Crenezumab

Roche’s biggest bet in the Alzheimer’s disease space may be its phase 3-stage amyloid asset, but the Swiss pharma is keeping its options open when it comes to targeting tau. Tau is one of two proteins that become deformed and clumped in Alzheimer’s, with the other, amyloid beta, already successfully targeted by Biogen’s Leqembi and Eli Lilly’s Kisunla. At the Alzheimer’s Association International Conference in London this week, Biogen shared phase 2 data from diranersen, an antisense oligonucleotide that blocks the production of tau by binding to its precursor mRNA. Speaking to Fierce on the sidelines of the conference, Roche’s Head of Early Development in Neuroscience and Rare Diseases, Luka Kulic, M.D., Ph.D., namechecked Biogen’s data as “interesting,” while also raising “ some questions .” Roche once had an anti-tau antibody in the works, courtesy of a $120 million upfront pact with UCB for bepranemab, but the Swiss pharma walked away in 2024 , just days before the therapy failed a phase 2 test. While Roche doesn’t have a tau candidate in the clinic at the moment, the company “of course remains very interested” in tackling tau, Kulic said. “That's a key major player in Alzheimer's disease.” Roche didn’t arrive at AAIC empty-handed. The drugmaker presented interim data from an open-label extension (OLE) of its phase 1b/2a Brainshuttle AD study of its anti-amyloid-beta antibody trontinemab. The readout included patients for whom a PET scan had shown they still had beta-amyloid plaques in the brain. These patients received monthly doses of trontinemab until the PET scan no longer showed any plaques, at which point they switched to maintenance therapy. Meanwhile, patients who began with a negative scan were put on a maintenance regimen straight away. “What we observed was … sustained—and even the further continued—reduction in amyloid,” Kulic said. “Even in the group that was amyloid PET negative at OLE baseline, there is some movement of the amyloid PET signal.” “So we are able, from a very low signal to [go] even lower,” he added. The “most exciting part” was noting a concurrent reduction in certain plasma biomarkers, such as the inflammation marker glial fibrillary acidic protein, he added. “It continues to further decrease, so to further trend towards normalization, which was an exciting finding for us,” said Kulic, who acknowledged that the small size of the study means there remain caveats at this stage. It was also good news on the safety front, with just two cases of amyloid-related imaging abnormality-edema (ARIA-E)—a swelling of the brain associated with anti-amyloid beta antibodies—among the 126 patients. But with two amyloid drugs for Alzheimer’s already on the market in the form of Leqembi and Kisunla, does Roche think trontinemab has what it takes to stand out? “The profile that we have seen in the Brainshuttle AD study, with the really very low incidence of ARIA-E and the very rapid and profound amyloid lowering, gives us confidence that we might have a differentiated profile,” Kulic said. The positive signs from the data so far are a welcome validation for Roche’s decision to stick with the amyloid hypothesis, despite a couple of high-profile setbacks. Back in 2022, the company saw gantenerumab and crenezumab fail to prove they could reduce the rate of cognitive decline in various phase 3 studies, leading Roche to abandon both programs . “What we have learned is that … we need to really get rid of a substantial amount of amyloid to see clinical efficacy,” Kulic said. The real proof of whether trontinemab has resolved this issue will be a pair of phase 3 studies, which began recruiting patients in September 2025 and are expected to read out in 2028. Those trials, called Trontier 1 and 2, are enrolling patients with early symptomatic Alzheimer's disease. Still, Roche is set to launch a separate late-stage study in the next couple of months focusing on patients who aren’t yet cognitively impaired despite showing evidence of amyloid pathology. Further back in the pipeline is a gamma secretase modulator that’s designed as a potential oral treatment for the neurodegenerative condition, with phase 2 data set to read out next year. There’s also a phase 1-stage drug aimed at down-regulating the production of apolipoprotein E, which Kulic described as a “major genetic player in Alzheimer's disease.” One thing that sets trontinemab apart is that it’s the most advanced anti-beta antibody to use brain shuttle technology. Specifically, the therapy harnesses transferrin receptor (TFR) to drive the drug into the brain. The TFR pathway is also being explored in Alzheimer’s prospects from the likes of Novartis , and Kulic acknowledged that transfferin “obviously works very well for amyloid-lowering antibodies.” However, Roche is also on the lookout for other routes to shuttle drugs to the brain, a strategy that culminated in a $2 billion biobucks pact with Manifold Bio last year. “One of the ideas behind the Manifold deal is to identify novel receptors,” he said. “For different targets, different indications, different modalities, there might be other receptors that might be more suitable.” “So we remain open to exploring different ways to overcome the blood-brain barrier,” Kulic added.

十六年、16亿美元、2300余次引用、146项临床试验全部归零。 这几组震撼的核心数据，指向同一场席卷全球的学术闹剧：2006年《Nature》一张蛋白条带图，将Aβ56淀粉样蛋白寡聚体推上神坛，彻底带偏阿尔茨海默病研究整整16年。 2006年假说封神、2022年丑闻曝光、2024年官方确认撤稿。十六年间，全球学界、药企、科研人员集体为一场人造的“完美科研突破”买单。 2022年《Science》重磅调查彻底揭穿真相：那张相关性高达0.98的经典条带图，存在严重人为篡改，条带拼接、数据筛选美化，所谓的精准致病关联完全不成立。历经两年拉扯，2024年论文通讯作者Karen Ashe承认图像篡改、同意撤稿，但核心一作Sylvain Lesné始终沉默拒认错，至今仍任职原岗位、申领科研经费。 这场顶级学术造假如何意外曝光？ 一张假图为何能绑架全球十六年的科研资源？ 普通研究者如何避开“完美科研故事”的陷阱？ Aβ是什么？Aβ56为何能一夜封神？阿尔茨海默病研究的核心——Aβ淀粉样蛋白。我们可以将大脑神经元比作精密车间，APP蛋白是常规原料，经β、γ两种“分子剪刀”剪切后，会脱落形成Aβ蛋白碎片。 Aβ蛋白自带黏附性，零散碎片会逐渐聚集：小型聚合体为寡聚体，大量堆积则形成特征性淀粉样斑块。二十年来，淀粉样蛋白假说始终是领域主流，学界普遍认为Aβ寡聚体是损伤神经元、引发认知衰退的元凶，且有临床佐证：患者脑内Aβ含量显著偏高，清除Aβ可轻微延缓病情恶化。 但该领域一直存在无法回避的争议：部分患者清除斑块后症状无改善，不少老人脑内斑块堆积却头脑清晰。Aβ的致病地位始终存疑，而2006年这篇论文，看似完美终结了所有争议。 研究团队提出的Aβ56，是56kDa的可溶性寡聚体，介于单体Aβ与成熟斑块之间。实验显示，小鼠脑内Aβ56表达量与记忆损伤高度同步，相关系数高达0.98。在充满个体差异、实验波动的神经生物学领域，这个近乎满分的数据，搭配完整的分子—细胞—行为实验逻辑，瞬间征服了整个学界。 彼时整个领域急需明确研究方向，这套“零瑕疵”的成果，成了所有人追捧的科研锚点。无人质疑数据过于完美、无人深究条带规整得违背实验常识，这篇论文迅速被写入综述、基金标书、药企研发规划，十六年的科研赛道被彻底锁死。 2300次高引，16亿美金砸出一场空依托顶刊光环与权威背书，这篇造假论文成为领域标杆，累计被引超2300次，是本世纪阿尔茨海默病研究最高引的奠基文献之一，持续引领全球科研风向。 学术热度直接撬动巨额资金倾斜。美国NIH对Aβ寡聚体方向的专项资助从近乎为零，飙升至年度16亿美元，十六年累计投入16.5亿美元，占据全美该病研究总预算的半壁江山。 罗氏、礼来、强生等药企纷纷重仓该赛道，落地上百条研发管线，累计启动146项临床试验。仅礼来solanezumab、罗氏crenezumab两大核心项目，十年累计投入超80亿美元，最终全部临床失败。海量民间实验室、初创企业的投入无从统计，近半数顶尖药企因持续亏损无果，彻底退出阿尔茨海默病研发领域。 事后复盘可见，0.98的完美数据并非实验所得，而是通过筛选阳性结果、舍弃阴性数据、拼接篡改图像人工打造的成果。2300余次引用中，绝大多数是跟风叠加研究，真正独立验证、可复现的扎实成果寥寥无几。 学术自查全部失效，造假靠商业审计曝光最讽刺的是，这场持续十六年的顶级造假，从未被期刊审稿、同行评审、高校自查发现，最终的真相曝光，只是一场偶然的商业副业审计。 2021年，神经科学家Matthew Schrag受投资机构委托，核查药企阿尔茨海默病候选药物的临床数据。排查过程中，他在PubPeer平台看到针对2006年Lesné论文的零星质疑，随即开展系统性复核。 由于不属于该领域核心圈层，Schrag无利益捆绑、无思维桎梏，通过专业图像工具核查发现致命漏洞：论文Western blot条带存在大量复制泳道、跨实验组拼接、对比度刻意优化等造假行为，整套核心数据完全不具备可重复性。 2022年7月《Science》正式曝光丑闻，《Nature》才对论文挂上风险警示。拖延两年后，通讯作者认错同意撤稿，但一作全程沉默、拒不担责，至今安然在岗。一套完整的学术监督体系，败给了一场商业风险排查，暴露了科研圈严重的自查短板。 一代人的科研青春，被畸形生态耽误百亿资金的损耗肉眼可见，但一代科研人的职业生涯、固化的行业顽疾，是更难修复的深层创伤。 我认识的一名博后，六年科研生涯全部押注Aβ56靶点。他反复调整实验条件、更换试剂，反复尝试复现经典结果，最终只剩满满一硬盘的阴性数据。长期自我怀疑、实验屡屡失败、投稿持续被拒，让他延期毕业，最终彻底放弃深耕多年的科研领域，转行企业任职。 这样的研究者遍布全球。在当时的学术生态下，追随主流完美假说，就能拿课题、发高分文章、稳职称；而质疑权威、报送阴性结果，只会面临论文被拒、基金落选、职业受阻的困境。 久而久之，领域内形成集体沉默的畸形闭环，无人敢纠错、无人愿质疑，学术自净机制彻底失灵。一次次药物研发失败，也不断消耗患者与公众对科研的信任，让行业公信力持续受损。 五步极简科研自查清单，低成本规避数据风险： 这场十六年的闹剧，不止是个人造假的问题，更是体系漏洞与跟风心态的缩影。制度完善需要时间，但一线研究者可通过简单规范，避开同类科研陷阱。这里整理了五步极简科研自查清单，低成本规避数据风险： 1.原始数据多层存档：文件按规范命名多备份，保留原始实验底片，绝不直接修改原文件； 2.图像处理全程留痕：仅做基础线性调整，所有操作记录在册，杜绝拼接、复制、选择性修图； 3.实验预注册前置：实验前锁定假设与分析方案，避免事后筛选数据、迎合结论； 4.盲法交叉复核：关键实验数据、条带分析交由他人盲审，规避主观偏差； 5.阴性结果同等归档：完整留存所有实验数据，不刻意隐瞒失败结果。 长期来看，期刊需推行注册报告模式、强制上传原始图像；资助机构需设立复现专项基金，将数据可重复性纳入评审标准，从根源打破“重阳性、轻真实”的行业乱象。 科学的底色从来不是完美，而是求真与证伪。包容失败、鼓励质疑、坚守真实，才是科研最该有的样子 免责申明：本文来源网络，仅用于学术分享，文章、图片等版权归原作者享有，如涉著作权事宜请联系小编修改或删除。