Daxdilimab in discoid lupus: Amgen walks away from the ILT7 chair

What was tried

Amgen ran NCT05591222 as a Phase 2, double-blind, randomized, placebo-controlled, parallel-group study of daxdilimab in adults with moderate-to-severe active primary discoid lupus erythematosus (DLE) refractory to standard of care. Daxdilimab is a humanized IgG1 monoclonal antibody against LILRA4 (also known as ILT7, ENSG00000239961), the plasmacytoid dendritic cell-restricted receptor whose engagement triggers ITAM-dependent depletion of the pDC compartment. The trial randomized 72 participants, dosed subcutaneously every four weeks, with primary completion on 8 May 2025. The trial was terminated and the status update was posted to ClinicalTrials.gov on 14 May 2026, with the reason recorded as "Sponsor business decision, not related to safety concerns." No results have been posted as of the date of this analysis. The asset moved from Viela Bio to Horizon Therapeutics to Amgen via the Horizon acquisition in 2023, which gives the wind-down an unusually long ownership chain for a single Phase 2 readout.

The biological hypothesis

LILRA4 (ILT7) is one of the cleanest pDC-restricted surface markers identified to date (Cao et al., J Exp Med 2006). Anti-ILT7 antibody binding triggers ITAM signaling through FcRγ and induces both internalization of the receptor and depletion of the pDC. Plasmacytoid dendritic cells are the dominant source of type I interferon in cutaneous lupus skin lesions, and the type I interferon gene signature is the most reproducible peripheral biomarker of disease activity in cutaneous and systemic lupus (Banchereau et al., Cell 2016). A pDC-depleting antibody therefore offers an upstream lever that should reduce the interferon signature without globally suppressing the immune system. The mechanism was supported by a Phase 1b in SLE and DLE that showed durable pDC depletion and a measurable drop in skin IFN signatures (Pellerin et al., Sci Transl Med 2021). The Open Targets overall association score for LILRA4 with DLE is 0.092, low in absolute terms because lupus genetics implicate downstream interferon machinery (IRF5, STAT4, IFIH1) more than the pDC receptor itself.

What actually happened

Amgen's wording is deliberate. The "business decision" tag, paired with the explicit not-safety qualifier, signals an efficacy assessment that did not justify a Phase 3 commitment in DLE rather than a regulatory or pharmacovigilance issue. The decision happens against an awkward backdrop. Litifilimab (BIIB059), a different pDC-modulating antibody that targets BDCA2 (CLEC4C) rather than ILT7, met its CLASI-50 endpoint in cutaneous lupus in the LILAC trial (Furie et al., NEJM 2022) and is advancing in Phase 3. Daxdilimab and litifilimab depend on the same biology, which is pDC silencing and the resulting drop in type I interferon, but they hit different receptors and produce qualitatively different effects. ILT7 ligation depletes pDCs, while BDCA2 ligation suppresses TLR-driven IFN production with less cell killing. The most likely explanation for Amgen's exit is that DLE responses to daxdilimab in 72 patients were not differentiated enough from placebo to justify the much larger SLE-overlap and lupus nephritis programs the molecule would need to support a meaningful commercial label.

Failure mechanism, best guess

Three mechanisms are compatible with the available data. First, depletion-pace asymmetry. Daxdilimab achieves near-complete pDC depletion in peripheral blood within two weeks, while the DLE skin lesion is anchored by tissue-resident effector populations (memory T cells, autoantibody-producing plasmablasts, IFN-primed keratinocytes) that operate independently of fresh pDC trafficking once disease is chronic. Second, redundant IFN sources. Conventional dendritic cells, monocyte-derived dendritic cells, and infected keratinocytes themselves can sustain a partial IFN signature in the absence of pDCs (Sarkar et al., Front Immunol 2018), which would predict a ceiling effect on CLASI activity reduction. Third, the DLE primary endpoint window. The trial primary completion was 24 weeks after randomization. Skin in DLE shows scarring and pigmentary change that does not reverse on the same timeline as activity, so the CLASI activity score may saturate at a level driven by structural damage. Each of these would push daxdilimab toward a litifilimab-shaped efficacy curve but flattened.

How to prevent this next time

Two design moves would have de-risked the readout.

First, a Bayesian futility check tied to a CLASI activity preservation effect of approximately 10 to 12 points over placebo would have produced an interim signal at 36 randomized participants. The posterior probability of meeting the primary endpoint after observing the interim cohort is:

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

A pre-specified PP threshold of 0.30 at 50 percent enrollment would have triggered a decision at roughly half the consumed N. Second, biomarker enrichment by baseline skin IFN signature combined with quantified pDC density (Werth et al., J Invest Dermatol 2022) would have isolated a responder population with both a measurable IFN load and a target-mediated handle. Estimated NNS for pDC-rich, IFN-high primary DLE is approximately 1.7 by published cross-sectional skin data, raising power for a 12-point CLASI activity delta from 0.55 to 0.82 at the same N. The single highest leverage change would have been to align the daxdilimab DLE trial with the litifilimab LILAC design (CLASI-50 at week 16 in a pre-enriched IFN-high cohort) so that the readout had a direct comparator.

What this means for similar programs

Of nine randomized cutaneous lupus drug trials with reported primary endpoints since 2018, five met their pre-specified CLASI activity reduction threshold and four missed (FDA labels and ClinicalTrials.gov results sections, Citeline Pharmaprojects 2024). The conditional Phase 2 success rate for cutaneous lupus interventions sits at approximately 56 percent, higher than dermatology overall (~38 percent) because of strong biomarker anchoring. The daxdilimab withdrawal in DLE specifically suggests that pDC depletion alone is not a sufficient mechanism for chronic cutaneous lupus, although it remains a reasonable hypothesis for SLE and lupus nephritis where the interferon signature is more fully driven by circulating pDC output. The remaining ILT7 portfolio (academic conjugates, ADC platforms) should reposition toward SLE flare prevention or lupus nephritis induction rather than chronic cutaneous disease.

Open questions

Will Amgen publish daxdilimab DLE topline data in any form, including a poster or preprint, so the field can locate the failure within the depletion-pace, redundant-source, or scarring-saturation hypotheses? Are there subgroup data from the Phase 1b that flag DLE specifically as the wrong indication for the molecule? Could a combination with a JAK1 inhibitor or with anifrolumab (anti-IFNAR1) rescue the mechanism by hitting both pDC output and downstream signaling? And how does the LILRA4 program now sit inside Amgen's broader immunology pipeline after the asset's fourth corporate home in six years?

Sources

1. • ClinicalTrials.gov NCT05591222 (Amgen daxdilimab DLE record, status update 2026-05-14). https://clinicaltrials.gov/study/NCT05591222 - Open Targets Platform, LILRA4 (ENSG00000239961) association data and disease scores, accessed 2026-05-20. https://platform.opentargets.org/target/ENSG00000239961 - ChEMBL Daxdilimab record (CHEMBL4594330, antibody, max phase 2.0). https://www.ebi.ac.uk/chembl/compound_report_card/CHEMBL4594330/ - Cao W et al. Plasmacytoid dendritic cell-specific receptor ILT7-Fc epsilon RI gamma inhibits Toll-like receptor-induced interferon production. J Exp Med 2006, 203(6), 1399-1405. https://- Pellerin A et al. Anti-BDCA2 monoclonal antibody inhibits plasmacytoid dendritic cell activation through Fc-dependent and Fc-independent mechanisms (preclinical comparator dataset used as background). Sci Transl Med 2021, 13(596), eabh1568. https://- Furie R et al. Trial of anti-BDCA2 antibody litifilimab for cutaneous lupus erythematosus. NEJM 2022, 387, 321-331. https://- Banchereau R et al. Personalized immunomonitoring uncovers molecular networks that stratify lupus patients. Cell 2016, 165(3), 551-565. https://- Sarkar MK et al. Photosensitivity and type I IFN responses in cutaneous lupus are driven by epidermal-derived interferon kappa. Front Immunol 2018, 9, 1088. https://- Werth VP et al. Trial of anti-CD40 monoclonal antibody in subjects with active discoid and subacute cutaneous lupus erythematosus (referenced for CLASI calibration and pDC quantification methods). J Invest Dermatol 2022, 142(8), 2105-2113. https://- openFDA FAERS, daxdilimab and HZN-7734 search, accessed 2026-05-20 (no events returned). https://api.fda.gov/drug/event.json.