Pathos借助AI Agents获得两款有望进入临床开发的癌症药物
Pathos Picks Up Two Cancer Drugs Poised for Clinical Development With AI Agents
Pathos AI是一家开发人工智能赋能技术、用于癌症药物研究的初创公司。该公司已获得两种分子的权利，这两种分子的临床测试将借助一个旨在让试验更快、更高效的专有平台推进。
总部位于纽约的Pathos周一晚间表示，将预付1.25亿美元，以获得由Alphamab开发的JSKN016的相关权利。这款潜在的同类首创药物旨在针对两种癌症靶点，即TROP2和HER3蛋白。虽然目前已有分别作用于这两个靶点的FDA批准药物，但目前还没有同时针对两者的疗法。
总部位于中国的Alphamab将JSKN016设计为一种双特异性抗体药物偶联物（ADC），可与肿瘤上的TROP2和/或HER3结合，阻断这些驱动癌症的蛋白质所参与的信号通路。该分子还会释放拓扑异构酶I抑制剂，这是一类常被用作ADC药物载荷的化疗药物，从而提供另一种对抗癌症的机制。
在Alphamab主导下，JSKN016的静脉输注制剂已在中国进入针对三阴性乳腺癌患者的III期试验，这些患者的疾病对至少两线既往全身化疗未产生应答。该药物的皮下注射版本正在中国的一项Ib期研究和澳大利亚的一项I期研究中接受评估。
该协议授予Pathos在大中华区以外开发和商业化JSKN016的独家权利，Alphamab保留大中华区内的权利。如果该药物达到相关里程碑，Pathos可能向Alphamab支付最高21亿美元，此外，如果该药物进入市场，还需支付销售特许权使用费。
Pathos表示，JSKN016是借助Foundry这一用于肿瘤药物开发的初创公司专有AI平台技术完成识别并确定开发优先级的。据Pathos介绍，Foundry使用数千个AI Agents并行工作，分析生物学、临床和真实世界数据。这种分析可以识别出具有较高潜力的药物。
另一项交易使Pathos参与AstraZeneca管线中一款候选癌症药物的合作开发。这种小分子药物AZD4241属于被称为蛋白质降解剂的一类疗法。在某些癌症中，突变形式的雌激素受体（ER）会驱动疾病发展。AZD4241旨在降解这一受体。该分子目前正处于针对ER阳性、HER2阴性乳腺癌的临床前开发阶段。
与AstraZeneca的协议使Pathos负责AZD4241的早期临床开发。双方未披露财务条款。
在Pathos的主导下，JSKN016和AZD4241的临床开发都将借助Foundry推进。公司表示，除用于药物发现外，该技术还可以识别符合临床试验条件的患者并设计研究方案。Foundry还可以优化给药剂量，并将患者匹配到合适的疗法。Pathos AI首席执行官Iker Huerga在一份准备好的声明中表示，癌症药物研发的瓶颈不在于发现分子，而在于证明这些药物对正确的患者有效。
“AZD4241具有令人信服的作用机制，”他在一份准备好的声明中说。“Foundry的任务是设计能够证明这一点的试验——将这款药物与其生物学特征决定需要该药物的患者进行匹配。这就是我们压缩时间的方式。”Pathos表示，其管线中的所有药物均已由Foundry完成识别、评估或加速开发。最先进的项目是P-500，这是一种可穿过血脑屏障、抑制一种名为PRMT5的酶的小分子抑制剂。
该药物于2024年从Prelude Therapeutics获得许可，此后已进入针对晚期实体瘤的临床开发中期阶段，其中包括高级别胶质瘤和葡萄膜黑色素瘤。
Pocenbrodib是一种抑制CBP/p300蛋白的小分子抑制剂，最初由Forma Therapeutics开发；Forma Therapeutics于2022年被Novo Nordisk收购。Pathos于2023年获得该药物的权利；目前，该药物正处于针对前列腺癌、乳腺癌以及多发性骨髓瘤的早期临床开发阶段。
5月，Pathos收购了DeuterOncology的多数股权。DeuterOncology是一家总部位于比利时的生物技术公司，目前正处于早期临床开发阶段，其研发项目DO-2是第三代MET抑制剂，正处于针对MET改变型非小细胞肺癌的早期临床开发阶段。DO-2由Foundry识别。
插图：Getty Images
Pathos AI, a startup that developed artificial intelligence-enabled technologies for cancer drug research, has secured rights to two molecules whose clinical testing will proceed with a proprietary platform designed to make trials faster and more efficient.
New York-based Pathos said late Monday it is
paying $125 million up front for rights to JSKN016, developed by Alphamab. This potential first-in-class drug is designed to go after two cancer targets, the proteins TROP2 and HER3. While there are FDA-approved drugs that hit each target, there are currently no therapies that address both.
China-based Alphamab designed JSKN016 as a bispecific antibody drug conjugate (ADC) that binds to TROP2 and/or HER3 on tumors, blocking signaling pathways of these cancer-driving proteins. This molecule also releases topoisomerase I inhibitor, a type of chemotherapy often used as an ADC drug payload, offering another mechanism to fight cancer.
By Michael Blackman, MD, MBA Chief Medical Officer, Greenway Health®
Under Alphamab, an intravenously infused formulation of JSKN016 has reached Phase 3 testing in China in patients with triple-negative breast cancer that has failed to respond to at least two prior lines of systemic chemotherapy. A subcutaneously injected version of the drug is being evaluated in a Phase 1b study in China and in a Phase 1 study in Australia.
The agreement grants Pathos exclusive rights to develop and commercialize JSKN016 outside of Greater China, where Alphamab retains rights. Pathos could pay Alphamab up to $2.1 billion if the drug achieves milestones, plus royalties from sales if the drug reaches the market.
Pathos said JSKN016 was identified and prioritized for development with Foundry, the startup’s proprietary AI platform technology for oncology drug development. According to Pathos, Foundry uses thousands of AI agents that work in parallel to analyze biological, clinical, and real-world data. This analysis can identify drugs with high potential.
A separate deal brings Pathos into a collaboration on a cancer drug candidate from the pipeline of AstraZeneca. This small molecule, AZD4241, is part of the class of therapies called protein degraders. In some cancers, mutated forms of the estrogen receptor (ER) drive the disease. AZD4241 is designed to degrade this receptor. This molecule is in preclinical development for ER-positive, HER2-negative breast cancer.
The
agreement with AstraZeneca makes Pathos responsible for early clinical development of AZD4241. No financial terms were disclosed.
Under Pathos, clinical development of both JSKN016 and AZD4241 will proceed with Foundry. Beyond its applications in drug discovery, the company says this technology can also identify patients eligible for a clinical trial and design the study. Foundry can also optimize dosing and match patients to the right therapies. In a prepared statement, Pathos AI CEO Iker Huerga said the bottleneck in cancer drug R&D is not in finding molecules, but rather in proving that they work in the right patients.
“AZD4241 has a compelling mechanism,” he said in a prepared statement. “Foundry’s job is to design the trial that proves it — matching this drug to the patients whose biology demands it. That is how we compress time.” Pathos says all of the drugs in its pipeline have been identified, evaluated, or accelerated by Foundry. The most advanced program is P-500, a brain-penetrating small molecule inhibitor of an enzyme called PRMT5.
Licensed from Prelude Therapeutics
in 2024, this drug has since reached mid-stage clinical development for advanced solid tumors, including high-grade glioma and uveal melanoma.
Pocenbrodib, a small molecule inhibitor of a protein called CBP/p300, was initially developed by Forma Therapeutics, which was acquired by Novo Nordisk in 2022. Pathos licensed rights to this drug in 2023; it is currently in early clinical development for prostate and breast cancers as well as multiple myeloma.
In May, Pathos
acquired a majority stake in DeuterOncology, a Belgium-based biotech in early clinical development with DO-2, a third-generation MET inhibitor in early clinical development for MET-altered non-small cell lung cancer. DO-2 was identified by Foundry.
Illustration: Getty Images