key: cord-1029352-zrthvayp
authors: Damoiseaux, Jan; Dotan, Arad; Fritzler, Marvin J.; Bogdanos, Dimitrios P.; Meroni, Pier Luigi; Roggenbuck, Dirk; Goldman, Michel; Landegren, Nils; Bastard, Paul; Shoenfeld, Yehuda; Conrad, Karsten
title: Autoantibodies and infection with SARS-CoV2 infection: The spectrum from association to clinical implicationreport of the 15th Dresden symposium on autoantibodies
date: 2021-12-09
journal: Autoimmun Rev
DOI: 10.1016/j.autrev.2021.103012
sha: 5dde28127277ac5be56b0b3955ee2874cfbe198f
doc_id: 1029352
cord_uid: zrthvayp

The relation between infections and autoimmune diseases has been extensively investigated. Multiple studies suggest a causal relation between these two entities with molecular mimicry, hyperstimulation and dysregulation of the immune system as plausible mechanisms. The recent pandemic with a new virus, i.e., SARS-CoV-2, has resulted in numerous studies addressing the potential of this virus to induce autoimmunity and, eventually, autoimmune disease. In addition, it has also revealed that pre-existing auto-immunity (auto-Abs neutralizing type I IFNs) could cause life-threatening disease. Therefore, the topic of the 15th Dresden Symposium on Autoantibodies was focused on autoimmunity in the SARS-CoV-2 era. This report is a collection and distillation of the topics presented at this meeting.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection leading to clinical manifestations associated with coronavirus disease 2019 (COVID -19) has been characterized by immune dysregulation evidenced as cytokine, T-cell and B-cell abnormalities (reviewed in [1] [2] [3] [4] ). Since the onset of the pandemic, research and observational studies are currently captured to more than 2000 publications per week with a grand total in excess of 175,000 publications! Included in this plethora of publications is considerable debate about the significance of the heightened autoinflammatory responses in severe COVID-19 and evidence that the observed immune dysregulation leads to systemic autoimmune rheumatic diseases (SARD) [5, 6] , and references therein). In addition, a number of reports suggest that some COVID-19 patients continue to develop de novo clinical signs and symptoms of SARD during recovery and what is now called multi-inflammatory syndrome, Kawasaki-like disease and Long-COVID [2] . Mechanisms potentially leading to autoimmunity in COVID-19 include increased release of self-antigens because of tissue damage, neutrophil activation and NETosis, molecular mimicry from homologous sequences of SARS-COV-2 with human proteins and activation of autoreactive immune cells [1] [2] [3] . In 2020, inborn errors of type I system, )2( excessive neutrophil extracellular trap formation with neutrophil-associated cytokine responses, )3( and molecular mimicry between self-components of the host and the virus.

Since the SARS-CoV-2 had infected an enormous number of individuals worldwide, it is hard to estimate the long-term effects on global health in terms of autoimmunity. Nevertheless, it is essential to remember that the development of autoantibodies could be regarded as the preclinical stage of autoimmune diseases; thus, the long-term autoimmune implications of SARS-CoV-2 are remained to be seen. As the presentation of autoantibodies is primarily found in severely ill COVID-19 patients, whether or not they will be persistent for years to come is still unknown. It is of great importance to recognize those autoimmune manifestations of COVID-19 in order to properly cope with their outcomes in the ongoing pandemic and the long-term post-pandemic period.

A significant limitation of many published studies that reported the emergence of autoantibodies and SARD in COVID-19 has been the lack of contemporaneous disease controls with similar clinical characteristics and longitudinal data monitoring the development of SARD over time. With these issues in mind, the autoantibody profiles were characterized in critically ill COVID-19 patients and it was explored if the observational cohort of adult COVID-19 patients admitted to an intensive care unit with acute respiratory failure was different from contemporaneous, similarly ill non-COVID-19 patients [1] [2] [3] [4] . At the time of this study, no COVID-19 specific interventions had been administered. The presence of autoantibodies was analyzed longitudinally (up to 5 separate time points) using a HEp-2 indirect immunofluorescence assay (HEp-2 IFA) and autoantigen-based multiplexed immunoassays that typify autoreactivity observed in SARD, assays for anti-phospholipid antibodies (aPL), as well as a multiplexed array for detection of anti-cytokine autoantibodies.

worse clinical severity scores. APLA were predominantly IgG anti-cardiolipin (aCL; 48%) followed by IgM aCL (21%), with a tendency toward a higher frequency among the COVID + patients. However, aCL antibodies were not associated with surrogate markers of thrombosis, but IgG aCL was strongly associated with worse disease severity and higher antinuclear antibody (ANA) titers, regardless of COVID-19 status. An association between aCL and anti-cytokine autoantibodies tended to be higher among the COVID + group. However, there were no statistically significant differences between COVID + and COVIDfor any of the autoantibodies tested. This was confirmed using Bayesian analysis using the credible estimates of the posterior probabilities compatible with our results.

In conclusion, severe COVID + patients have similar humoral autoimmune features as comparably ill COVIDpatients, suggesting that autoantibodies are a feature of critical illness regardless of COVID-19 status. This study provided evidence that, when observed longitudinally, severe COVID-19 patients have a similar autoantibody prevalence as a comparator cohort of critically ill patients. Taken together, the data suggest that autoantibody production is a feature of immune dysfunction associated with acute systemic illness rather than a specific SARS-CoV2 driven immunopathology. This should not to be taken to infer that SARS-CoV2-specific autoantibodies will not be found.

Molecular mimicry has been considered a valid mechanism to account for the induction of SARS-CoV-2 induced autoimmune phenomena [14] . Emerging data stemming from bioinformatics studies demonstrating a plethora of molecular mimics, i.e., sets of viral and human antigens which share extensive amino acid homology, have supported this notion [15] . Bioinformatic approaches meticulously assessing the extent of protein-protein amino acid similarity have been focused on identifying short mimicking sequences at two levels: a) between viral and human sequences of any kind and origin, irrespective of whether such sequences stem from known viral epitopes recognized by antibodies and T-cells during natural infection and post-vaccination [15] , b) searches focused on the identification of mimics between known vial antigenic and disease-specific or disease-related autoantigen regions, which are frequently targeted by the respective antibodies and autoantibodies as has been documented by epitope mapping studies. Discovery of such mimics is superior to J o u r n a l P r e -p r o o f the discovery of viral/self-amino acid homology of any kind, which could be accidental and unlikely to bear pathophysiological significance, even if the extent of amino acid similarity is relatively high.

Bioinformatic analysis and identification of 5-mers, 6-mers or even 7-mers shared by known viral and self-antigens is not a sufficient argument to support the notion that molecular mimicry is a likely cause of the induction of autoimmunity and autoimmune diseases for several reasons [16] : a) mimics documented as sequential amino acid homologies is a Of relevance to these studies, another indirect approach in support of the likely existence of a molecular mimicry in motion, is to focus on commercial purified polyclonal or monoclonal anti-SARS-CoV-2 antibodies or monoclonal antibodies against the respective autoantigens.

To confirm the original findings, these monospecific viral antibodies did not react with human autoantigens and vice versa. A recent study by Vojdani et al. has, however, provided exciting data in support of the presence of immunologic cross-reactivity involving such antibodies. They were able to show by ELISA that anti-spike SARS-CoV-2 monoclonal antibodies are targeting various autoantigens, such as GAD-65, mitochondria, phospholipids, and liver microsomes [18] . Also, in the same study anti-nucleoprotein SARS-CoV-2 monoclonal antibodies were reacting with a plethora of human autoantigens [18] . This evidence of human autoantigen recognition by anti-viral antibodies had led the authors to suggest that molecular mimicry may account for the observed anti-viral related human autoantibody reactivity. Bogdanos et al. were unable to replicate these data. Neither 4-5 of β 2 GPI was limited to only 3/58 (5.2%) tested sera for each domain and did not correlate with aCL/anti-β 2 GPI, nor with thrombosis [22] .

In conclusion, while medium/high aPL levels with D1 specificity are associated with vascular events in APS, low antibody titers with reactivity against  2 GPI epitope(s) different from D1 or D4,5 can be found in COVID-19 [23] . Such a difference may explain the lack of association with thrombotic events in COVID-19. The lack of β 2 GPI-dependent aPL or aPS/PT antibodies does not support the hypothesis that aPL can be responsible for LA phenomenon or the prolonged aPTT in these patients.

COVID-19 patients suffer from a systemic inflammation with complement activation, which may be responsible for high density of  2 GPI on the activated endothelium [24] [25] [26] [27] . In this context, even low titers of aPL may become pathogenic, thus potentiating or even triggering thrombus formation, especially when anticoagulation is suspended. Hence, while transitory aPL are likely to be clinically irrelevant in COVID-19 patients as in other infections, detection of aPL may be useful for identifying patients potentially at risk of thrombosis after the hospital discharge.

With respect to the association between thrombotic events observed in COVID-19 patients and the APS, this may not be restricted to the aPL included in the classification criteria, but may also entail non-criteria aPL. However, it is a well-established fact that infection-induced non-criteria aPL could occur in a transient manner and may constitute a non-pathogenic epiphenomenon. Notwithstanding, aPL IgG to prothrombin (aPT) extracted from SARS-CoV-2 infected patients was shown to trigger an accelerated hypercoagulation through the activation of innate immune mechanisms encompassing neutrophils and the corresponding release of neutrophil extracellular traps (NETs) [28] .

To investigate the relationship between criteria and non-criteria aPL and the strength of the acid, phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol, phosphatidylnositol, PS, PT, and annexin V (AnV), respectively) [30] . Additionally, the tripartite automated blood immunoassay technology was used to gauge the humoral response to SARS-CoV-2 by detecting IgG to spike ectodomain (S), receptor-binding domain (RBD) and nucleocapsid protein (NC) in these cohorts [29] .

Significant distributional changes between non-infected and SARS-CoV-2-infected individuals were ascertained only for IgM positivity against AnV, β 2 GPI, and PT as well as IgM levels to β 2 GPI, and PT (Fisher's exact and Wilcoxon rank sum test with Benjamini-Hochberg correction, p < 0.01, respectively). Consequently, for further regression analysis by fixed-and mixed-effect models, quantitative data of IgM levels against β 2 GPI, and PT were employed. To predict the occurrence of β 2 GPI and PT IgM levels depending on the humoral response to SARS CoV-2, ordinary least square regression model analysis with the addition of variables as fixed and as mixed effects in multiple linear regressions was run. While β 2 GPI IgM levels were correlated with the strength of the anti-viral IgG response only, IgM against PT was best predicted by the strength of the IgG response against SARS-CoV-2, but also by the patient's sex as well as disease severity [29] .

In summary, these findings highlight a correlation of IgM aPL, and here particularly IgM Since TTP associated with anti-PF4 antibodies was found to develop after vaccination with adenovirus-based, but not mRNA vaccines, it seems plausible that the adenovirus vector is J o u r n a l P r e -p r o o f involved in this rare adverse reaction. As vaccine adenoviruses infect endothelial cells upon intramuscular injection, it can be speculated that they might represent a source of SARS-CoV-2 spike protein [32] . Heparan sulfate proteoglycan expressed on the luminal side of endothelial cells could then bind either membrane-bound or soluble spike protein. Indeed, Kowarz et al. recently repeated that alternative splicing of the spike protein gene conveyed by the adenovirus could lead to truncated soluble spike protein variants [33] . Spike proteins could induce the release of PF4 through platelet activation via ACE-2 dependent and ACE-2 independent mechanisms. PF4 released by activated platelets could become immunogenic after binding heparan sulfate proteoglycan and recapitulate the sequence of events described above.

Ongoing efforts to decipher the common mechanisms involved in PF4-related TTP developing in different settings might provide important insights for our understanding of the roles of infection and vaccines in triggering autoimmunity [34] .

The SARS-CoV-2 pandemic has revealed two novel clinical entities. First, a disease that mimics Kawasaki disease and typically manifests in children appears to be associated with autoantibodies. Second, acquired immunodeficiency due to the presence of pre-existing autoantibodies against type I interferons (IFN) results in more severe clinical manifestations of COVID-19. These special conditions will be discussed in this section.

Infections by SARS-CoV-2 are typically mild or asymptomatic in children, but can, in rare cases, trigger a severe uncontrolled inflammatory response that has features in common with Kawasaki disease. The multisystem inflammatory syndrome in children (MIS-C) with COVID-19 typically presents 4-6 weeks after infection, with high fever, organ dysfunction, and elevated markers of inflammation. In a collaboration between research groups in Sweden and Italy a systems immunology approach was applied to characterize MIS-C as compared to children with Kawasaki disease, children with mild SARS-CoV-2 infection, and healthy children [35] . We profiled immune cell compositions and cytokines in blood, and we J o u r n a l P r e -p r o o f employed a previously demonstrated approach to proteome-scale autoantibody screening using microarrays of 9000 full-length human proteins [36] . In conclusion, this study revealed several autoantibody targets with a putative role in the development of MIS-C. The involvement of autoantibodies in MIS-C has been further substantiated in later studies with similar methodological approaches in other cohorts [38, 39] .

Type I interferons (IFNs) are anti-viral cytokines and are the first line of defense against many viruses. Surprisingly, neutralizing autoantibodies against type I IFNs have been known since the 1980's in patients with systemic lupus erythematosus, in patients treated with IFN- or IFN-β, and were even reported in one patient with a severe varicella zoster virus infection. These autoantibodies were, nevertheless, thought to be clinically silent.

Interestingly, their production can begin early in infancy, and are found in all patients with J o u r n a l P r e -p r o o f autoimmune polyendocrine syndrome type-1 (APS-1), due to germline mutations of AIRE.

They are also found in patients with hypomorphic mutations of RAG1 or RAG2, in men with mutations of FOXP3 and immunodysregulation polyendocrinopathy enteropathy X-linked (IPEX), in women with heterozygous null mutations of X-linked NEMO and incontinentia pigmenti [7] , in thymoma, and in patients with myasthenia gravis. Given the anti-viral role of type I IFNs and the finding that inborn errors of type I IFN immunity could underlie lifethreatening COVID [8] , it was anticipated that autoantibodies to type I IFNs might be causal to the development of severe COVID-19, even in patients without APS-1 or other genetic cause underlying these autoantibodies.

First, a large international cohort of patients was tested in 2020 for autoantibodies neutralizing IFN-2 and/or -. Surprisingly, at least 10% of patients with life-threatening COVID-19 pneumonia carried these neutralizing autoantibodies, while none were found in the individuals with asymptomatic or pauci-symptomatic infection [7] . These autoantibodies were found mostly in men (95%) and in the patients over 65 years old. These findings were later replicated world-wide. APS-1 patients are at very high risk of developing severe or critical COVID-19 pneumonia, although with incomplete penetrance, and should benefit from early vaccination and prompt treatment in case of infection before vaccination [40] .

Next, it was examined if more patients might have lower neutralizing titers of autoantibodies. New assays were set-up to test lower titers of autoantibodies and neutralization, in plasma diluted 1:10, against 100 pg/mL of type I IFNs. It appeared that 13.6% of patients of all ages were positive for neutralizing autoantibodies against type I IFNs.

Of note, some of them were only detectable by the neutralization assay. The prevalence increased with age with >20% in individuals older than 80 years, and in about 20% of all deceased individuals. The odds-ratios (OR) of having the autoantibodies showed that they confer a very high risk of having severe disease. Indeed, the highest odds ratios were those of having autoantibodies neutralizing IFN-2 and IFN- at 10 ng/mL and 100 pg/mL (67, P < 7.8x10 -13 and 54, P < 10 -13 ), while the presence of autoantibodies against IFN-2 (45 at 10 ng/mL, P < 7.8x10 -13 and 23 at 100 pg/mL, P < 10 -13 ) and against IFN-2 and/or IFN-, or IFN-Importantly, in all patients tested, the autoantibodies against type I IFNs were present before SARS-CoV-2 infection, as in patients with APS-1 [8, 40] .

Finally, it was investigated if the increase in the elderly was also seen in the uninfected population. We thus recruited a much larger cohort of uninfected adult individuals, of all ages. Strikingly, the prevalence of autoantibodies in the general population neutralizing 10 ng/mL (and 100 pg/mL) of type I IFNs, increases importantly and significantly with age, with 0.17% (1.1%) of positive individuals before the age of 70 years, and more than 1.4% (4.4%) positive individuals after the age of 70 years. These autoantibodies were most likely clinically silent until SARS-CoV-2 infection. Interestingly, these autoantibodies to type I IFNs can also underlie severe adverse events following vaccination with the yellow-fever live-attenuated vaccine [41] .

Overall, autoantibodies to type I INFs underlie life-threatening complications in a fifth of individuals over 80 years old and in a fifth of fatal COVID-19. They can be detected before infection, including in convalescent plasma which could then be excluded from donation [42] . Positive individuals should be vaccinated as early as possible, although not with a live attenuated vaccine [41] and should be managed promptly in case of infection. It is also likely that these autoantibodies neutralizing type I IFNs underlie other viral diseases, especially in the elderly.

In this review multiple key presentations given at the 15 th Dresden Symposium on Autoantibodies are summarized. With respect to the plethora of autoantibodies that have been associated with COVID-19, there are many remaining questions. For most autoantibodies it is not known if these autoantibodies already pre-existed before SARS-CoV-2 infection, it is not known if they persist after recovering from the disease, and if so, whether they will cause autoimmune disease upon follow-up. It can be anticipated that if millions of people become infected within a relatively short time-span, a substantial number of infected individuals will simultaneously develop autoimmune diseases. Furthermore, patients in intensive care units appear to have widespread autoimmune reactions, irrespective of being infected by the SARS-CoV-2. Therefore, a solid causal relation needs to be established. Molecular mimicry might explain such a causal relation [2, 43] , but as 4. The temporal appearance of a SARD that developed during the pandemic in an individual should not be equated to causality. An important consideration for assessing autoimmune phenomena among critically ill patients is the need for longitudinal sampling and clinical follow-up because the development of autoantibodies is time dependent. Therefore, a "snapshot" at arbitrary times (e.g., cross-sectional studies) is unable to capture this process. 5 . Confounding factors such as medications, pre-existing autoimmune illnesses, and other comorbidities need to be taken into consideration.

6. Many publications identify 'limitations' of their study but this typically appears near the end of the discussion. Readers must pay close attention to the limitations of the study, preferably before reading the entire manuscript. Indeed, if the limitations are taken in the context of the discussion of the results, much of the discussion may be moot.

In conclusion, the 15 th Dresden Symposium on Autoantibodies was an excellent podium to exchange current knowledge on autoimmunity in the SARS-CoV-2 era. Unfortunately, this review could not cover the whole spectrum of presentations on this topic because the research data underlying some of the presentations were not yet published. This included presentations on the potential role of autoantibodies directed against G-protein coupled receptors (GPCR) [44] and autoantibodies to the angiotensin converting enzyme (ACE)2 [45, 46] . It can be speculated that the first are involved in the clinical manifestations associated with Long-COVID because clinical manifestations resemble diseases, like fibromyalgia and silicon-induced autoimmunity, due to an autoantibody-mediated dysregulation of the autonomic nervous system [47, 48] . The autoantibodies to cell-surface bound ACE2 may either interfere with the binding to and infection of airway epithelial cells by SARS-CoV-2 or result in disturbance of the renin-angiotensin system that may be associated with the intrinsic effects of SARS-CoV-2 infection. Altogether, it is evident that further research is needed to answer the many remaining questions regarding the association between a plethora of autoantibodies and COVID-19. Nevertheless, SARS-CoV-2 is considered a strong stimulator of both the innate and adaptive immune system and, Figure 1 

On behalf of the COVID-19 chapter of the "Longitudinal Biomarkers in Lung Injury" study groupCollaborators. COVID-19-associated autoimmunity as a feature of acute respiratory failure

The SARS-CoV-2 as an instrumental trigger of autoimmunity

Autoantibodies in COVID-19 -a model for viral induced autoimmunity

Anticardiolipin and other antiphospholipid antibodies in critically ill COVID-19 positive and negative patients

Autoantibodies in COVID-19: frequency and function

New-onset IgG autoantibodies in hospitalized patients with COVID-19

Autoantibodies against type I IFNs in patients with life-threatening COVID-19

Life-Threatening COVID-19: Defective Interferons Unleash Excessive Inflammation

Covid-19 and autoimmunity

The mosaic of autoimmunity: genetic factors involved in autoimmune diseases

The mosaic of autoimmunity: hormonal and environmental factors involved in autoimmune diseases

The mosaic of autoimmunity: prediction, autoantibodies, and therapy in autoimmune diseases

Hyperstimulation of adaptive immunity as the common pathway for silicone breast implants, autoimmunity, and lymphoma of the breast

Molecular mimicry between SARS-CoV-2 spike glycoprotein and mammalian proteomes: implications for the vaccine

Homology between SARS CoV-2 and human proteins

Bioinformatic and immunological analysis reveals lack of support for measles virus related mimicry in Crohn's disease BMC Med

Microbial mimics are major targets of crossreactivity with human pyruvate dehydrogenase in primary biliary cirrhosis

Reaction of Human Monoclonal Antibodies to SARS-CoV-2 Proteins With Tissue Antigens: Implications for Autoimmune Diseases

Autoimmunity roots of the thrombotic events after COVID-19 vaccination

Lupus Anticoagulant and Abnormal Coagulation Tests in Patients with Covid-19

Antiphospholipid antibodies and COVID-19 thrombotic vasculopathy: one swallow does not make a summer

Anti-phospholipid antibodies in COVID-19 are different from those detectable in the anti-phospholipid syndrome

New insight into antiphospholipid syndrome: antibodies to β2glycoprotein I-domain 5 fail to induce thrombi in rats

In vivo distribution of β2 glycoprotein I under various pathophysiologic conditions

Complement activation and endothelial perturbation parallel COVID-19 severity and activity

Neutrophils and contact activation of coagulation as potential drivers of COVID-19

Multiple-organ complement deposition on vascular endothelium in COVID-19 patients

Prothrombotic autoantibodies in serum from patients hospitalized with COVID-19

Anti-prothrombin autoantibodies enriched after infection with SARS-CoV-2 and influenced by strength of antibody response against SARS-CoV-2 proteins

Antiphospholipid antibodies detected by line immunoassay differentiate among patients with antiphospholipid syndrome, with infections and asymptomatic carriers

Serological and viral genetic features of patients with COVID-19 in a selected German patient cohort-correlation with disease characteristics

Thrombotic thrombocytopenia associated with COVID-19 infection or vaccination: Possible paths to platelet factor

Vaccine-Induced Covid-19 Mimicry Syndrome: Splice reactions within the SARS-CoV-2 Spike open reading frame result in Spike protein variants that may cause thromboembolic events in patients immunized with vectorbased vaccines. Preprint (Version 1) available at Research Square

SARS-CoV-2, the autoimmune virus

The immunology of Multisystem Inflammatory Syndrome in Children with COVID-19

Proteome-wide survey of the autoimmune target repertoire in autoimmune polyendocrine syndrome type 1

The Global Phosphorylation Landscape of SARS-CoV-2 Infection

Mapping Systemic Inflammation and Antibody Responses in Multisystem Inflammatory Syndrome in Children (MIS-C)

The autoimmune signature of hyperinflammatory multisystem inflammatory syndrome in children

Preexisting autoantibodies to type I IFNs underlie critical COVID-19 pneumonia in patients with APS-1

Auto-antibodies to type I IFNs can underlie adverse reactions to yellow fever live attenuated vaccine

Convalescent plasma in Covid-19: Possible mechanisms of action

Molecular mimicry between SARS-CoV-2 and the female reproductive system

What Makes Antibodies Against G Protein-Coupled Receptors so Special? A Novel Concept to Understand Chronic Diseases

IgM autoantibodies recognizing ACE2 are associated with severe COVID-19. medRxiv