英矽智能发现AI驱动的WRN抑制剂，靶向高度微卫星不稳定性癌症

· 在高度微卫星不稳定（MSI-H）癌细胞中，WRN被确定为合成致死靶点，对于不响应现有疗法的MSI-H肿瘤患者具有治疗潜力。

· 英矽智能Chemistry42下属创新模块Alchemistry利用非平衡切换方法，可以准确且快速地估算蛋白质-配体复合物的结合自由能，简化并加速了ISM2196的优化过程。

· 在临床前研究中，ISM2196在多种MSI-H癌症模型中表现出强大的体内抗肿瘤疗效，并具有良好的选择性、安全性和ADMET（吸收、分布、代谢、排泄和毒性）特性。

由生成式人工智能（AI）驱动的临床阶段生物医药科技公司英矽智能宣布提名临床前候选药物ISM2196，这是一款由AI驱动发现的、潜在“同类最佳”（best-in-class）WRN抑制剂，旨在通过合成致死策略治疗晚期转移性微卫星不稳定性（MSI）癌症。

DNA解旋酶Werner (WRN）是RecQ解旋酶家族成员，在维持基因组完整性和促进DNA损伤修复中起关键作用。最近的研究发现，WRN缺陷会选择性杀伤MSI-H癌细胞，同时不影响微卫星稳定（MSS）细胞生存，这使WRN成为MSI-H肿瘤的合成致死靶点。以结直肠癌、子宫内膜癌、胃癌为代表，超过20种不同类型的癌症中均观察到MSI-H状态，全球范围内每年MSI-H癌症的确诊病例数达到数十万。

ISM2196是一款创新强效WRN抑制剂，其设计与优化过程由Chemistry42下属模块Alchemistry赋能。Alchemistry利用非平衡切换方法，准确且快速地估算蛋白质-配体复合物的结合自由能，为药物发现与研发过程提供助力。在WRN项目的开发中，多数AI预测的化合物表现出优于参考化合物的活性。此后，英矽智能研发团队进一步整合并分析了来自Alchemistry的排序结果、ADMT活性及候选化合物的PK/PD属性，进而快速筛选出有前景的候选分子用于后续合成和测试。

ISM2196在临床前实验中显示了具有前景的研究结果，在多种癌症模型中，包括结直肠癌、子宫内膜癌和胃癌等，均表现出强大的体内抗肿瘤疗效。此外，它还展示了理想的类药性，包括出色的体外ADMET特性、良好的体内暴露、低血浆清除率和在多种物种模型中较好的口服生物利用度。这些发现突显了ISM2196作为潜在同类最佳WRN抑制剂治疗MSI-H肿瘤的潜力。英矽智能已在2024年的AACR年会上披露了关于这个项目的部分临床前关键数据。

英矽智能联合首席执行官兼首席科学官任峰博士表示，“尽管免疫检查点抑制剂已被批准用于治疗MSI-H肿瘤，但仍有相当一部分患者对现有疗法无应答，也有一部分患者饱受现有药物毒性的困扰。WRN抑制剂有望为他们带来新的希望，我们正在加紧推进ISM2196的IND-enabling研究，期待与经验丰富的合作伙伴一起，推进该候选药物的临床验证，为患者提供创新的治疗选择。”

英矽智能创始人兼首席执行官Alex Zhavoronkov博士表示，“我们从未停止过对Pharma.AI平台的开发和研究工作，集成了前沿技术的新AI模块正在展现出巨大的潜力。在WRN项目中，这些模块优化苗头化合物系列的经验，经验再次彰显了AI技术推动药物发现的可能性，我们期待尽快为那些具有未被满足医疗需求的患者带来下一代的创新疗法。”

2016年，英矽智能全球首次在同行评审期刊上阐述了使用生成式人工智能设计新型分子的概念，为涵盖生成生物学、化学和医学等领域的商业化Pharma.AI 平台奠定了基础。自2021年以来，英矽智能在自有人工智能平台Pharma.AI的支持下，建立了超过30条丰富的自研管线组合，并从中提名了19款临床前候选项目，其中9款化合物获得临床试验许可。

在近期举办的IMGAIA（Insilico Medicine Generative AI Action）网络研讨会上，英矽智能发布了自有Pharma.AI平台更新，重点介绍了首次亮相的创新应用，包括用于隐私计算的Biology42: PandaOmics Box硬件、用于虚拟数据生成和生物医学研究的Precious-3 GPT，以及用于科研文件起草的研究助理Science42: DORA。以此为基点，英矽智能期待引领更具社会责任、更加可持续的技术突破。如您有兴趣试用上述平台，请联系 BD@insilicomedicine.com。

· WRN was identified as a synthetic lethal target in microsatellite instability-high (MSI-H) cancer cells, holding treatment potential for MSI-H patients not responding to available therapies.

· Alchemistry, one of the latest modules of Chemistry42, streamlined and accelerated the optimization of ISM2196, by leveraging non-equilibrium switching to accurately and rapidly compute binding free-energy estimates for protein-ligand complexes.

· In preclinical studies, ISM2196 has demonstrated potent in vivo anti-tumor efficacy in multiple MSI-H cancer models, as well as favorable selectivity, safety and ADMET profiles.

Insilico Medicine, a clinical stage generative artificial intelligence (AI)-driven drug discovery company, today announced the nomination of ISM2196, an AI-driven, potential best-in-class small molecule WRN inhibitor targeting advanced metastatic microsatellite instability (MSI) cancers guided by synthetic lethality strategy.

Werner helicase (WRN) is a member of the RecQ helicase family, which plays a critical role in maintaining genome integrity and promoting DNA damage repair. Recent studies identified that WRN deficiency selectively impairs the viability of MSI-H but not microsatellite stable (MSS) cancer cells, which indicates WRN is a synthetic lethal target for MSI-H tumors. MSI-H condition is observed in more than 20 different types of cancers, especially in colorectal cancer, endometrial cancer, gastric cancer and others, with hundreds of thousands of cases diagnosed globally each year.

ISM2196 is a novel, potent WRN inhibitor, designed with the assistance of Alchemistry, one of the latest modules of Chemistry 42. Alchemistry enhances drug R&D by leveraging non-equilibrium switching to accurately and rapidly compute binding free-energy estimates for protein-ligand complexes. In developing of the WRN program, most predicted compounds demonstrated superior activity to the reference compounds. Insilico's scientists further integrated and analyzed the ranking results from Alchemistry, ADMT activity and PK/PD properties of the candidates, enabling the rapid selection of promising molecules for synthesis and testing.

ISM2196 has shown promising results in preclinical studies, demonstrating potent in vivo anti-tumor efficacy across multiple cancer models, including colorectal, endometrial, and gastric cancers. Additionally, it showcased good drug-like properties, including excellent in vitro ADMET profiles, favorable in vivo exposure, low clearance, and optimal oral bioavailability across multiple preclinical species. These findings highlight the potential of ISM2196 as a best-in-class WRN inhibitor treating MSI-H tumors. Insilico disclosed parts of the key preclinical data during AACR 2024.

“Although immune checkpoint inhibitors have been approved for the treatment of MSI-H tumors, a significant proportion of patients do not respond to available therapies and a subset of patients suffer from drug toxicity problems. WRN inhibitors are expected to be promising novel treatment options,” says Feng Ren, Ph.D., Co-CEO and Chief Scientific Officer of Insilico Medicine. “We are conducting the IND-enabling study of ISM2196 and look forward to working with experienced partners to advance the candidate to clinical validation and providing innovative treatment options to patients.”

“We have never stopped developing the Pharma.AI platform, which is demonstrating even greater potential with the integration of new modules powered by cutting-edge technology,” says Alex Zhavoronkov, Ph.D., Founder and CEO of Insilico Medicine. “The experience of AI optimized drug candidates in the WRN program once again underscores the possibilities of AI technology to transform drug discovery and deliver next-generation therapeutics to patients with unmet medical needs."

In 2016, Insilico first described the concept of using generative AI for the design of novel molecules in a peer-reviewed journal, which laid the foundation for the commercially available Pharma.AI platform. Since then, Insilico keeps integrating technical breakthroughs into Pharma.AI platform, which is currently a generative AI-powered solution spanning across biology, chemistry and clinical development. Powered by Pharma.AI, Insilico has nominated 19 preclinical candidates in its comprehensive portfolio of over 30 assets since 2021 and has received IND approval for 9 molecules.

During the recent Insilico Medicine Generative AI Action (IMGAIA) webinar, updates on the Pharma.AI platform were presented, highlighting its latest features, including Biology42: PandaOmics Box hardware for confidential computing, Precious-3 GPT for virtual data generation and biomedical research, and Science42: DORA for drafting scientific documents. These enhancements underline Insilico’s commitment to pioneering breakthroughs responsibly and sustainably. Those who are interested in trial versions of the abovementioned platforms are encouraged to contact BD@insilicomedicine.com.