Major depressive disorder is a prevalent and debilitating mental health condition contributing to a growing global burden. Late-life depression (LLD), affecting individuals over 60 years of age, is further associated with elevated risks for cardiovascular diseases, cognitive decline, and dementia. Treatment responses vary widely, potentially due to underlying neurodegeneration and cellular senescence. We aimed to explore blood-based biomarkers related to Alzheimer’s disease and senescence-associated secretory phenotype (SASP) proteins, seeking to identify biological underpinnings of LLD and their association with response to psychotherapy.

We performed a secondary analysis of the Cognitive Behavioral Therapy for Late-Life Depression (CBTlate) trial in 228 participants aged 60 years and older with a diagnosis of LLD. Depression trajectories were compared using clustering. In participants with available plasma samples, biomarker data were generated post hoc. We assessed associations between biomarkers and depression trajectories, biomarker dynamics, and their ability to predict treatment response.

Two depression trajectories were identified: persistently high stable Geriatric Depression Scale (GDS) scores (hsGDS) and decreasing scores over time (dGDS). The hsGDS group had more severe baseline depression (p = 2.88 × 10−6), anxiety (p = 4.39 × 10−4), and sleep disorders (p = 1.09 × 10−3), and was more likely to have a history of major depression (p = 0.01) and mild cognitive impairment (p = 0.01). Biomarker analysis revealed elevated baseline plasma neurofilament light chain (NfL, p = 2.51 × 10−2) and reduced C-X-C Motif Chemokine Ligand 5 (CXCL5, p = 2.83 × 10−2) in the hsGDS group. Including CXCL5 in predictive models improved trajectory differentiation (p = 3.94 × 10−3).

Cellular aging biomarkers like CXCL5 may improve understanding of LLD and guide personalized therapeutic interventions.