Considerations for infection risks identified in target safety review.

Target Safety Review (TSR) is a document that triggers the Target Safety Assessment and is designed to predict and identify on-target safety risks. However, with regard to the anticipated on-target risks extracted by TSR, nonclinical toxicity studies sometimes have limitations in predicting clinical safety risks. In this study, we investigated whether nonclinical toxicity studies could predict clinical infection risks expected in TSR. Using the OFF-X database, we identified 942 targets and 1323 drugs associated with infection risks. Through successive filters, i.e., applying "drug alert", "on-target", "causality with severity", "class score", and "evidence levels", biologics became the focus, resulting in six targets and 17 drugs approved by PMDA. From PMDA reports, package inserts, and interview forms, we extracted immune-related findings in nonclinical toxicity studies and infection risks in clinical settings to analyze predictivity by target. Our survey shows clear immune-related toxicity in nonclinical studies of CD20 inhibitors (obinutuzumab, ofatumumab, rituximab), TNF-alpha inhibitors (adalimumab, golimumab), and IL-6R antagonists (sarilumab, tocilizumab). Indicators included B cell depletion, thymus atrophy, and decreased neutrophils. For IL-17, IL-23, and IL-1beta inhibitors, nonclinical studies did not always predict the infection risks observed in clinical settings. Although nonclinical toxicity studies sometimes provided insights into potential infection risks which can be anticipated by its target, their predictive power for specific mechanism of actions and infections appeared limited. Understanding the MOAs and biological background of the drug through the TSR and identifying and mitigating risks early in the project are essential for the safe development of new therapeutic agents.