Oxytocin receptor (OXTR) is not associated with optimism in the Nurses’ Health Study

Abstract

The oxytocin system is implicated in complex social behavior such as socioemotional functioning, social recognition and bonding, and in modulating biological responses to stress.1,2 The oxytocin receptor (OXTR) gene encodes the OXTR; a G-protein-coupled receptor that mediates the effects of the neurohormone oxytocin. 3,4 Prior studies have reported associations between an intronic single-nucleotide polymorphism (rs53576) of OXTR and various stress-related and psychological traits.4,5 Most recently, Saphire-Bernstein et al6 reported that rs53576 A allele carriage conferred lower optimism relative to G homozygotes, among 344 men and women aged 18–36 years of various ethnicities. We sought to replicate the association between OXTR rs53576 and optimism in 1229 women from the Nurses’ Health Study (NHS). The NHS was established in 1976 when 121 700 women registered as nurses aged 30–55 years from 11 US states completed a mailed questionnaire on medical history and lifestyle characteristics.7 Every 2 years, follow-up questionnaires are sent. The 2004 and 2008 questionnaires included information on antidepressant use and the revised life orientation test,8 the same dispositional optimism measure used in the Saphire-Bernstein et al.6 study. We generated five different optimism scores to provide a thorough investigation of the previously reported association. Year-specific optimism scores were derived by taking the mean of all non-missing items in 2004 (α =0.77) and also in 2008 (α =0.75). Individuals with >2 items missing were excluded (up to 8% of analytic sample). Using both assessments, an overall mean optimism score was also derived, as were two subscale scores reflecting an ‘optimistic’ or ‘pessimistic’ outlook.8 Genotypes for the current analysis were available from two case-control nested genome-wide association studies, initially designed to assess kidney stone disease and glaucoma (total n= 1294).9,10 Samples were genotyped for OXTR rs53576 at the Broad Center for Genotyping and Analysis using the Illumina 610Q or 660Q array (Illumina, San Diego, CA, USA). Genotyping success rate for rs53576 was > 98%.9,10 Principal components analyses using genome-wide data were conducted to assess self-reported race. Genetically inferred non-Caucasian samples were too few (<3% of analytic sample) for meaningful analysis and therefore excluded. The 2004 and 2008 optimism scores were moderately correlated (r = 0.65) in the 1229 women with complete genetic and phenotype data (2004 mean age = 71.3 years, s.d. = 6.7). Overall optimism scores ranged from 1.33 to 4.25 (overall mean = 4.13, s.d. = 0.68). Frequencies (%) of the GG, GA and AA genotypes were 491 (40%), 559 (45%) and 179 (15%), respectively, and did not depart from Hardy–Weinberg equilibrium (P = 0.33). The A allele frequency was 38% in NHS versus 28% among white participants (n = 87) in the Saphire-Bernstein et al.6 study. Table 1 indicates no significant differences in means across genotypes for any optimism scores (all PADD > 0.35). Optimism scores also did not differ between G homozygotes and A carriers (all PDOM > 0.17). Linear regression models adjusting for age, genotype platform, and case-control status yielded no significant association between rs53576 and any optimism score, regardless of genetic model (all P > 0.16). The NHS sample had substantial power ( > 99%) to detect the effect size reported previously: β = −0.168 under a dominant model.6 Table 1 Mean (s.d.) optimism scores by OXTR genotype Unlike our sample, Saphire-Bernstein et al.’s6 sample consisted of men and women recruited for a stress and coping study, was relatively young (mean age 21) and had mixed ethnic ancestry (for example, 36% Asian). However, although the frequency of the G allele and mean optimism scores differed somewhat across ethnic and gender groups in the Saphire-Bernstein et al.6 study, the pattern of associations did not. Recognizing other differences between the two samples, we conducted several sensitivity analyses (Table 1), considering effects stratified by age or education, excluding women reporting antidepressant use, or with disease (that is, kidney stone or glaucoma). Associations between rs53576 and optimism were null in all analyses. We were limited in power to restrict analysis to even younger women. However, a formal test for OXTR–age interaction was marginally significant (P = 0.10), suggesting age may be an important modifier for future exploration. In summary, we did not replicate the association between rs53576 and optimism in our sample of Caucasian women. Our findings do not exclude a possible role for OXTR in psychosocial resources. The link between rs53576 and optimism may be applicable to a population with characteristics similar to individuals in Saphire-Bernstein et al’s6 study. However, our findings suggest caution is warranted when considering the association between OXTR and optimism, and additional research is needed to determine the generalizability of the results reported by Saphire-Bernstein et al.6