Impaired glymphatic system in genetic frontotemporal dementia: a GENFI study

Abstract

The glymphatic system is an emerging target in neurodegenerative disorders. Here, we investigated the activity of the glymphatic system in genetic frontotemporal dementia with a diffusion-based technique called diffusion tensor image analysis along the perivascular space. We investigated 291 subjects with symptomatic or presymptomatic frontotemporal dementia (112 with chromosome 9 open reading frame 72 [C9orf72] expansion, 119 with granulin [GRN] mutations and 60 with microtubule-associated protein tau [MAPT] mutations) and 83 non-carriers (including 50 young and 33 old non-carriers). We computed the diffusion tensor image analysis along the perivascular space index by calculating diffusivities in the x-, y- and z-axes of the plane of the lateral ventricle body. Clinical stage and blood-based markers were considered. A subset of 180 participants underwent cognitive follow-ups for a total of 640 evaluations. The diffusion tensor image analysis along the perivascular space index was lower in symptomatic frontotemporal dementia (estimated marginal mean ± standard error, 1.21 ± 0.02) than in old non-carriers (1.29 ± 0.03, P = 0.009) and presymptomatic mutation carriers (1.30 ± 0.01, P < 0.001). In mutation carriers, lower diffusion tensor image analysis along the perivascular space was associated with worse disease severity (β = -1.16, P < 0.001), and a trend towards a significant association between lower diffusion tensor image analysis along the perivascular space and higher plasma neurofilament light chain was reported (β = -0.28, P = 0.063). Analysis of longitudinal data demonstrated that worsening of disease severity was faster in patients with low diffusion tensor image analysis along the perivascular space at baseline than in those with average (P = 0.009) or high (P = 0.006) diffusion tensor image analysis along the perivascular space index. Using a non-invasive imaging approach as a proxy for glymphatic system function, we demonstrated glymphatic system abnormalities in the symptomatic stages of genetic frontotemporal dementia. Such measures of the glymphatic system may elucidate pathophysiological processes in human frontotemporal dementia and facilitate early phase trials of genetic frontotemporal dementia.

Keywords: DTI-ALPS; frontotemporal dementia; frontotemporal lobar degeneration; genetic; glymphatic system.

Graphical Abstract

Figure 1

Distribution of DTI-ALPS across disaggregating by disease stage (A) or by genetic mutation and disease stage (B). DTI-ALPS, diffusion tensor image analysis along the perivascular space. Comparisons of interest were ‘young NC versus presymptomatic’, ‘old NC versus symptomatic’ and ‘presymptomatic versus symptomatic’ in A (i.e. three pairwise comparisons in total) and ‘young NC versus all presymptomatic groups’, ‘old NC versus all symptomatic groups’ and ‘presymptomatic versus symptomatic within the same mutation groups’ within the three genetic groups in B (i.e. nine pairwise comparisons in total). Statistics values: F(3) = 6.05, P < 0.001 for A and F(7) = 3.46, P = 0.001 for B. Post hoc pairwise comparisons of interest were adjusted using the false discovery rate correction.

Figure 2

Association of DTI-ALPS with disease severity (i.e. CDR-FTLD) (A), expected years of onset (B), plasma NfL (C) and plasma GFAP (D) in mutation carriers. DTI-ALPS, diffusion tensor image analysis along the perivascular space; CDR-FTLD, CDR® Dementia Staging Instrument plus NACC behaviour and language domains; NfL, neurofilament light; GFAP, glial fibrillary acidic protein. Plasma NfL and GFAP values were log-transformed. Statistics values: F(1) = 12.48, P < 0.001 for A; F(1) = 2.80, P = 0.096 for B; F(1) = 3.49, P = 0.063 for C; and F(1) = 0.78, P = 0.379 for D.

Figure 3

Association between DTI-ALPS at baseline and longitudinal worsening in disease severity. DTI-ALPS, diffusion tensor image analysis along the perivascular space; CDR-FTLD, CDR® Dementia Staging Instrument plus NACC behaviour and language domains. Baseline DTI-ALPS scores of mutation carriers were converted into Z-scores using mean and standard deviation of healthy controls, and mutation carriers were classified as having ‘high’ (Z-score > 1), ‘average’ (Z-score between −1 and 1) or ‘low’ (Z-score < −1) baseline DTI-ALPS. Statistics values: χ²(2) = 10.39, P = 0.006. Post hoc comparisons: ^a > ^b.