MLN-8054

Gastrointestinal nematodes (GINs) are among the most common parasites of humans, livestock, and companion animals. GIN parasites infect 1-2 billion people worldwide, significantly impacting hundreds of millions of children, pregnant women, and adult workers, thereby perpetuating poverty. Two benzimidazoles with suboptimal efficacy are currently used to treat GINs in humans as part of mass drug administrations, with many instances of lower-than-expected or poor efficacy and possible resistance. Thus, new anthelmintics are urgently needed. However, screening methods for new anthelmintics using human GINs typically have low throughput. Here, using our novel screening pipeline that starts with human hookworms, we screened 30,238 unique small molecules from a wide range of compound libraries, including ones with generic diversity, repurposed drugs, natural derivatives, known mechanisms of action, as well as multiple target-focused libraries (e.g., targeting kinases, GPCRs, and neuronal proteins). We identified 55 compounds with broad-spectrum activity against adult stages of two evolutionary divergent GINs, hookworms (Ancylostoma ceylanicum) and whipworms (Trichuris muris). Based on known databases, the targets of these 55 compounds were predicted in nematode parasites. One novel scaffold from the diversity set library, F0317-0202, showed good activity (high motility inhibition) against both GINs. To better understand this novel scaffold's structure-activity relationships (SAR), we screened 28 analogs and created SAR models highlighting chemical and functional groups required for broad-spectrum activity. These studies validate our new and efficient screening pipeline at the level of tens of thousands of compounds and provide an important set of new GIN-active compounds for developing novel and broadly active anthelmintics.

Predicting drug-target binding affinity is critical for drug discovery, as it helps identify promising drug candidates and predict their effectiveness. Recent advancements in deep learning have made significant progress in tackling this task. However, existing methods heavily rely on training data, and their performance is often limited when predicting binding affinities for new drugs and targets. To address this challenge, we propose a novel Generalized Feature Learning (GFLearn) model for drug-target binding affinity prediction. By integrating Graph Neural Networks (GNNs) with a self-supervised invariant feature learning module, our GFLearn model can extract robust and highly generalizable features from both drugs and targets, significantly enhancing prediction performance. This innovation enables the model to effectively predict binding affinities for previously unseen drugs or targets, while also mitigates the common issue of prediction performance degrading due to shifts in data distribution. Extensive experiments were conducted on two diverse datasets across three challenging scenarios: new drugs, new targets, and combinations of both. Comparisons with state-of-the-art methods demonstrated that our GFLearn model consistently outperformed others, showcasing its robustness across various prediction tasks. Additionally, cross-dataset evaluations and noise perturbation experiments further validated the model's generalizability across different data distributions. Case studies on two drug-target pairs, Canertinib-PIK3C2G and MLN8054-FLT1, provided further evidence of GFLearn's ability to make accurate binding affinity predictions, offering valuable insights for drug screening and repurposing efforts.