Integrated genome-wide methylation and expression analyses reveal functional predictors of response to antidepressants

Chelsey Ju, Laura M. Fiori, Raoul Belzeaux, Jean Francois Theroux, Gary Gang Chen, Zahia Aouabed, Pierre Blier, Faranak Farzan, Benicio N. Frey, Peter Giacobbe, Raymond W. Lam, Francesco Leri, Glenda M. MacQueen, Roumen Milev, Daniel J. Müller, Sagar V. Parikh, Susan Rotzinger, Claudio N. Soares, Rudolf Uher, Qingqin LiJane A. Foster, Sidney H. Kennedy, Gustavo Turecki

Research output: Contribution to journalArticlepeer-review

27 Scopus citations


Major depressive disorder (MDD) is primarily treated with antidepressants, yet many patients fail to respond adequately, and identifying antidepressant response biomarkers is thus of clinical significance. Some hypothesis-driven investigations of epigenetic markers for treatment response have been previously made, but genome-wide approaches remain unexplored. Healthy participants (n = 112) and MDD patients (n = 211) between 18–60 years old were recruited for an 8-week trial of escitalopram treatment. Responders and non-responders were identified using differential Montgomery-Åsberg Depression Rating Scale scores before and after treatment. Genome-wide DNA methylation and gene expression analyses were assessed using the Infinium MethylationEPIC Beadchip and HumanHT-12 v4 Expression Beadchip, respectively, on pre-treatment peripheral blood DNA and RNA samples. Differentially methylated positions (DMPs) located in regions of differentially expressed genes between responders (n = 82) and non-responders (n = 95) were identified, and technically validated using a targeted sequencing approach. Three DMPs located in the genes CHN2 (cg23687322, p = 0.00043 and cg06926818, p = 0.0014) and JAK2 (cg08339825, p = 0.00021) were the most significantly associated with mRNA expression changes and subsequently validated. Replication was then conducted with non-responders (n = 76) and responders (n = 71) in an external cohort that underwent a similar antidepressant trial. One CHN2 site (cg06926818; p = 0.03) was successfully replicated. Our findings indicate that differential methylation at CpG sites upstream of the CHN2 and JAK2 TSS regions are possible peripheral predictors of antidepressant treatment response. Future studies can provide further insight on robustness of our candidate biomarkers, and greater characterization of functional components.

Original languageEnglish (US)
Article number254
JournalTranslational psychiatry
Issue number1
StatePublished - Dec 1 2019
Externally publishedYes

ASJC Scopus subject areas

  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Biological Psychiatry


Dive into the research topics of 'Integrated genome-wide methylation and expression analyses reveal functional predictors of response to antidepressants'. Together they form a unique fingerprint.

Cite this