DNA Methylation Patterns in Relation to Acute Severity and Duration of Anxiety and Depression

Abstract

Depression and anxiety are common mental disorders that often occur together. Stress is an important risk factor for both disorders, affecting pathophysiological processes through epigenetic changes that mediate gene-environment interactions. In this study, we explored two proposed models about the dynamic nature of DNA methylation in anxiety and depression: a stable change, in which DNA methylation accumulates over time as a function of the duration of clinical symptoms of anxiety and depression, or a flexible change, in which DNA methylation correlates with the acute severity of clinical symptoms. Symptom severity was assessed using clinical questionnaires for anxiety and depression (BDI-II, IDS-C, and HAM-A), and the current episode and the total lifetime symptom duration was obtained from patients' medical records. Peripheral blood DNA methylation levels were determined for the BDNF, COMT, and SLC6A4 genes. We found a significant negative correlation between COMT_1 amplicon methylation and acute symptom scores, with BDI-II (R(22) = 0.190, p = 0.033), IDS-C (R(22) = 0.199, p = 0.029), and HAM-A (R(22) = 0.231, p = 0.018) all showing a similar degree of correlation. Our results suggest that DNA methylation follows flexible dynamics, with methylation levels closely associated with acute clinical presentation rather than with the duration of anxiety and depression. These results provide important insights into the dynamic nature of DNA methylation in anxiety and affective disorders and contribute to our understanding of the complex interplay between stress, epigenetics, and individual phenotype.

Keywords: BDNF; COMT; DNA methylation; SLC6A4; anxiety; candidate genes; depression; epigenetics.

Figure 1

Distribution of all variables used in the models. The following dependent (clinical) variables were used: IDS-C, BDI-II, HAM-A, dur_episode (duration of the last episode in months), and dur_total (lifetime duration of the disorder). The following independent (epigenetic) variables were employed in the models: mBDNF_1, mBDNF_2, mBDNF_3, mBDNF_4, mBDNF_5, mBDNF_6, mBDNF_7, mBDNF_8, mBDNF_9, mCOMT_1, mCOMT_2, mCOMT_3, mSLC6A4_1, and mSLC6A4_2. The outliers are represented as points.

Figure 2

The correlation between mCOMT_1 and the clinical variables (BDI-II, HAM-A, and IDS-C) (A–C). (A) BDI-II × mCOMT_1. A significant regression equation was found (F(1,22) = 5.17, p = 0.033), with an R² of 0.190, indicating that 19% of the variance in Beck-BDI could be predicted using mCOMT_1. The patient’s average BDI-II score decreased by 2.19 for each unit of mCOMT_1. (B) IDS-C × mCOMT_1. A significant regression equation was found (F(1,22) = 5.47, p = 0.029), with an R² of 0.199, indicating that almost 20% of the variance in IDS-C could be predicted using mCOMT_1. The patient’s average IDS-C score decreased by 2.26 for each unit of mCOMT_1. (C) HAM-A × mCOMT1. A significant regression equation was found (F(1,22) = 6.59, p = 0.018), with an R² of 0.231, indicating that 23% of the variance in HAM-A could be predicted using mCOMT_1. The patient’s average HAM-A score decreased by 1.81 for each unit of mCOMT_1. (D) PC1 × mCOMT_1. A significant regression equation was found (F(1,22) = 7.80, p = 0.011), with an R² of 0.262, indicating that 26% of the variance in PC1 could be predicted using mCOMT_1. The patient’s average PC1 score decreased by 0.291 for each unit of mCOMT_1; PC1 (first principal component).

Figure 3

(A) Principal component PC1 (Dimension 1) as a summary variable of 3 clinical variables (IDS-C, BDI-II, and HAM-A). (B) Contributions of the clinical variables to PC1. Abbreviations: Beck-BDI: Beck Depression Inventory (BDI-II); IDS_C: Inventory of Depressive Symptomatology, Clinician version (IDS-C); and HAM_A: Hamilton Anxiety Rating Scale (HAM-A).

Figure 4

Boxplot of participants’ mean methylation level of COMT_1 amplicon separated by gender. There was no significant difference in the methylation levels (t(23) = 1.79, p = 0.086).