In Alzheimer’s disease (AD), pathological changes begin decades before symptom onset. The preclinical pathophysiology of AD is not limited to the neuronal compartments; in fact, immune system-mediated actions (neuroinflammation, characterized by activation of microglia and astrocytes) may contribute as much (or more) to AD disease pathogenesis as do amyloid plaques and neurofibrillary tangles 1, 2. For example, a recent histopathological study showed that whereas brains from AD patients had both amyloid beta (Aβ) plaques and robust microglia activation, those from preclinical patients with Aβ plaques did not have microglia activation 3, suggesting that neuroinflammation occurs in preclinical AD and predicts disease progression. Although neuroinflammation in AD patients can be detected by positron emission tomography, this method is expensive, radiative, and prone to problems of genetic heterogeneity, poor signal-to-noise ratio, and difficult interpretation of results 4, 5. To address these limitations, we developed a novel magnetic resonance imaging technique – diffusion basis spectrum imaging (DBSI) – that can specifically image neuroinflammation 6-9. Here, we propose to develop DBSI as a new method to, for the first time, noninvasively track neuroinflammatory changes that occur in preclinical AD patients.

Participants: 175 cognitively normal participants participants were selected from a larger population enrolled at Knight Alzheimer’s Disease Research Center (Knight ADRC) of the Washington University School of Medicine (St Louis, MO, USA), in longitudinal studies of memory and aging (Table 1).

CSF Collection and Analysis: CSF (20-30 mL) was collected by lumbar puncture as previously described 10. Samples were analyzed by ELISA (Innotest; Innogenetics, Ghent, Belgium) after one freeze-thaw for beta amyloid 42 (Aβ42), total tau, and tau phosphorylated at threonine-181 (ptau181) and by ELISA (Quidel, San Diego, CA) after two freeze-thaw cycles for VILIP-1 and YKL-40.

NIA-AA preclinical AD stages: Cognitively normal participants were divided into preclinical stage 0 and stage 1 groups according to CSF measures and research criteria proposed by the National Institute on Aging and the Alzheimer’s Association (NIA-AA)11.

Imaging Acquisition and Processing: Diffusion MRI images were collected (2x2x2mm voxels, TR=14500ms, TE=102ms, multi-bvalue scheme, 23 directions and bmax = 1400s/mm2). Data were collected in two 6 minute runs on 3T Trio scanner (Siemens, Erlangen, Germany). DTI and DBSI derived mean diffusivity, fractional anisotropy, radial diffusivity and axial diffusivity were computed for all datasets. The whole brain voxel-wise DTI-indices was analyzed using Tract Based Spatial Statistics (TBSS) (available in FSL).

DTI-derived axial and mean diffusivity was significantly lower (Fig. 1 A&B) in 31 White matter (WM) regions in stage 1 than in stage 0 participants, suggesting neuronal injury in stage 1. DTI findings contradict with the negative CSF ptau181 finding (no neuronal injury) in stage 1. Strikingly, we found that DBSI-derived cell fraction (restricted isotropic component) was higher in stage 1 participants than in stage 0 (p < 0.07 corrected for multiple comparisons). Importantly, the WM regions with the increased cell fraction (Fig. 1C) largely overlapped with the regions showing decreased DTI-derived axial and mean diffusivity (Fig. 1 A&B). We also examined the WM regions with significant cell fraction increase (p <0.05) before multiple comparison (Fig. 1D) and found high similarity with the WM regions showing increased trend after the correction of multiple comparison, confirming that the presence of inflammatory microglial activation/accumulation 12, instead of white matter damages, caused the reduced DTI-derived axial and mean diffusivities in the preclinical stage 1 participants. To assess the accuracy of DBSI cell fraction, we derived DBSI total cell fraction by adding up cell fraction from all voxels on the WM skeleton (green lines in Fig.1). Similarly, DBSI total cell and extracellular water fraction is derived by adding up cell fractions and extracellular fractions. YKL-40 strongly correlates with DBSI total cell fraction (r = 0.65, p<0.005) and DBSI total cell plus extracellular fraction (r = 0.7, p <0.005) for all stage 0 and stage 1 subjects, confirming the accuracy of DBSI neuroinflammation index (cell fraction). Statistics controlled for age, gender, education, ApoE4 genotype, and family history. Increased inflammatory cell infiltration in preclinical stage 1 subjects strongly suggested that neuroinflammation follows Aβ deposition but precedes tau pathology. DBSI detected neuroinflammation correlate with invasive CSF measures of neuroinflammation. This exciting preliminary finding suggests that development of DBSI neuroinflammation biomarkers could: 1) improve our understanding of the role of neuroinflammation in AD progression, 2) provide noninvasive, safe, low-cost, and accurate measures of preclinical AD pathologies for early diagnostics, 3) be used to identify asymptomatic subjects at highest risk of developing dementia, and 4) lead to effective new disease-modifying therapies targeting immune pathologies in AD.