CSF neurodegeneration biomarkers predict MTL atrophy over time in cognitively healthy older adults

Background Structural decline in the medial temporal lobe is an unspecific biomarker of neurodegeneration, assessed in normal older adults (OA) and Alzheimer’s Disease (AD). We investigated, in OA, whether CSF biomarkers of neurodegeneration (t‐tau, NFL, Ng, FABP3) predict longitudinal brain atrophy, whether this association is moderated by core AD biomarkers of beta‐amyloid (Aβ) and phosphorylated tau (p‐tau) and whether it is related to specific cellular and molecular mechanisms by using neurogenetic profiles. Method Longitudinal structural MRI up to 7 years and baseline CSF biomarkers levels were collected in 189 OA (mean age = 74.8 years). Hippocampus (HC) and entorhinal cortex (EC) were selected as regions of interest. Linear mixed and generalized additive mixed models (GAMM) were used to assess the effects of Biomarker, Time, and Time×Biomarker on EC thickness and HC volume. Similar models were used to assess the interaction with core AD biomarkers. Correlation between cortical decline maps, associated with CSF biomarkers, and gene‐expression maps (Allen Human Brain Atlas https://human.brain‐map.org/) for specific cellular components, was assessed with BrainSMASH package (https://brainsmash.readthedocs.io/en/latest/) Result Higher NFL, FABP3, and t‐tau CSF values were associated with steeper EC thinning and more HC atrophy (p<.05) over time (Figure 1a). GAMM showed significant non‐linear interactions with Time (except NG): NFL and FABP3 on EC (edf = 3.15, 3.81) and NFL and t‐tau on HC (edf = 3.09, 2.77) (Figure 1b). Aβ + status (<550 pg/mL) did not significantly moderate the association with any biomarker (all tests p>.08) (Figure 2a). The relationships between CSF NFL with EC thinning (t = ‐3.27, p = 0.001) and HC atrophy (t = ‐2.91, p = 0.004), and FABP3 with EC thinning (t = ‐1.98, p = .05) remained significant, independently of p‐tau. The spatial pattern of cortical atrophy associated with the CSF biomarkers of neurodegeneration (Figure 2b) overlapped with distinct neurogenetic profiles: t‐tau with axonal growth cone (r = ‐.50, p unc = .002); NFL with cortical cytoskeleton (r = ‐.61, p fdr <.001) and Ng with the dendritic shaft (r = ‐.40, p unc <.009) components. Conclusion CSF NFL and FABP3 predict HC atrophy and EC thinning independently of Aβ/p‐tau. The neurodegenerative biomarkers may improve the accuracy of individual predictions of neurodegenerative changes in OA and identify those with a higher risk of future cognitive decline.