Autoimmune diseases (AID) are hard to manage chronic inflammatory diseases. Management of these diseases relies on non-selective therapies with limited clinical efficacies and significant side effects. Janus Kinase 1 (JAK1) and Chemokine Receptor 4 (CXCR4) inhibitors have shown promising results in treating AID prompting us to model these interesting targets. Towards this end, we implemented elaborate ligandbased and structure-based computational workflows to explore the structural features required for potent JAK1 inhibitors. In our ligand-based route we coupled extensive pharmacophore exploration with genetic function algorithm (GFA), k-nearest neighbor (kNN) or multiple linear regression (MLR) analyses to generate predictive QSAR models based on optimal combinations of pharmacophores and physicochemical descriptors. On the other hand, we implemented the structure-based methodology; docking-based Comparative Intermolecular Contacts Analysis (db-CICA), to identify the best possible docking settings required to fit the training set into the binding pocket of JAK1. Optimal db-CICA models were converted into corresponding pharmacophore models. Pharmacophore models developed using structure-based and ligand-based techniques were used as virtual search queries to scan the National Cancer Institute (NCI) database for new promising JAK1 inhibitory leads. Six micromolar hits were captured: 266, 293, 294, 314, 326, and 336. The most active hit 326 has an IC50 of 1.04 µM. the best pharmacophore model was subsequently used to guide synthesis of new JAK1 inhibitors. The best synthetic analogue 253 has IC50 value of 16 μM.

Similarly, CXCR4 inhibitors were modeled via ligand-based computational workflow to explore the structural features required for potent CXCR4 inhibitors. Pharmacophore models were used as virtual search queries to scan the National Cancer Institute (NCI) database for new promising CXCR4 inhibitory leads. Two leads were identified: 339and 353. The most active hit 339 has an IC50 of 24 µM

​​