Shared TCR Vβ21.3+ T cell immunological signature between MIS-A and MIS-C

In the spring of 2020, case reports emerged describing an unusual severe hyperinflammatory syndrome temporally associated with SARS-CoV-2 in previously healthy children and adolescents (1). This syndrome shared some features with Kawasaki disease but differed in terms of older age (with a median of 8 years) and frequent associations with vasoplegia and prominent gastrointestinal involvement, whereas coronary aneurisms observed in Kawasaki disease were rare (2). This life-threatening condition was named multisystem inflammatory syndrome in children (MIS-C). Although extremely rare, a similar presentation was subsequently reported in adults and termed as multisystem inflammatory syndrome in adults (MIS-A).

As in children, MIS-A onset is estimated to occur 2–6 wk after exposure to SARS-CoV-2. While MIS-A is less common than MIS-C, both conditions share similar clinical and biological features, including persistent fever, systemic hyperinflammation, and multiorgan involvement (3, 4). Particularly, gastrointestinal symptoms and cardiovascular complications are predominantly seen across patients. Severe cases present with multiple organ failure and shock requiring intensive care, but most patients recover with immunosuppressive treatment and supportive care (4, 5, 6, 7).

The pathophysiology of MIS is insufficiently understood. Previous studies and case reports on MIS-C suggested a possible role for inborn errors of immunity (8, 9) and a multitude of immune abnormalities such as elevated levels of innate alarmins and systemic inflammatory proteins, dysregulated innate response, autoantibodies, and aberrant activation of the adaptive response with increased cytotoxicity (10, 11, 12, 13). Specifically, a polyclonal expansion of Vβ21.3+ T cells was observed almost exclusively in MIS-C patients compared to healthy controls, Kawasaki disease, toxic shock syndrome, and mild and severe SARS-CoV-2–infected subjects (13, 14, 15).

To date, and due to the rarity of its occurrence, MIS-A remains poorly studied, particularly at the immunological level. Here, we specifically investigate the T cell response in two European (Danish and French) MIS-A cohorts to determine whether MIS-A shares the characteristic T cell receptor (TCR) Vβ repertoire skewing observed in MIS-C.

We first explored a Danish cohort of five MIS-A patients, two males (33 and 50 years old) and three females (23, 38, and 35 years old), referred to as MIS-A 1 through MIS-A 5. All of them were admitted to two major university hospitals in Denmark, between December 2020 and February 2022. Prior to their MIS-A episode, all patients were healthy and had no history of recurrent infections, autoimmunity, or malignancies. All five individuals were infected with SARS-CoV-2 up to 7 wk prior to their hospitalization. Four patients (MIS-A 1–4) were not previously vaccinated against SARS-CoV-2, while MIS-A 5 had received two doses of BNT162b2 mRNA vaccine, 5 and 4 mo before the MIS-A episode (Fig. 1 A).

The patients presented with persistent fever, signs of hyperinflammation, gastrointestinal symptoms, and myocardial dysfunction. Other manifestations included throat irritation, myalgia, severe hypotension, respiratory and circulatory failure, and mucocutaneous involvement. Treatments included intravenous immunoglobulins (IVIGs), corticosteroids, fluid resuscitation, and antibiotics. All patients recovered and were discharged 1 to 2 wk after MIS-A onset. Patients’ characteristics and clinical features are summarized in Table 1.

Regarding the patients’ biological features, laboratory results from the first 3 days of hospital admission showed high leukocyte counts and elevated levels of the inflammation marker C-reactive protein. We also found high levels of D-dimer, brain-type natriuretic peptide, and troponin, revealing cardiac involvement, when measured in several patients (Table 2).

For immunological explorations, we isolated peripheral blood mononuclear cells (PBMCs) from all MIS-A patients within the first month following the beginning of the hyperinflammatory syndrome. Sampling time varied among patients and ranged from 2 to 21 days after MIS-A onset (Fig. 1 A and Table 1). We then analyzed the TCR Vβ repertoire distribution and T cell immunophenotype using spectral flow cytometry. Our results showed an expansion of Vβ21.3+ T cells in at least one of the three T cell subsets, CD3+, CD4+, or CD8+ T lymphocytes, in three MIS-A patients (MIS-A 1, MIS-A 3, and MIS-A 4) (Fig. 1 B). In these individuals, we also found an upregulation of the expression of activation markers CD38 and HLA-DR and of exhaustion markers TIM-3 and PD-1 on Vβ21.3+ cells compared to Vβ21.3− cells within CD3+, CD4+, and CD8+ T cell populations (Fig. 1 C). Additionally, we observed a higher abundance of effector memory (EM) T cells (CD45RA− CCR7−) within the Vβ21.3+ T cell compartment compared to the Vβ21.3− T cells. Together, these data mirror our previous immunological findings in MIS-C (Fig. S1, A and B).

For patients MIS-A 2 and MIS-A 5, from whom PBMCs were harvested at day 14 and 11 after MIS-A onset, respectively, we were not able to detect an overt expansion of the Vβ21.3 family. However, we found a similar immunophenotype of the Vβ21.3+ T cell subset as in Vβ21.3− expanded patients. Specifically, an upregulation of activation and exhaustion marker expression as well as an increase in EM T cells were observed within the Vβ21.3+ population compared to Vβ21.3− cells (Fig. 1 C). Similarly, these data suggest an involvement of the Vβ21.3 T cells in the inflammatory response in MIS-A 2 and 5.

We sought to confirm our findings from the Danish MIS-A cases by analyzing the TCR Vβ repertoire in a bigger French cohort of eleven MIS-A patients (referred to as MIS-A 6 through MIS-A 16). For comparison, we included three control groups not meeting the MIS-A case definition criteria: (1) COVID-19-associated myocarditis (n = 8), (2) post anti-SARS-CoV-2 vaccination (BNT162b2 mRNA vaccine)–associated myocarditis (n = 2), and (3) influenza-associated myocarditis (n = 4). All patients were admitted to the intensive care unit (ICU) at the university hospital La Pitié-Salpêtrière in Paris between January 2021 and January 2023. Briefly, prior to their hyperinflammatory episode, all MIS-A patients had a history of SARS-CoV-2 infection (proven by positive nasopharyngeal test or serology), and none of them had received any COVID-19 vaccine. MIS-A clinical symptoms included fever, abdominal and cardiac involvement, hypotension, and shock. Detailed clinical and biological features of these MIS-A patients were previously published (16) and are summarized in Tables 1 and 2. Peripheral whole blood was collected from all patients (MIS-A and controls) upon admission to the ICU, and TCR Vβ repertoire was analyzed by conventional flow cytometry. In line with our results from the Danish MIS-A cohort, we observed an expansion of Vβ21.3+ T cells in CD4⁺ and/or CD8⁺ subsets mainly in MIS-A patients (6/11), and to a much lesser extent in COVID-19–associated myocarditis (2/8) and post COVID-19 vaccination–associated myocarditis (1/2) (Fig. 2, A and B). No Vβ21.3 expansion was detected in the influenza-associated myocarditis group (0/4). Consistent with prior reports in MIS-C (14, 17), we also observed other Vβ expansions in certain MIS-A cases. These expansions, however, were not consistent across patients, with the notable exception of the TCR Vβ9+ subset, which was commonly enriched across all four patient groups (Fig. 2, A and B).

MIS-A is a severe life-threatening complication of SARS-CoV-2 infection in adults and shares with its pediatric counterpart, MIS-C, the same clinical and biological features. Multiple studies characterized the immune dysregulation in MIS-C and highlighted a unique immune signature defined by the expansion of TCR Vβ21.3+ T cells (13, 14, 15, 17). However, the immune response in MIS-A remains poorly explored, mainly due to its exceptionally rare occurrence. Here, we retrospectively investigated the TCR Vβ repertoire in two European cohorts of MIS-A. In the Danish cohort, we found an expansion of Vβ21.3+ T cells in at least one of the three T cell subsets (CD3+, CD4+, or CD8+ T lymphocytes). This expansion was associated with an upregulation of activation and exhaustion markers and an increase in EM T cell frequency, specifically within the Vβ21.3+ T cell population. Interestingly, we also observed an activated profile of Vβ21.3+ T cells in all MIS-A patients, including those lacking Vβ21.3 expansion. The expansion of activated T cell subsets is normally followed by a homeostatic contraction phase during which the frequencies of the expanded populations return back to baseline (18). Consequently, depending on the sampling time from onset of the acute inflammatory episode, skewing of the TCR Vβ repertoire toward a specific Vβ expansion may be attenuated or even lost for late analysis time points. We therefore postulate that the lack of Vβ21.3 expansion in two out of the five Danish MIS-A patients might be related to a delayed timing of sampling after the onset of the inflammatory episode. We then extended our investigations to a larger French cohort where Vβ21.3 expansion was predominantly observed in MIS-A patients, compared to patients affected with infectious or inflammatory myocarditis not related to MIS-A. Of note, no Vβ21.3 expansion was found in influenza-associated myocarditis patients. This observation highlights a difference between SARS-CoV-2 and influenza virus in triggering an inflammatory response in infected subjects. We hypothesize that the differences in viral antigen composition, conformation, or distribution, as well as the nature of the immune response elicited by those antigens, might differentially shape the characteristics of the superantigen-like expansion of T cell subsets. Besides TCR Vβ21.3, we found other Vβ expansions in some MIS-A cases from both cohorts that were not consistent across patients. Such scattered expansions were previously reported in MIS-C, irrespective of concomitant Vβ21.3 expansion (14, 17). We hypothesize that co-expanded Vβ specificities may share homologous binding paratopes with Vβ21.3, and therefore could be activated by similar ligands. Alternatively, these expanded T cells might be responding to different stimuli or reflect bystander activation. Among the MIS-A co-expanded subsets in the French cohort, TCR Vβ9 expansion was almost as frequent as Vβ21.3. However, it should be noted that TCR Vβ9 expansion was also detected in patients from all control myocarditis groups. Endomyocardial lymphocytic infiltrates have previously been reported in myocarditis, associated or not with COVID-19 (19, 20). Thus, it would be of interest to investigate whether Vβ9+ T cells might play a pathogenic role in the myocardial inflammatory process. Finally, it would be important to determine whether, like in MIS-C, Vβ21.3+ T cell expansion in MIS-A is polyclonal, hence suggesting an underlying superantigen effect.

Our study has some limitations, mainly due to its retrospective nature and the rarity of MIS-A cases, which is reflected by the small number of patients in our cohorts. In the Danish one, blood sampling after the onset of MIS-A was performed at different time points among patients and happened only after treatment initiation in three out of five patients. This might have potentially impacted the TCR Vβ repertoire data, particularly by attenuating the observed TCR Vβ skewing. The lack of PBMCs at time of hospital admission hampered our ability to assess the possible effect of treatment administration on TCR Vβ distribution and expanded Vβ specificities in these patients. In the French cohort, we could not perform deep immunophenotyping to assess the activation status of both Vβ21.3+ and Vβ21.3− T cells due to lack of sample availability.

Collectively, the major findings from the present study highlight a skewing of the TCR Vβ repertoire toward a Vβ21.3 response in MIS-A, similar to what was previously reported in MIS-C. Thus, in addition to shared clinical and biological features, the immunological signature of expanded and activated Vβ21.3+ T cells is also common between MIS-C and MIS-A, therefore providing further evidence that these two medical conditions are highly similar and may reflect manifestations of the same immune dysregulation in children and adults, respectively.

Skewing of the TCR Vβ repertoire was very well studied in the context of bacterial and viral superantigens that bypass the antigen specificity of the TCR, thereby inducing a potent polyclonal activation of different T cell Vβ subsets (21). In MIS, the mechanisms underlying the Vβ21.3 response are yet to be determined. Two computational studies suggested the presence of a unique superantigenic motif in the spike protein of SARS-CoV-2 that shares similarities with staphylococcal enterotoxin B and might potentially trigger the polyclonal expansion of Vβ21.3+ T cells and the development of MIS-C (22, 23). However, we have recently unraveled a series of pediatric cases who developed MIS-C in the absence of a SARS-CoV-2 exposure (24). Moreover, a recent study has established a link between TGF-β overproduction and Epstein-Barr virus (EBV) reactivation in MIS-C, suggesting that Vβ21.3+ cells could be EBV specific (25). Further investigations are required to identify potential drivers of the superantigenic-like Vβ21.3 response in MIS-A. Understanding the contribution of Vβ21.3+ T cells to MIS pathophysiology may inform future targeted therapies and help improve clinical management of affected patients.

This study was approved by the Danish National Committee on Health Ethics (1-10-72-80-20), the Data Protection Agency, and the Institutional Review Board. French cohort study was conducted in accordance with the Declaration of Helsinki using the database registered at the Commission Nationale de l’Informatique et des Libertés (registration no. 1950673), in agreement with the ethical standards of La Pitié-Salpêtrière hospital’s Institutional Review Board, the Committee for the Protection of Human Subjects, and French law.

Danish and French patients were included from two major university hospitals in Denmark (Aarhus University Hospital and Hvidovre Hospital) and the university hospital La Pitié-Salpêtrière in Paris, respectively, between March 2020 and January 2023, after oral and written informed consent were obtained and in accordance with the Helsinki Declaration and national ethics guidelines. Adult patients with a clinical diagnosis of MIS-A could be included in the study if they fulfilled the following criteria (in accordance with Centers for Disease Control and World Health Organization guidelines): documented SARS-CoV-2 infection, prolonged fever, marked inflammation, evidence of clinically severe illness requiring hospitalization, with multisystem (>2) organ involvement, including cardiac involvement and exclusion of other plausible diagnoses, and were treated with a MIS-A–specific therapy. For the Danish cohort, detailed medical history and individual clinical and biological features can be found in the following section and in Tables 1 and 2, respectively. For comparison, MIS-C data provided by previous study (14) were used. For the French cohort, clinical and biological features of MIS-A and control patients were previously published (16) and summarized in Tables 1 and 2.

For PBMC isolation, blood sampling for Danish MIS-A patients was performed between 2 and 21 days after MIS-A onset, as indicated on Fig. 1 A and in Table 1. PBMCs were purified from heparinized blood by Ficoll density centrifugation and stored in liquid nitrogen.

For the Danish cohort, cryopreserved PBMCs were thawed, washed, and rested for 2 h in complete RPMI medium (RPMI 1640 with GlutaMAX [Gibco] supplemented with 1 mM sodium pyruvate [Gibco], 10 mM HEPES buffer solution [Gibco], 1X penicillin-streptomycin [Gibco], and 10% fetal bovine serum [Eurobio Scientific]). Cells were first incubated with LIVE/DEAD fixable dead cell stain (Invitrogen) for 15 min at room temperature (RT) in PBS (Gibco), then washed, and surface staining antibodies were added for 30 min at RT in PBS supplemented with 0.5% bovine serum albumin (Sigma-Aldrich). Finally, cells were washed and fixed with BD Cytofix/Cytoperm solution (BD Biosciences) for 20 min at 4°C according to the manufacturer’s instructions.

For TCR Vβ repertoire screening, PBMCs (Danish cohort) or whole blood (French cohort) were stained with Vβ-specific monoclonal antibodies from Beckman Coulter (Kit IM3497) according to the manufacturer’s instructions. Vβ expansion was defined by value higher than the mean plus 2 standard deviation in healthy adults (reference values are provided by Beckman Coulter for the Kit IM3497 and are obtained from a cohort of 85 healthy controls except for Vβ4, Vβ7.2, and Vβ13.2 values, which were obtained from 46 controls).

Danish cohort data were acquired on Cytek Aurora spectral flow cytometer and analyzed with FlowJo software (version 10.9.0). French cohort data were acquired on Beckman Coulter DxFLEX conventional flow cytometer and analyzed with CytExpert software (Beckman Coulter, version 2.0.2.18).

MIS-A 1 was a 33-year-old previously healthy Caucasian male who 6 wk after an acute infection with SARS-CoV-2 presented with persistent fever, dry cough, dyspnea, abdominal discomfort, and diarrhea. The patient was admitted to the hospital 1 wk later with a universal rash resembling erythema multiforme and respiratory and circulatory failure, requiring immediate respiratory support, fluid resuscitation, and vasopressor therapy in the ICU. Empirical antibiotic therapy with clarithromycin and ceftriaxone was initiated, together with hydrocortisone.

The patient showed biochemical signs of inflammation, liver injury, and renal failure. A chest x-ray showed bilateral pulmonary interstitial infiltrates. Transthoracic echocardiography (TTE) after intubation showed global hypokinesia (left ventricular ejection fraction [LVEF] ∼20%). Troponin T (TnT) and pro-brain-natriuretic peptide (NT-proBNP) levels were increased. A left upper lobe segmental lung embolus was detected on the chest CT. During admission, the patient was positive for both SARS-CoV-2 antibodies and SARS-CoV-2 PCR oropharyngeal swab. No other underlying etiology was found after extensive microbiological evaluation. Transesophageal echo showed no signs of infective endocarditis.

Hemodialysis was initiated due to renal failure. Simultaneously, the patient was treated with IVIG on suspicion of MIS-A and improved immediately thereafter. The patient was hospitalized a total of 6 days.

MIS-A 2 was a 50-year-old Chilean man only known with familial hypercholesterolemia. 3 wk prior to admission, he was tested positive for SARS-CoV-2, with mild flu-like symptoms. He was admitted to the hospital 18 days later with fever, headache, sore throat, aching muscles and joints, abdominal pain, swollen lymph nodes on the left side of neck, and a universal rash. Biochemical signs of inflammation were also observed. Contrast-enhanced CT confirmed enlarged cervical lymph nodes and an enlarged left tonsil, suggesting retropharyngeal abscess. This was later disproved by surgical exploration and histological examination of the removed left tonsil.

Initially, empirical piperacillin-tazobactam administration and fluid resuscitation were initiated. This was later switched to meropenem, clindamycin, and metronidazole without any effect. On the fifth day of hospitalization, the patient still had a fever and increased levels of inflammatory biomarkers. He was admitted to the ICU with respiratory distress. Chest-CT showed bilateral infiltrations without signs of pulmonary embolism. TTE showed decreased left ventricular systolic function (LVEF 40–45) without valvular lesions or pericardial effusion. No other underlying etiology was found after extensive microbiological evaluation.

Upon suspicion of MIS-A, the patient was treated with hydrocortisone and IVIG after 8 days of hospitalization. Thereafter, the fever and inflammatory marker levels rapidly decreased, and the patient fully recovered.

MIS-A 3 was a 23-year-old Caucasian female who tested positive for SARS-CoV-2 with mild lethargy. Approximately 5 wk later, she developed high fever, myalgia, throat irritation, cough, expectoration, retrosternal chest pain, conjunctival injection, and chapped lips. She was admitted to the hospital 1 wk later and initially treated with penicillin and ciprofloxacin for suspected pneumonia.

The patient showed biochemical signs of inflammation, together with increased levels of cardiac enzymes and NT-proBNP. The chest x-ray was normal. Electrocardiogram showed sinus tachycardia with T-wave inversion on all precordial leads. No arrhythmias were detected during telemetry. The TTE was normal. Cardiac MRI showed signs of myocarditis. No other underlying etiology was identified after microbiological evaluation.

On suspicion of MIS-A, IVIG, methylprednisolone, and acetylsalicylic acid were initiated on the third day of hospitalization. Thereafter, the patient recovered quickly.

MIS-A 4 was a previously healthy, 38-year-old Lebanese woman. She developed fever, sore throat, and diarrhea 6 days prior to admission to the hospital. The patient tested positive for SARS-CoV-2 on hospital admission and presented with persistent fever and hypotension (80/60 mm Hg). The chest x-ray was normal. Biochemical signs of inflammation were observed, together with increased TnT and NT-proBNP levels. TTE showed a decreased LVEF (∼40%).

Initially, piperacillin/tazobactam was administered along with fluid resuscitation. The microbiological evaluation did not identify other causative pathogens. Therefore, IVIG was initiated on the third day when MIS-A was suspected. The fever and inflammation subsided shortly thereafter, and she fully recovered.

MIS-A 5 was a 35-year-old previously healthy Somali woman who tested positive for SARS-CoV-2 3 wk prior to admission. During hospitalization, she presented with persistent high fever, abdominal pain, vomiting, and myalgia (especially in the lower extremities). The patient had biochemical signs of hyperinflammation, together with increased TnT and NT-proBNP levels. No other underlying etiology was found after microbiological evaluation.

Due to labile blood pressure and multiorgan failure, she was transferred to the ICU. During hospitalization, the patient was treated with piperacillin/tazobactam and clindamycin. She fully recovered.

Fig. S1 shows the expansion and activation status of Vβ21.3+ T cells in MIS-C, thus illustrating the shared Vβ21.3 signature between MIS-C and MIS-A.

Raw data underlying main and supplemental figures are available from the corresponding author upon request.

We thank the patients for participating in our study. We also thank the medical staff who helped collecting, processing, and storing the biological samples. We thank the SFR Biosciences (Université Claude Bernard Lyon 1, CNRS UAR3444, INSERM US8, ENS de Lyon, France), specifically Estelle Devêvre and Sébastien Dussurgey from the SFR Biosciences AniRA-cytometry facility for technical assistance. We thank the MIS-A & COVID-19 Taskforce clinicians’ group.

This work was supported by the Institut National de la Santé et de la Recherche Médicale and by French and European grants managed by the Agence Nationale de la Recherche (ANR-21-CE17-0064 [SOCSIMMUNITY]), ANR-21-RHUS-0008 (COVIFERON) from the ANR–Recherche Hospitalo-Universitaire Program, and by the Horizon Europe (HORIZON-HLTH-2021-DISEASE-04-07, 101057100 [UNDINE]), and by the Centre de référence des rhumatismes inflammatoires, des interféronopathies et des maladies autoimmunes. T.H. Mogensen was also funded by NOVO Nordisk Foundation (NNF21OC0067157 and NNF20OC0064890). Work in the Gorochov laboratory was also supported by the French Foundation for Medical Research, Paris, France (FRM EQU202203014622), by the Département Médico-Universitaire de Biologie et Génomique Médicales et Hygiène, AP-HP, Paris, France, and by the Paris Region – DIM ITAC.

Author contributions: Liliane Khoryati: conceptualization, data curation, formal analysis, investigation, methodology, project administration, validation, visualization, and writing—original draft, review, and editing. Signe Bech Soerensen: investigation. Christophe Parizot: data curation, formal analysis, and writing—original draft, review, and editing. Raphaelle Lautraite: data curation. Marc Pineton de Chambrun: data curation, supervision, validation, and writing—review and editing. Guy Gorochov: conceptualization, funding acquisition, project administration, supervision, and writing—review and editing. Trine H. Mogensen: conceptualization, formal analysis, funding acquisition, investigation, project administration, resources, supervision, validation, and writing—original draft, review, and editing. Alexandre Belot: conceptualization, data curation, funding acquisition, investigation, project administration, resources, supervision, validation, and writing—review and editing.