Basic Science and Pathogenesis.

Background: Mechanisms driving cerebrovascular decline during Alzheimer's disease and related dementias (ADRD) are poorly understood. Methylenetetrahydrofolate reductase (MTHFR) is an enzyme in the folate/methionine pathway. Variants in MTHFR, notably 677C>T, are associated with ADRD. ∼30% of individuals carry at least one copy of MTHFR, causing a 50% decrease in MTHFR enzyme efficiency. Reduced efficiency can lead to elevated homocysteine in blood, resulting in vascular inflammation and increased risk for cerebrovascular damage. To examine vascular risk in a late-onset Alzheimer's disease (LOAD) model, MODEL-AD created a mouse carrying known ADRD risk factors with the Mthfr variant. While plasma-based protein biomarkers hold promise for identifying ADRD pathology, additional markers are necessary. Studies support eye exams as a synergistic strategy due to overlapping mechanisms of damage in retina and brain, so retinal phenotyping has been incorporated in our study.

Methods: The LOAD2.Mthfr model was created by the IU/JAX/PITT MODEL-AD Center by incorporating the Mthfr allele with the LOAD2 alleles (hAbeta/APOE4/Trem2*R47H). Cerebrovascular health is assessed in 4, 12, 18 and 24 month-old mice by immunofluorescence to examine barrier integrity in cells that make up the cerebrovascular unit: endothelium, smooth muscle, pericytes, basement membrane, tight junctions, and extracellular deposition of blood products. Differential expression of vascular related genes/proteins is being examined by RNA-sequencing and proteomics. In vivo retinal imaging, including optical coherence tomography and fluorescence angiography, and ex vivo retinal IF was performed to assess and compare neuronal and vascular health in the retina to the brain.

Results: All mice have been aged, and in vivo assays completed. Compared to LOAD2 controls, LOAD2.Mthfr mice do not show significant changes in frailty, activity, or cognition. At 24 mos, transcriptomic and proteomic data suggest LOAD2 and LOAD2.Mthfr show some expression changes similar to those seen in human AD. By immunofluorescence, like LOAD2, Mthfr mice do not develop ThioS-positive amyloid plaques, but assessment of vascular-specific changes is in progress. Plasma and retinal biomarker analysis is also in progress.

Conclusions: MODEL-AD continues to create strains that develop endophenotypes of LOAD for preclinical testing. Ultimately, this work will aid in the identification of novel therapeutic approaches that reduce cerebrovascular pathology in ADRDs.