PhD thesis defense in biomedical and pharmaceutical sciences by Margaux Mignolet

Antibody-mediated autoimmunity as a biological basis for neuro-long COVID

Candidate

Margaux Mignolet

Promoters

Prof. Charles Nicaise, UNamur, Department of medicine, Molecular Physiology Research Unit (URPhyM), Laboratory Neurodegeneration and Regeneration (LNR) - promoter

Prof. Nicolas Gillet, UNamur, Department of veterinary medicine, Integrated Veterinary Research Unit (URVI) - co-promoter

Prof. Pierre Bulpa, CHU UCL Namur, Intensive Care Unit - co-promoter

Jury

• Prof. Charles Nicaise, UNamur (promoter)
• Prof. Nicolas Gillet, UNamur (co-promoter)
• Prof. Pierre Bulpa, CHU UCL Namur (co-promoter)
• Prof. Benoît Muylkens, UNamur (President)
• Prof. Diego Castanares Zapatero, UNamur; Belgian Health Care Knowledge Centre (KCE)
• Prof. Johan Van Weyenbergh, KULeuven
• Prof. Bénédicte Machiels, ULiège

Summary

Background: Approximately 6–10% of individuals infected with SARS-CoV-2 develop persistent symptoms, collectively referred to as long COVID. Among the most frequently reported manifestations are fatigue (28%), post-exertional malaise (30%), postural orthostatic tachycardia syndrome (31%), and cognitive impairment (27.6%). Cognitive and neuropsychiatric manifestations include memory impairment (14%), anxiety (13.4%), depression (18%), and pain (8.4%). Although the pathophysiological mechanisms underlying long COVID remain poorly understood, several hypotheses have been proposed, including dysregulation of the immune response leading to the development of pathogenic autoantibodies.

Objectives: The overall objective of this thesis was to determine whether immunoglobulin G (IgG) from patients with neurological long COVID (neuro-long COVID) contributes to their neurological manifestations. Participants were selected according to the 2021 WHO consensus definition and were required to present at least two manifestations of neuro-long COVID. Following a neuropsychological assessment, blood samples were collected. IgG were purified from serum and passively transferred to C57BL/6J mice. Behavioral assessments were performed to investigate cognitive, affective, and pain-related outcomes. The involvement of IgG in the observed behaviors was assessed by comparing these results with those obtained following administration of IgG-depleted serum or papain-digested IgG. The ability of IgG to recognize structures of the central and peripheral nervous systems was subsequently investigated by immunohistochemistry and immunofluorescence in murine and human post-mortem tissues. Transcriptomic analysis of dorsal root ganglia (DRG) was also performed to identify molecular signatures associated with the transfer of patient-derived IgG. Finally, an antigen microarray was used to assess IgG reactivity against a broad panel of human autoantigens.

Results: Passive transfer of IgG from patients with neuro-long COVID did not induce alterations in cognitive, anxiety-related, or depressive-like behaviors. In contrast, mice developed transient mechanical and thermal hypersensitivity, which was absent following administration of IgG-depleted serum or enzymatically digested IgG, demonstrating the involvement of IgG in this phenotype. Tissue analyses did not reveal human IgG in the brain parenchyma or spinal cord, whereas human IgG were detected on the cell bodies of sensory neurons in the DRG, predominantly within A-fiber and, to a lesser extent, C-fiber subpopulations. This binding was not associated with neuroinflammation. Transcriptomic analysis of the DRG identified 311 significantly upregulated and 21 downregulated transcripts in mice receiving patient-derived IgG. The enriched pathways were mainly associated with inflammatory responses, whereas underrepresented pathways were related to synaptic transmission. Antigen microarray screening did not identify an antigenic target specifically associated with long COVID but revealed several targets showing trends toward increased reactivity. Finally, analysis of human post-mortem DRG confirmed the ability of patient-derived IgG to bind to human sensory neurons.

Conclusion: These findings demonstrate that IgG from patients with neuro-long COVID can induce transient mechanical and thermal hypersensitivity following passive transfer into mice, associated with their preferential binding to sensory neurons in the DRG. The absence of central nervous system involvement suggests that this effect is primarily mediated through a peripheral mechanism. These data therefore support a potential contribution of IgG to the pathophysiology of long COVID, particularly in a subset of patients presenting with pain-related manifestations.