Distinct Autoimmune Antibody Signatures Between Hospitalized Acute COVID-19 Patients, SARS-CoV-2 Convalescent Individuals, and Unexposed Pre-Pandemic Controls

Abstract Increasing evidence suggests that autoimmunity may play a role in the pathophysiology of SARS-CoV-2 infection during both the acute and "long COVID" phases of disease. However, an assessment of autoimmune antibodies in convalescent SARS-CoV-2 patients has not yet been reported. Methodology: We compared the levels of 18 different IgG autoantibodies (AABs) between four groups: (1) unexposed pre-pandemic subjects from the general population (n = 29); (2) individuals hospitalized with acute moderate to severe COVID-19 (n = 20); (3) convalescent SARS-COV-2-infected subjects with asymptomatic to mild viral symptoms during the acute phase with samples obtained between 1.8 and 7.3 months after infection (n = 9); and (4) unexposed pre-pandemic subjects with systemic lupus erythematous (SLE) (n = 6). Total IgG and IgA levels were also measured from subjects in groups 1-3 to assess non-specific pan-B cell activation. Results: As expected, in multivariate analysis, AABs were detected at much higher odds in SLE subjects (5 of 6, 83%) compared to non-SLE pre-pandemic controls (11 of 29, 38%) [odds ratio (OR) 19.4,95% CI, 2.0 - 557.0, p = 0.03]. AAB detection (percentage of subjects with one or more autoantibodies) was higher in SARS-CoV-2 infected convalescent subjects (7 of 9, 78%) [OR 17.4, 95% CI, 2.0 - 287.4, p = 0.02] and subjects with acute COVID-19 (12 of 20, 60%) compared with non-SLE pre-pandemic controls, but was not statistically significant among the latter [OR 1.8,95% CI, 0.6 - 8.1, p = 0.23]. Within the convalescent subject group, AABs were detected in 5/5 with reported persistent symptoms and 2/4 without continued symptoms (p = 0.17). The multivariate computational algorithm Partial Least Squares Determinant Analysis (PLSDA) was used to determine if distinct AAB signatures distinguish subject groups 1-3. Of the 18 autoantibodies measured, anti-Beta 2-Glycoprotein, anti-Proteinase 3-ANCA, anti-Mi-2 and anti-PM/Scl-100 defined the convalescent group; anti-Proteinase 3-ANCA, anti-Mi-2, anti-Jo-1 and anti-RNP/SM defined acute COVID-19 subjects; and anti-Proteinase 3-ANCA, anti-Mi-2, anti-Jo-1, anti-Beta 2-Glycoprotein distinguished unexposed controls. The AABs defining SARS-COV-2 infected from pre-pandemic subjects are widely associated with myopathies, vasculitis, and antiphospholipid syndromes, conditions with some similarities to COVID-19. Compared to pre-pandemic non-SLE controls, subjects with acute COVID-19 had higher total IgG concentration (p-value=0.006) but convalescent subjects did not (p-value=0.08); no differences in total IgA levels were found between groups. Conclusions: Our findings support existing studies suggesting induction of immune responses to self-epitopes during acute, severe COVID-19 with evidence of general B cell hyperactivation. Also, the preponderance of AAB positivity among convalescent individuals up to seven months after infection indicates potential initiation or proliferation, and then persistence of self-reactive immunity without severe initial disease. These results underscore the importance of further investigation of autoimmunity during SARS-CoV-2 infection and its role in the onset and persistence of post-acute sequelae of COVID-19.