Schrödinger Ends CDC7 Inhibitor Program Following Two Patient Deaths in AML Trial
Schrödinger, founded in 1990 and known for its physics-based computational platform used in drug and materials discovery, has halted development of its experimental CDC7 inhibitor SGR-2921 after the drug was linked to two patient deaths in a phase 1 trial for relapsed or refractory acute myeloid leukemia (AML) and high-risk myelodysplastic syndromes. The decision removes a candidate the company had once described as the most potent CDC7 inhibitor reported, while it continues advancing other oncology, immunology, and neurology programs in early-stage development.
CDC7, a kinase involved in DNA replication initiation, has been a target of oncology research for over two decades, but no CDC7 inhibitors have progressed to late-stage clinical trials. SGR-2921 had shown early monotherapy activity, and Schrödinger had planned to explore combination regimens with standard-of-care agents. However, the company now believes the path to developing the agent in combination “would be difficult to pursue”.
With SGR-2921 dropped, Schrödinger’s current oncology pipeline includes:
- SGR-1505 (MALT1 inhibitor) – A small-molecule allosteric inhibitor of mucosa-associated lymphoid tissue lymphoma translocation protein 1 (MALT1), in Phase 1 trials for relapsed or refractory B-cell lymphomas and chronic lymphocytic leukemia. The compound has orphan drug and fast track designations in mantle cell lymphoma and Waldenström macroglobulinemia, respectively. Initial Phase 1 data showed a favorable safety profile and preliminary efficacy signals.
- SGR-3515 (Wee1/Myt1 inhibitor) – A dual inhibitor targeting kinases involved in cell cycle regulation and DNA damage repair. Phase 1 studies are ongoing in advanced solid tumors, with preclinical data indicating sustained target inhibition and activity as both monotherapy and in combination regimens.
- SGR-4174 (SOS1 inhibitor) – A small molecule designed to disrupt SOS1–KRAS interactions, targeting the most frequently mutated oncogene in human cancers. Preclinical data showed high selectivity for SOS1 over SOS2 and other kinases, with strong tumor growth inhibition as monotherapy and in combination with MEK or KRAS inhibitors. Potential indications include lung adenocarcinoma and RASopathies such as Neurofibromatosis Type 1.
- EGFRC797S inhibitors – Multiple identified compounds aimed at treating advanced non-small cell lung cancer patients who progress after first-line EGFR inhibitor therapy, targeting the C797S resistance mutation. These agents are intended for use in combination regimens to achieve deeper and more durable responses.
In addition to oncology, Schrödinger reports having investigational programs in immunology and neurology, also developed using its computational drug discovery platform.
Image credit: Schrödinger’s Pipeline
Schrödinger’s proprietary computational platform combines molecular modeling, simulation, and AI-assisted workflows. At its core is Free Energy Perturbation (FEP+), a physics-based method for predicting protein-ligand binding affinities with accuracy close to experimental measurements (about 1 kcal/mol). The technology is used across a wide chemical space to evaluate potency, selectivity, and solubility, supporting work from early structure prediction through to protein engineering.
In hit discovery, FEP+ can rescore virtual screening results, identify novel chemical cores through core hopping, and conduct large-scale in silico fragment screens. During hit-to-lead and lead optimization, it is applied to refine on-target potency, improve selectivity across gene families, and adjust ADMET properties while maintaining activity. In protein engineering, it is used to refine antibody candidates, predict peptide binding and stability, and design enzymes with specific substrate preferences. Active learning methods allow the workflow to scale to large compound libraries, reducing the need for extensive experimental screening.
Topics: Clinical Trials
