Animal models of Alzheimer’s disease have suggested that tau pathology propagation, facilitated by amyloid pathology, may occur along connected pathways. To investigate these ideas in humans, we combined amyloid scans with longitudinal data on white matter connectivity, hippocampal volume, tau positron emission tomography and memory performance in 256 cognitively healthy older individuals. Lower baseline hippocampal volume was associated with increased mean diffusivity of the connecting hippocampal cingulum bundle (HCB). HCB diffusivity predicted tau accumulation in the downstream-connected posterior cingulate cortex in amyloid-positive but not in amyloid-negative individuals. Furthermore, HCB diffusivity predicted memory decline in amyloid-positive individuals with high posterior cingulate cortex tau binding. Our results provide in vivo evidence that higher amyloid pathology strengthens the association between HCB diffusivity and tau accumulation in the downstream posterior cingulate cortex and facilitates memory decline. This confirms amyloid’s crucial role in potentiating neural vulnerability and memory decline marking the onset of preclinical Alzheimer’s disease.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This work was supported in part by the Athinoula A. Martinos Center for Biomedical Imaging, P41 EEB015896 and shared instrumentation grants S10RR021110, S10OD010364, S10RR023401, S10RR023043 and 1S10RR019307. The research was supported in major part by the Harvard Aging Brain Study (P01 AG036694). We thank all the participants of the Harvard Aging Brain Study. H.I.L.J. received funding from Alzheimer Nederland (WE.15-2014-06). T.H. received funding from NIH grant K01 AG040197, P01 AG036694, P50 AG005134, R01 AG053509 and R01 AG034556. J.S. received funding from NIH grant K23EB019023. K.P. is funded by NIA grant K23 AG053422-01 and the Alzheimer’s Association. K.J. received funding from NIH grants R01 EB014894, R21 AG038994, R01 AG026484, R01 AG034556, P50 AG00513421, U19 AG10483, P01 AG036694, R13 AG042201174210, R01 AG027435 and R01 AG037497 and the Alzheimer’s Association grant ZEN-10-174210. R.S. receives research support from the following grants: P01 AG036694, U01 AG032438, U01 AG024904, R01 AG037497, R01 AG034556, K24 AG0350007, P50 AG005134, U19 AG010483, R01 AG027435, Fidelity Biosciences, Harvard NeuroDiscovery Center and the Alzheimer’s Association.
Integrated supplementary information
. For each measurement, the number of observations is depicted per time point. For tau PET, the first measurement was acquired throughout the study, depending on the time the subject entered the study. Associations between amyloid burden and hippocampal volume where tested cross-sectionally using both baseline measures and longitudinal using amyloid burden at baseline and hippocampal volume over time. Associations between hippocampal volume and diffusion were tested using hippocampal volume at baseline and diffusion over time (for the reversed association, we used diffusion at baseline and hippocampal volume over time). Associations between diffusion and tau pathology were tested using diffusion at baseline and tau pathology over time. Finally, the associations with memory were investigated with memory longitudinally and diffusion at baseline and biomarker data dichotomized (using the Gaussian Mixture Modeling approach). Abbreviations: F/U = Follow-up; Hipp = Hippocampus; PET = Positron Emission Tomography.
Supplementary Figure 2 Associations between hippocampal volume and diffusivity of the HCB and UF at baseline.
. Associations between hippocampal volume and tract diffusivity of the hippocampal cingulum bundle (top row, n=256 independent participants) and uncinate fasciculus (bottom row, n=253 independent participants) for left (first two columns) and right sides (last two columns). Abbreviations: FA = Fractional Anisotropy; MD = Mean Diffusivity.
Supplementary Figure 3 Spaghetti plot of the relationship between HCB diffusivity over time and hippocampal volume at baseline for amyloid-positive and amyloid-negative individuals.
. Plots show longitudinal subject-specific values for the association between MD values of the HCB over time and hippocampal volumes (adjusted for intracranial volume). Cooler colors represent lower hippocampal volumes (right hemisphere: top panel, left hemisphere: bottom panel). Amyloid positive individuals (n=61 independent participants) are shown in dashed lines, while amyloid negative individuals (n=183 independent participants) in solid lines. Values are unadjusted, raw values. Abbreviations: MD: mean diffusivity, HCB: Hippocampal cingulum bundle.
Supplementary Figure 4 Associations between change in hippocampal volume and change in mean diffusivity in the HCB and UF.
. The line plots (top row = left, bottom row = right hemisphere) show that change in right hippocampal volume over time predicted increased MD of the right HCB over time (left bottom corner; n=256 independent participants). No significant associations were found between change in hippocampal volume and changes in MD of the UF (n=253 independent participants). In all line plots, estimated marginal means of the moderation by change in hippocampal volume are plotted at the mean and ± 1 standard deviation, but analyses were done continuously using mixed-effects linear models. Shaded areas around the fit lines show the 95% confidence interval. All p-values are two-sided and unadjusted for multiple comparisons.
Supplementary Figure 5 Spaghetti plot of the relationship between PCC tau binding over time and baseline HCB diffusivity for amyloid-positive and amyloid-negative individuals.
. Plots show longitudinal subject-specific values for the association between PCC tau binding over time and baseline MD of the hippocampal cingulum bundle (HCB) Cooler colors represent lower MD values (right hemisphere: top panel, left hemisphere: bottom panel). Amyloid positive individuals (n=36 independent participants) are shown in dashed lines, while amyloid negative (n=103 independent participants) individuals in solid lines. Values are unadjusted, raw values. Abbreviations: MD: mean diffusivity.
. Note: Multiple Gaussian distributions (i.e. mixtures) were fit to the PCC tau (top) or IT tau (bottom) signal of the baseline data (n=141 independent participants). The number of mixtures is plotted against the Bayesian Information Criterion (BIC, higher is better). Left panels show the distribution assuming unequal variances (red) or equal variances (blue). For the PCC: BIC=131.23, LRTS for 2 mixture components=16.95, p=0.025; LRTS for 3 mixture components =6.40, p=0.506 with a BIC=130.66. For the IT BIC=87.56, LRTS for 2 mixture components=11.11, p=0.044; LRTS for 3 mixture components =6.82, p=0.25 with a BIC=77.57. Right panels show the probability density function for the estimated Gaussian distribution superimposed on the baseline subject density histogram for PCC (top) or IT (bottom) tau values. The green vertical line indicates the optimal probability threshold (PCC tau = 1.28 SUVr partial volume corrected; IT tau = 1.73 SUVr partial volume corrected). Abbreviations: PCC = Posterior cingulate cortex, IT = Inferior Temporal cortex,LRTS= Likelihood ratio test statistic.
Supplementary Figure 7 Spaghetti plot of the relationship between memory performance over time and baseline HCB diffusivity for individuals with low versus high PCC tau.
Plots show longitudinal subject-specific values (n=141 independent participants) for the association between memory performance over time and baseline MD values of the hippocampal cingulum bundle (HCB). Cooler colors represent lower MD values (right HCB: top panel, left HCB: bottom panel). Individuals with high PCC tau values are shown in dashed lines, while individual with low PCC tau binding in solid lines. Values are unadjusted, raw values. Abbreviations: MD: mean diffusivity.