{"id":2600,"date":"2026-08-18T12:00:00","date_gmt":"2026-08-18T16:00:00","guid":{"rendered":"https:\/\/www.insilens.com\/?p=2600"},"modified":"2026-08-18T19:31:34","modified_gmt":"2026-08-18T23:31:34","slug":"fmi-and-epfl-map-the-druggable-crbn-interactome","status":"publish","type":"post","link":"https:\/\/www.insilens.com\/?p=2600","title":{"rendered":"FMI and EPFL Map the Druggable CRBN Interactome"},"content":{"rendered":"

Institutions:<\/strong> Friedrich Miescher Institute & EPFL · Event Type:<\/strong> Peer-Reviewed Publication (Nature Biotechnology) · Technology:<\/strong> Proteome-Wide Molecular Glue Interactome Mapping · Target:<\/strong> Cereblon (CRBN) · Publication Date:<\/strong> August 13, 2026<\/p>\n

\"FMI<\/p>\n

Summary<\/h4>\n

Researchers from the Friedrich Miescher Institute, EPFL and collaborators mapped a latent cereblon molecular-glue interaction space by combining high-throughput protein-complementation screening with AI-based protein-surface mimicry. GluePCA identified approximately 210 zinc-finger domains that bind CRBN–pomalidomide, and a proteome-scale computational\/experimental workflow validated 43 previously unreported CRBN binders alongside six known proteins. The work expands chemically plausible starting points for molecular-glue discovery but does not demonstrate that most binders can be degraded, modulated safely or converted into drug candidates.<\/p>\n

What Happened<\/h4>\n

The investigators built a library spanning thousands of human zinc-finger reporters and measured drug-induced CRBN association in parallel. They then used MaSIF-mimicry, a geometric deep-learning method, to query protein surfaces across AlphaFold-derived structural models and prospectively selected non-zinc-finger candidates for experimental testing.<\/p>\n

The central conceptual finding is that CRBN’s apparent substrate promiscuity can be organized as recurring surface patterns rather than only sequence motifs. Binding-focused discovery recovered known neo-substrates and exposed a larger latent interactome. Data, structural models and analysis code were deposited publicly, improving reproducibility and follow-on discovery access.<\/p>\n

Deep Analysis<\/h4>\n

Molecular glues create or stabilize a protein-protein interface between an E3 ligase and a neo-substrate, leading to ubiquitination and proteasomal degradation when geometry, lysine accessibility and cellular context are favorable. CRBN is clinically validated by immunomodulatory drugs and CELMoDs, but new glue discovery remains constrained by the difficulty of predicting productive ternary complexes.<\/p>\n

This study separates target recruitment from degradation. That distinction is important: a protein can bind CRBN–compound yet escape ubiquitination because the ligase cannot access suitable lysines, the complex orientation is unproductive, the protein resides in the wrong compartment, or resynthesis offsets degradation. The 43 new binders are therefore starting points, not validated therapeutic targets.<\/p>\n

The hybrid design addresses a search-space problem. High-throughput interaction assays provide empirical anchors, while surface-based AI extends beyond obvious sequence families. Translation still requires medicinal chemistry to tune ternary-complex cooperativity, permeability, pharmacokinetics and selectivity; quantitative proteomics to establish degradation; and disease models to connect target loss with a therapeutic window.<\/p>\n

Competitive Displacement<\/h4>\n

The platform could improve early target enumeration relative to serendipitous phenotypic glue discovery and sequence-only degron rules. It may complement — not replace — chemoproteomics, structure-guided ternary-complex design, DNA-encoded libraries, covalent screening and phenotypic approaches. Its advantage will depend on hit conversion rates and whether the latent binding map yields targets that were inaccessible to existing platforms. The work also creates potential partnering and startup value because the same workflow may transfer to other E3 ligases with structural information.<\/p>\n

Company and Product Background<\/h4>\n

The study is an academic platform effort led by protein-degradation and computational-design groups rather than a clinical company program. CRBN is a substrate receptor within the CUL4–RBX1–DDB1 E3 ubiquitin ligase complex. Pomalidomide and related glutarimides bind CRBN and redirect it toward non-native proteins, including transcription factors relevant to multiple myeloma. Targeted protein degradation can remove scaffolding and transcriptional functions that conventional occupancy-driven inhibitors may not control, balanced by risks of off-target neo-substrate loss, teratogenicity, tissue exposure and resistance through pathway mutation.<\/p>\n

Signal Extraction<\/h4>\n\n\n\n\n\n\n\n\n
Signal<\/th>\nVerified Evidence<\/th>\nCurrent Limit<\/th>\n<\/tr>\n<\/thead>\n
Zinc-finger map<\/td>\n~210 domains bound CRBN–pomalidomide in GluePCA<\/td>\nBinding does not equal degradation<\/td>\n<\/tr>\n
Proteome expansion<\/td>\n43 novel and six known binders validated<\/td>\nTherapeutic relevance not established<\/td>\n<\/tr>\n
Computational method<\/td>\nAI surface mimicry searched structurally distant proteins<\/td>\nProspective enrichment rate requires broader benchmarking<\/td>\n<\/tr>\n
Translational state<\/td>\nPublic data, models and code enable follow-up<\/td>\nNo optimized candidate, PK\/PD or in-vivo efficacy<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Reading the Signal<\/h4>\n

Bull case:<\/strong> The latent interactome materially expands rational molecular-glue target discovery, supported by proteome-scale search, prospective experimental validation and recovery of known biology. This would be upgraded by high conversion of binders into selective degraders with disease-relevant in-vivo activity.<\/p>\n

Bear case:<\/strong> The map is biologically informative but may have limited drug-discovery yield because ternary-complex productivity and compound optimization remain the dominant bottlenecks, given the many post-binding requirements and absence of therapeutic candidates. This would be strengthened by poor portability across target classes and ligases.<\/p>\n

InSilens Take<\/h4>\n

This is a 4\/5 technology signal with uncertain direction. The study offers a disciplined route to map CRBN-recruitable surfaces and can reduce early discovery randomness, but it has not yet shortened the full translational chain from binding to degradation, exposure, safety and efficacy. The key falsifier is low binder-to-degrader conversion. The key upgrade is reproducible generation of selective, drug-like degraders against previously inaccessible disease targets.<\/p>\n

Signal Assessment<\/h4>\n

Importance:<\/strong> 4\/5 — platform-level expansion of a clinically validated E3 ligase. Direction:<\/strong> Uncertain — discovery breadth is clear; therapeutic conversion is untested. Confidence in facts:<\/strong> High — peer-reviewed paper with deposited data and code. Confidence in interpretation:<\/strong> Moderate — drug-discovery yield and portability remain unknown.<\/p>\n","protected":false},"excerpt":{"rendered":"

Researchers from the Friedrich Miescher Institute, EPFL and collaborators mapped a latent cereblon molecular-glue interaction space by combining high-throughput protein-complementation screening with AI-based protein-surface mimicry. GluePCA identified approximately 210 zinc-finger domains that bind CRBN\u2013pomalidomide,…<\/p>\n","protected":false},"author":1,"featured_media":2607,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[4],"tags":[379,378,380],"class_list":["post-2600","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technology-modalities","tag-epfl","tag-friedrich-miescher-institute","tag-molecular-glue"],"blocksy_meta":[],"_links":{"self":[{"href":"https:\/\/www.insilens.com\/index.php?rest_route=\/wp\/v2\/posts\/2600","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.insilens.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.insilens.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.insilens.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.insilens.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=2600"}],"version-history":[{"count":1,"href":"https:\/\/www.insilens.com\/index.php?rest_route=\/wp\/v2\/posts\/2600\/revisions"}],"predecessor-version":[{"id":2614,"href":"https:\/\/www.insilens.com\/index.php?rest_route=\/wp\/v2\/posts\/2600\/revisions\/2614"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.insilens.com\/index.php?rest_route=\/wp\/v2\/media\/2607"}],"wp:attachment":[{"href":"https:\/\/www.insilens.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=2600"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.insilens.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=2600"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.insilens.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=2600"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}