/maternal separation along with the dexamethasone treatment. It is possible that individuals continued to upregulate 5-HTT from PM until the following morning, regardless of treatment. This stress effect may have obscured any genotype/rearing group differences in DEX potency. Our final prediction that differences in cortisol regulation between rearing/genotype groups would mirror differences in 5-HTT expression between groups was not borne out. Despite a significant 5-HTT upregulation, cortisol levels in MR infants did not change significantly in the 5 h between the AM and PM samples. We can only conclude that upregulation of 5-HTT that we observed in MR individuals is certainly possible without concomitant changes in cortisol availability within this time frame. Without comprehensive support for all three of our hypotheses, we cannot deduce that GC availability is important for 5-HTT transcriptional regulation at all time points and likely does not mediate individual differences in 5-HTT expression.
Our data illustrate an exciting alternative to the notion that early experience leads to stable individual differences in peripheral 5-HTT function: regulation of 5-HTT was influenced by early experience at an early stage in development. Although we predicted that 5-HTT expression would be stable and ‘trait-like’ as are other measures of 5-HT function (blood 5-HT, Kinnally et al. 2006; CSF 5-HIAA, Higley & Linnoila 1997; brain 5-HTT availability, Ichise et al. 2006), we did not observe differences in gene expression in the initial blood sample based on genotype or rearing. Instead, we found that MR infants upregulated 5-HTT in response to a stressor, while NR infants did not. While it is true that the initial sample was not a perfect ‘resting’ measure, as the 2 h prior to sampling were likely stressful to the infant and may have been an adequate amount of time for changes in gene expression, we suggest that the dynamics of the serotonin transporter system, rather than its baseline, may be trait-like at this early stage in development. The relationship of developmental regulation of 5-HTT in response to a stressor and adult 5-HTT function would be of great interest.
The significance of individual differences in 5-HTT regulation in tissue that is not directly involved with brain function or behavior is not clear. As there is some evidence that platelet and lymphocyte 5-HT activity may mirror central 5-HT activity, it is tempting to speculate that early experience impacts neural 5-HTT expression in a similar manner. 5-HTT upregulation in response to a psychological challenge, which we observed in MR individuals, may lead to an increased capacity for serotonergic reuptake and reduced serotonergic firing during or immediately following a stressor. Such neural events have previously been associated with greater 5-HT function and reduced emotional dysfunction in a variety of species (Adell et al. 1997; Hranilovic et al. 2005; Racca et al. 2005; Senkowski et al. 2003), consistent with the notion that 5-HTT regulation in response to stress may be an important factor in neurobehavioral and psychiatric outcomes. Future research that generalizes these findings to the brain could lend great insight into the molecular basis of neurobehavioral phenotypes that persist across the lifespan.
We gratefully acknowledge the contributions of Linda Fritts, Dr David Wolf, Genesio Karere, Laura DelRosso and Laura Agostonelli to this work. We also appreciate the comments of several anonymous reviewers on earlier drafts of this manuscript. Funded by National Institutes of Health (NIH) NCRR17584 (to L.A.L.), NIH RR019970 (to J.P.C.). The authors have no biomedical financial interest in or conflict of interest with the findings reported.