Immunophenotyping of circulating leukocytes reveal non-specific activation of innate and adaptive immune systems in Multi-system Inflammatory Syndrome of Childhood Temporally Associated with SARS-Cov-2 Infection: Descriptive cohort study

1 Immunophenotyping of circulating leukocytes reveal non-specific activation of innate and adaptive immune systems in Multi-system Inflammatory Syndrome of Childhood Temporally Associated with SARS-Cov-2 Infection: Descriptive cohort study Michael J. Carter1,2 # PhD; Matthew Fish3,4,5 # MSc; Aislinn Jennings3,4,# MSc; Katie J. Doores4 PhD; Paul Wellman2 BSc, Jeffrey Seow4 PhD; Sam Acors4 PhD; Emma Timms5, Julia Kenny1,2 PhD; Stuart Neil4 PhD; Michael H. Malim4 DPhil; Shane M. Tibby2* MD; Manu Shankar-Hari3,4* PhD # Denote equal contributions *Corresponding authors

Although there are similarities between MIS-C and Kawasaki disease such as acute mucocutaneous inflammation and shock, there are major epidemiological and clinical differences. Epidemiologically, Kawasaki disease has a peak incidence around 10-months of age, is more prevalent in Asian children and in Northeast Asian countries such as Japan, with seasonal variation 13,14 . In contrast, MIS-C is reported in older children, and thus far, almost exclusively in Europe and the United States [1][2][3][4][5][6][7][8][9][10][11][12] . Clinical findings such as gastrointestinal symptoms, temporal association with the SARS-CoV-2 pandemic, leukopenia, thrombocytopenia, excess cytokines and increase in biomarkers of heart failure (such as natriuretic peptides) reported in MIS-C 1-12 , are unusual in Kawasaki disease 13,14 . Therefore, MIS-C is hypothesized as a distinct immunopathogenic illness, resulting from either altered interferon responses or antibody dependent enhancement (ADE), associated with SARS-CoV-2 infections 15 . We surmised that understanding the leukocyte alterations in MIS-C would contribute towards understanding these immunopathogenic mechanisms and, would inform clinical care of this new illness.
In this context, we report a cohort study, describing the acute alterations in the circulating leukocytes in children with a diagnosis of PIMS-TS 16 , which is referred to as MIS-C in this manuscript. After detailed description of clinical and laboratory characteristics similar to previous reports 1-12 , we describe the immunophenotype of circulating leukocytes, at three clinically relevant illness phases: acute (within 72 hours of hospitalization), resolution (at clinical improvement defined as improved respiratory status or cardiac support and C-reactive protein (CRP) <100mg/L) and convalescent (at first outpatient follow-up after recovery).
Amongst the eight seronegative children, six had a clear history of either prior symptoms suggestive of SARS-CoV-2 infection, close household contact with confirmed cases of SARS-CoV-2 infection, presence at mass gatherings or parents who were health care workers. In seropositive patients, only IgG to N protein was detected, IgG for S and RBD were higher than IgM and positively correlated

Neutrophil activation and impaired antigen presentation in antigen presenting cells (APCs)
Absolute neutrophil, monocyte, DC and natural killer (NK) cell counts were similar during the acute, resolution and convalescent phases, and were similar to healthy controls ( Figure  positive correlation between DC HLA-DR MFI and the co-stimulatory CD86 MFI also implies impaired antigen presentation function.

Differential T cell depletion with evidence of activation
Pan-T cell lymphopenia observed in the acute phase, were more consistent in the seropositive patients and returned to normal by convalescence ( Figure- Figures, represents the

Discussion
Our study is the first report of cellular immune changes in children with a diagnosis of MIS-C. The median age of our cohort was comparable to reported literature 4,6,8 , though older than some other cohorts 1,2,5, [9][10][11] . There were more boys in our cohort, which is consistent with larger cohort studies 6,11 , but higher than other cohorts 1,2,4,8-10 . White ethnicity was more common in our cohort, which is different to other studies 2,7 . Although the proportions of seroconverted patients in our cohort (68%) was consistent with largest cohort study 11 , it is lower than the reported prevalence between 80% 1,6,8 to 90% 9,10 . However, it is recognized that seroconversion may not occur in asymptomatic infections 30 .
Other clinical features of our cohort were similar to the published literature 1-12 , including acute inflammation, cardiac dysfunction, procoagulant state, prevalence of coronary artery aneurysms, use of intravenous immunoglobulins, use of corticosteroids and other immunomodulatory agents (eTable-

5; eTable-6).
Kawasaki Disease is characterized by neutrophilia, eosinophilia, and activation of the interleukin-1 pathways 13,14 . In contrast we observed, that the neutrophils and monocytes were activated. The APCs (monocytes, DCs, B cells) had reduced HLA-DR expression, due to raised IL-6.
There was pan lymphopenia. The CD4+CCR7+ T cells and gamma-delta T cells were the only activated T cell subsets. The natural effector B cells were reduced, and the class-switch response was observed in the resolution phase. There was no discernable grouping of the immune cell changes by serological status. The major differences in the immunological changes between the different phases of illness probably represents transmuted 31  This is a small, single center study, from a tertiary pediatric center, serving an ethnically diverse population. To state that MIS-C is truly distinct from Kawasaki disease, we require concurrent immunophenotyping of children with Kawasaki disease. We highlight that most MIS-C patients have evidence of SARS-CoV-2 infection, before they become acutely unwell, but did not undertake stool virology testing to rule out viral replication occurring in the gastrointestinal tract 42 . Although an argument could be made that sensitive tests could result in misclassification of seroconversion and virology 43 , none of the immune abnormalities we describe were skewed by seropositive status. We have not explored the mechanisms for the immune cell abnormalities; as, functionality of immune cells is not assessable with cells preserved in leukocyte stabilizing medium.
In conclusion, based on our cohort characteristics and the immune cell changes we observed, MIS-C is likely to be a distinct immunopathogenic illness associated with profound changes to the innate and adaptive immune cells, in children infected with SARS-CoV-2 ≥2 weeks before the onset of acute illness. Our data highlights that the value of targeted immunomodulation based on cytokine profile.

Online Methods
Please see online supplement for additional description and data

Setting and approvals
Evelina London Children's Hospital is a major tertiary referral center for pediatric infectious diseases, cardiology and intensive care, and coordinates retrieval for 2 million children in South London and Southeast England, a region which had the highest number of COVID-19 confirmed cases in the UK during the pre-recruitment and recruitment window of our cohort (eFigure-7). Our institution identified and reported one of the first cases series of MIS-C 2 . The current cohort study was approved by the UK Health Research Authority (20/HRA/1714) 44 . Informed consent was obtained (from parents or guardians of children < 16 years of age) by trained health professionals.

Case definitions, blood sampling and clinical data
We recruited patients to this prospective cohort study between 27 th April 2020 and 25 th May 2020.
Children included in this cohort have not previously been reported by our group 6

Statistical analysis
Continuous data were summarized as median and interquartile range (IQR) and categorical data presented as frequency and percentage (%). We did not perform a-priori sample size calculations.
Due to limited sample size, the statistical significance reported should only be interpreted as indicative of the direction of change in biological signals. To compare differences between healthy controls and MIS-C illness phases (T1, T2 and T3), we represented data using box and whisker plots showing all the data points using Tidyverse package 52 and used Wilcoxon signed-rank test with Bonferroni correction for multiple comparisons. The strength and direction of association between variables was assessed using Spearman's correlation. All analyses were performed using R studio interface 53 and R 54 .