Mignolet 2026 — Long-COVID IgG passive transfer reproduces sensory (but not cognitive) phenotype in mice

One-paragraph summary

This paper extends the Goebel 2021 FM-IgG passive-transfer paradigm to neurological long COVID (LC). Total IgG from LC patients meeting the 2021 WHO consensus definition was injected intraperitoneally into C57Bl/6 mice for four consecutive days. Recipient mice showed transient mechanical allodynia and thermal hypersensitivity during the first post-injection week, abolished when IgG-depleted serum or papain-digested IgG was substituted. The IgG accumulated specifically in lumbar dorsal root ganglia and colocalized with proprioceptive and nociceptive sensory neuron somata, without inducing local neuroinflammation or astrogliosis. Critically, mice showed no difference from controls on anxiety, depression, or short/long-term spatial memory tasks — only the sensory phenotype transferred. On human post-mortem DRG tissue, patient-derived IgG showed immunoaffinity for sensory neuron somata. The cleanly differential phenotype (sensory yes, cognitive no) is the second key finding: it suggests the autoantibody mechanism is responsible specifically for the pain/sensory component of LC neurological symptoms, while brain-fog/depression/anxiety phenotypes require a separate (non-IgG-mediated, non-DRG-targeting) mechanism.

Claims as triples

• IgG_LC — causes → mechanical_allodynia_in_mice [evidence: passive transfer; confidence: emerging]
• IgG_LC — causes → thermal_hypersensitivity_in_mice [evidence: passive transfer; confidence: emerging]
• IgG_LC — present_in → dorsal_root_ganglion [evidence: immunostaining of recipient mice; confidence: emerging]
• IgG_LC — binds → sensory_neuron_soma [evidence: human post-mortem DRG immunoaffinity; confidence: emerging]
• IgG_LC — does_not_cause → cognitive_dysfunction_in_mice [evidence: behavioral assays; confidence: emerging — NEGATIVE FINDING]
• IgG_LC — does_not_cause → anxiety_in_mice [evidence: behavioral assays; confidence: emerging — NEGATIVE FINDING]
• post_covid_syndrome — bridges → fm_autoimmune [via IgG_LC mechanism mirroring IgG_fm mechanism; confidence: emerging]
• autoantibody_mediated_pain — operates_in → post_covid_syndrome [evidence: this paper; confidence: emerging — NEW]

Methods note

LC patients (full sample size from search summary not specified — full text needed; appears to be a modest patient cohort) were enrolled by 2021 WHO consensus definition with formal neuropsychological assessment. Patients with prior autoimmune or neurological history were excluded. Age- and sex-matched asymptomatic individuals served as healthy controls. Total IgG was purified by protein G chromatography. IgG was injected i.p. into C57Bl/6/J mice on four consecutive days. Behavioral testing over the subsequent two weeks measured: paw withdrawal threshold (mechanical), thermal hypersensitivity, spatial working memory, depression-like behavior, anxiety-like behavior. Negative controls: IgG-depleted serum + papain-digested IgG (both abolished the phenotype, confirming Fc/whole-IgG dependence). DRG immunostaining co-localized patient IgG with proprioceptive and nociceptive sensory neuron markers. Human post-mortem DRG tissue used for direct binding assay.

Limitations

• Sample sizes of patient cohort, control cohort, and mice per group are not reported in the search summary; full text required for replication-level evaluation.
• Sensory phenotype is transient (first post-injection week); does not capture the durable LC pain phenotype seen in patients lasting months to years. This is consistent with the Goebel 2021 FM paradigm (also transient in mice) and is interpreted as a model artifact (passive IgG is metabolized over ~3 weeks; durable LC patients are presumed to continuously produce the antibody).
• DRG colocalization is at the cellular level (sensory neuron somata) rather than antigen-specific level — the specific molecular target is not identified. This contrasts with Seefried 2025 in FM, where 9 binding clusters and a FABP7+ satellite glial population were identified.
• The negative findings (no anxiety, depression, memory) cannot rule out other LC-specific mechanisms for cognitive symptoms — only that the IgG → DRG → pain pathway does not extend to the cognitive domain. Brain fog likely operates through a separate mechanism (possibly the predictive-coding / interoceptive-inference axis per Strube 2026 / Herman 2026, or via microglial activation per Coluzzi 2025).
• LC patient inclusion based on consensus clinical criteria, not biomarker stratification — heterogeneity in the IgG pool likely averages effects.

Open questions raised

• Q-NEW (Q78): Does the LC-IgG sensory mechanism operate through MRGPRX2 / mast cells (Sanchez 2025) or through direct sensory-neuron binding (or both)? The DRG immunostaining suggests direct neuronal targeting, but MC mediators (released near DRG) could also drive sensitization without direct neuron binding.
• Q-NEW (Q79): Are LC-pain patients distinguished from LC-brain-fog patients by IgG status? Mignolet's clean phenotype split (sensory IgG-driven, cognitive non-IgG-driven) predicts that anti-DRG IgG positivity should be a discriminator within the LC syndrome.
• Do LC-IgG-positive patients overlap with FM-anti-SGC-IgG-positive patients at the antigen level? Cross-reactivity testing on the Seefried 2025 9 binding clusters using LC patient sera would directly resolve.
• Does FcRn blockade (efgartigimod) attenuate LC pain symptoms? Q48 extends directly.
• Does plasmapheresis + IVIG (Oesch-Régeni 2025 ME/CFS pilot protocol) work in LC-pain patients?

Triangulation notes

• Closes a major portion of Q7 (does post-COVID FM cluster in autoimmune subtype?) — affirmative on the mechanism level. The same passive-transfer-IgG-to-DRG-causes-pain paradigm operates in both FM and LC. Treats post-COVID FM as an instance of the autoimmune-IgG subtype of FM rather than a parallel mechanism.
• Strengthens B1 bridge (Long COVID ↔ FM via autoantibody-mediated SFN) — was the cleanest unclosed bridge in the project; now anchored on direct passive-transfer evidence on both sides. The bridge is now load-bearing for translational claims about FM and LC sharing autoimmune mechanism.
• Parallels Sanchez 2025 mechanistic question: both LC-IgG and FM-IgG produce pain via passive transfer to mice and bind DRG. Does the LC-IgG mechanism run through MRGPRX2 → mast cells (Sanchez 2025), or through direct sensory neuron binding (this paper)? Critical Q78. Co-investigation in Goebel/Sanchez/Dong lab plausible given the conference-abstract framing.
• Differentiates from generic-LC literature: the negative findings (no cognitive/anxiety/depression behavioral effects from IgG) are as important as the positive findings. Generic-LC literature treats brain fog and pain as one syndrome; this paper supplies the first mechanistic dissociation.
• Compatible with the predictive-coding / interoceptive-inference framework (Strube 2026, Herman 2026) as a separate, parallel mechanism for the cognitive component of LC. The two mechanisms may co-exist within a single patient.

Bridges

• B1 (Long COVID ↔ FM via autoantibody-mediated SFN) — promoted in evidentiary status. Direct cross-condition passive-transfer evidence on both sides. Status: open → closing.
• B3 (CRPS ↔ FM via pathogenic IgG) — strengthening. CRPS, FM, and LC all have IgG → pain passive-transfer evidence; if the antibody target overlaps across all three, B3 closes alongside B1.

Cure-path implications

LC pain (and post-COVID FM subset) joins FM-autoimmune as a candidate target for the same four-level intervention hierarchy established by Sanchez 2025:

• Plasma-cell depletion (daratumumab — Fluge/Mella 2025 ME/CFS pilot)
• Circulating IgG removal (FcRn blockers, plasmapheresis + IVIG — Oesch-Régeni 2025)
• Receptor antagonism (MRGPRX2 antagonists if mechanism overlaps — Sanchez 2025)
• Downstream MC stabilization (cromolyn — Christoforou 2026)

Stratification by post-COVID onset history + anti-DRG IgG positivity would identify the candidate-responder subset within the broader LC population.