Anxiety symptoms are a common mental problem among adolescents worldwide (1). Exposure to anxiety symptoms in adolescence exerts a long-term negative impact on cognitive, behavioral, and social abilities (2, 3). Adolescents with anxiety symptoms are more likely to develop anxiety disorders in subsequent adulthood (4), imposing a heavy economic burden on families, and society due to treatment (5). Previous studies showed that 10–30% of adolescents reported anxiety symptoms globally (4, 6), and about 27.3% of Chinese middle school students suffered from anxiety symptoms (7). Although anxiety symptoms are multifactorial (1), compelling evidence has demonstrated that adverse life events, including childhood maltreatment, are significantly associated with anxiety symptoms (8, 9). Moreover, a history of childhood maltreatment increases an individual's vulnerability to the development of anxiety (10).

Childhood maltreatment includes physical, emotional, and sexual abuse as well as physical and emotional neglect. The experience of childhood maltreatment exerts a negative influence on the development of cognitive and emotional regulation and leads to dysfunction of the hypothalamus-pituitary-adrenal (HPA) axis (9, 11). HPA axis, which regulates the synthesis and release of glucocorticoids, is one of the key stress response systems. Once an individual is continuously exposed to an environmental stressor, it can cause changes in the ability of the HPA axis to respond to stress, even resulting in dysfunction (12, 13). The dysregulation in this system has been proved as a risk factor for stress-related psychiatric disorders, such as anxiety and depression (14). Besides, previous studies have indicated that adolescents exposed to childhood maltreatment are more likely to emerge anxiety symptoms (10, 15).

Substantial literature supports an association between adverse life events and poor mental health outcomes (15, 16). However, the molecular mechanism that accounts for such a solid and lasting effect has not been adequately studied. Evidence suggests that sustained exposure to adverse life events such as childhood maltreatment may lead to dysfunction of the stress response system. To adapt to such changes, the organism will undergo stable and lasting epigenetic modifications, such as DNA methylation, which can affect gene expression (13, 17). DNA methylation is an epigenetic modification that regulates gene expression without changing the DNA sequence. A growing body of studies indicates that abnormal DNA methylation is associated with a higher risk of stress-related mental disorders (18, 19). Additionally, altered methylation may affect the expression of genes related to the stress response system, suggesting that it plays a vital role in the pathogenesis of mental disorders (20). FK506 binding protein 5 (FKBP5) gene is an essential regulatory in the stress response system whose product is involved in the regulation of glucocorticoid receptor (GR) activity in the intracellular ultra-short negative feedback loop (21). Previous studies have shown that alteration in DNA methylation of the FKBP5 gene is associated with mental disorders such as anxiety or depression (22, 23). Besides, evidence suggested that individuals with childhood maltreatment showed lower levels of methylation of the FKBP5 gene compared to those without maltreatment (21, 24). Consequently, the above evidence provides some information that DNA methylation of the FKBP5 gene may play an important role in the association between childhood maltreatment and anxiety symptoms.

We thus conducted this prospective study among Chinese adolescents to explore (1) the longitudinal association between childhood maltreatment and anxiety symptoms, (2) the association between childhood maltreatment and DNA methylation of the FKBP5 gene, and (3) the association of DNA methylation of the FKBP5 gene with anxiety symptoms at follow-up.

Our study indicated that there were significant differences between adolescents with and without anxiety symptoms in living arrangements, HSS, academic pressure, classmate relationships, teacher-classmate relationships, and current drinking. The confounding effects of these variables need to be considered when exploring the association between childhood maltreatment and anxiety symptoms. Besides, these results may help us identify adolescents at a higher risk of developing anxiety symptoms.

The current study provided evidence that childhood emotional abuse was associated with anxiety symptoms at follow-up after adjusting for age, gender, BMI, living arrangement, HSS, academic pressure, classmate relationships, teacher-classmate relationships, current drinking, and anxiety symptoms at baseline, which was consistent with previous studies. Banducci et al. (45) reported that individuals exposed to childhood emotional abuse were more likely to have a higher level of anxiety symptoms in adulthood. In a cross-sectional study in Finland, we could observe that severe emotional abuse was an important predictor of anxiety symptoms in adulthood (46). Similarly, a longitudinal study in the Netherlands suggested that people with emotional abuse in childhood were more likely to develop more unfavorable personality characteristics (such as neuroticism and hopelessness) and adopt poorer coping styles, which were negatively associated with the remission of anxiety symptoms (47). The experience of childhood abuse disturbed the development of individual emotional regulation. People who had been abused were lacking in emotional support, and were challenging to alleviate their negative emotions, which eventually led to the onset of anxiety symptoms and even developed into anxiety disorders subsequently (11, 48). Another explanation from the neurobiological mechanism was that the experience of childhood maltreatment could influence the normal function of the HPA axis. Sustained exposure to adverse life events such as childhood maltreatment may leave the stress response system in a hypersensitive state. To adapt to such changes, the organism will elevate the secretion of corticotropin-releasing hormone (CRH), leading to the hyperreactivity of the HPA axis as well as the autonomic nervous system (24, 49). The dysfunction of the HPA axis has been shown to be associated with an increased risk of poor mental outcomes (50).

Furthermore, one significant finding of this study was that a history of childhood physical and emotional neglect was negatively associated with DNA methylation in the promoter region of the FKBP5 gene after adjusting for the covariates. Although the association did not withstand the multiple hypotheses, compelling evidence suggested that individuals who were exposed to physical and emotional neglect were more likely to get a lower methylation level of FKBP5 DNA methylation. Klengel and colleagues demonstrated that individuals with childhood abuse exhibited significant demethylation of intron 7 of the FKBP5 gene compared to those without abuse (21). Moreover, a prospective longitudinal study reported that early life stress, especially childhood abuse, was a significant predictor of altered DNA methylation of an intron of the FKBP5 gene (49). Considering that childhood maltreatment occurred during a critical period of brain development when brain plasticity was enhanced, children were vulnerable to the neuropsychological effects of adverse life events, resulting in an increased likelihood of epigenomic alterations (51). FKBP5 gene plays a vital role in the stress response system, and childhood maltreatment as an environmental stressor may be more likely to cause alterations in FKBP5 gene methylation, and the alteration in DNA methylation of the FKBP5 gene has been shown to be a risk factor for mental disorders (22, 24, 49). These findings recommended the necessity for schools and families to be aware of the dangers of childhood maltreatment and provide emotional support to adolescents who have experienced childhood maltreatment.

Additionally, our final multiple linear regression models showed that DNA methylation of the promoter region of the FKBP5 gene showed an association with anxiety symptoms at follow-up after adjusting for the covariates. However, the association did not withstand the multiple hypothesis testing, consistent with previous literature. In the Detroit Neighborhood Health Study, Bustamante et al. reported that DNA methylation in the promoter region of the FKBP5 gene was not a significant predictor of mental health (29). However, extensive studies have highlighted the importance of DNA methylation of the FKBP5 gene in predicting mental health outcomes (44, 51). FKBP5 plays an important role in the intracellular ultra-short negative feedback loop that regulates GR activity. Moreover, demethylation of FKBP5 increases the expression of FKBP5, resulting in diminished negative feedback and reduced sensitivity to GR, which leads to prolonged glucocorticoid exposure and excessive stress response (22, 24). Nevertheless, this study did not confirm that FKBP5 DNA methylation was associated with anxiety symptoms at follow-up, a plausible explanation was that the effect of methylation of an individual CpG unit was weak, and most CpG units could not survive the multiple hypothesis testing (44), thereby the combined effect should be considered in later studies.

Our study has several limitations. First, we used a retrospective self-report scale to examine childhood maltreatment history, which can not preclude the recall bias. Moreover, adolescents might not be reluctant to respond to the experience of childhood maltreatment, resulting in an underestimate of maltreatment. Second, our study sample only derived from adolescents in school on the day of the survey and did not include those who were absent from school. The experience of childhood maltreatment and anxiety symptoms might be more prevalent in those students. Third, we only measured methylation in peripheral blood, and it was not clear whether the result accurately reflected DNA methylation in the brain. However, using peripheral blood is a widely accepted method in detecting DNA methylation. This study only included methylation of the promoter region of the FKBP5 gene, and methylation of other regions should be considered in further studies. Fourth, it is possible that our study was underpowered to detect the associations between FKBP5 methylation and anxiety symptoms. However, a post-hoc power calculation using in G^*Power software indicated that the sample size in the current study was adequately powered (>90%) to detect medium effects sizes. Despite these limitations, the strength of the current study is that our study utilized a prospective nested case-control study design and has a larger sample size than previous studies. The study population is focused on adolescents aged 13–18 years old, and we are concerned with the subclinical state of anxiety disorders, aiming to identify early biomarkers of mental health in adolescents for early intervention.

In summary, our study demonstrates that childhood emotional abuse could be an important risk factor for subsequent anxiety symptoms; childhood maltreatment is not significantly associated with FKBP5 DNA methylation; DNA methylation of FKBP5 gene promotor could not predict the subsequent anxiety symptoms. Based on the current study results, more attention should be paid to the mental health of adolescents with childhood maltreatment. Besides, the establishment of a nationwide surveillance system on adolescent mental health is strongly recommended.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by Sun Yat-sen University School of Public Health Institutional Review Board. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

The study was originally conceived and designed by CL and LG. Investigation and data collection were performed by WLa, WLi, XD, YG, and WW. Data analysis was performed by WLa, WLi, and WW. WLa and WLi wrote the first draft of the manuscript, with a number of edits from CL and LG. All authors were in agreement with the final submitted manuscript.

This work was supported by the National Natural Science Foundation of China (Grant Nos. 81761128030 and 81903339) and Natural Science Foundation of Guangdong Province (Grant No. 2019A1515011091).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.