Non-depleting anti-CD19 did not move gadolinium-enhancing lesions in relapsing multiple sclerosis

What was tried

Eli Lilly ran NCT06220669, an adaptive Phase 2a/2b, randomised, double-blind, placebo-controlled study of LY3541860 in adults with relapsing multiple sclerosis. A Phase 2a stage tested two intravenous dose levels against placebo, and a Phase 2b stage tested two further dose levels against placebo. The primary endpoint was the cumulative number of new T1 gadolinium-enhancing lesions at 12 weeks. Secondary endpoints covered total T1 gadolinium-enhancing lesions, new or enlarging T2 lesions, and plasma trough concentrations.

Eligibility required a relapsing course by the 2017 revision of the McDonald criteria, with at least one relapse in the previous year, two in the previous two years, or one active gadolinium-enhancing lesion within 180 days of screening. The Expanded Disability Status Scale score had to be 5.5 or lower. Primary progressive and non-active secondary progressive disease were excluded.

The study opened on 19 March 2024 and enrolled 63 participants against a planned 2027 completion. Actual primary completion was 18 October 2025 and the record closed on 13 July 2026 as terminated. The sponsor gave one reason: the interim analysis data failed to demonstrate the intended treatment effect. No results have been posted.

The biological hypothesis

LY3541860 is a non-depleting anti-CD19 monoclonal antibody. CD19 is a co-receptor for the B cell antigen receptor complex that lowers the threshold for downstream signalling, and Open Targets describes exactly this function. The design intent, set out in preclinical work published in JCI Insight in 2023, was to bind CD19 and suppress B cell receptor signalling without triggering antibody-dependent or complement-dependent cytotoxicity. A Nature Reviews Rheumatology commentary the same year framed the appeal directly, which is that non-depleting inhibition might deliver the benefits of B cell targeting while leaving humoral immunity intact.

Multiple sclerosis was a reasonable place to test that idea because B cell targeting already works there. What it lacks is a human genetic anchor to CD19 itself. Open Targets scores the CD19 and multiple sclerosis association at 0.098, composed of a literature datatype score of 0.345 and a clinical datatype score of 0.144, with no genetic association datatype contributing. The clinical component records what drugs have been tried, not evidence that CD19 variation causes the disease.

What actually happened

The trial stopped at an interim analysis for failure to show the intended effect on new gadolinium-enhancing lesions. Fierce Biotech reported on 4 February 2026 that LY3541860 was one of three clinical-stage programmes Lilly dropped, with a company spokesperson saying the asset did not meet Lilly's high bar for continued development. The rheumatoid arthritis programme, NCT06859294, is already in the Claidex graph as an efficacy failure.

No effect sizes were released, so there is nothing to compare directly. What can be compared is the bar. In the Phase 2 ocrelizumab trial published in the Lancet in 2011, anti-CD20 depletion cut gadolinium-enhancing lesions at week 24 by 89 percent at 600 mg, 95 percent confidence interval 68 to 97, and by 96 percent at 2000 mg, interval 89 to 99. The HERMES rituximab trial reported reduced gadolinium-enhancing lesion counts at weeks 12 through 24 and a relapse rate of 14.5 percent versus 34.3 percent at week 24. This endpoint reads fast and with high contrast when the mechanism works.

Failure mechanism, best guess

This reads as an efficacy failure rooted in pharmacology rather than in trial conduct. The evidence that B cell targeting controls new inflammatory lesions in relapsing disease comes almost entirely from agents that remove B cells from circulation. Anti-CD20 antibodies deplete. Inebilizumab, an approved anti-CD19 antibody, also depletes, with approvals in neuromyelitis optica spectrum disorder, generalised myasthenia gravis and IgG4-related disease. CD19-directed CAR T cells deplete deeply, including in tissue.

LY3541860 was built to do something different. Functional inhibition of a signalling co-receptor reduces the responsiveness of a B cell that remains present. If lesion-forming activity depends on the physical presence of B cells where they act, as antigen-presenting cells, cytokine sources or precursors of tissue-resident populations, then silencing the receptor without clearing the cell would not be expected to reproduce the anti-CD20 result. The same molecule reading out negative in rheumatoid arthritis points at the pharmacological class rather than at anything specific to the central nervous system.

How to prevent this next time

No endpoint-level data were released, so no quantitative reanalysis is possible here and none is offered. The levers available before the first participant was dosed are qualitative and they are the useful ones.

The first is a mechanism-matched benchmark. Every prior success on this endpoint came from depletion. A programme proposing functional inhibition instead was implicitly claiming that depletion is not the active ingredient, and that claim was testable more cheaply than in a 63-participant lesion trial. A short pharmacodynamic study of B cell receptor signalling suppression in circulating and, where accessible, tissue-resident B cells, read against a depleting comparator, would have priced that assumption first.

The second is base-rate adjustment. The Open Targets record for CD19 and multiple sclerosis carries no genetic association datatype, and the clinical datatype score of 0.144 is generated partly by the drug programmes being counted, which makes it a weak independent prior. A crowded target with a literature-driven association and a novel pharmacology should carry a higher prior probability of failure than a target entering with genetic support.

The third is a red-team question the rheumatoid arthritis programme was positioned to answer. Both indications ran in parallel. Sequencing them, or gating the multiple sclerosis start on an interim from the arthritis study, would have let one negative result inform the other rather than producing two.

The single highest leverage change would have been running a mechanism-matched pharmacodynamic study against a depleting anti-CD19 or anti-CD20 comparator before committing to a lesion endpoint that only depletion has ever moved.

What this means for similar programs

The Claidex mechanism risk score for CD19 rises to 75 out of 100, in the red band, after this insert. The four documented failures are not one signal. Two are strategic reprioritisations of the KYV-101 CAR T in lupus nephritis and diffuse cutaneous systemic sclerosis, closed without efficacy readouts. Two are efficacy failures of the same non-depleting antibody in two diseases. The second pair is the informative one. Non-depleting CD19 inhibition has not yet shown a clinical effect in an autoimmune indication, while depleting CD19 agents hold three approvals.

Programmes pursuing functional inhibition of B cell surface targets should treat depletion as the default comparator rather than as an alternative strategy, and should expect to fund a head-to-head pharmacodynamic comparison. Programmes pursuing CD19 depletion are not implicated here.

Open questions

Which dose levels were tested, and whether the interim preceded the Phase 2b stage, cannot be answered from the public record. Whether the antibody achieved the receptor occupancy and signalling suppression seen preclinically is unknown, and that distinction separates a mechanism that does not work from a molecule that did not engage. Posting the Phase 2a data would matter, because a negative lesion result with confirmed target engagement is a more valuable record entry than a termination notice.

Sources

1. • ClinicalTrials.gov, NCT06220669, LY3541860 in relapsing multiple sclerosis, record accessed 24 September 2026: https://clinicaltrials.gov/study/NCT06220669 - ClinicalTrials.gov, NCT06859294, LY3541860 in rheumatoid arthritis: https://clinicaltrials.gov/study/NCT06859294 - A nondepleting anti-CD19 antibody impairs B cell function and inhibits autoimmune diseases, JCI Insight, 2023,- Non-depleting anti-CD19 B cell inhibition, Nature Reviews Rheumatology, 2023,- Kappos L and colleagues, Ocrelizumab in relapsing-remitting multiple sclerosis: a phase 2 randomised placebo-controlled multicentre trial, Lancet, 2011,- Hauser SL and colleagues, B-cell depletion with rituximab in relapsing-remitting multiple sclerosis, New England Journal of Medicine, 2008,- Inebilizumab for the treatment of neuromyelitis optica spectrum disorder (N-MOmentum), Lancet, 2019,- Masson G, Lilly drops 3 pipeline drugs, including gene therapy from $1B Prevail buyout, Fierce Biotech, 4 February 2026: https://www.fiercebiotech.com/biotech/lilly-drops-3-pipeline-drugs-including-gene-therapy-1b-prevail-buyout - Open Targets Platform, CD19 (ENSG00000177455) and multiple sclerosis (MONDO_0005301) association and drug candidate records, queried 24 September 2026: https://platform.opentargets.org.