Deucravacitinib in hidradenitis suppurativa: a TYK2 pilot that lost its funding at seven patients

What was tried

Deucravacitinib (BMS-986165, marketed as Sotyktu) is an oral allosteric inhibitor that binds the regulatory pseudokinase domain of TYK2 rather than its catalytic site, giving selectivity over JAK1, JAK2, and JAK3. It is approved for moderate-to-severe plaque psoriasis. NCT05997277 was an investigator-sponsored Phase 2 pilot run by Beth Israel Deaconess Medical Center with Bristol-Myers Squibb as collaborator, testing deucravacitinib against placebo in moderate-to-severe hidradenitis suppurativa. The design was randomized, double-blind for participants and investigators, with the median change from baseline in inflammatory lesion count at Week 16 as the primary endpoint (ClinicalTrials.gov NCT05997277).

The study began on 30 November 2023 and reached primary completion on 25 February 2025. It enrolled seven participants, five on deucravacitinib and two on placebo, before being terminated. The recorded reason is verbatim: funding was withdrawn by the sponsor due to funding priorities and the study could not continue without it (ClinicalTrials.gov NCT05997277).

The biological hypothesis

TYK2 sits downstream of the IL-23, IL-12, and type I interferon receptors. Blocking it dampens the IL-23 to IL-17 axis without the broad cytokine shutdown of pan-JAK inhibition. The human genetics are unusually supportive for autoimmune skin disease: the TYK2 P1104A loss-of-function variant reduces risk across several immune-mediated conditions, which is the genetic basis for targeting the pathway. Deucravacitinib delivered on this in psoriasis, beating placebo and apremilast in the POETYK PSO-1 and PSO-2 Phase 3 trials (Journal of the American Academy of Dermatology 2023, and), and showed activity in a Phase 2 systemic lupus study (Arthritis and Rheumatology 2023).

Hidradenitis suppurativa is a different problem. Its inflammation is driven heavily by neutrophils, TNF, and IL-17, with a less clean dependence on the IL-23 to IL-17 relay that TYK2 modulates. Open Targets gives TYK2 an overall association of 0.37 with hidradenitis suppurativa, and that score is carried by clinical evidence (0.60) with weak literature support (0.05) and no listed genetic association for this specific disease (Open Targets Platform, ENSG00000105397). The rationale here was a reasonable mechanistic extension from psoriasis, not a genetically anchored hidradenitis hypothesis.

What actually happened

The registry shows posted results for the seven enrolled patients. The primary endpoint, median change from baseline in inflammatory lesion count at Week 16, was -2.0 with an interquartile range of -4 to -2 on deucravacitinib and -9 with an interquartile range of -14 to -4 on placebo (ClinicalTrials.gov NCT05997277). On the surface placebo reduced lesions more, but with five active and two placebo patients this comparison carries essentially no information. No serious adverse events and no deaths were recorded in either arm. The study did not fail on a prespecified efficacy analysis. It ended because the money stopped.

Failure mechanism, best guess

This is a structural failure rather than a biological one. An investigator-sponsored pilot that depends on a single industry funding source has no buffer when that sponsor reprioritizes, and hidradenitis suppurativa was never the asset's lead indication. The trial was sized as a pilot, so even a clean run would not have settled the efficacy question. The numerically worse active arm at n equals five is a sampling artifact, not evidence of harm or futility. The deeper issue is that the program committed enrollment and patient exposure to a question it was never powered to answer, in an indication where the TYK2 rationale is weaker than the psoriasis precedent implies given the neutrophil and TNF dominated biology of hidradenitis.

How to prevent this next time

Two quantitative tools would have protected the patients enrolled and the information generated. First, a power calculation should have set the minimum informative sample before opening. To detect a difference of roughly 5 inflammatory lesions with a standard deviation near 6 at 80% power and a two-sided alpha of 0.05 requires on the order of 24 patients per arm, so a 5-versus-2 allocation had close to no chance of yielding an interpretable result regardless of funding. Pairing that with a Bayesian predictive probability of success would have made the same point in decision terms:

PP(success)=∫θ​P(final success∣θ,interim data)⋅π(θ∣interim data)dθ

with a prespecified floor below which the study should not open. Second, a competitive and funding red-team would have flagged the single-source dependency and the indication-prioritization risk before first patient in, and pointed enrichment toward the subset of hidradenitis with measurable IL-23 to IL-17 pathway activation, where a TYK2 effect is most plausible. The single highest leverage change would have been securing fully committed funding for a properly powered cohort, on the order of 24 patients per arm, before enrolling a single participant.

What this means for similar programs

Repurposing an approved targeted drug into a second dermatologic indication is attractive because safety is known, but the genetics and cytokine biology do not always travel. TYK2 has strong genetic backing in psoriasis and broader autoimmunity, weaker support in hidradenitis suppurativa. Programs extending mechanism into neutrophil and TNF dominated diseases should plan confirmatory enrollment and funding around the weaker rationale, not the stronger source indication, and should avoid pilots so small that they generate exposure without an answer.

Open questions

Would a fully enrolled, adequately powered cohort have separated deucravacitinib from placebo in hidradenitis suppurativa? Is there a biomarker-defined subgroup with active IL-23 to IL-17 signaling where TYK2 inhibition works in this disease? Will the broader TYK2 class, now eight or more programs deep, test hidradenitis again with proper sizing? None of these can be answered from seven patients.

Sources

1. • ClinicalTrials.gov, NCT05997277, study record with posted results, enrollment 7, termination reason citing withdrawn funding, primary endpoint values and adverse event counts. - Armstrong et al., POETYK PSO-1, Journal of the American Academy of Dermatology 2023,- Strober et al., POETYK PSO-2, Journal of the American Academy of Dermatology 2023,- Morand et al., deucravacitinib in systemic lupus erythematosus Phase 2, Arthritis and Rheumatology 2023,- Open Targets Platform, target TYK2 (ENSG00000105397), hidradenitis suppurativa association 0.37. - OpenFDA FAERS, Sotyktu, 2607 reports, top terms DRUG INEFFECTIVE, ACNE, PRURITUS, RASH. - ChEMBL, deucravacitinib CHEMBL4435170, small molecule, max phase 4. - Claidex Mechanism Risk Score, TYK2, computed live from the Claidex claims table. Available from: https://clinicaltrials.gov/study/NCT05997277.