从酿成悲剧到启迪变革：这个分子如何用60年时间，催生了一类全新的抗癌药物 | Bilingual

来源：市场资讯

（来源：药明康德 药明康德）

编者按：沙利度胺曾是酿成现代药物史上惨痛悲剧的分子，却因科学家对其机制的深入探索，逐渐成为分子胶药物研究的重要起点之一，这背后也折射出分子胶降解剂从偶然发现走向理性设计的发展趋势。在这一趋势的推动下，该领域近期迎来了实质性突破：首款cereblon E3连接酶调节剂（CELMoD）药物的获批，标志着该技术正在加速迈向临床和商业化转化；与此同时，分子胶降解剂的应用潜力也从血液肿瘤逐步扩展至实体瘤、炎症及免疫疾病，展现出多元化的发展态势。

尽管前景广阔，但分子胶降解剂的发现、开发和生产仍面临诸多挑战。充分释放其治疗潜力，离不开创新生态系统的协同努力。药明康德生物学业务平台（WuXi Biology）围绕分子胶药物发现建立了系统性且一体化的研究体系，可为合作伙伴提供从苗头化合物识别到早期验证与优化的综合赋能能力。DNA编码化合物库（DEL）平台构建了覆盖百万至数十亿级的小分子与多肽化合物库，并开发升级专属“分子胶”靶向库，以更广的化学空间和更深的靶点挖掘能力，精准定位活性苗头化合物。

在苗头化合物验证及优化阶段，WuXi Biology构建了“Direct-to-Biology（D2B）”快速迭代平台，将药物化学、计算机辅助药物设计（CADD）、高通量合成和生物学评价紧密衔接，实现设计—合成—测试—分析（DMTA）的高效循环。借助纳摩尔级自动化合成与384或1536孔板高通量检测体系，可在约2–3周内完成一轮超过上千个化合物的结构—活性关系（SAR）探索，加快分子胶从初始苗头化合物到先导化合物的优化过程。通过一体化能力体系，WuXi Biology能够为分子胶药物研发提供高效率、易扩展的一站式解决方案，帮助合作伙伴加速探索传统上被认为“不可成药”的靶点并推动创新疗法的发现与转化。

20世纪60年代初，多个国家陆续报告新生儿出现严重肢体畸形，其中包括海豹肢畸形（phocomelia）。调查随后发现，大量相关病例与母亲在孕期服用沙利度胺存在关联。此后，沙利度胺在多个国家和地区陆续撤市。

但沙利度胺并未就此从医学视野中消失。1964年，以色列医生Jacob Sheskin在治疗麻风结节性红斑患者时，意外发现沙利度胺能够明显改善症状，相关研究于1965年发表；1999年，一项2期研究进一步显示，沙利度胺对难治性多发性骨髓瘤具有抗肿瘤活性。

同一个分子，一边造成了现代药物史上的惨痛悲剧，另一边却在炎症和肿瘤领域显现出治疗曙光。研究者由此推测，沙利度胺表面上的“矛盾性”背后，可能藏着某种尚未揭示的调控机制。而要揭开这层神秘面纱，首先要回答一个核心问题：沙利度胺的分子靶点是什么？

▲沙利度胺的分子结构（图片来源：PubChem）

偶然发现的分子胶降解剂——沙利度胺

这个问题悬而未决了近半个世纪，直到2010年发表在Science上的一项研究才给出了答案：沙利度胺的主要直接靶点蛋白之一是cereblon（CRBN）。CRBN是由CUL4、RBX1和DDB1等组成的CRL4 E3泛素连接酶复合物中的底物受体，负责识别特定底物，并将其招募至E3泛素连接酶复合物，随后促进底物泛素化，并使其经蛋白酶体降解。简单来说，沙利度胺及其类似物通过结合CRBN，改变E3泛素连接酶的底物选择性。

靶点已找到，新的问题也随之而来：CRBN能识别哪些新的底物？

这一谜底在4年后被揭开。两篇同样发表于Science的研究发现，来那度胺（沙利度胺的衍生物）能够诱导CRBN招募并降解两种转录因子：IKZF1（Ikaros）和IKZF3（Aiolos）。这两种蛋白是多发性骨髓瘤细胞赖以生存的关键因子，一旦被降解，肿瘤细胞便难以维持生存。这一发现揭示了来那度胺等免疫调节药物抗多发性骨髓瘤作用的重要分子机制。

而这一机制，正是后来被称为分子胶降解剂的经典范式：像胶水一样，将E3泛素连接酶和靶蛋白连接在一起，让靶蛋白被泛素化，进而降解。

▲分子胶降解剂诱导靶蛋白泛素化示意图（图片来源：参考文献[1]）

从偶然发现到理性设计

尽管分子胶降解剂已在研究中展现出潜力，但历史上大多数分子胶都是偶然发现的，这种模式难以满足临床转化对靶向范围、降解效率和成药性的更高要求。

幸运的是，结构生物学研究的进展带来了转机。

CRBN E3泛素连接酶复合物及其与分子胶形成的复合物晶体结构相继被阐明，让科学家从原子层面认识了这类分子胶的工作原理。许多经典CRBN类分子胶降解剂保留共同的关键药效团——戊二酰亚胺环，正是这个结构单元负责嵌入CRBN的疏水口袋中，并通过疏水接触和氢键网络稳定结合。而药物分子的其他部分则参与塑造新的蛋白质-蛋白质相互作用界面，是决定新底物招募范围和效率的关键因素之一。

在此思路下，科学家们开始尝试进行理性设计，有针对性地改造分子，以期实现更充分、更持久的蛋白质降解。围绕这一目标，多种前沿设计策略不断涌现：例如在保留戊二酰亚胺环的基础上，通过结构优化开发新型分子胶；也可对非降解性靶蛋白配体进行化学修饰，引入“共价手柄”，探索将其转化为共价分子胶降解剂等。

CELMoD正是这一理性设计浪潮下的代表性成果。

CELMoD是在沙利度胺这类免疫调节药物（IMiD）骨架基础上，通过系统性结构优化而设计出的新型分子胶降解剂。两类分子均含有戊二酰亚胺环这一共同的CRBN结合“抓手”，但在骨架其他结构模块上存在差异。与传统IMiD相比，以iberdomide等为代表的CELMoD引入额外的结构延伸，可增加与CRBN的相互作用，并更有效地诱导有利于底物招募的构象变化。因此，CELMoD与传统IMiD虽然共享CRBN依赖性机制，但在结合亲和力、底物降解效率及细胞效应等方面呈现差异。

以近期获FDA加速批准的首款CELMoD药物——百时美施贵宝（Bristol Myers Squibb）的Zenbexus（iberdomide）为例。Iberdomide含有额外的苯基和吗啉基团，可增加与CRBN及相关底物的相互作用，其对CRBN的结合亲和力较来那度胺和泊马度胺显著提高，并且对来那度胺和泊马度胺耐药细胞依然具有抑制活性。

这些努力都指向同一目标，即让经验驱动的分子胶降解剂发现逐步迈向可预测、可拓展的系统化开发。

新型分子胶的“星辰大海”

如果说Zenbexus的获批标志着CELMoD这类分子胶降解剂从理性设计迈入临床应用与商业化阶段的关键跨越，那么不断拓展的疾病版图、靶点空间与机制内涵，则铺陈出分子胶更为广阔的星辰大海。

一方面，分子胶降解剂持续在血液肿瘤领域获得积极结果。百时美施贵宝另一款在研CELMoD药物mezigdomide联合卡非佐米和地塞米松的新药申请（NDA）目前正在接受FDA审评，其PDUFA目标日期为2027年5月13日。

另一方面，分子胶降解剂也开始拓展应用至实体瘤、炎症性疾病以及免疫疾病等领域。实体瘤领域，SEED Therapeutics在研RBM39靶向分子胶降解剂ST-01156已进入1/1b期临床研究，正在晚期实体瘤患者中进行评估；Neomorph在研分子胶降解剂NEO-811正在局部晚期或转移性、不可切除透明细胞肾细胞癌患者中开展1/2期临床试验。炎症性疾病领域，Monte Rosa Therapeutics靶向NEK7的分子胶降解剂MRT-8102，正在探索用于治疗由NLRP3炎症小体及IL-1、IL-6通路驱动的炎症相关疾病，并公布了积极的1期临床中期数据；免疫疾病领域，Monte Rosa Therapeutics靶向VAV1的分子胶降解剂MRT-6160，可将外周血T细胞中的VAV1水平降低超过90%。

值得一提的是，分子胶的机制并不局限于单一“降解”。经典的分子胶降解剂通过招募E3泛素连接酶来诱导靶蛋白降解，但分子胶还可以不依赖泛素-蛋白酶体系统发挥作用。这类被称为“非降解型分子胶”（non-degrading molecular glues）的分子，通过稳定瞬时的蛋白质间相互作用（PPI）来调控蛋白功能，且不诱导靶蛋白降解，为药物开发提供了一种颇具前景的策略。如非降解型分子胶CLEO4-88，可促进CTLH E3泛素连接酶复合物的底物受体GID4与过氧化物酶体硫解酶ACAA1之间的相互作用，并抑制ACAA1的酶活性。

Revolution Medicines近日获得FDA批准的Rasonque（daraxonrasib）也体现了非降解型分子胶的药物开发策略，它首先与亲环素A（CypA）结合，形成的二元复合体进一步结合活性状态的RAS蛋白，形成由RAS、药物分子和CypA组成的三元复合体，从而阻断RAS与下游效应蛋白的相互作用。

从这些进展和趋势可以看出，分子胶领域的发展正在迈向一种更广义的新药设计模式。随着结构生物学与化学生物学的深度融合，可预测的精准设计也将继续拓宽分子胶药物的应用边界，为疾病治疗带来更多可能。

From Tragedy to Transformation: How One Molecule Helped Give Rise to a New Class of Anticancer Drugs Over 60 Years

Though once associated with a tragic episode in pharmaceutical history, thalidomide was later repurposed through decades of mechanistic investigation, ultimately becoming an important starting point for molecular glue drug research. Its journey reflects the broader evolution of molecular glue degraders—from serendipitous discovery to increasingly rational design.

Recent developments underscore how far the field has progressed. The approval of the first cereblon E3 ligase modulator (CELMoD) agent marks an important milestone in translating this therapeutic concept into clinical practice and commercialization. At the same time, the investigational landscape for molecular glue degraders is expanding beyond hematologic malignancies into solid tumors, inflammatory conditions, and immune-mediated diseases, reflecting a rapidly diversifying pipeline.

Despite promising potential, the discovery, development, and manufacturing of molecular glue degraders still face numerous challenges. Fully unlocking their therapeutic potential requires collaborative efforts across an innovative ecosystem. WuXi Biology has established a comprehensive and integrated platform for molecular glue drug discovery, providing partners with end-to-end support from hit identification to early validation and optimization.

The DNA-encoded library (DEL) platform has established small-molecule and peptide libraries on the scale of millions to billions of compounds, and has developed and continuously refined a dedicated “molecular glue”–focused library. By accessing a broader chemical space and achieving in-depth target-specific screening capabilities, it enables the precise identification of active hit compounds.

To accelerate hit confirmation and lead optimization, WuXi Biology has developed a Direct-to-Biology (D2B) rapid iteration platform that tightly connects medicinal chemistry, computer-aided drug design (CADD), high-throughput synthesis, and biological evaluation within a seamless design–make–test–analyze (DMTA) workflow. Supported by nanomole-scale automated synthesis and 384- or 1536-well high-throughput assay systems, the platform enables a full cycle of structure–activity relationship (SAR) exploration across thousands of compounds in approximately two to three weeks, significantly accelerating the progression of molecular glue hits toward lead compounds. Together, these integrated capabilities provide an efficient and scalable one-stop solution for molecular glue discovery, enabling partners to explore previously intractable targets and advance innovative therapeutic programs.

Image source: 123RF

In the early 1960s, several countries reported severe limb malformations in newborns, including phocomelia. Subsequent investigations linked many of these cases to maternal thalidomide use during pregnancy, prompting the drug to be withdrawn from markets across multiple countries and regions.

Yet thalidomide’s story did not end there. In 1964, Israeli physician Jacob Sheskin serendipitously observed that thalidomide markedly alleviated symptoms of erythema nodosum leprosum (ENL), with his findings published the following year. Decades later, in 1999, a Phase 2 study provided further evidence that thalidomide had antitumor activity in refractory multiple myeloma.

The same molecule therefore came to embody a striking paradox: it had been associated with a devastating chapter in pharmaceutical history, yet it also demonstrated therapeutic potential in inflammatory disease and cancer. Researchers suspected that this apparent contradiction reflected an as-yet-unknown regulatory mechanism. To understand it, they first needed to answer a fundamental question: What is thalidomide’s molecular target?

▲The molecular structure of thalidomide（Image source: PubChem）

Thalidomide: A Serendipitously Discovered Molecular Glue Degrader

The answer remained elusive for nearly half a century. Then, in 2010, a study published in Science identified cereblon (CRBN) as one of the major direct protein targets of thalidomide. CRBN functions as a substrate receptor within the CRL4 E3 ubiquitin ligase complex, working with the CUL4-RBX1-DDB1 core machinery to recognize specific substrates and recruit them for ubiquitination and subsequent proteasomal degradation. Put simply, thalidomide and its analogs bind to CRBN and alter the substrate specificity of the E3 ligase complex.

Once the target had been identified, another question immediately followed: Which proteins were now being recruited by CRBN?

Four years later, two studies published in Science provided an important part of the answer. They showed that lenalidomide, a thalidomide derivative, induces CRBN to recruit and degrade two transcription factors, IKZF1 (Ikaros) and IKZF3 (Aiolos). Both proteins are important for the survival of multiple myeloma cells, and their degradation compromises tumor cell viability. These findings revealed a key molecular mechanism underlying the anti-myeloma activity of immunomodulatory drugs such as lenalidomide.

This mechanism became a classic paradigm for molecular glue degraders: rather than simply occupying and inhibiting a target, a small molecule promotes a new interaction between an E3 ubiquitin ligase and a target protein, leading to target ubiquitination and subsequent degradation.

▲Schematic representation of molecular glue degrader-induced protein-of-interest ubiquitination (Image source: reference [1]）

From Serendipity to Rational Design

Although molecular glue degraders have demonstrated considerable therapeutic potential, many of the earliest examples were discovered serendipitously. Reliance on chance discovery, however, makes it difficult to systematically expand the target space, optimize degradation efficiency, and improve drug-like properties to meet the demands of clinical development.

Structural biology provided an important turning point.

The elucidation of crystal structures of the CRBN E3 ubiquitin ligase complex and its complexes with molecular glues gave researchers an atomic-level view of how these compounds function. Many classical CRBN-based molecular glue degraders share a glutarimide ring as a core pharmacophore. This structural motif fits into a hydrophobic pocket in CRBN, where it is stabilized through hydrophobic interactions and a network of hydrogen bonds.

Other portions of the molecule help shape a new protein–protein interaction surface. These structural features play a critical role in determining which neosubstrates can be recruited and how efficiently degradation occurs.

Armed with these insights, researchers began moving from serendipitous discovery toward rational design, deliberately modifying molecular structures to achieve more efficient and sustained protein degradation. A range of strategies has emerged, including optimizing molecular glue scaffolds while retaining the glutarimide core and chemically modifying non-degrading target ligands with a “covalent handle” to explore their conversion into covalent molecular glue degraders.

CELMoDs are a notable outcome of this rational design effort. CELMoDs are a new class of molecular glue degraders developed through systematic optimization of immunomodulatory drug (IMiD) scaffolds such as those derived from thalidomide. CELMoDs and classical IMiDs share the glutarimide ring as a common CRBN-binding motif but differ in other structural features.

Compared with traditional IMiDs, representative CELMoDs incorporate additional structural extensions that can strengthen interactions with CRBN and promote conformations that favor neosubstrate recruitment. Although both classes act through CRBN-dependent mechanisms, they can differ substantially in binding affinity, substrate degradation efficiency, and downstream cellular effects.

A recent example is Bristol Myers Squibb’s Zenbexus (iberdomide), the first FDA-approved CELMoD agent. Iberdomide contains additional phenyl and morpholine groups that enhance interactions with CRBN and recruited substrates. It binds CRBN with substantially greater affinity than lenalidomide and pomalidomide and retains antitumor activity in multiple myeloma cells resistant to those earlier IMiDs.

Together, these advances point toward a common goal: transforming molecular glue degrader discovery from an empirical process into a more predictable, scalable, and systematic discipline.

Expanding the Frontier of Molecular Glue Therapeutics

The approval of Zenbexus represents an important step in bringing rationally designed CELMoDs into clinical use. Yet the broader potential of molecular glue therapeutics may lie in the continuing expansion of their disease indications, target space, and underlying mechanisms.

In hematologic malignancies, molecular glue degraders continue to generate encouraging clinical progress. Bristol Myers Squibb’s investigational CELMoD mezigdomide, in combination with carfilzomib and dexamethasone, is currently under FDA review for relapsed or refractory multiple myeloma, with a Prescription Drug User Fee Act (PDUFA) target date of May 13, 2027.

At the same time, development is expanding into solid tumors, inflammatory diseases, and immune-mediated disorders. SEED Therapeutics’ RBM39-targeting molecular glue degrader ST-01156 has entered a Phase 1/1b trial in patients with advanced solid tumors. Neomorph’s NEO-811, meanwhile, is being evaluated in a Phase 1/2 trial in patients with locally advanced or metastatic, unresectable clear-cell renal cell carcinoma.

In inflammatory diseases, Monte Rosa Therapeutics’ NEK7-targeting molecular glue degrader MRT-8102 is being investigated for conditions driven by the NLRP3 inflammasome and IL-1/IL-6 signaling pathways, with positive interim Phase 1 results already reported. In immune-mediated diseases, the company’s VAV1-targeting degrader MRT-6160 has achieved greater than 90% degradation of VAV1 in peripheral blood T cells.

Importantly, the potential of molecular glues extends beyond protein degradation itself. Classical molecular glue degraders recruit E3 ubiquitin ligases to promote proteasomal degradation of a target protein, but other molecular glues can modulate protein function without inducing degradation.

These so-called non-degrading molecular glues stabilize otherwise transient protein–protein interactions, creating new functional states without eliminating the target protein. This expands the molecular glue concept beyond targeted degradation and opens an additional therapeutic strategy for modulating previously difficult-to-drug proteins.

CLEO4-88 provides one example. The compound promotes an interaction between GID4, a substrate receptor of the CTLH E3 ligase complex, and the peroxisomal thiolase ACAA1. Rather than triggering ACAA1 degradation, formation of this induced complex suppresses ACAA1 enzymatic activity.

Revolution Medicines’ recently FDA-approved Rasonque (daraxonrasib) also exemplifies a non-degrading molecular glue strategy. Rasonque first binds cyclophilin A (CypA), forming a binary complex that subsequently engages RAS proteins in their active state. This interaction creates a ternary complex comprising RAS, daraxonrasib, and CypA, thereby blocking interactions between RAS and downstream effector proteins.

Taken together, these advances suggest that molecular glues are evolving into a broader drug design paradigm—one that extends beyond a single mechanism or therapeutic area. As structural biology and chemical biology become increasingly integrated, more predictable and precise design strategies could further expand the range of proteins and biological functions accessible to molecular glue therapeutics, opening new possibilities for drug discovery and disease treatment.

Key Takeaways:

• Molecular glue drug discovery is shifting from serendipity toward rational design. Decades of research into thalidomide and CRBN have revealed how small molecules can reshape protein interactions and selectively recruit new substrates, providing a foundation for more predictable molecular glue design.

• The first FDA-approved CELMoD agent marks an important milestone for the field. The approval of Zenbexus (iberdomide) demonstrates how rationally designed CRBN modulators can translate molecular glue biology into clinically validated therapies, while additional CELMoDs continue to advance through development.

• The molecular glue concept is expanding beyond both hematologic cancers and protein degradation. Emerging programs are targeting solid tumors, inflammatory and immune-mediated diseases, while non-degrading molecular glues such as Rasonque (daraxonrasib) illustrate how induced protein interactions can modulate disease-relevant proteins without triggering their degradation.

• WuXi Biology provides an integrated platform to accelerate molecular glue discovery from hit identification through early optimization. Its capabilities combine molecular glue-focused DNA-encoded libraries, high-throughput screening, Direct-to-Biology (D2B), CADD, rapid synthesis, and biological evaluation within an integrated DMTA workflow, helping partners explore broader chemical space and advance molecular glue hits toward lead compounds.

参考资料：

[1] Tan X, Huang Z, Pei H, et al. Molecular glue-mediated targeted protein degradation: A novel strategy in small-molecule drug development. iScience. 2024 Aug 21;27(9):110712. doi: 10.1016/j.isci.2024.110712.

[2]Sheskin J. Thalidomide in the treatment of lepra reactions. Clinical Pharmacology & Therapeutics. 1965;6:303-306.

[3]Singhal S, Mehta J, Desikan R, et al. Antitumor activity of thalidomide in refractory multiple myeloma. New England Journal of Medicine. 1999;341(21):1565-1571

[4]Ito T, Ando H, Suzuki T, et al. Science. 2010;327(5971):1345-1350.

[5]Krönke J, Udeshi ND, Narla A, et al. Lenalidomide causes selective degradation of IKZF1 and IKZF3 in multiple myeloma cells. Science. 2014;343(6168):301-305.

[6]Lu G, Middleton RE, Sun H, et al. The myeloma drug lenalidomide promotes the cereblon-dependent destruction of Ikaros proteins. Science. 2014;343(6168):305-309.

[7]Oleinikovas V, Gainza P, Ryckmans T, et al. From Thalidomide to Rational Molecular Glue Design for Targeted Protein Degradation. Annu Rev Pharmacol Toxicol. 2024 Jan 23;64:291-312. doi: 10.1146/annurev-pharmtox-022123-104147.

[8]Fischer ES, Böhm K, Lydeard JR, et al. Structure of the DDB1-CRBN E3 ubiquitin ligase in complex with thalidomide. Nature. 2014;512(7512):49-53.

[9]Chamberlain PP, Lopez-Girona A, Miller K, et al. Structure of the human Cereblon-DDB1-lenalidomide complex reveals basis for responsiveness to thalidomide analogs. Nature Structural & Molecular Biology. 2014;21(9):803-809.

[10]Lu W, Zhang J, et al. Exploring non-degrading molecular glues for protein–protein interactions. Trends in Biochemical Sciences. 2025.

[11]Darling S, Bajrami I, West SC. From serendipity to strategy: molecular glue degraders in cancer therapeutics. Crit Rev Biochem Mol Biol. 2025 Aug-Dec;60(4-6):200-230. doi: 10.1080/10409238.2025.2564068. Epub 2025 Oct 1. PMID: 41032400.

[12]Chana, C.K., Ben Makhlouf, I., Kim, J. et al. The molecular glue CLEO4-88 inhibits the ACAA1 thiolase by induced binding to GID4. Nat Chem Biol (2026). https://doi.org/10.1038/s41589-026-02183-4.

[13]Monte Rosa Therapeutics Announces Positive Interim Phase 1 Data of MRT-8102 Demonstrating Profound CRP Reductions in Elevated CVD-risk Subjects. Retrieved January 7, 2026 from Monte Rosa Therapeutics Announces Positive Interim Phase 1 Data of MRT-8102 Demonstrating Profound CRP Reductions in Elevated CVD-risk Subjects. Retrieved January 7, 2026, from https://ir.monterosatx.com/news-releases/news-release-details/monte-rosa-therapeutics-announces-positive-interim-phase-1-data

[14]吴雄，焦宇，吕仕铭，等. CELMoDs类IKZF1/3 降解剂的研究进展. 药学进展. 2023, ·47 (1): 66-74.

[15]田艾琳 ,刘晓, 孙睿婧, 等. 降解型分子胶研究进展：靶向锌指类转录因子IKZF1-IKZF3. 亚太医学. 2026-06-17.

[16]孙杰, 顾明希, 臧兰兰, 等. 分子胶的开发: 从偶然发现到合理设计. 药学学报. 2025, 60(11): 3345−3354