- Review
Late-life depression (LLD) is increasingly recognized as a complex multisystem disorder associated with accelerated biological aging, impaired physiological resilience, frailty, and an increased risk of cognitive decline and dementia. Although growing evidence supports the close relationship between depression, frailty, and neurodegeneration, the molecular mechanisms underlying this continuum remain incompletely understood. In particular, conventional clinical and imaging biomarkers capture established disease but provide limited insight into the systemic biological processes that precede neurodegeneration. Proteomics has emerged as a powerful systems biology approach capable of characterizing the dynamic molecular networks involved in aging, neuropsychiatric disorders, and neurodegenerative diseases. Recent advances in high-throughput mass spectrometry have identified alterations in proteins associated with chronic inflammation, complement activation, mitochondrial dysfunction, oxidative stress, lipid metabolism, synaptic integrity, neurotrophic signaling, and cellular senescence across LLD, frailty, cognitive impairment, and Alzheimer’s disease (AD). These convergent proteomic signatures suggest that these conditions may share common biological pathways consistent with progressive loss of physiological resilience rather than independent pathological entities. In this review, we examine the current evidence supporting shared proteomic mechanisms linking LLD, frailty, cognitive frailty (CF), and neurodegeneration. We discuss the potential integration of large-scale proteomic profiling with established AD biomarkers within the AT(N) framework to improve biological characterization, early risk stratification, and prediction of cognitive decline. Particular emphasis is placed on CF as a transitional and potentially modifiable phenotype in which systemic molecular dysregulation precedes overt neurodegeneration. By integrating evidence from geroscience, neuropsychiatry, and proteomics, we propose that proteomic profiling can redefine LLD and frailty as manifestations of a common biological vulnerability continuum. This systems-level perspective may facilitate the identification of individuals at high risk for dementia, support the development of multimodal biomarker panels, and advance precision neuropsychiatry through earlier diagnosis and personalized preventive interventions.
17 November 2026




![SDS-PAGE analysis of BSA preparations loaded at 5 and 1 µg. BSA preparations from: 1: Amersham/GE Health Care; 2: Merck-Sigma; 3: Sigma-Aldrich; 4: VWR. The analysis shows a clear predominance of the expected MW of 66.4 kDa and, for 5 µg BSA, faint bands at lower MW likely arising from proteolytic products of BSA in addition to higher MW of BSA oligomers, in agreement with previous findings [8]. Raw uncropped images of the stain-free and G-250 stained gels can be found in Supplementary Material, Figure S1.](https://mdpi-res.com/proteomes/proteomes-14-00059/article_deploy/html/images/proteomes-14-00059-ag-550.jpg)



![3D structures of the human mechanosensitive ion channel PIEZO1 ((A–C); PDB ID: 8YEZ; ref. [6]) and mouse PIEZO2 ((D–F); PDB ID: 6KG7; ref. [7]) determined by cryo-electron microscopy. Top, side, and bottom projections are shown in plots (A) and (D), (B) and (E), and (C) and (F), respectively. Structural information is not available for human PIEZO2; therefore, corresponding data were retrieved for its mouse homologue. Dashed lines correspond to missing electron density regions: residues 1–569, 645–677, 713–788, 880–914, 957–969, 1059–1995, 1122–1153, 1234–1282, 1371–1407, 1424–1512, 1569–1644, 1748–1778, 1804–1939, and 1981–2000 in PIEZO1 and residues 1–7, 143–200, 305–331, 359–490, 614–674, 824–933, 1383–1426, 1515–1556, 1578–1667, 1728–1947, 2056–2074, 2112–2235, and 2283–2298 in PIEZO2. The three PIEZO protomers are shown in orange, green, and purple for structural clarity.](https://mdpi-res.com/proteomes/proteomes-14-00041/article_deploy/html/images/proteomes-14-00041-g001-550.jpg)

