CSF biomarkers for frontotemporal dementia and its pathological subtypes

Background Frontotemporal dementia (FTD) is caused by frontotemporal lobar degeneration (FTLD) and is characterized mainly by the presence of aggregates of the proteins tau (FTLD‐Tau) or TDP43 (FTLD‐ TDP) in the brain, which likely require distinct pharmacological therapy. However, the clinical presentation of the pathological subtypes is heterogeneous and overlapping, making this diagnostic subtyping at present virtually impossible. Thus, there is a strong unmet need to identify body‐fluid biomarkers to discriminate FTD and its pathological subtypes. Most biomarker studies performed to date have analyzed pathologically heterogeneous populations. The few studies analyzing antemortem cerebrospinal fluid (CSF) with known underlying neuropathology have revealed several candidate biomarkers (e.g. p/tTau ratio). Despite these promising results, their specificity is still not optimal, most of the identified markers are awaiting further validation and their diagnostic accuracy remains to be evaluated. Method We have recently taken advantage of a novel high‐multiplex and sensitive antibody‐based proteomic technology to analyze >600 proteins in a large multicenter cohort providing CSF samples from controls (n=195) and FTD patients (n=199). For a subset of FTD cases, the underlying specific pathology was known (FTLD‐Tau=34: 18 autopsy‐confirmed and 27 MAPT mutation carriers; FTLD‐ FTLD‐TDP=54: 27 autopsy‐confirmed and 18 C9orf72 and 9 GRN mutation carriers). Differences in the protein expression profile were analyzed by GlobalTesting. Generalized linear modeling was used to identified classification protein signatures. Result We identified CSF protein biomarker signatures for specific diagnosis of FTD (8‐9 proteins, AUC>0.85). The proteins identified were related to mechanisms involved in immune system, extracellular remodeling and cell killing. There were no between‐group differences in the CSF protein profiles of the FTLD groups. However, differences in the absolute levels of some proteins were detected when autopsy‐confirmed cases and mutation carriers were analyzed separately. Conclusion The similarity of the CSF profiles between FTLD groups suggests shared pathological mechanisms. Alternatively, each subtype may be heterogeneous, which possibly results from analyzing together familial and autopsy‐confirmed cases. Taken together, the FTLD studies performed to date indicate a highly complex disorder challenging the development of efficient diagnostic and therapeutic markers.