Protein aggregation is both diagnostic and causal for neurodegenerative diseases, but also increases with age in the brain and other organs. Constituent proteins may contribute structurally to aggregates due to their misfolding and adhesion to other proteins; however, some also play roles in clearance so their sequestration reduces abundance or availability. We analyzed protein compositions and internal architecture of aggregates from Alzheimer’s disease (AD) brain tissue. Using a novel cross-linking protocol and neural-network tools to analyze aggregate “contactomes”, we identified AD-specific protein interactions that are disrupted by RNA-interference (RNAi) knockdown of either partner, or by drugs that target their interfaces, in human-cell and C. elegans models of neuropathic aggregation. We identified tau/14-3-3G and hexokinase/14-3-3G among influential protein interactions, based on neural-network and other machine-learning analyses. These interfaces were screened for stable binding of FDA-approved drugs, as follows. (i) We first performed a 3-tiered in-silico screening of molecular-structure libraries, progressively increasing stringency. (ii) Top candidates from those predictive screens were assessed in biological systems: cultured human-cell and C. elegans models of AD-like aggregation. The best drug candidates, prioritized by efficacy in suppressing aggregation across these model systems, will now be pursued in mice. In Aim 1, we will characterize the impact of 4 drugs, two for each target interface, by analyzing aggregates from human cell lines: Aß1–42 amyloid from SY5Y-APPSw neuro-blastoma cells; and tau tangles from SY5Y-tau cells. Each cell line will be exposed to each drug at several concentrations, or vehicle only (controls), and aggregates will be captured on magnetic beads coated with antibodies to Aß or tau. Aggregates insoluble in a strong detergent (sarcosyl) will be cross-linked with aggregate-permeant “click” reagents, and the linked peptide pairs are identified to define the internal architecture of these large complexes. By comparing the proteins and their direct contacts between treated and untreated cells, we will identify proteins, modifications, and interactions most depleted by drug exposure. In Aim 2, drugs targeting these same interfaces will be assessed for their ability to reduce and/or postpone cognitive impairment in 3xTg-fAD mice, a model incorporating 3 mutations underlying familial AD pedigrees. As they age, these mice accrue cerebral amyloid plaque followed by tau tangles. Drugs (at 2 well-tolerated doses) will commence at either 4 or 12 months of age, with intervals between treatments determined by bioavailability and clearance times. We will assess cognitive traits just prior to the first doses, and at 4-month intervals thereafter, based on novel-object recognition (Y-maze format) and 2 IntelliCageTM learning and recall tasks. The younger mouse groups will be sacrificed at 16 months, while mice commencing treatment at 12 months will be sacrificed at 20 months; at these ages, untreated controls display significant cognitive decline and cerebral Aß and tau aggregates. Aggregate interactomes will be analyzed as in Aim 1, from cerebra isolated postmortem. By comparing controls (vehicle only) to mice receiving drugs, we will learn whether drugs disrupted their intended targets. We thus can seek multiple correlates of cognitive decline, with adequate power to detect differences that depend on age, drug, or dose, and any interactions among variables.
These studies, by focusing on AD-specific interactions between aggregate constituents, will open up a new realm of drug targets not previously explored — protein interfaces that lack any beneficial functions and are absent from tissues of healthy, young individuals. Drugs targeting these protein interfaces offer a unique potential to avoid off-target binding and thus to minimize deleterious drug effects. These novel therapeutic agents could provide effective and well-tolerated interventions to prevent and/or treat AD. Because many aggregate constituents are shared by diverse neuropathologies (and indeed by many other age-dependent diseases), they may have broad utility to alleviate a wide range of aggregation-associated pathologies.

I01 BX006445

2024-10-01

2028-09-30