CSF TMEM106B Levels Track Frontotemporal Dementia Severity and Brain Atrophy

Frontotemporal lobar degeneration (FTLD) is a leading cause of early-onset dementia, yet reliable fluid biomarkers that track disease severity across its heterogeneous subtypes remain scarce. TMEM106B — a lysosomal protein encoded by a gene on chromosome 7 — has emerged as a major genetic susceptibility factor for FTLD, and TMEM106B protein aggregates have been identified across aging brains and multiple neurodegenerative conditions. This study is the first large-scale investigation of CSF TMEM106B as a quantifiable biomarker in clinical FTLD populations.

The investigators enrolled participants through two independent multicenter cohorts. The discovery cohort (n = 271; 51% women; median age 59 years) comprised individuals with neuropathology-confirmed sporadic FTLD, presymptomatic or symptomatic carriers of pathogenic variants in C9orf72, GRN, or MAPT genes, and healthy controls. The validation cohort (n = 383; 48% women; median age 64 years) included clinically diagnosed sporadic FTD, Alzheimer disease (AD), and controls. CSF TMEM106B was quantified using aptamer-based proteomics (SomaScan v3.0 for discovery, v4.1 for validation). Brain MRI volumes and longitudinal clinical severity scores over two years were integrated with the CSF measurements.

Lower CSF TMEM106B was significantly associated with greater clinical disease severity across both cohorts (β = −0.15; 95% CI, −0.24 to −0.04; P = .003). Crucially, this association was independent of neurofilament light chain (NfL), a well-validated neurodegeneration marker, suggesting TMEM106B captures a distinct biological dimension of disease burden. Lower CSF TMEM106B was also strongly linked to smaller frontotemporal brain volumes on MRI (β = 0.42; 95% CI, 0.24–0.61; P < .001), and participants with lower levels at baseline showed faster clinical progression over the following two years (β = −2.21; 95% CI, −3.70 to −0.72; P = .001).

A striking genetic finding emerged: the TMEM106B rs1990622 genotype — previously associated with FTLD risk — directly influenced CSF TMEM106B levels. Individuals carrying the protective G/G genotype had significantly lower CSF TMEM106B compared to those with the risk A/A genotype, regardless of their underlying neuropathological diagnosis or disease-causing mutation. This genotype effect was observed across both familial (C9orf72, GRN, MAPT) and sporadic FTLD groups, pointing to a genetically regulated mechanism that modulates protein levels in the CNS compartment.

Notably, CSF TMEM106B did not discriminate between FTLD subtypes (e.g., behavioral variant FTD versus primary progressive aphasia) nor between FTLD and Alzheimer disease, limiting its diagnostic specificity. The authors interpret this as consistent with TMEM106B being a general marker of lysosomal or neuronal stress that spans neurodegenerative proteinopathies rather than a disease-specific indicator. This finding also opens intriguing questions about whether interventions targeting lysosomal function could modulate TMEM106B levels. The study's cross-sectional design for most analyses and the use of different SomaScan assay versions across cohorts are important methodological caveats that warrant cautious interpretation of the quantitative comparisons.