Genetic and Clinical Characteristics of Patients in the Middle East With Multisystem Inflammatory Syndrome in Children

Key Points Question What are the clinical, genetic, and laboratory characteristics of Middle Eastern patients with multisystem inflammatory syndrome in children (MIS-C)? Findings In this cohort study of 45 patients with MIS-C of primarily Arab and Asian origins, an enrichment of rare, likely deleterious immune-related genetic variants was found, with a possible association between genetic findings and MIS-C onset and resistance to treatment. Meaning These findings suggest that comprehensive genetic profiling of patients with MIS-C of diverse ethnicities is essential to characterize the genetic contribution to this disease.


Introduction
Multisystem inflammatory syndrome in children (MIS-C) is a critical and potentially life-threatening complication of COVID-19 in pediatric settings.][9][10] Epidemiologic, clinical, and immunologic investigations have revealed that MIS-C has phenotypic similarities to Kawasaki disease. 11However, immune response in patients with Kawasaki disease is typically characterized by proinflammatory signatures, including increases in interleukin (IL) 17 and relatively less macrophage activation syndrome-like cytokine profile levels, whereas patients with MIS-C have high levels of IL-15 and interferon (IFN) γ in severe cases, and more than 50% of patients with MIS-C have a macrophage activation syndrome-like cytokine phenotype. 12me possible indications of the pathophysiologic mechanisms behind MIS-C have been identified.Cytokine profiling of patients with MIS-C revealed elevated inflammatory markers (IL-18 and IL-6), activation of myeloid and lymphocytic chemokines (CCL3, CCL4, and CDCP1), and dysregulation of mucosal immune markers (IL-17A, CCL20, and CCL28). 13Furthermore, B-cell receptor sequencing and autoantibody assays of patients with MIS-C identified dysregulated B-cell responses, autoantibody production, and complement-and myeloid cell-mediated inflammation. 14rthermore, immunologic profiling of children with MIS-C and children with COVID-19 revealed distinct features of lymphocyte activation, mainly CX3CR1 + CD8 + T-cell activation, in patients with MIS-C. 15 a recent study, 16 exome sequencing of 2 patients with MIS-C delineated a role of 1 gene, SOCS1 (OMIM 603597), in the IFN pathway in predisposing individuals to infection-associated autoimmune cytopenias.Similarly, another exome sequencing study 17 identified defects in XIAP (OMIM 300079) and CYBB (OMIM 300481) in 3 of 18 sequenced patients with MIS-C, whereas RNA sequencing of patients with MIS-C revealed dysregulated cytotoxic lymphocyte responses to SARS-CoV-2 infection as major drivers. 18These genetic studies suggest dysregulation of the inflammatory response as a characteristic of MIS-C, although they were limited to a small number of patients who were mostly of European origin.
We performed whole exome sequencing in 45 patients, primarily of Middle Eastern origin, with MIS-C and 25 healthy children who were infected with SARS-CoV-2 but remained asymptomatic or developed only mild clinical symptoms.We hypothesized that rare, deleterious genetic variants affecting immune-related genes would be enriched in patients of Middle Eastern origin with MIS-C compared with matched controls and that such variation would contribute to disease onset, symptoms, clinical outcomes, and/or management.

Study Design and Participants
This cohort study, conducted from September 1, 2020, to August 31, 2021, was approved by the Dubai Scientific Research Ethics Committee-Dubai Health Authority and the institutional review board of the Specialty Hospital, Jordan.Patients (and their guardians) recruited in Dubai or Jordan provided written informed consent for their deidentified data to be used for research, and this study was performed in accordance with the relevant laws and regulations that govern research in both countries.This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. 19 recruited 70 children between 1 day and 18 years of age and divided them into 2 groups for this study.The first group included 45 patients diagnosed with MIS-C.Inclusion criteria for this group were consistent with the case definition of MIS-C set by the World Health Organization or Center for Disease Control and Prevention (eTable 1 in the Supplement).All patients had evidence of SARS-CoV-2 infection by RT-PCR, SARS-CoV-2 antibodies, or recent exposure to a confirmed case of COVID-19.At least 2 organs were affected by the illness, and the blood profiles of these patients revealed increased inflammatory markers.Exclusion criteria were applied to all patients with another diagnosis that could affect their disease course, such as congenital heart disease, failure to thrive, or other syndromes.The second group was recruited at the same time and included 25 healthy children (control group) (eFigure 1 in the Supplement) who had a SARS-CoV-2 infection confirmed by RT-PCR but were asymptomatic or experiencing mild symptoms.Individuals in the control group were followed up for 12 weeks to ensure that no signs of MIS-C disease were detected.

Clinical and Demographic Information
Deidentified patient information was extracted from electronic medical records at Al Jalila Children's Hospital (Dubai, United Arab Emirates), Latifah Hospital (Dubai, United Arab Emirates), and the Specialty Hospital (Jordan), where those patients were treated and recruited for the study.Variables included demographic information, signs and symptoms on admission, inflammatory markers with cytokine profile, cardiac manifestation, course of illness, admission to the pediatric intensive care unit (PICU), treatment, and outcome.Patient's country of origin was determined through hospital medical records.This study focused exclusively on patients from Arab countries, which we define as all Middle Eastern (Arab-speaking) countries except Iran and Turkey.

Whole Exome Sequencing
Whole exome sequencing was performed in the genomic laboratory at Al Jalila Children's Hospital (Figure 1).DNA was extracted from peripheral blood cells using standard DNA extraction protocols (Qiagen).After fragmentation by ultrasonication (Covaris), genomic DNA was processed to generate sequencing-ready libraries of short fragments (300-400 bp) using the SureSelect kit (Agilent).RNA baits targeting all coding regions were used to enrich for whole exome regions using the SureSelect Clinical Research Exome V2 kit (Agilent).The enriched libraries underwent next-generation sequencing (2 × 150 bp) using the SP flow cell and the NovaSeq platform (Illumina).

Bioinformatics Analysis and Variant Filtration
Sequencing data were processed using an in-house custom-made bioinformatics pipeline to retain high-quality sequencing reads with greater than 10 × coverage across all coding regions.High-quality variants with read depths greater than 10, genotypic quality greater than 30, and mapping quality greater than 60 were retained for downstream analysis. 202][23][24] We used 3 filters to retain (1)   21,22 (eTable 2 in the Supplement) and gnomAD allele frequency less than 0.5%.We applied similar filters to whole exome sequencing data from controls (n = 25) and compared these variants between patients with MIS-C and controls.

Enrichment Analysis
Enrichment of rare, likely deleterious genetic variants in patients with MIS-C was determined by comparing the proportion of individuals with at least 1 functional allele (nonsense, frameshift, missense, and canonical +1 and +2 residues at the 5′ donor splice site and −1 and −2 residues at the 3′ acceptor splice site) in the MIS-C group with that in controls.P values were calculated using the Fisher exact test (GraphPad Prism, version 9.2.0;GraphPad Software Inc).Similarly, the number of individuals with more than 1 heterozygous variant (double heterozygotes) were compared between the MIS-C and control groups as another measure of the burden of genetic variation in patients.
Furthermore, to rule out the possibility that the genes affected with truncating or LoF variants in patients with MIS-C were simply tolerant to such variation, we quantified the fractions of such variants in those genes in general populations, namely gnomAD, 25 Middle East Variome database 26 (created in-house by assembling sequencing data from Qatar 27 and the Greater Middle East

Pathway and Protein-Protein Interaction Analysis
String 29 was used for pathway enrichment and protein-protein interaction (PPI) network analysis.
The network nodes correspond to proteins, whereas the edges represent confidence based on available experimental evidence and expert-curated databases.We applied an interaction score of highest confidence (0.9) to avoid any false-positive interactions among proteins.

Statistical Analysis
We divided patients into the 2 groups based on the presence or absence of genetic variation and     28 and a Qatari cohort (N = 1005), 27 or in our internal cohort of healthy individuals (N = 124) (Figure 3A).Finally, using a similar filtration strategy in the control group, we detected only 3 truncating variants (eTable 6 in the Supplement) across all 186 genes (vs 20 in patients with MIS-C), further confirming the higher burden of such variants in the MIS-C group (Figure 3B).Furthermore, 6 of the 14 patients with MIS-C with truncating variants harbored 2 such variants in 2 different genes (Figure 3B and  ).
We performed a similar enrichment analysis on the set of patients with MIS-C (n = 26) and controls (n = 24) from Arab countries of origin.There was a higher burden of LoF variants in Arab patients with MIS-C relative to other population databases (Figure 3C).More importantly, there was an enrichment of LoF and missense variants in patients with MIS-C when compared with Arabmatched controls, ruling out a possible genetic background effect (Figure 3D).

Pathway and Protein Network Analysis
Genes with enriched variants belonged to the Toll-like receptor signaling, retinoic acid-inducible gene I-like receptor signaling, interferon-mediated immune response, and NOD-like receptor signaling pathways as the main pathways altered in patients with MIS-C (Table ).We

Genetic Associations With Clinical and Demographic Characteristics of Patients With MIS-C
We tested the association between the genetic status of patients with MIS-C and their demographic, clinical, and laboratory findings (Figure 1).On the basis of age distribution of patients with MIS-C  (eFigure 4 in the Supplement), we observed that patients who were younger than 3 years were more likely to harbor a rare variant in the immune-related genes (7 patients with genetic findings [36.8%; 95% CI, 19.1%-58.9%]vs 2 without genetic findings [7.7%; 95% CI, 2.1%-24.1%]were younger than 3 years at onset; P = .02,Fisher exact test) (Figure 4B).No significant clinical association was found between genetic findings and symptoms (eTable 7 in the Supplement).Most laboratory markers were similarly abnormal in all patients with MIS-C; those with a genetic finding had higher fibrinogen levels, but this finding was not statistically significant (eTable 5 in the Supplement).Similarly, no associations were obtained when analysis was focused only on patients with MIS-C who tested positive for SARS-CoV-2 (n = 36) while excluding those with only exposure to COVID-19 (eTables 8 and 9 in the Supplement).

Management, Hospital Course, and Outcomes
All patients recovered after a mean (SD) length of stay of 6.

Discussion
In this cohort study, we characterize the genomic landscape, phenotypic features, inflammatory and cellular markers, and clinical management and outcomes of the largest prospective cohort to date of patients with MIS-C from the Middle East.To our knowledge, this study is also the largest MIS-C cohort involving whole exome sequencing data.Our cohort included patients who were mostly from the Middle East (66.7%),including 26 from Arab countries and 9 from Asian countries.Patients of such backgrounds have long been underrepresented in genetic studies, emphasizing the importance of our study in characterizing the genetic landscape of MIS-C disease in this cohort.Although clinical presentations and laboratory markers in this cohort were consistent with recently described MIS-C cohorts elsewhere, 7 our analysis revealed significant enrichment of rare, likely deleterious variants mainly affecting the Toll-like receptor signaling, retinoic acid-inducible gene I-like receptor, NOD-like receptor signaling, and interferon-mediated immune response pathways in patients relative to ageand genetically matched controls and to the general population of diverse origins (including our internal cohort of healthy individuals).Those pathways overlap with the currently characterized immunologic profile in patients with MIS-C discussed in the Introduction. 12,13 addition, transcriptomic profiling of patients with severe COVID-19 found upregulation of cytokines and chemokines, including CCL2, CCL22, CXCL9, CCL2/MCP-1, CXCL10/IP-10, CCL3/MIP-1A, and CCL4/MIP1B and CXCL12, and certain interferons and interleukins, including IFIH1, IFI44, IFIT1, and IL10. 33,34Similarly, single-cell transcriptomic profiling of older individuals with SARS-CoV-2 identified type 1 and 2 interferon deficiencies and reduced expression of antiviral defense genes in certain cell types, indicating decreased immune ability in response to SARS-CoV-2 infection in older individuals. 35All these studies [33][34][35] and our current genetic findings indicate possible mechanistic convergence between MIS-C and severe COVID-19 on key immune-related pathways.Nonetheless, although our genetic analysis revealed some overlap with the type 1 IFN pathway, shown to be altered in patients with severe COVID-19, 18 most of the identified variants, especially truncating ones, in patients with MIS-C affect genes that do not overlap with those in patients with severe COVID-19.
These results suggest that MIS-C and severe COVID-19 have distinct genetic determinants while possibly altering similar inflammatory pathways.
Our genetic analysis also revealed that 7 of the 19 patients (36.8) had more than 1 variant, including 2 patients with 4 and 3 variants, respectively, further highlighting the significant burden of genetic variation in this cohort.Despite the relatively high rate of consanguineous marriage in this part of the world, our analysis did not identify any rare homozygous variants that might predispose individuals to MIS-C.It is possible that such variants might lead to severe immune-related disorders irrespective of SARS-CoV-2 infection.Patients with such genetic variants or disorders would have been excluded from our study because almost all our patients did not have significant medical history before viral infection.

Limitations
This study has some limitations.Our analysis showed that patients with genetic findings tend to have earlier onset of disease (younger than 3 years) and possible resistance to IVIG treatment; however, no significant associations were found between inflammatory markers or clinical symptoms and genetic findings.Such associations cannot be ruled out for several reasons.First, our sample size might not be empowered to detect such associations, and our patient population was primarily of Middle Eastern origin.Second, our study was designed to mainly capture rare, relatively large effect variants in selected genes (N = 186).Therefore, genetic contribution from noncoding variants or deleterious variants outside the 186 studied genes or attributable to polygenic, small-effect variants across the genome cannot be ruled out.Larger sample sizes are still needed to capture such smalleffect genetic contributions.
Another limitation of this study is the lack of functional analyses to characterize the mechanism(s) through which the mutated genes contribute to MIS-C disease onset and/or progression.However, most variants are highly concentrated to genes in the IFN and Toll-like receptor pathways, suggesting that disturbance of such pathways might underlie the cytokine storms and dysregulated inflammatory markers in those patients.Protein-protein network analysis further confirms the significant interaction and convergence of most mutated genes in this study.

JAMA Network Open | Genetics and Genomics
28 ), and an internal local cohort (exome sequencing data from healthy parents in the United Arab Emirates), and compared these data with those for patients with MIS-C.The aggregate LoF allele fraction was obtained by summing the total LoF allele fraction in each of the above genes (total LoF allele counts divided by maximum allele number in the database).The Fisher exact test was performed using GraphPad Prism software, version 9.2.0.

Figure 1 .
Figure 1.Graphical Representation of Study Design, Participants, Sequencing Protocol, Bioinformatic Analysis, and Genomic and Clinical Characterization of Patients With Multisystem Inflammatory Syndrome in Children (MIS-C) 4 [4.0]years) were prospectively recruited into the cohort (eTable 1 in the Supplement).Thirty-six patients (80.0%) had evidence of SARS-CoV-2 infection by RT-PCR or antibody testing, whereas 9 (20.0%) had exposure to the virus through contact with patients with COVID-19.The mean (SD) duration of fever was 5 (2.4) days (eTables 3 and 4 in the Supplement), and all patients had laboratory evidence of inflammation, including lymphopenia and/or significantly elevated inflammatory markers, such as C-reactive protein, erythrocyte sedimentation rate, D-dimer, ferritin, fibrinogen, and IL-6 (eTable 5 in the Supplement).

Figure 2 .a
Figure 2. Age, Sex, and Country of Origin of Patients With Multisystem Inflammatory Syndrome in Children obtained a PPI network, based on experimental evidence and expert-curated databases, with 10 nodes representing proteins and 23 edges representing confidence level, with a PPI enrichment P < 1.0 −16 .Interactions among IFNB1, IFNA21, IFNA4, IFNA6, IFNAR2, TRAF3, IRF3, and TLR3 indicate the role of IFN-mediated immune responses (eFigure 3 in the Supplement).

Figure 3 .
Figure 3. Burden of Immune-Related Loss of Function (LoF) and Missense Variants in Patients With Multisystem Inflammatory Syndrome in Children (MIS-C)

Figure 4 .
Figure 4. Genetic Findings and Associations With Age and Response to Treatment

eTable 2 . 5 . 7 . 8 .
Immune-Related Genes Used to Filter Rare Coding Variants (Yellow-Highlighted Genes Were Recently Implicated in Severe COVID-19) eTable 3. Clinical Characteristics of MIS-C Cohort (N = 45) eTable 4. All MIS-C Cases in This Study Along With Evidence of SARS-Cov-2 Infection, Fever, and Inflammatory Markers eTable Inflammatory Markers in MIS-C Patients on Admission eTable 6. Variants Detected in Control Group eTable Clinical and Pathological Findings eTable Clinical and Pathological Findings of MIS-C Patients With Positive SARS-CoV-2 Status Only (n=36) eTable 9. Inflammatory Markers of MIS-C Patients With Positive SARS-CoV-2 Status Only (n=36) eTable 10.Treatment, Management, and Outcomes

JAMA Network Open | Genetics and Genomics
truncating or loss of function (LoF) variants in the 186

Downloaded From: https://jamanetwork.com/ on 09/24/2023
Categorical variables were compared using the Fisher exact test.Tests were 2-tailed, and P < .05 was considered statistically significant.Given the potential for type I error as a result of multiple comparisons, findings should be interpreted as exploratory.

Table .
Genetic Findings in Patients With Multisystem Inflammatory Syndrome in Children

Table .
Genetic Findings in Patients With Multisystem Inflammatory Syndrome in Children (continued)