Growth factor polygenic disposition shapes associations of childhood adversity with gray matter and clinical profiles in depression: A two-cohort imaging-genetics study.
Abstract
Aim
Adverse childhood experiences (ACE) are associated with greater symptom severity and treatment resistance in major depressive disorder (MDD), but their effects vary markedly across individuals. Growth factor polygenic profiles may contribute to this differential susceptibility by influencing stress responsivity, neuroplasticity, and structural brain variation.
Methods
In an inpatient cohort (n = 88) and a population-based UK Biobank cohort (n = 226), we tested whether polygenic scores (PGSs) for growth factors moderated the associations of ACE with symptom severity and treatment-resistant depression (TRD) in depressed patients. The moderating effects of growth factor PGSs were modeled individually (single-pathway moderation) and jointly using principal component analysis (PCA). We further tested whether the identified associations were mediated by gray matter volumes (GMVs) in a moderated mediation framework.
Results
In the clinical cohort, the associations between ACE and symptom severity were moderated by PGSs for stem cell factor [SCF/KITLG (t = 3.92, pFDR = 0.002)] and hepatocyte growth factor [HGF/MET (t = -2.81, pFDR = 0.03)]. Moderated-mediation models identified conditional indirect effects involving GMV in parahippocampal and posterior cingulate regions (indices of moderated mediation: 0.17 and -0.33, respectively). A PCA-derived growth factor profile showed a convergent pattern with the single-PGS findings and moderated the associations of ACE with TRD (z = -2.29, P = 0.022). In the UK Biobank cohort, the moderating effect of ACE × SCF/KITLG PGS was replicated (t = 2.27, P = 0.024).
Conclusion
Growth factor polygenic disposition related to SCF/KITLG and HGF/MET may contribute to differential susceptibility to ACE in MDD. Structural brain findings were observed in medial temporal and cingulate regions and warrant confirmation in independent samples.