Identifying novel therapeutic targets for Alzheimer’s disease is a high priority. The two approved drugs share the same target—beta amyloid aggregates—and provide only limited therapeutic benefit to a relatively small percentage of the patient population. Molecules and brain cells involved in the innate immune response—the body’s first line of defense against harmful substances—are of particular interest given the strong evidence implicating neuroinflammation in Alzheimer’s disease. Indeed, many known genetic risk factors for late-onset Alzheimer’s are primarily expressed by brain immune cells (astrocytes and microglia). Two of these genes, APOE and TREM2, are involved in the transition of microglia from a homeostatic sentry state to a DAM (disease-associated microglia) state in response to beta amyloid and cell debris. Both also play key roles in cholesterol and lipid handling. As strategies targeting APOE and TREM2 make their way through the clinical drug discovery and trial pipeline, the field continues to search for other promising candidates that can positively intervene in this complicated immune response.
Drs. Cashikar and Elgendy suggest that another DAM gene called cholesterol 25-hydroxylase (CH25H) might be one of those promising candidates. While working with RLG member Dr. Dave Holtzman and SAB member Dr. Steve Paul, they previously found that levels of CH25H are higher in human brain tissue from Alzheimer’s patients compared to healthy controls and in both amyloid and tau mouse models. CH25H primarily modifies cholesterol but also plays other roles in lipid processing and proinflammatory immune responses. Its role in Alzheimer’s disease pathology is still relatively unstudied; however, prior studies link this gene and related signaling molecules to other nervous system disorders characterized by inflammation and following viral influenza infections. In their preliminary studies, the Cashikar and Elgendy teams found that deleting the CH25H gene in a mouse model of tau pathology (PS19) prevented the migration of peripheral T cells into the brain, reduced glial activation, and improved neurodegeneration. They thus hypothesize that targeting CH25H will be protective against neuroinflammation and neurodegeneration.
The team is pursuing a drug discovery effort that leverages cutting-edge computational (in silico) modeling to address the lack of structural information about the target. Molecules affect a protein’s activity by physically interacting with them, so knowing the structure of a protein and how it changes its conformation in its biological environment is vital to predicting what molecules might be able to access and bind to different positions on the protein. Historically, the techniques to determine protein structure were expensive and difficult, and very few proteins had been well defined. Scientists had to repeatedly combine a protein of interest with as many of its potential binding partners as the lab could afford to test and observe the outcomes to find pairs that interacted, a low-potential and tedious approach given the immense number of potential partners. In silico modeling provides an alternative approach to studying these proteins with limited structural information.
In their first year, the Cashikar and Elgendy team, in collaboration with Dr. Lamees Hegazy’s lab, used Alphafold—an AI-driven tool to predict protein structures that won its developers the 2024 Nobel Prize in Chemistry—to establish a well-predicted experimental physical structure for CH25H. After the refinement of the structure, they performed virtual (in silico) modeling of the structure and its molecular dynamics in relevant biological conditions (e.g., embedded in bi- and monolayer cellular membranes). They then performed in silico screening of over one million candidate molecules for likely binding affinity to CH25H’s predicted structure and available binding sites. Thanks to these computational methods, what would have taken years with old methods (if even possible) was completed in just the first year of this project. In parallel preparation for the eventual biological validation of the in silico hits, the team also developed assays to quantify both the amount of CH25H and the enzymatic activity of CH25H in primary mouse microglia. These assays will be critical to assessing the efficacy of candidate compounds to inhibit CH25H and will be used during the remaining screening and validation experiments.
The team is now synthesizing and experimentally validating the binding activity of the lead candidates from the screening in cell culture, including measuring which ones best inhibit CH25H activity. They will also be evaluated for target specificity, drug-like properties, and likely brain penetrance. Building on their earlier in silico and in vitro work, they will then advance to in vivo studies that will bring the work closer to human biology. These studies will test whether their top candidate can delay the onset of neuroinflammation and neurodegeneration in a tau mouse model. The team will measure several outcomes in the mouse brain, including markers of inflammation in microglia and astrocytes, tissue loss, and pathological forms of tau. Drs. Cashikar and Elgendy ultimately seek to leverage these data to apply for NIH funding to support further small molecule drug discovery efforts, the filing of an Investigational New Drug application, and phase I clinical testing.
