Comparison of Safety Biomarkers for Chemotherapeutics

Toxicity is one of the leading causes of attrition in drug development.¹ In this context, safety biomarkers are invaluable as they enable the likelihood, presence, or extent of toxicity as an adverse effect to be determined.² Safety biomarkers can therefore reduce the risk and cost associated drug development, permitting pharma companies to optimize their resource allocation and ultimately accelerate the delivery of safe, effective drugs to market.

However, biomarker discovery is challenging. A reliable safety biomarker must be both specific and sensitive to drug-induced organ toxicities, requiring extensive research and validation. Furthermore, sifting through 1.1M+ biomarker papers on PubMed to find relevant molecular signatures is time-consuming and not comprehensive. Here, we show how AI can navigate the vast literature to uncover relevant safety biomarkers of rituximab, a monoclonal antibody used to treat various types of cancers and autoimmune diseases.³

By machine-reading the literature, Causaly extracted 68 biomarkers used to monitor rituximab resistance. Using the advanced filtering capabilities, a holistic view of more than 400 biomarkers affected by rituximab administration in clinical studies were identified.

Using the timeline view, recently reported biomarkers affected by the administration of rituximab in clinical studies can be visualized. This provides a representation of the latest findings in chronological order, based on when the relationship of each biomarker and rituximab was first identified. As an example, PROC gene was selected as a recently reported biomarker (within this criteria) for further exploration.

A 2023 study reported an interaction between CD27, NCAM1, PROC and the target protein of rituximab in membranous nephropathy.⁴ Using the multi-hop module, we can break down the drug-biomarker relationship, to uncover potential mediators of rituximab on PROC and underlying mechanisms. This uncovered 4 genes and proteins as potential mediators of the drug-gene relationship.

We can then perform a comparative analysis, to investigate if a biomarker is affected by a variety of drugs. In this example, we compared rituximab to chemotherapeutic drugs 5-fluorouracil, cytarabine and cisplatin. 150 biomarkers were common between all four drugs, while 226 biomarkers were found to be unique to rituximab.

The identification and utilization of safety biomarkers plays a key role in mitigating toxicity risks and reducing costs in drug development, thereby accelerating the delivery of safe and effective drugs to patients. AI can streamline the identification of relevant biomarkers from the ever-growing biomedical literature, offering insights into drug resistance and toxicity.

Want to discover more about how Causaly can accelerate your biomarker research?