Susan Abushakra MD, Luc Bracoud MS, Joël Schaerer, Aidan Power MD, John Hey PhD, David Scott PhD,  Joyce Suhy PhD, Martin Tolar MD PhD & the Alzheimer Disease Neuroimaging Initiative (ADNI)

The apolipoprotein ε4 allele (APOE4) is a major genetic risk factor for Alzheimer’s disease (AD), with APOE4 carriers (homozygotes and heterozygotes) comprising ~65% of AD patients. APOE4 increases beta amyloid (Aβ) monomer production, diminishes clearance and increases their aggregation into soluble neurotoxic oligomers. Homozygotes have shown high rates of fibrillar amyloid pathology, high Aβ oligomer burden, and early cognitive decline. Homozygotes are thus an optimal target population for drugs that inhibit Aβ oligomer formation, such as tramiprosate or its pro-drug ALZ-801 (Abushakra 2017, Kocis 2017, Hey 2017).

To optimize imaging biomarker selection for a planned ALZ-801 trial in Early to Mild AD, we analyzed datasets from ADNI-1 and the subset of a tramiprosate Phase 3 study that underwent serial volumetric MRI. We evaluated atrophy rates of hippocampus and other MRI measures, and their correlations to clinical decline in APOE4/4 and APOE3/3 subjects.

ADNI-1 study enrolled 722 subjects, comprising 255 Cognitively Normal (CN), 301 Late Mild Cognitive Impairment (LMCI) and 166 Mild Alzheimer’s Disease (AD) individuals. LMCI group had 228 APOE3/3 non-carriers and 73 APOE4/4 carriers, while the AD group had 101 APOE3/3 and 65 APOE4/4 subjects. We analyzed data from the subgroup with MRIs at baseline, 12 and 24 months (, that included: LMCI (93 APOE3/3, 29 APOE4/4) and AD (29 APOE3/3, 21 APOE4/4). Clinical scores (MMSE, ADAScog13, CDR-SB) were collected at Baseline and Months 3, 6, 9, 12, 18 and 24. The 3D T1-weighted MRI collection consisted of MP RAGE (Siemens), 3D TFE (Philips) and 3D Fast SPGR (General Electric) pulse sequences, with 1.25×1.25×1.2 mm3 voxel resolution in sagittal orientation. The tramiprosate dataset, the Mild subgroup of APOE4/4 patients (MMSE 20-26) with  MRI data at screening and Week 78, included 15 APOE4/4 subjects in placebo arm. The 3DT1 protocol consisted of MP RAGE (Siemens), 3D TFE (Philips) and 3D Fast SPGR (General Electric) pulse sequences, with 1.0×1.0×1.6 mm3 voxel resolution in coronal orientation. Data were processed centrally with fully-automated methods: FreeSurfer v5.3 for brain segmentation at Baseline. Volumes of whole brain (WBV), lateral ventricles (LVV) and hippocampus (HCV= L+R) were derived. Volume changes at follow-up timepoints were assessed using Boundary Shift Integral. Cortical thickness was measured using FreeSurfer, Mayo AD signature ROI (Jack 2017) was calculated at Baseline, changes were analyzed by a Jacobian-based method.

Baseline volumetric measures were adjusted for age, years of education, and head size. Clinical score changes were estimated by fitting a linear model for each subject. Correlations between baseline MRI measures, clinical decline, and volumetric changes at M24 were analyzed by Pearson’s correlations. Comparative analyses were focused on APOE3/3 versus APOE4/4 subjects in LMCI and AD groups.

Demographics and baseline scores in the LMCI and AD groups were similar except for APOE4/4 subjects being significantly younger than APOE3/3 with a smaller HCV (p<0.001). In LMCI but not AD group, APOE4/4 subjects also had higher ADAScog than APOE3/3 (p=0.016). Across the 4 subgroups at baseline, the smallest HCV was in APOE4/4 AD subjects, followed by APOE4/4 LMCI, APOE3/3 AD, then APOE3/3 LMCI.

Hippocampal volume changes: HCV showed significantly greater % decline at 24 months in APOE4/4 AD subjects than the APOE3/3, and in APOE4/4 LMCI than APOE3/3. The order of decline was greatest in APOE4/4 AD, followed by APOE3/3 AD subjects, then APOE4/4 LMCI and smallest in APOE3/3 LMCI. In APOE4/4 subjects, annualized decline rate at 24 months was ~8% in AD, ~5.5% in LMCI, while tramiprosate APOE4/4 Mild AD subjects showed ~6% decline at Week 78. In ADNI-1, cortical thickness in APOE4/4 and APOE3/3 groups showed a similar pattern with largest decline in APOE4/4 AD subjects.

Correlation of baseline volumetrics to clinical changes: In APOE4/4 LMCI, baseline cortical thickness (Mayo index) correlations to MMSE worsening was moderately strong and significant (r=0.62). Other correlations were weaker or not significant.

Correlation of volumetric changes to clinical changes: In the APOE4/4 group, significant correlations of HCV, WBV, and cortical thickness changes to ADAScog and MMSE worsening over 24 months were observed for LMCI group, but were not significant in the AD group. In the LMCI group, r-values for ADAScog were 0.54-0.64, and for MMSE were 0.38-0.62 (all p< 0.05).

APOE4/4 subjects with Mild AD show high rates of hippocampus atrophy over 12 and 24 months. APOE4/4 Mild AD subjects from the tramiprosate Phase 3 study show a similar but slightly lower HCV decline than the corresponding ADNI dataset. APOE4/4 homozygotes at the LMCI stage (Early AD) also show high rates of HCV atrophy which correlate well with worsening of cognitive scores (ADAScog, MMSE). Confirmation of this hippocampal atrophy-cognitive decline relationship in larger studies, may allow use of HCV as a surrogate outcome in Early AD trials. These data also support the important role of HCV as an imaging  biomarker outcome in future ALZ-801 trials in Early to Mild AD.