هنگامی که سوگ باعث بیماری شما می شود: سوگ ناشی از التهاب سیستمیک یک سوال از ژنوتیپ است

Abstract Although bereavement is associated with increased morbidity and mortality in the surviving spouse, some widow(er)s remain healthy. Genetic variability in expression of inflammatory markers in response to stress may be the key to this observation. The present study compares bereaved vs. married/partnered older adults, investigating the impact of bereavement status, pro-inflammatory cytokine single nucleotide polymorphisms (SNPs) on circulating markers of inflammation and hypothesizing a gene by environment (GxE) effect. The study sample included 64 older adults, of which 36 were widow(er)s. Circulating levels of inflammatory markers IL-6, IL-1RA and sTNFRII were measured. Participants were genotyped for SNPs in the IL-6 gene (IL-6 −174 and −572), the IL-1β gene (IL-1β −511), and TNF-α gene (TNF-α −308). Grief severity was assessed with the Inventory of Complicated Grief. Bereaved participants had higher circulating levels of IL-1RA and IL-6. This increase could not be explained by pro-inflammatory genotype frequency differences, or Complicated Grief diagnosis. However, a GxE effect with the IL-6 −174 SNP moderated individual vulnerability to higher circulating levels of inflammation resulting from bereavement exposure. These results suggest a possible mechanism for the increase in morbidity and mortality in the surviving spouse. Genetic variability interacts with an environmental stressor, leading to increased inflammatory markers in genetically susceptible subjects only. For these patients, clinical interventions for bereavement-related stressor reduction might be crucial for overall health.

1. Introduction The death of a spouse is one of the most distressing life events, and is associated with an increase in morbidity and mortality risk, independent of a host of covariates (Boyle et al., 2011). Stressful life events are linked to dysregulation of the immune system and increased cellular inflammatory signaling (Irwin and Cole, 2011). Because it is our long-term goal to determine the mechanism linking bereavement and morbidity/mortality, and because substantial evidence shows that increases in inflammatory markers such as interleukin (IL)-6 are associated with mortality risk (Ershler and Keller, 2000), we investigated whether responses to spousal bereavement are impacted by markers of inflammation. Chronic stress as a result of caregiving (i.e., those taking care of a severely ill family member) or social isolation is associated with the up-regulated gene expression and production of systemic markers of inflammation (Kiecolt-Glaser et al., 2003 and McDade et al., 2006), especially of IL-6, IL-1 and TNF-alpha (Dantzer, 2001). Further data indicate genetic variability in the expression of inflammatory markers in response to stress (Cole et al., 2010 and Cole et al., 2011). For example, the presence of the guanine/cytosine (G/C) single nucleotide polymorphism (SNP) in the promoter of the IL-6 gene (IL-6) −174 bp upstream of the transcription start site affects the binding of a β-adrenergic-sensitive transcription factor, GATA-1 ( Cole et al., 2010). Following an in vitro β-adrenergic stimulus, the IL-6 −174G SNP leads to increased IL-6 production, while the IL-6 −174C SNP does not. In vivo, an association between depression and increased mortality risk was found only in IL-6 −174GG homozygous patients ( Cole et al., 2010), but levels of inflammatory markers were not characterized. Thus, it is not known whether genetic variability in cytokine gene polymorphisms alters the influence of life stress on the production of inflammatory markers in humans. Given evidence that some widow(er)s show an increase in morbidity and mortality, but others remain healthy and thus appear to be protected from this “widowhood effect” (Buckley et al., 2011), the present study hypothesizes that bereavement may be a socio-environmental stress that interacts with pro-inflammatory genetic variation to produce modulated levels of inflammation. This gene by environment (GxE) interaction predicts that among widow(er)s, those with genotypes that are associated with higher expression of inflammatory markers would exhibit greater systemic signs of inflammation, e.g. higher circulating levels of inflammatory markers. Conversely, widow(er)s with low pro-inflammatory genotypes are hypothesized to have similar circulating levels as non-bereaved individuals. The work reported here compares bereaved and healthy married/partnered older adults, first investigating the impact of bereavement status on circulating levels of inflammatory markers IL-6, IL-1RA as a marker of IL-1 activity (Arend et al., 1998) and soluble tumor necrosis factor receptor II (sTNFRII) as a surrogate marker of TNF-α levels (Diez-Ruiz et al., 1995 and Schuld et al., 1999). Second, corresponding to our circulating measures, we investigated the impact of pro-inflammatory cytokine SNPs in the IL-6, IL-1β, and tumor necrosis factor alpha (TNF-α genes on the mentioned inflammatory markers and a possible interaction with bereavement status. Given evidence of in vitro data for the GxE pathway for the IL-6 −174 SNP, we primarily focused on the interaction between bereavement stress and this polymorphism.

. Results Group comparisons on demographic characteristics using ANOVAs and Chi-square tests are shown in Table 3. There were no significant group differences between bereaved and non-bereaved study participants. Table 3. Demographic characteristics. Bereaved (N = 36) Non-bereaved (N = 28) F-value/X2 P-value Mean/N SD/% Mean SD/% Age 72.9 5.8 72.4 4.2 0.17 0.69 Gender (female) 23 64% 15 54% 0.7 0.4 Ethnicity (non-Caucasian) 10 28% 5 18% 0.86 0.35 Employment (retired) 23 64% 15 54% 0.7 0.4 Education (post graduate) 14 39% 13 46% 0.37 0.55 Years married/partnered 35.8 17.6 41.2 11.2 2 0.16 Body mass index 26.6 5.2 27.4 5.6 0.31 0.58 Alcohol (drinks per week) 2.1 4.3 1.7 2.9 0.18 0.67 Continuous variables: mean (±SD), F-value (ANOVA); categorical variables: N (%), X2 value (chi square test). Table options 3.1. Group differences in levels of circulating markers of inflammation Body mass index (BMI) and age have been shown to affect inflammatory marker production and therefore were included with bereavement status in regression analyses (O’Connor et al., 2009). Controlling for BMI and age, bereavement was a significant predictor of higher levels of IL-6 (2.8 ± 2.2 pg/ml vs. 1.9 ± 0.9 pg/ml, β = .26, p = .04), and IL1-RA (305 ± 167 pg/ml vs. 227 ± 94 pg/ml, β = .33, p = .01), but not for sTNFRII (2903 ± 1066 pg/ml vs. 2750 ± 725 pg/ml, β = .08, p = .55). In order to rule out that differences between the groups could be attributed to more current stressors other than bereavement, we correlated levels of pro-inflammatory cytokines with PSS score, SRRS-r score and time since bereavement. Linear correlations were all not significant. 3.2. Group differences in genotype frequencies Genotype and allele frequencies are presented in Table 2. In spite of the ethnic diversity, genotype frequencies met Hardy–Weinberg equilibrium and were consistent with the literature (see HapMap allele frequencies at SNP database of the National Center for Biotechnology Information, http://www.ncbi.nlm.nih.gov/snp/). None of the grouped genotypes differed in gender, but significant differences in ethnicity (Caucasian vs. non-Caucasian) were observed for the IL-6 −572 SNP (Fisher’s exact test, one-sided p < .01), and IL-1β −511 SNP (Fisher’s exact test, one-sided p = .01). Therefore ethnicity was controlled for in all analyses involving IL-6 −572 and IL-1β −511 SNPs. As the reported differences in pro-inflammatory cytokine levels between bereaved and non-bereaved participants might be moderated by the difference of frequency of a high/low producing cytokine-genotype we investigated proportional genotype distribution between the two groups. Analyses revealed a significant difference in distribution between the groups only for TNF-α −308 (Fisher’s exact test, one-sided p = .04). Twenty-eight percent bereaved versus 7% non-bereaved participants were identified as carriers of the high-producing A-allele. Although this might have been due to sampling bias, as described above, circulating levels of sTNFRII did not differ between groups, so the unbalanced distribution of genotype appeared not have a detectable effect on circulating level. 3.3. Gene by Environment (GxE) effect Given evidence of in vitro data for the GxE pathway for the IL-6 −174 SNP, we primarily focused on this polymorphism. To determine whether genotype moderated IL-6 as a function of group, simple regressions were computed separately for bereaved/non-bereaved ( Fig. 1 and Table 4). Despite the death of a spouse, bereaved IL-6 −174C allele carriers had virtually the same mean level of IL-6 as non-bereaved −174C carriers (controlling for BMI β = .06, p = .75). In contrast, bereaved IL-6 −174G homozygotes had over twofold higher mean levels of IL-6 compared to non-bereaved −174G homozygotes (controlling for BMI β = .37, p = .03). Using regression analyses to investigate the relationship between bereaved/non-bereaved groups and IL-6 −174 genotype revealed an interaction term that approached significance for predicting IL-6 levels (β = −0.38, p = 0.09), over and above the main effects of bereavement group and genotype. Gene (IL-6 −174) by environment (bereavement) effect on circulating IL-6. ... Fig. 1. Gene (IL-6 −174) by environment (bereavement) effect on circulating IL-6. aSignificant difference in circulating IL-6 between bereaved and non-bereaved subjects with IL-6 −174GG genotype. F (1,30) = 5.09, p = .03. Figure options Table 4. Gene (SNP) by environment (bereavement) effect on inflammatory marker levels. N IL-6 pg/ml SD β P-value IL-6 −174C carrier Bereaved 17 2.2 1 Non-bereaved 15 2.1 1.1 0.06 0.75 IL-6 −174GG Bereaved 19 3.4 2.7 Non-bereaved 13 1.6 0.5 0.37 0.03 IL-6 −572GG Bereaved 29 2.7 1.9 Non-bereaved 23 2 0.9 0.18 0.2 IL-6 −572C carrier Bereaved 7 3.6 3 Non-bereaved 5 1.3 0.3 0.56 0.09 N IL-1RA pg/ml SD β P-value IL-1β −511G carrier Bereaved 26 311.1 189 Non-bereaved 24 239.6 94.8 0.27 0.05 IL-1β −511AA Bereaved 10 289.7 91.5 Non-bereaved 4 149.7 29.8 0.63 0.01 N sTNFRII pg/ml SD β P-value TNF-α −308GG Bereaved 26 2665 738 Non-bereaved 26 2764 750 −0.06 0.7 TNF-α −308A carrier Bereaved 10 3522 481 Non-bereaved 2 2561 178 0.28 0.38 β and P-values from linear regressions, controlling for BMI; biomarker data were log-transformed for statistical analyses. Table options Analyses were performed to investigate whether a similar GxE effect exists for the other polymorphisms (Table 4). First, bereaved IL-6 −572G allele (low IL-6 producing) homozygotes had similar mean levels of IL-6 as non-bereaved −572G homozygotes (controlling for BMI β = .18, p = .20). In contrast, bereaved IL-6 −572C allele carriers trended toward higher levels of IL-6 compared to non-bereaved −572C allele carriers (controlling for BMI β = .56, p = .09). Controlling for ethnicity in each of the preceding analyses that had sufficient power (n > 20, see Table 4) did not change the results. Second, bereaved IL-1β −511G allele (low IL-1 producing) carriers had different mean levels of IL-1RA as non-bereaved −511G carriers (controlling for BMI β = .27, p = .05). Also, bereaved IL-1β −511A homozygotes had significantly higher mean levels of IL-1RA compared to non-bereaved −511A homozygotes (controlling for BMI β = .63, p = .01). Again, controlling for ethnicity in each of the preceding analyses that had sufficient power did not change the results. Third, no GxE effect could be seen for TNF-α −308 SNP. Bereaved TNF-α −308G allele (low TNF producing) homozygotes had similar mean levels of sTNFRII as non-bereaved −308G homozygotes (controlling for BMI β = .06, p = .70). Similarly, bereaved TNF-α −308A allele carriers had equal levels of sTNFRII compared to non-bereaved −308A allele carriers (controlling for BMI β = .28, p = .38). 3.4. Effects of grief severity (Complicated Grief) The intensity of grief symptoms may influence the extent of the stress response and thus could also have an impact on cytokine production. We assessed all bereaved participants to determine whether they met criteria for Complicated Grief, a syndrome characterized by persistent separation distress with feelings of overwhelming yearning and preoccupation with the deceased (Prigerson et al., 2009 and Shear et al., 2011). Regressing grief severity (ICG score) on levels of inflammatory markers did not reach significance for any of the markers, controlling for age and BMI (IL-6: F (3, 31) = 1.76, p = .18; IL-1RA: F (3, 31) = .49, p = .69; sTNFRII: F (3, 31) = 1.91, p = .15). Thirteen of the 36 bereaved subjects met diagnostic criteria for Complicated Grief. Not surprisingly, regressing the dichotomous subgroups Complicated Grief and Non-complicated Grief on the cytokine markers, and controlling for age and BMI, revealed that group did not predict any of the markers. Genotype frequencies between Complicated Grief and Non-complicated Grief subgroups were not significantly different, which may be due to lack of power due to small sample sizes rather than to biological differences. In addition, BDI-II data was available on a subset of the sample: 23 bereaved subjects (9 Complicated Grief) and 15 non-bereaved subjects. Since we excluded for MDD the overall BDI scores were very low. We could not find a GxE effect for the IL-6 −174 SNP and depressive symptoms. This was also true for the other SNPs.