Cancer drug development is notoriously complex, typically taking over a decade from discovery to market approval, with costs scaling into billions of dollars. The process is fraught with scientific uncertainties, high failure rates in clinical trials, and the need for extensive safety evaluations. However, repurposing, which involves finding new anticancer uses for medications already approved for other indications, leverages known safety profiles, pharmacokinetics, and manufacturing processes. This drastically shortens development cycles and enhances the feasibility of testing drugs across diverse cancer types.

Sajwani and colleagues detail the molecular underpinnings that enable such repurposing, explaining how drugs designed for non-oncological targets may inadvertently affect cancer cell survival pathways. For instance, medications primarily utilized in metabolic disorders, immune modulation, or infectious diseases have demonstrated off-target effects that inhibit tumor growth or sensitize cancer cells to conventional chemotherapy. These mechanisms include interference with signaling cascades, epigenetic modulation, and disruption of tumor microenvironment interactions.

The article underscores the pivotal role of computational biology and high-throughput screening in identifying repurposing candidates. Advanced in silico models analyze vast datasets from genomic, proteomic, and pharmacological studies to predict drug-cancer interactions with remarkable precision. Such integrative approaches bypass traditional trial-and-error methods, enabling researchers to shortlist the most promising compounds for experimental validation rapidly.

Furthermore, the study presents multiple case examples where drug repurposing has yielded significant clinical promise. Drugs like metformin, initially an antidiabetic agent, have exhibited antiproliferative effects in several cancers including breast and pancreatic tumors. Likewise, certain antipsychotics and anti-inflammatory agents display potential by modulating intracellular signaling pathways critical for tumor growth and metastasis. These instances illuminate the untapped reservoir of pharmacological tools awaiting oncological application.

Regulatory agencies have also begun to adapt frameworks to facilitate faster approval of repurposed drugs. Since safety data already exist, new indications can often be granted following smaller, focused clinical trials, diminishing the barriers to patient access. Sajwani et al. emphasize that harmonizing regulations with scientific advances is crucial to maximize the impact of repurposed therapies.

Nonetheless, the authors caution that challenges persist. Intellectual property issues can diminish pharmaceutical companies’ incentive to invest in repurposing, given limited patent protection on older drugs. Additionally, the heterogeneity of tumors requires personalized approaches wherein repurposed drugs must be matched to particular genetic or molecular cancer profiles, necessitating companion diagnostics.

To confront these challenges, the research advocates for multi-disciplinary collaboration, integrating oncologists, pharmacologists, computational scientists, and regulatory experts. This ecosystem fosters innovation by combining deep mechanistic understanding with clinical insights and regulatory know-how, ensuring repurposed drugs transition smoothly from bench to bedside.

The report also highlights the role of real-world data analytics and patient registries in monitoring the long-term efficacy and safety of repurposed drugs in diverse populations. These post-market surveillance strategies provide critical feedback, informing iterative improvements in treatment protocols.

Importantly, repurposing expands therapeutic access not only by accelerating development but by lowering costs, enabling broader distribution in low-resource settings. This democratization of cancer care aligns with global health imperatives, addressing disparities exacerbated by high drug prices and scarcity.

Finally, Sajwani et al. envision a future where drug repurposing operates synergistically with other emerging modalities such as immunotherapy and targeted gene editing. Combining repurposed drugs with cutting-edge treatments could potentiate efficacy and overcome resistance mechanisms that bedevil cancer therapy.

In summary, drug repurposing marks a paradigm shift in oncology drug development, deftly navigating around traditional obstacles to deliver treatments faster, cheaper, and more effectively. The compelling evidence and sophisticated methodologies presented by Sajwani and colleagues herald a new epoch in cancer therapeutics, one where innovation meets pragmatism, and hope is rekindled for patients worldwide.

Subject of Research: Drug repurposing strategies in cancer therapy and their scientific, clinical, and regulatory implications.

Article Title: Drug repurposing in oncology: a path beyond the bottleneck.

Article References:
Sajwani, N., Suchitha, G.P., Keshava Prasad, T.S. et al. Drug repurposing in oncology: a path beyond the bottleneck. Med Oncol 42, 443 (2025). https://doi.org/10.1007/s12032-025-02994-w

Image Credits: AI Generated

FGFR-targeted therapies have historically suffered from limited specificity, particularly because the kinase domains of FGFR1, FGFR2, and FGFR3 share over 95% homology within their ATP-binding pockets. This formidable similarity renders the development of selective drugs exceedingly difficult, often leading to pan-FGFR inhibitors that inadvertently suppress FGFR1 and FGFR4 activity. Such broad inhibition results in dose-limiting toxicities, notably hyperphosphatemia and gastrointestinal disturbances, which curtail therapeutic benefit and patient tolerability. Moreover, the clinical efficacy of current FGFR inhibitors is further compromised by emerging resistance mutations predominantly within FGFR2 and FGFR3, underlining the urgent need for next-generation selective inhibitors.

The Insilico team embarked on a rational drug design campaign leveraging their Chemistry42 AI-driven molecular design platform. Initial steps involved meticulous structural analysis which highlighted the kinase hinge region, referred to as core A, as the crucial site for selective ligand engagement. Employing these insights, a pharmacophore model was constructed emphasizing essential hydrogen bond donors and acceptors, while simultaneously incorporating spatial features predictive of selectivity and potency toward FGFR2/3 over other family members. This foundational model laid the groundwork for deep chemical space exploration.

Capitalizing on the generative capabilities of Chemistry42, researchers synthesized an extensive virtual library exceeding 10,000 candidate molecules featuring diverse heterocyclic cores and linkers. Rigorous in silico screening employed the protein-ligand interaction (PLI) scoring function, designed to quantify predicted binding affinities and interaction quality within the kinase domain. This scoring paradigm, alongside filters for drug-like physicochemical properties, enabled the identification of promising scaffolds potentially capable of overcoming resistance mutations by displaying binding flexibility in the hinge region.

Among the top candidates emerged a unique amide-based scaffold, designated core3 (C3), which demonstrated superior predicted selectivity and potency profiles. The adaptive binding characteristics of the C3 core suggested an enhanced capacity to tolerate conformational changes induced by mutant variants of FGFR2/3, a key advantage in circumventing acquired drug resistance. Subsequent computational refinement involved ADMET (absorption, distribution, metabolism, excretion, and toxicity) profiling and rigorous free energy perturbation calculations via Chemistry42’s Alchemistry module, refining the molecular candidates for optimized binding thermodynamics and favorable pharmacokinetics.

The computationally designed molecule ISM7594 embodies the culmination of these efforts—a covalent dual inhibitor with structural innovations centered on a distinctive hinge-binding motif and the novel C3 core. In vitro characterization revealed nanomolar inhibitory potency against both FGFR2 and FGFR3 kinases, paired with an extraordinary selectivity margin exceeding 100-fold relative to FGFR1 and FGFR4 isoforms. Crucially, ISM7594 maintained robust inhibitory activity against clinically relevant FGFR2/3 mutants that commonly drive resistance to existing treatments.

Functional cellular assays underscored the antiproliferative efficacy of ISM7594 in cancer cell lines harboring pathologic FGFR2/3 alterations, while exhibiting minimal cytotoxicity in cells lacking receptor aberrations. These results point to a therapeutic window that could maximize tumor targeting while sparing normal tissues. Furthermore, preclinical animal models demonstrated that ISM7594 yields significant tumor growth suppression with a markedly improved safety profile compared to currently approved pan-FGFR inhibitors, reinforcing its potential clinical value.

This pioneering study, published in the Journal of Medicinal Chemistry, not only exemplifies the power of AI-augmented drug design but also affirms the pivotal role of integrative computational-experimental workflows in expediting the discovery of precision medicines. Dr. Xiao Ding, Head of Chemistry & DMPK and Senior Vice President of Medicinal Chemistry at Insilico Medicine, emphasized the synergy of computational innovation and experimental validation as critical to translating in silico hypotheses into tangible therapeutic candidates.

The research advances beyond the initial discovery phase; in early 2025, Insilico detailed further structure-activity relationship explorations through the discovery of pyrrolopyrazine carboxamide derivatives exhibiting enhanced selectivity and mutant resistance profiles. This iterative optimization reflects the dynamic capabilities of the Chemistry42 platform to not only generate but rapidly refine drug-like molecules through successive design cycles.

Since its inception in 2014, Insilico Medicine has grown into a leading clinical-stage AI biotechnology company, consistently propelling the frontiers of drug discovery by fusing deep generative modeling with reinforcement learning and transformer architectures. Their innovative platforms enable holistic approaches spanning novel target identification, molecular generation, and predictive modeling of pharmacodynamics and pharmacokinetics. This confluence of biology, chemistry, and cutting-edge AI has positioned Insilico among the top global institutions contributing to biological and natural sciences, underscoring their influence on the future landscape of precision medicine.

The successful design and validation of ISM7594 highlight a compelling paradigm shift—where artificial intelligence becomes an indispensable partner in the rational creation of highly selective therapeutics capable of addressing complex challenges such as target homology and drug resistance. This breakthrough heralds a new era in cancer treatment development, promising safer, more effective interventions tailored to molecular vulnerabilities with unprecedented speed and precision.

Subject of Research: Development of selective FGFR2/3 inhibitors overcoming resistance mutations in cancer therapy

Article Title: Rational Design and Identification of ISM7594 as a Tissue-Agnostic FGFR2/3 Inhibitor

News Publication Date: 25-Jun-2025

Web References:
https://doi.org/10.1021/acs.jmedchem.5c00928
https://insilico.com/

References:
[1] Wang, Y. et al. (2025) ‘Rational design and identification of ISM7594 as a Tissue-Agnostic FGFR2/3 inhibitor,’ Journal of Medicinal Chemistry [Preprint]. https://doi.org/10.1021/acs.jmedchem.5c00928

Image Credits: Insilico Medicine

Keywords

Pharmaceuticals, Fibroblast Growth Factor Receptors, FGFR2/3 Inhibitors, Cancer Therapy, AI-Driven Drug Design, Molecular Modeling, Chemistry42, Drug Resistance, Kinase Inhibitors, Precision Medicine, Medicinal Chemistry

The cyclin-dependent kinases 12 and 13 are integral to the regulation of the DNA damage response (DDR) pathway, a crucial mechanism that maintains genomic stability. Their role in tumorigenesis and the emergence of resistance to various antitumor therapies underscores the need for effective inhibitors that can selectively and potently target these proteins. Historically, the development of such inhibitors has encountered significant obstacles, primarily due to issues related to toxicity and ineffectiveness linked to previous non-covalent and covalent inhibitors.

To overcome these barriers, the researchers at Insilico Medicine leveraged their state-of-the-art generative AI platforms, particularly PandaOmics and Chemistry42. These powerful tools enabled a comprehensive analysis of potential therapeutic targets, allowing CDK12 to emerge as a top candidate from extensive multiomic datasets. The AI’s ability to process vast amounts of biological data not only facilitated the identification of promising targets but also aided in the prioritization of indications for cancer types most likely to benefit from this therapeutic approach.

Following the identification of CDK12, the research team utilized AI-guided structure-activity relationship (SAR) methodologies. This approach allowed them to design new compounds with optimized properties, focusing on enhancing the selectivity and stability of the inhibitors while minimizing off-target effects. The resultant compound series demonstrated improved oral bioavailability and pronounced inhibitory activity against CDK12 and CDK13.

During the preclinical evaluation phase, extensive in vitro and in vivo examinations of compound 12b revealed remarkable potency, achieving nanomolar concentrations across various cancer cell lines. The compound not only exhibited favorable pharmacokinetic properties but also displayed substantial efficacy in targeted cancer models, including breast cancer and acute myeloid leukemia (AML). Importantly, these positive effects were observed without inducing intolerable side effects, addressing a significant hurdle in the development of novel cancer therapeutics.

The implications of this research extend beyond mere findings, as Dr. Hongfu Lu, the co-lead author of the study and Senior Director of Chemistry at Insilico Medicine, articulated the company’s vision of revolutionizing the drug discovery process using advanced AI technologies. Dr. Lu emphasized the potential of CDK12/13 dual inhibitors to effectively target treatment-resistant tumors, marking a crucial step toward enhancing cancer treatment outcomes and patient care.

The use of AI in drug discovery represents a paradigm shift in the way therapeutic agents are developed. By harnessing the power of deep learning and generative models, Insilico Medicine is not only accelerating the discovery of new compounds but is also optimizing existing ones to make significant improvements in efficacy and safety. This intersection of technology and drug development is paving the way for breakthroughs that could redefine treatment options available to oncologists and their patients.

The study’s findings are particularly germane in light of the growing recognition of cancer’s complexity and the necessity for tailored therapeutic strategies. The insights gained from the extensive computational analyses and biological evaluations provide a compelling framework for future studies aimed at expanding the applicability of CDK12/13 inhibitors. As research continues, there is optimism that these findings will transition to clinical trials, potentially revolutionizing therapeutic approaches for patients with difficult-to-treat cancers.

The urgency for novel cancer therapies has never been more pressing, particularly as the rise of treatment-resistant tumors poses an ever-growing challenge. Insilico Medicine’s work not only illuminates the path forward for targeted therapies but also reinforces the potential of AI to catalyze advancements in oncology. The collaborative efforts amongst biologists, chemists, and data scientists are crucial as they work towards delivering innovative treatments that leverage cutting-edge technology.

As this field of research advances, it is essential to maintain a dialogue about the implications of AI in drug discovery, particularly regarding ethical considerations, data security, and the need for robust regulatory frameworks. The promise of generative AI-driven drug development offers unprecedented opportunities, yet it comes with responsibilities that must be navigated carefully to ensure that therapeutic breakthroughs benefit all segments of the population.

In conclusion, Insilico Medicine’s publication shines a light on a promising avenue in cancer therapeutics, highlighting the role of AI in overcoming longstanding challenges in drug discovery. The introduction of orally available covalent CDK12/13 dual inhibitors signifies a remarkable achievement toward developing therapies that could reshape the landscape of cancer treatment. As the field evolves, continued research, collaboration, and innovation will be essential to bring these promising compounds from the lab to the clinic, ultimately changing the lives of patients battling cancer.

Subject of Research: CDK12/13 dual inhibitors as a treatment for refractory cancers
Article Title: Design, synthesis, and biological evaluation of novel orally available covalent CDK12/13 dual inhibitors for the treatment of tumors
News Publication Date: 13-Feb-2025
Web References: www.insilico.com
References: Lu, H., et al. (2025). Journal of Medicinal Chemistry. DOI: 10.1021/acs.jmedchem.4c01616
Image Credits: Not provided

Keywords

Generative AI, Molecular targets, Breast cancer, Discovery research, Colorectal cancer, Ovarian cancer, Computational chemistry