Association between Epstein-Barr virus serological reactivation and psychological distress: a cross-sectional study of Japanese community-dwelling older adults

Psychological distress, stress-induced immunosuppression, and EBV serological reactivation

Reactivation of EBV has been shown to occur as a result of impaired cellular immune responses evoked by psychosocial stress in old age. Older people may experience a variety of psychosocial stressors, including marital problems [21], social exclusion because of living alone [22], and interpersonal stress related to refusal or being abandoned [23]. A previous study showed that patients with breast cancer who experienced adversity in childhood, such as the death of a parent or serious problems between their parents, had higher antibody titers for EBV than those who did not experience such events [24]. Evidence suggests that social stress for young people (e.g., student examinations) also alters their cellular immunity [25, 26]. A case-control study showed that EBV reactivation, as defined by EBV DNA in saliva, among astronauts who participated in the Space Shuttle mission was increased during the mission compared with the control group [27]. Therefore, a possible explanation for our results is that predisposition to psychological distress may impair the host’s cellular immune response, which leads to EBV serological reactivation [28].

A prospective observational study reported that depressive symptoms in late-gestation mothers were associated with prenatal EBV serological reactivation, as defined by elevation of EA-IgG levels (adjusted OR 2.74, 95% CI: 1.23, 6.08) [20]. In addition, a case-reference study involving pregnant women in the United States with/without pre-pregnancy depression found that depressed women had a higher frequency of EBV serological reactivation (as defined by the EA-IgG antibody) compared with reference participants (48% vs. 30%, P = 0.01) [29]. These results suggested that psychological distress may increase the risk for EBV serological reactivation during pregnancy. Our results were consistent with the results of these previous studies.

Sex difference in the association between EBV serological reactivation and psychological distress

We found that women accounted for the majority (71%, 328/463) of the EBV serological reactivation group. After adjusting for sex, EBV serological reactivation was only significant in the group of older women. Consistent with our study, previous studies only found an association between EBV reactivation and psychological distress in women [19, 30]. However, an explanation for this sex difference remains to be established. Sex differences in inflammatory, immunological, or endocrinal (e.g., sex hormones) physiology may explain this discrepancy. This finding highlighted the importance of considering sex disparities in further investigations to elucidate the mechanisms underlying the association. As our sample included fewer men than women, we cannot rule out potential random error because of lack of power.

Another possible explanation for our results is a sex difference in cellular immunity modulated by insulin-like growth factor I (IGF-1). Serum IGF-I levels decrease with age [31]. However, in a cohort of healthy adults in the United States, the decline in IGF-I levels was more pronounced in females than in males [31]. Individual susceptibility to mental diseases after experiencing psychosocial stress may also differ between men and women. Mouse experiments showed that the stress response to traumatic brain injury increased the transfer of serum IGF-1 to the brain [32]. In addition, serum IGF-1 was negatively correlated with anxiety behavior in mice measured 1 week after stress from the elevated plus maze [32]. In a cohort of 2,686 Dutch adults, IGF-1 was associated with stressful life event experiences [32]. An observational cohort study involving 8,791 Japanese pregnant women found that women with high serum IGF-1 levels in early pregnancy were less likely to develop postpartum depression than women with low serum IGF-1 levels [33]. These results implied IGF-1 is involved in mood homeostasis and has a protective role in psychological distress. IGF-1 regulates the immune function in various aspects of T cells, B cells, and monocytes through binding to the IGF-1 receptor and may therefore act protectively in EBV serological reactivation. In older age, women have a more pronounced decline in IGF-1 levels than men. Therefore, the stress-induced immune response modified by IGF-1 may explain the sex difference in the association between psychological distress and EBV serological reactivation.

We also found a strong association in older women after age stratification, which supported the EBV serological reactivation mechanism being due to psychological distress. Older participants tended to have lower IGF-1 levels and were more sensitive to stress, which may have increased their risk for altered immunity modulation. However, no such reaction was observed in men, and the association was strong in women, which supported the existence of a clear sex difference.

Study strengths and limitations

A major strength of this study was the large number of participants (N = 2,821). Another strength was our use of EBV EA-IgG instead of EBV viral capsid antigen (VCA) IgG. The seropositivity of EBV VCA IgG for children aged 5–9 years exceeds 90% in east Asia, and is especially high in Japanese children (100% in 1968) [34]. A recent study also showed that the VCA IgG was almost 100% positive among patients with inflammatory bowel disease aged ≥60 years [35]. In terms of the seropositivity difference among generations, our study cohort comprised a generation born between 1924 and 1959. Compared with Western populations, Asian children tend to be infected early in life (i.e., before 1 year of age) and >90% of children aged 5–9 years are infected [36]. In addition, the recent time trend of EBV VCA IgG seropositivity among children in Japan showed a gradual decrease, but epidemiology data for Japan suggested this was expected to be ≥90% between 1924 and 1959 [36]. Because EBV VCA IgG is ubiquitous among Japanese older people, the detection of EBV serological reactivation using EBV VCA IgG titers may have had a ceiling effect, meaning that most participants among the older population remained positive. Therefore EBV EA-IgG may be a worthwhile marker to include in further studies of EBV in older adults.

Some limitations of this study need to be considered. First, our cross-sectional design meant we were unable to establish cause–effect relationships. Second, there is potential for random error because of the sample size, which might have led to the sex difference in the association between EBV serological reactivation and psychological distress. Third, there might have been some diagnostic uncertainty. We defined psychological distress using an interview based on the K6, not a self-administered questionnaire. Fourth, we had no information about diagnosis or treatment of mental diseases. These factors may be potential residual confounders. Fifth, although we defined EBV serological reactivation using EBV EA-IgG, we could not assess EBV DNA to detect reactivation. Sixth, we cannot completely rule out the existence of primary EBV infection in our study. However, because of our large sample size and the large proportion of older adults infected in early life in Japan [36], the presence of potential cases of primary EBV infection is unlikely to change our conclusions. Finally, we cannot rule out psychological distress due to chronic somatic diseases, especially autoimmune diseases such as systemic lupus erythematosus and antiphospholipid syndrome. However, our results were adjusted for chronic illnesses, including rheumatoid arthritis.

Study setting and participants

We conducted a cross-sectional survey in Goto City, which is located in the Goto Islands, which comprise the western archipelago of Nagasaki Prefecture. The population of Goto City was 37,327 people in 2015, of which 36.8% were aged 65 years and over. In Japan, screening and treatment for non-communicable diseases is conducted based on the Health and Welfare for the Aged Act, and Goto City conducts health examinations for adults aged 40 years and over who live in the community. These health examinations are held at community centers in all districts of Goto City from April to September every year.

The Nagasaki Islands study was conducted with the cooperation of Goto City, and targeted chronic diseases such as atherosclerosis, cerebrovascular diseases, rheumatoid arthritis, frailty, and sarcopenia. We distributed flyers to all family units in the study areas, which explained our study as an additional medical checkup free of charge to the population aged 60 years or older. The participant recruitment period was from 2017 to 2019. We approached all adults who underwent a medical examination. We excluded 578 individuals because they were younger than 60 years (n = 516) or had missing data for EBV EA-IgG (n = 58), K6 score (n = 2), medication history of diabetes mellitus (n = 1), or low-density lipoprotein cholesterol (n = 1). This left 2,821 participants (1,049 men and 1,772 women) aged 60–94 years for inclusion in this study.

Examinations

The researchers and trained research assistants administered an interview to obtain information on: participants’ past medical history of stroke (yes or no), ischemic heart disease (yes or no), diabetes mellitus (yes or no), dyslipidemia (yes or no), joint pain (yes or no), history of hospitalization in the past 1 year (yes or no), marital status (married/unmarried/other or bereaved/divorced), and number of household members.

Body weight and height were measured with participants wearing light-weight clothing and without shoes, and the BMI of each participant was calculated. Blood samples were collected using heparin sodium and a siliconized tube at the time of clinical examination. Most participants may have fasted because they were instructed to fast before undergoing their health checkup. Serum concentrations of low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglycerides, aspartate aminotransferase, alanine aminotransferase, anti-cyclic citrullinated peptide (anti-CCP) antibodies, and HbA1c were measured by standard laboratory procedures.

Several variables were defined or calculated based on the questionnaire and laboratory results. Diabetes mellitus was defined as an HbA1c concentration ≥6.5% or current use of hypoglycemic drugs. Rheumatoid arthritis was defined as an anti-CCP antibodies concentration ≥4.5 U/mL and presence of joint pain. Liver disease was defined as an aspartate aminotransferase or alanine aminotransferase concentration ≥40 U/L. Dyslipidemia was defined as a low-density lipoprotein cholesterol concentration ≥140 mg/dL, high-density lipoprotein cholesterol concentration <40 mg/dL, triglycerides concentration ≥150 mg/dL, or use of lipid-lowering drugs. As an indicator of chronic illnesses, we used the modified Charlson Comorbidity Index to obtain the sum of the number of chronic illnesses (stroke, ischemic heart disease, diabetes mellitus, rheumatoid arthritis, and liver disease) [37].

Measurement of EBV EA-IgG

We defined EBV EA-IgG titers as an indicator of EBV serological reactivation with the below-mentioned rationale. EBV enters memory B cells and escapes the immune response by restricting the expression of the virus’s own genes [9]. EBV maintains latent infection on infected B cells and reactivates during the lysis cycle. The EBV EA-IgG response is considered indirect evidence of the initiation of lytic replication [38]. In this process, EBV infection elicits an IgG response to VCA, followed by an IgG response to early antigen. EBV EA-IgG increases during the first 3–4 weeks and is detected from a few months to up to 2 years after infection, but high titers have also been seen during reactivation [13, 39]. In contrast to EBV EA-IgG, VCA IgG remains generally detectable for life after acute EBV infection, and antibodies against EBV nuclear antigen remain expressed during the latent phase of the infection [40]. Therefore, in the present study, the seropositivity of EBV EA-IgG was defined as an indicator of serological reactivation of the lytic cycle of EBV infection.

An enzyme immunoassay kit (Denka Inc., Niigata, Japan) was used for EBV EA-IgG detection. This kit detects anti-EA antibodies that reflect a diffuse pattern: EA (D) IgG. Standard calibrators were used in each assay to calculate index values/optical density ratios, which served as a semiquantitative measure of antibody levels. All assays met predetermined quality control measures based on positive, negative, and blank controls. The results were based on index cutoff values: ≤0.5 was considered negative, 0.5–0.9 was considered equivocal, and ≥1.0 was considered positive according to the manufacturer’s instructions. We defined EBV serological reactivation using the EBV EA-IgG index cutoff value of 1.0.

Kessler 6 (K6)

The K6 comprises six items based on the item response theory and is used for screening psychological distress [41, 42]. The test has good reliability and validity, and the validity of the Japanese version has been established [43, 44]. Participants respond to the question: “During the past 30 days, about how often did you feel: 1) nervous, 2) hopeless, 3) restless or fidgety, 4) so depressed that nothing could cheer you up, 5) that everything was an effort, and 6) worthless?” Responses are on a 5-point scale (0 = none of the time to 4 = all of the time). We calculated the sum of the reported scores (range: 0–24) to obtain participants’ K6 scores. A higher score indicates greater severity. Based on a previous study that screened psychological distress in a Japanese population [42], we defined a K6 score of ≤4 as no psychological distress, and a score ≥5 as psychological distress.

Statistical analyses

Differences in mean values or proportions of variables by EBV serological reactivation were analyzed using Student’s t-tests for continuous variables (age, BMI, modified Charlson Comorbidity Index), and McNemar chi-square tests for categorical variables (sex, dyslipidemia, stroke, ischemic heart disease, diabetes mellitus, rheumatoid arthritis, liver disease, history of hospitalization in the past 1 year, marital status, living alone, and psychological distress). There were no missing values. We performed a Pearson’s bivariate correlation analysis to examine associations between EBV serological reactivation and other variables, and psychological distress and other variables. A scatterplot showing EBV EA-IgG and age for all participants and by sex was prepared.

Next, we performed logistic regression analysis and calculated ORs and 95% CIs for EBV serological reactivation and psychological distress. In the multivariable logistic regression analysis, adjustments were made a priori for variables based on previous studies [7, 45–49]; age (continuous variable), sex (dichotomous variable), BMI (continuous variable), dyslipidemia (dichotomous variable: no, yes), modified Charlson Comorbidity Index (continuous variable), history of hospitalization in the past 1 year (dichotomous variable: no, yes), marital status (dichotomous variable: married/unmarried/other or bereaved/divorced; 0, 1, respectively), and living alone (dichotomous variable: no, yes). We conducted logistic regression analysis stratified by sex to examine any sex-based differences. Because the correlation coefficient between EBV serological reactivation and age was positive, we considered it was worth investigating associations linked to age. Therefore, we stratified participants into younger and older groups by the median age (72.4 years). We used logistic regression analysis to examine EBV serological reactivation and psychological distress stratified by age. All analyses were two-tailed, and P-values < 0.05 were regarded as statistically significant. Statistical analyses were performed using STATA® version 14.0 (StataCorp, College Station, TX, USA).