On this page
Insight
Vepdegestrant's approval in May 2026, the first for a PROTAC, was a milestone for targeted protein degradation and closed a twenty-five-year gap between concept and licensed medicine. In the two years before it arrived, pharma put more than $28 billion of potential value behind the field, whose central design problem, building molecular glues beyond CRBN chemistry, is still unsolved.
The FDA approval of vepdegestrant (Veppanu; Arvinas Operations, Inc.) on May 1, 2026, gave targeted protein degradation (TPD) its first approved PROTAC and closed a twenty-five-year loop between scientific concept and licensed medicine. That gap was not a failure of nerve or regulatory caution. It was five sequential scientific barriers, each requiring the previous to be resolved before meaningful progress was possible. Understanding that history is the prerequisite for evaluating the next frontier, rationally designed molecular glue degraders, and for reading the commercial activity around it.
Major pharmaceutical companies put more than $28 billion of aggregate potential deal value, upfronts and milestones combined, behind TPD platforms and assets in the twenty-four months from June 2024 to June 2026. That figure is the sum of the disclosed transactions in the record below. The approval of vepdegestrant was anticipated years in advance; the deals reflect that anticipation, not its arrival.
The 2001 starting point
The PROTAC concept was formally established by Sakamoto, Kim, Kumagai, Mercurio, Crews, and Deshaies in Proceedings of the National Academy of Sciences in 2001. Protac-1, a phosphopeptide derived from the IκBα degron fused to a MetAP-2 ligand, recruited an E3 ubiquitin ligase to a target protein, induced ubiquitination, and directed the target to the 26S proteasome for destruction. The mechanism was correct. The molecule was a peptide with no cell permeability.
Cationic cell-penetrating peptide sequences could force membrane entry but introduced toxicity and selectivity liabilities incompatible with drug development. The field iterated on peptide scaffolds for nearly a decade without solving the exposure problem. The barrier was not regulatory or commercial; it was chemical.
The 2010 CRBN discovery
The field's genuine enabling breakthrough was not a PROTAC publication. Ito, Ando, Suzuki, and colleagues reported in Science in 2010 that cereblon (CRBN), a substrate receptor of the CRL4 E3 ubiquitin ligase complex, is the primary molecular target of thalidomide teratogenicity. Thalidomide had been marketed since 1957; its molecular target was unknown for fifty-three years.
The practical consequence was immediate. Thalidomide and its IMiD derivatives were small, cell-permeable, orally bioavailable compounds with well-defined binding conformations and established medicinal chemistry. They were E3 ligase ligands. Attaching them to target-binding warheads via chemical linkers could, in principle, generate small-molecule PROTACs with drug-like properties.
The 2012 VHL ligand and the 2015 burst
The second structural enabler came from Buckley, Van Molle, Gareiss, Ciulli, Crews, and colleagues, who reported the first small-molecule ligands for the von Hippel-Lindau (VHL) E3 ubiquitin ligase in the Journal of the American Chemical Society in 2012. Two validated E3 ligase tools, CRBN and VHL, with crystallographic structural data and tractable medicinal chemistry, now existed. In May and June 2015, dBET1 (Winter et al., Science) and catalytic VHL-recruiting PROTACs (Bondeson et al., Nature Chemical Biology) established modern PROTAC drug discovery and the sub-stoichiometric, event-driven pharmacology that gives degraders their advantage over occupancy-based inhibitors.
The twenty-five-year gap decomposes cleanly into five phases: cell permeability (2001 to 2010), E3 ligase structural tools (2010 to 2012), ternary complex optimization (2012 to 2015), IND-enabling translation (2015 to 2019), and clinical validation (2019 to 2026). Vepdegestrant was approved May 1, 2026.
Figure 1. The twenty-five-year arc, and the five sequential barriers that had to fall in order.
Small-molecule TPD relies on the induced proximity of a protein of interest (POI) to an E3 ubiquitin ligase complex via two distinct architectures. Heterobifunctional PROTACs are tripartite molecules that bridge the POI and an E3 ligase, permitting ubiquitin transfer and proteasomal degradation; their liabilities are elevated molecular weight (700 to 1200 Da) and the hook effect, competitive inhibition of ternary complex formation at high concentrations. Monovalent molecular glue degraders (MGDs) are low-molecular-weight, Rule-of-Five-compliant compounds that bind an E3 ligase substrate receptor and create a de novo interface that recruits a neo-substrate.
Of the more than 600 E3 ubiquitin ligases in the human genome, the clinical TPD pipeline is heavily concentrated around the CRL4-CRBN complex and, to a lesser extent, the CRL2-VHL complex. This concentration imposes mutational escape risks and structural vulnerabilities beyond the hook effect, including the ongoing challenge of accessing novel E3 ligase substrate receptor handles.
To bypass the proteasome's size and localization constraints, next-generation architectures route targets to the lysosome or autophagosome. Lysosome-targeting chimeras (LYTACs) engage cell-surface receptors to degrade extracellular and membrane proteins. Advanced lysosomal architectures (McR-TACs, CPPTACs, HerTACs, PolyTACs) extend this logic and remain preclinical. Autophagy-targeted degraders (AUTOTACs) recruit the autophagy receptor p62/SQSTM1 for ubiquitin-independent clearance of larger payloads. Degrader-antibody conjugates (DACs) apply antibody delivery to PROTAC payloads; the Roche/C4 Therapeutics DAC partnership (April 2026, $20 million upfront, up to $1 billion-plus in milestones) is the clearest capital commitment to testing that hypothesis.
Molecular glue degraders are already FDA-approved. Thalidomide, lenalidomide, and pomalidomide are all IMiD-class glues developed through empirical chemical modification of a scaffold whose mechanism was unknown at the time. Lenalidomide's anti-myeloma mechanism was elucidated by Krönke and colleagues in Science in 2014, eight years after the drug's approval. The question for the next twenty years is not whether IMiD-class drugs work. It is whether a molecular glue can be designed from first principles for a specific therapeutic target using structural and computational reasoning rather than phenotypic screening of IMiD scaffold variants.
A PROTAC has an intrinsic design handle: linker modification. A molecular glue is monovalent. The designer must identify which neo-substrate a glue-remodeled E3 surface will recruit, determine which chemical modifications improve that interface, and do so in a system where the ternary complex exists only when the glue is present. The training data, characterized ternary complex structures, is almost entirely confined to CRBN-IMiD chemistry.
The Petzold Science 2025 paper
The pivotal scientific publication in the field between 2024 and 2026 is Petzold, Gainza, Annunziato, and colleagues (all Monte Rosa Therapeutics), published in Science on July 3, 2025. Using structure-based computational matchmaking, the authors predicted more than 1,600 CRBN-compatible G-loop proteins, identified 184 carrying a structurally differentiated helical G-loop motif, and experimentally validated 22 representative neo-substrates. The significance is equivalent to the 2012 VHL ligand discovery for PROTACs: it provides the structural grammar for rational neo-substrate selection within CRBN chemistry. It does not resolve the non-CRBN E3 ligase design problem.
AI adds structural hypotheses but not yet prospective design. A 2025 benchmark of AlphaFold3, Boltz-1, Chai-1, Protenix, and RoseTTAFold All-Atom found the models insufficient for prospective molecular glue design involving new glue and neo-substrate pairs. Insilico Medicine reported in Nature Communications in 2025 the first generative-AI design of a first-in-class PROTAC targeting PKMYT1 de novo; for molecular glues, the CRBN-biased training data remains the binding constraint. The BMS CELMoD franchise (mezigdomide in Phase 3, golcadomide in Phase 2) is the furthest advanced next-generation glue pipeline, the direct legacy of the 2019 $74 billion Celgene acquisition.
Vepdegestrant: reading the data carefully
On May 1, 2026, the FDA approved vepdegestrant for adults with ER-positive, HER2-negative, ESR1-mutated advanced or metastatic breast cancer detected by an FDA-authorized test, after progression on at least one line of endocrine therapy including a CDK4/6 inhibitor. In the 270-patient ESR1-mutant subpopulation of VERITAC-2, median PFS was 5.0 versus 2.1 months on fulvestrant (HR 0.57; 95% CI 0.42 to 0.77; P < 0.001), with ORR 19% versus 4%. In the 624-patient ITT population, PFS did not reach significance (HR 0.83). Approval is limited to the ESR1-mutant population, a companion diagnostic-defined label; framing the asset as a broad success overstates its commercial potential, which depends on biomarker-stratified selection.
Androgen receptor and BTK programs
Bavdegalutamide (ARV-110), the first AR PROTAC in the clinic (2019), showed a PSA50 of approximately 46% in patients with AR T878X/H875Y mutations. ARV-766 (luxdegalutamide) extends the class to wild-type AR and clinically relevant AR LBD mutants including L702H, H875Y, and T878A. ASCO 2024 data showed a PSA50 response rate of 50.0% in 28 PSA-evaluable patients with AR LBD mutations, with no dose-limiting toxicities and no MTD reached in Phase 1. AstraZeneca (AZD9750) and Qilu (QLH12016) have also entered the AR degrader space.
BTK degradation addresses an unmet need that covalent inhibitors cannot fully meet: acquired C481S resistance does not impair degraders working through induced proximity. Bexobrutideg (NX-5948; Nurix) is the furthest advanced BTK degrader clinically. Phase 1a data in CLL (48 patients) reported an objective response rate of 83.0%, with median PFS of 22.1 months at 16.6 months of follow-up, as presented at ASH 2025. On June 8, 2026, Roche announced a co-development and co-commercialization agreement: $700 million upfront and up to $1.6 billion in milestones, $2.3 billion in total potential value, with Roche bearing 60% of development costs. This is the largest single upfront payment for a clinical-stage PROTAC asset to date and positions bexobrutideg against Lilly's pirtobrutinib in inhibitor-refractory CLL.
A risk underappreciated in much of the commercial commentary: acquired BTK A428D mutations confer resistance to BTK degrader therapy in CLL, documented by Wong and colleagues in a 2024 Leukemia publication. This substitution disrupts the degrader's capability to induce a productive ternary complex with cereblon. Degradation shifts the evolutionary escape routes rather than eliminating them; late-stage implementation will require continuous molecular resistance monitoring.
The defining commercial feature of this period is not that pharmaceutical companies are paying for clinical-stage PROTAC assets. It is that three separate top-tier companies committed a combined $4.55 billion in potential milestones to Neomorph (San Diego), a company with no approved drug and no Phase 3 asset: Novo Nordisk (February 2024, up to approximately $1.46 billion, rare disease and cardiometabolic), Biogen (October 2024, ~$1.45 billion, neurology and immunology), and AbbVie (January 2025, up to $1.64 billion, oncology and immunology). The structural explanation is concentration: a small set of firms has assembled credible systematic molecular glue discovery for novel non-IMiD targets.
Monte Rosa executed two Novartis deals in twelve months, together the field's deepest single-partner commitment: $150 million upfront for VAV1 glue MRT-6160 (October 2024), then $120 million upfront against up to $5.7 billion for the QuEEN platform (September 2025). Total potential Novartis exposure is approximately $7.8 billion. QuEEN has produced one clinical candidate, which is a CRBN-based molecule; whether it warrants a $5.7 billion structure will be answered by non-CRBN chemistry data over the next three to five years.
| Transaction | Date | Upfront | Total potential value | Strategic focus |
|---|---|---|---|---|
| J&J / Halda Therapeutics | Nov–Dec 2025 | $3.05B (acquisition) | $3.05B | RIPTAC platform; HLD-0915 (Phase 1/2, mCRPC, Fast Track) |
| Novartis / Monte Rosa (Deal 2) | Sep 2025 | $120M + up to $60M options | ~$5.7B | QuEEN AI platform; immune-mediated diseases |
| Novartis / Monte Rosa (Deal 1) | Oct 2024 | $150M | ~$2.1B | VAV1 glue MRT-6160 (Phase 2); immunology |
| AbbVie / Neomorph | Jan 2025 | Undisclosed | $1.64B | Multi-target MGD platform; oncology, immunology |
| LEO Pharma / Gilead (STAT6) | Jan 2025 | $250M | ~$1.7B | Oral STAT6 degraders/inhibitors; atopic dermatitis, asthma, COPD |
| Biogen / Neomorph | Oct 2024 | Undisclosed | ~$1.45B | MGD platform; neurology, immunology |
| Novo Nordisk / Neomorph | Feb 2024 | Undisclosed | ~$1.46B | MGD platform; rare disease, cardiometabolic |
| Genentech / Orionis (2nd deal) | May 2025 | $105M | $2B+ | Allo-Glue monovalent MGD platform; oncology |
| Roche / Nurix (bexobrutideg) | Jun 2026 | $700M | ~$2.3B total | BTK degrader NX-5948; CLL/B-cell malignancies |
| Roche / C4 Therapeutics (DAC) | Apr 2026 | $20M | $1B+ | Degrader-antibody conjugates |
| Pfizer / TRIANA Biomedicines | 2024 | $49M | ~$1.5B | Molecular glue degraders; oncology |
| Kymera / Gilead CDK2 (option) | Jun 2025 | $85M aggregate | ~$750M | CDK2 molecular glue; oncology |
| Kymera / Gilead CDK2 (exercise) | Apr 2026 | $45M | (~$750M total) | CDK2 molecular glue; IND-enabling |
| Eli Lilly / Magnet Biomedicine | Feb 2025 | $40M (incl. equity) | ~$1.25B | TrueGlue platform; oncology |
| Eisai / SEED Therapeutics | Aug 2024 | $24M (equity) | ~$1.5B | E3 glue discovery; neurodegeneration, oncology |
| BMS / VantAI | May 2024 | Undisclosed | ~$674M | AI-driven non-CRBN/non-IMiD MGD design |
| Proxygen / MSD | Apr 2023* | Undisclosed | ~$2.55B | MGD discovery; multiple programs |
Figure 2. Disclosed targeted protein degradation transactions, June 2024 to June 2026. *The Proxygen/MSD deal closed April 2023, outside the window, and is shown for context.
Two transactions frame the extremes. Johnson & Johnson's acquisition of Halda Therapeutics (announced November 17, 2025; completed December 29, 2025) for $3.05 billion in cash is the period's sharpest valuation data point: a full acquisition of a Phase 1/2 asset (HLD-0915) built on the non-degradative RIPTAC mechanism, priced on visible data plus competitive tension. At the other end, on June 25, 2025, Sanofi discontinued Kymera's KT-474, a Phase 2b IRAK4 degrader, in favor of the next-generation KT-485. This was not a safety failure; KT-485 had superior potency, selectivity, and ADME. The episode carries a warning for BD professionals: medicinal chemistry iteration in this field is fast enough to obsolete a Phase 2b compound before it completes development.
Three categories of deals are backed by actual clinical evidence. Monte Rosa's first Novartis deal ($150 million upfront) is validated by Phase 1/2 MRT-6160 data. J&J/Halda ($3.05 billion cash) is premised on Phase 1/2 HLD-0915 data visible to diligence plus competitive tension. Roche/Nurix bexobrutideg ($700 million upfront) is anchored by strong Phase 1a CLL data and a Phase 2 pivotal study already enrolling, making it the best-evidenced upfront payment in the clinical PROTAC space.
Figure 3. Clinical evidence against capital committed. Placement, not size, carries the argument: the accented deals price optionality on discovery platforms with no Phase 3 asset.
The Neomorph three-deal portfolio represents $4.55 billion in potential milestones committed to a company with no Phase 3 asset. It is defensible as an option on scarce discovery capacity, but the milestone waterfalls assume clinical progression timelines that the scientific barriers above suggest are optimistic. The Kymera/Sanofi episode is the cautionary case those waterfalls have to survive: a milestone schedule that assumes a specific compound advances linearly to registration is exposed to the same fast medicinal-chemistry iteration that retired KT-474. Novartis's second Monte Rosa deal ($5.7 billion potential) prices what the QuEEN platform will do next, not what it has done. That question will be answered by non-CRBN chemistry data over the next three to five years.
The first PROTAC approval answers the question of feasibility. It does not answer the question of scale.
The clinical pipeline remains dominated by CRL4-CRBN compounds; the 600-plus other E3 ligases are both the field's greatest opportunity and its deepest bottleneck. Vepdegestrant's label is restricted to the ESR1-mutant population, and its durable commercial position depends on combination and sequencing data against oral SERDs. BTK A428D resistance will require molecular monitoring in Phase 3. The field has yet to demonstrate first-principles design of a monovalent glue on a novel, non-CRBN/non-VHL E3 ligase, and whether a DAC-liberated PROTAC payload can form a productive ternary complex before degradation remains undemonstrated in published in vivo data.
The field's trajectory over the next decade will likely be determined by three variables: whether AI-assisted neo-substrate prediction matures from confirmatory to prospective; whether non-CRBN E3 ligases yield tractable small-molecule handles; and whether clinical data from the current wave of trials justify the platform premiums being paid. As of yet, none have resolved. Until that changes, the $28 billion already committed is a wager on the field's direction, not a verdict on it.
Alacrita service
Alacrita's due diligence team evaluates assets and platforms like these for investors and acquirers, assessing E3 ligase dependency, resistance liabilities, companion-diagnostic exposure, and the achievability of milestone structures. Due Diligence.
A PROTAC is a two-headed molecule that physically bridges a target protein and an E3 ubiquitin ligase, so linker chemistry is a design handle but the molecule is large and hard to make drug-like. A molecular glue is a small, monovalent compound that binds one protein, usually the E3 ligase, and remodels its surface to recruit a new substrate. Glues are Rule-of-Five compliant, but they are far harder to design from first principles because the ternary complex exists only when the glue is present.
It proved feasibility: a PROTAC can become a licensed medicine. It did not prove breadth. The approval is limited to the ESR1-mutant population detected by a companion diagnostic, and the benefit in the overall intention-to-treat population was not statistically significant. The approval validates the mechanism, not a platform-wide premium.
Nearly all clinical degraders and almost all characterized ternary-complex structures are built on the CRL4-CRBN ligase. The 600-plus other E3 ligases in the human genome lack tractable small-molecule handles, so both design capacity and mutational-escape risk concentrate on a single ligase. Expanding beyond CRBN is the field's central unsolved design problem.
Some of it is. Deals such as J&J/Halda and Roche/Nurix are anchored to visible Phase 1/2 data. Others, including the Neomorph portfolio and the second Novartis/Monte Rosa deal, price optionality on discovery platforms with no Phase 3 asset. The record is a mix of evidence-backed conviction and bets on scarce discovery capacity.
Which E3 ligase the asset depends on, its resistance liabilities (for example BTK A428D in degrader-treated CLL), any companion-diagnostic dependency that narrows the label, and whether the milestone waterfall assumes linear clinical progression. Fast medicinal-chemistry iteration can obsolete a Phase 2b compound before it registers, as the Kymera/Sanofi IRAK4 program showed.
Anthony Walker, PhD
Anthony leverages more than 35 years of experience, including over a decade spent building & managing a biotechnology company and over 20 years as a management consultant to the pharmaceutical & biotech industries.