Cross‐modal associations between traditional and emerging CSF biomarkers and grey matter network disruption in autosomal dominant Alzheimer disease

Abstract

AbstractBackgroundThe grey matter covariance network of the brain, extracted from structural MR, becomes disrupted in neurodegenerative disorders such as Alzheimer disease (AD). While these disruptions ‐ as indicated by lower small‐world values ‐ relate to cognitive performance and can predict subsequent cognitive decline, the precise biological underpinnings of these network changes are unknown. Besides amyloid and tau, cerebrospinal fluid (CSF) biomarker levels can reflect synaptic damage, axonal degeneration and inflammatory processes. Here, we investigate whether CSF biomarker levels, reflecting a range of pathophysiological processes, are associated with grey matter network disruption in autosomal dominant AD (ADAD).MethodFrom the Dominantly Inherited Alzheimer Network (DIAN) Observational study, we included all participants with both T1‐weighted MRI and CSF collection at the same visit. We extracted individual‐level structural grey matter covariance networks [1], and calculated the small world coefficient summary statistic. CSF biomarkers included: Aβ42/40 ratio (amyloid aggregation), pTau181 (hyperphosphorylation), tTau and VILIP‐1 (neuronal injury and death), SNAP‐25 and neurogranin (synaptic damage), NfL (axonal injury), YKL‐40 and soluble TREM2 (neuro‐inflammation). We fitted univariate linear regression models with CSF markers as predictor and small world coefficient as outcome. To discern whether cross‐modal relationships were specific for AD, we assessed the interaction effects of mutation status.ResultAcross the whole group (219 mutation carriers/136 non‐carriers, 39±11 [mean±SD] years old, Table 1), CSF biomarker abnormality was associated with lower small‐world values, indicating disrupted network integrity (p<0.001), and with the highest effect size for NfL (β±SE= ‐0.72±0.05). Examining interactions with mutation status (p<0.05), we found that effects were driven by mutation carriers for pTau181 (β±SE= ‐0.58±0.05), tTau (β±SE= ‐0.55±0.06), SNAP‐25 (β±SE= ‐0.37±0.07), neurogranin (β±SE= ‐0.35±0.07), NfL (β±SE= ‐0.76±0.06)), and YKL‐40 (β±SE= ‐0.61±0.06, Figure 1), and to a smaller degree also present in non‐carriers for NfL (β±SE= ‐0.44±0.014) and YKL‐40 (β±SE= ‐0.32±0.08).ConclusionGrey matter network changes on MRI may involve multiple pathological processes. The strong relationship with NfL suggests that axonal loss contributes to disrupted grey matter networks as observed in AD. Studying cross‐modal relationships between MRI and CSF biomarkers can inform our understanding of underlying mechanisms of cognitive impairment due to neurodegeneration in AD. Grant: P50AG005681. Ref1:Tijms,2012 Cerebral Cortex.