Combined TNF-a & Solubility Prediction
Rheumatoid arthritis (RA) is a chronic autoimmune inflammatory disease that affects synovial joints by targeting the synovial membrane, articular cartilage, and bones. High levels of proinflammatory cytokines such as tumor necrosis factor alpha (TNF-α) and interleukin-1b (IL-1b) are associated in a variety of inflammatory diseases,3,4 such as RA, multiple sclerosis, inflammatory bowel disease, and Crohn’s disease (CD). The important role of TNF-α in the pathogenesis of RA was demonstrated both in experimental animal models and in RA patients. As a result, the blockade of TNF-α production may lead to the development of new anti-TNF-α therapies. They have been reported three drugs in use for treatment that block the activity of TNF-α: Infliximab (chimeric monoclonal antibody to human TNF), Adalimumab (human monoclonal antibody to TNF), and Etanercept (soluble TNF receptor construct). However, there are some drawbacks in the therapy of RA by these drugs including their high cost, inadequate clinical response, need of intravenous administration and several side effects such as increased risk of tuberculosis.
This novel multistep framework gives insight into the structural characteristics that affect the the inhibitory activity on TNF-α (NF-κΒ induction) and aqueous solubility. In addition, the simplicity of the proposed model provides expansion to its applicability such as in virtual screening procedures. The model can also be used to screen existing databases or virtual combinations to identify derivatives with desired activity and solubility. In this scenario, the classification model will be used to screen out “inactive” compounds, while the applicability domain will serve as a valuable tool to filter out “dissimilar” combinations.