NVG-2089 in CIDP: fourteen patients in seventeen months, and a target with no human anchor

What was tried

NVG-2089 is a recombinant IgG1 Fc-domain protein engineered to agonise type II Fc receptors. Nuvig Therapeutics tested it in INVGOR, NCT07027111, an open-label, non-randomised Phase 2 study in chronic inflammatory demyelinating polyradiculoneuropathy. The study ran from 25 April 2025 to 1 October 2026 and enrolled 14 participants. The registry lists one United States site. The entry was updated on 6 October 2026 with the reason for stopping given as "Business Decision".

The design had two sequential arms, treatment-naive and treatment-experienced. Diagnosis followed the EAN/PNS 2021 criteria and was confirmed by an independent adjudication committee. Treatment-naive participants needed an adjusted INCAT score of at least 2, treatment-experienced participants a score of 2 to 7 plus documented deterioration on interruption of immunoglobulin therapy or improvement on it. Both primary endpoints measured safety, specifically the incidence, nature and severity of treatment-emergent and serious adverse events at Week 14. Efficacy sat in the secondary endpoints as the percentage of treatment-naive participants with evidence of clinical improvement at Week 14. The study was reported at launch as targeting 60 participants across 40 sites.

The biological hypothesis

Intravenous immunoglobulin works in CIDP but is expensive, plasma-derived and supply-constrained. The account that has dominated the field for fifteen years holds that the anti-inflammatory activity of high-dose IgG runs through sialylated Fc, which induces the inhibitory receptor FcγRIIB and expands regulatory T cells. If that pathway is the active ingredient, a recombinant Fc protein engineered to engage it directly should reproduce the benefit at a fraction of the protein load.

NVG-2089 was built on that premise. Its published preclinical characterisation describes selective agonism of type II Fc receptors, Treg expansion and upregulation of FcγRIIB. In a randomised, blinded study in the spontaneous autoimmune peripheral polyneuropathy mouse model, NVG-2089 and high-dose immunoglobulin both stabilised motor and sensory performance, preserved compound muscle action potential amplitudes, and attenuated the distal latency prolongation associated with demyelination.

The human evidence for the target in this disease is thin. Open Targets scores the FCGR2B association with CIDP at 0.017, and the only contributing evidence type is literature mining at 0.143. No human genetic, clinical or animal model evidence is recorded for this pair. The same receptor scores 0.527 in systemic lupus erythematosus and 0.411 in ulcerative colitis, in both cases driven by genetic association components near 0.6, and it has reached Phase 3 elsewhere with obexelimab. It has simply never been genetically tied to CIDP.

A second problem is translational. The mouse evidence for the sialylated Fc to FcγRIIB axis has been contested for a decade, and a 2015 review set out in detail where the molecular and cellular components of that cascade differ between mice and humans and where extrapolation has been made incorrectly.

What actually happened

No outcome data from INVGOR has been reported. The study posted no results, and the sponsor says data from this study and from an ongoing immune thrombocytopenia study will be shared in the coming months.

What is available is the recruitment record. Fourteen participants enrolled over 524 days, or 0.81 per month. Holding that rate constant, reaching the planned 60 would have taken roughly 74 months from first patient in, about six years. The registry lists one activated site against 40 planned, which locates the shortfall in site activation rather than per-site accrual, though registry location lists for terminated studies are not always complete.

Nuvig chief executive David Woodhouse told Fierce Biotech on 6 October 2026 that "the cost and expected enrollment timeline in CIDP would limit our ability to move the program forward" and that "this was not a safety-related decision". The immune thrombocytopenia study continues, no layoffs were announced, and the remainder of the pipeline is preclinical.

The Phase 1 data presented at the 2025 Peripheral Nerve Society meeting is the other informative piece. Forty healthy volunteers received single doses from 9 to 500 mg/kg and 16 received three biweekly doses of 150 or 300 mg/kg or placebo. There were no serious adverse events, exposure was linear, and the half-life was 11 to 12 days. FcγRIIB was upregulated on conventional dendritic cells at all dose levels and activated CD39-positive regulatory T cells expanded at all dose levels, in both cases with no clear dose response.

Failure mechanism, best guess

The proximate cause is recruitment economics. A single-site study accruing less than one patient a month against a 60-patient target cannot finish on a biotech timeline, and the sponsor's own account matches that arithmetic.

The deeper issue is that the programme entered a slow-recruiting rare disease with no dose-selection information. A pharmacodynamic readout that saturates across a 55-fold dose range tells you the assay has a ceiling, or that the biology is all-or-none at low occupancy, but not which dose to carry forward. With both Phase 2 primary endpoints set on safety, the study was positioned to spend scarce CIDP patients without resolving either dose or effect size. That makes a trial expensive per unit of information, which is the condition under which slow accrual becomes fatal rather than inconvenient.

How to prevent this next time

Endpoint-level data is absent here, so a quantitative efficacy model is not supportable and none is offered. The qualitative levers are the ones that apply.

Base-rate adjustment: a 40-site activation plan should be validated against observed activation timelines before first-patient-in. The recruitment arithmetic above was computable from the first six months of accrual.

Biomarker enrichment: CIDP is not one disease. Nodal and paranodal antibody-defined subgroups respond differently to immunoglobulin therapy, and a mechanism explicitly designed to mimic immunoglobulin should prespecify which CIDP population it expects to help and recruit against that definition.

Red-team the target anchor: the question "what human evidence links FCGR2B to CIDP" has a short answer, which is literature mining alone. That does not forbid the programme, but it should have raised the bar on the dose-ranging and the enrichment strategy before a Phase 2 commitment.

The single highest leverage change would have been to require a dose-responsive pharmacodynamic marker from Phase 1 before committing to a 60-patient Phase 2 in a disease this slow to recruit.

What this means for similar programs

This is the first FCGR2B failure in the Claidex graph. The mechanism risk score for the target sits at 34 of 100, in the yellow band, and the shape of that score is unusual. The genetic deficit component contributes 14.74 of a possible 15, reflecting an association score of 0.017, while phase burden, archetype severity and saturation all stay low because only one programme has stopped. The score is driven almost entirely by absence of evidence rather than by accumulated failure.

Programmes that mimic immunoglobulin mechanisms should note that the receptor itself remains viable elsewhere. Direction matters too: a November 2025 preprint reports that deletion or targeted blockade of FcγRIIB impairs alpha-synuclein propagation in vivo, which is the opposite intervention on the same receptor in a different disease.

Open questions

Whether any of the 14 participants improved is unknown until Nuvig releases the data. Whether the flat Phase 1 pharmacodynamic response reflects assay saturation or genuine low-occupancy sufficiency has not been addressed publicly. How many of the 40 planned sites were ever activated is not stated. Whether the ongoing immune thrombocytopenia study produces the dose-response information CIDP never got will determine whether this molecule has a second indication.

Sources

1. • ClinicalTrials.gov. NCT07027111, INVGOR: A Phase 2, Open-label Study of Intravenous NVG-2089 in Participants With CIDP. Record updated 6 October 2026. https://clinicaltrials.gov/study/NCT07027111 - Otlu HG and colleagues. A recombinant IgG1 Fc-domain protein ameliorates inflammatory demyelinating peripheral neuropathy. Frontiers in Immunology, 2026.https://- Tjon AS and colleagues. Differences in Anti-Inflammatory Actions of Intravenous Immunoglobulin between Mice and Men: More than Meets the Eye. Frontiers in Immunology, 2015.https://- Anthony RM and colleagues. Intravenous gammaglobulin suppresses inflammation through a novel TH2 pathway. Nature, 2011.https://- Dalakas MC and colleagues. Intravenous immunoglobulin in chronic inflammatory demyelinating polyradiculoneuropathy: mechanisms of action and clinical efficacy. Expert Review of Neurotherapeutics, 2022.https://- Pascual-Goñi E and colleagues. Autoantibodies in chronic inflammatory demyelinating polyradiculoneuropathy. Current Opinion in Neurology, 2019.https://- Open Targets Platform. FCGR2B (ENSG00000072694) disease associations and clinical candidates, accessed 13 October 2026. https://platform.opentargets.org/target/ENSG00000072694 - Fierce Biotech. Big Pharma-backed Nuvig halts development of lead fusion protein in rare autoimmune disease, 6 October 2026. https://www.fiercebiotech.com/biotech/big-pharma-backed-nuvig-halts-development-lead-fusion-protein-rare-autoimmune-disease - NeurologyLive. Early Data Support Continued Development of NVG-2089 as IVIg Alternative in CIDP, 2025. https://www.neurologylive.com/view/early-data-support-continued-development-nvg-2089-ivig-alternative-cidp - bioRxiv preprint. Deletion or Targeted Blockade of FcγRIIb (CD32b) Impairs alpha-Syn Propagation In-Vivo, 18 November 2025. https://.