key: cord-0702647-7rfrvdyu
authors: Kumar, Deepak; Rostad, Christina A.; Jaggi, Preeti; Villacis Nunez, D. Sofia; Chengyu Prince,; Lu, Austin; Hussaini, Laila; Nguyen, Thinh H.; Malik, Sakshi; Ponder, Lori A.; Shenoy, Sreekala PV.; Anderson, Evan J.; Briones, Michael; Sanz, Ignacio; Prahalad, Sampath; Chandrakasan, Shanmuganathan
title: Distinguishing immune activation and inflammatory signatures of multisystem inflammatory syndrome in children (MIS-C) versus hemophagocytic lymphohistiocytosis
date: 2022-03-15
journal: J Allergy Clin Immunol
DOI: 10.1016/j.jaci.2022.02.028
sha: a2a8ae06768679848c2c712593a2cdf61018d06e
doc_id: 702647
cord_uid: 7rfrvdyu

Background Multisystem inflammatory syndrome in children (MIS-C) is a potentially life-threatening sequela of SARS-COV-2 infection characterized by hyperinflammation and multi-organ dysfunction. Though hyperinflammation is a prominent manifestation of MIS-C, there is limited understanding of how the inflammatory state of MIS-C differs from well characterized hyperinflammatory syndromes such as hemophagocytic lymphohistiocytosis (HLH). Objectives To compare the qualitative and quantitative inflammatory profile differences between MIS-C, COVID-19 and HLH patients. Methods Clinical data abstraction from patient charts, T cell immunophenotyping and multiplex cytokine and chemokine profiling were performed for MIS-C, COVID-19 and HLH patients. Results We found that both MIS-C and HLH showed robust T cell activation, markers of senescence, and exhaustion along with elevated Th1 and pro-inflammatory cytokines such as IFN-γ, CXCL9, and CXCL10. In comparison, the amplitude of T cell activation and the levels of cytokines/chemokines were higher in HLH when compared to MIS-C. Distinguishing inflammatory features of MIS-C included elevation in Th2 inflammatory cytokines such as IL-4 and IL-13 and cytokine mediators of angiogenesis, vascular injury, and tissue repair such as VEGF-A and PDGF. Immune activation and hypercytokinemia in MIS-C resolved at follow-up. In addition, when these immune parameters were correlated with clinical parameters, CD8+ T cell activation correlated with cardiac dysfunction parameters such as BNP and troponin and inversely correlated with platelet count. Conclusion Overall, this study characterizes unique and overlapping immunological features that help to define the hyperinflammation associated with MIS-C versus HLH.

S.C. serves on the advisory committee for SOBI. S.P. serves on a macrophage activation syndrome 48 adjudication committee for Novartis Pharmaceuticals and is an Op-Med fellow for Doximity for 49 2021-22. E.J.A has consulted for Pfizer, Sanofi Pasteur, Janssen, and Medscape, and his institution 50 receives funds to conduct clinical research unrelated to this manuscript from MedImmune, 51

Regeneron, PaxVax, Pfizer, GSK, Merck, Sanofi-Pasteur, Janssen, and Micron. He also serves on 52 a safety monitoring board for Kentucky BioProcessing, Inc. and Sanofi Pasteur. His institution has 53 also received funding from NIH to conduct clinical trials of Moderna and Janssen COVID-19 54 vaccines. C.A.R.'s institution has received funds to conduct clinical research unrelated to this 55 manuscript from BioFire Inc, GSK, MedImmune, Micron, Janssen, Merck, Moderna, Novavax, 56 PaxVax, Pfizer, Regeneron and Sanofi-Pasteur. The rest of the authors declare that they have no 57 relevant conflicts of interest. 58 5 

Peripheral blood mononuclear cells were used for immunophenotyping. Antibody information is 143 provided in supplemental data. Flow cytometry data was acquired on BD FACSymphony™ A5 144 and analyzed using FlowJo software v10. T cell activation was defined by co-expression of 145 HLADR + and CD38 + on effector memory (EM) CD4 + and CD8 + T cells [18] [19] [20] . Co-expression of PD-146 1 + and Tim3 + on EM CD4 + and CD8 + T cells were defined as markers for exhaustion 21 . T cell 147 senescence was assessed by expression of CD57 + on EM population of CD4 + and CD8 + T cells 22-148 25 . Definitions of T cell populations analyzed in this study is shown in Table E1 . 149

Plasma cytokine/chemokine profiling 150

Cytokine/ chemokine profiling of plasma samples from COVID-19 (n = 10), HLH (n = 8), MISC 151 (n = 19), MIS-C follow-up (n =10), and HCs (n = 19) was performed on a luminex platform. The 152 plasma samples for cytokine/ chemokine profiling were chosen randomly from a cohort of 69 153 MIS-C and 24 COVID-19 patients. A list of these analytes is provided in Table E2 . Soluble 154 interleukin-2 receptor (sIL-2R) and ferritin levels were done in clinical lab as standard of care. 155

Plasma levels of sCD163 were measured using human CD163 ELISA kit (Abcam, ab274394) as 156 per manufacturer's instructions. 157

Statistics 158

Cytokine/chemokine data was represented as dot plots showing all the data points using Graphpad 159

Prism version 9. One-way analysis of variance (ANOVA) with multiple comparison test was used 160

for calculating significant differences between groups. Frequencies of immune parameters were 161

represented with median and interquartile range (IQR). Fisher's exact test was used to compute 162 significance for categorical data. We applied a principal component analysis on all the cytokines 163

showing differences among patient groups based on PC1 and PC2 and created using Factoextra R 164

package. Heatmap of all the reported cytokines was generated in pheatmap R package. Spearman 165 correlation was used to find the correlation between clinical and flow-based parameters;correlation 166 coefficients and p-values were reported. Correlation matrix was made in "corrplot" package in R. 167 168

We enrolled 69 patients with MIS-C (Ages 2-19 years, Median = 11 years), 24 patients hospitalized 171

with COVID-19 but without MIS-C (Ages 2-17 years, Median = 11.5 years), 13 patients with HLH 172

(Ages 1 day-19.3 years, Median = 1.2 years), and 22 HCs (Ages 8-25 years, Median = 17 years). 173

The demographics, clinical parameters and standard of care laboratory tests for patients with MIS-174 C, COVID-19 and HLH are detailed in Table 1 and Table E3 . A follow-up immune evaluation was 175 performed for 31 patients with MIS-C, median follow-up of 2 months post diagnosis of MIS-C 176

(range 1 to 7 months) ( Fig E1) . 177 71% of patients with MIS-C required intensive care unit (ICU) care whereas 58.3% of patients 178

with COVID-19 were admitted to ICU. Most MIS-C patients were treated with steroids and IVIG 179 J o u r n a l P r e -p r o o f 6 ( Fig E2) . In our cohort, children with MIS-C had higher ferritin, CRP, neutrophil count and 180 thrombocytopenia compared to those with COVID-19 (Table 1) , consistent with the previously 181 published reports 6, 11, 26, 27 . 182

To investigate how qualitatively the inflammatory response of MIS-C differs from HLH patients 184 in plasma, we performed multiplex cytokine/chemokine profiling of 66 analytes in MIS-C (n = 185 19), MIS-C follow-up (n = 10), COVID-19 (n = 10), HLH (n = 8) and HCs (n = 19) (Fig 1, A) . 186 PCA analysis of all these cytokines showed that both HLH and MIS-C formed distinct clusters in 187 comparison to COVID-19, HCs, and follow-up MIS-C patients (Fig 1, B) . 188

We observed a cluster of 17 cytokines i.e., CXCL9, CXCL10, TNF, 15, IL-18, IL-27, IL-1α, IL-1RA, CXCL13, FLT-3L, M-CSF, MIP-1β and MCP-2, that were 190 significantly elevated in both MIS-C and HLH patients. A decreased level of MDC or CCL22 was 191 observed in patients with HLH and MIS-C. Out of these cytokines, concentrations of CXCL9, 6, IL-8, IL-18 and M-CSF were higher in HLH as compared to Next, we also looked at the unique cytokine signature in patients with 194 significant elevation of VEGF-A, sCD40L, IFN-α2, IL-4, IL-13, PDGF-AA and TARC were 195 found only in MIS-C but not in HLH. On the other hand, Eotaxin, GRO-α, MCP-1, RANTES, 196 I309, CTACK, MIP-1δ and IL-12p40 were significantly increased in patients with HLH but not 197 MIS-C. Cytokines such as MIP-1α, TGF-α, LT-α, Eotaxin-2, Eotraxin-3 and IL-16 were 198 significantly elevated in MIS-C as compared to controls, but we did not find significant difference 199 between HLH and MIS-C. Similarly, fractalkine, G-CSF and TPO were significantly elevated in 200 HLH as compared to HCs but did not show differences with respect to MIS-C. In children with 201 COVID-19, a limited number of cytokines such as IL-1RA, IL-6, PDGF-AA, MIP-1β and MCP2 202 were significantly elevated. 203

Next, we sought to investigate the cytokine families that were differentially regulated in these 204 cohorts. In general, patients with MIS-C showed prominent elevations in different cytokine 205 families (Fig 2, A-F). Cytokines and chemokines related to T cell activation such as 206 TNF, CXCL9 and CXCL10 were significantly elevated in both MIS-C and HLH groups. However, 207 the amplitude of these cytokines was higher in HLH when compared to MIS-C (Fig 2, A) . 208

Interestingly, cytokine mediators responsible for angiogenesis, vascular injury, and tissue repair 209 such as VEGF-A (p < 0.0001), PDGF-AA (p < 0.05), PDGF-AA/AB (p = 0.07) and FGF-2 (p = 210 0.06) were elevated in MIS-C when compared to HCs but not in HLH patients (Fig 2, E) . Increased 211 levels of these cytokines were also observed in some COVID-19 patients, although overall we did 212 not find significance between COVID-19 and HCs cohorts except for PDGF-AA (p = 0.04). Th2 213 inflammatory cytokines (i.e., IL-4, IL-13) were significantly elevated in patients with MIS-C but 214 not in patients with HLH. All the remaining cytokines are shown in Fig E3. As previously reported, 215

we also evaluated whether the values of TNF and IL-10 in combination can differentiate between 216

MIS-C and COVID-19 in our cohorts 28 . We found similar observations where sum of the plasma 217 levels of TNF and IL-10 were significantly elevated in MIS-C versus COVID-19 (Fig E3 ext. ) 218 7 Overall, these data demonstrate that patients with MIS-C and HLH have some overlap in the 219 inflammatory milieu, however the amplitude of the inflammation is much higher in HLH. 220

Additionally, differential expression of certain inflammatory cytokine and chemokines in MIS-C 221

suggest unique inflammatory pathways that are active in MIS-C but not in HLH. 222

Increased T cell activation has been reported in both pediatric . A  224 profound elevation of IFN-γ and its induced chemokines i.e., CXCL9 and CXCL10 32-34 and other 225 inflammatory markers such as IL-6 and TNF in patients with HLH when compared to patients with 226 MIS-C suggests higher T cell activation in HLH in comparison to MIS-C. In order to further 227 strengthen these observations and to investigate how T cell activation in compares to that of HLH, we assessed the expression of activation markers on CD4 + and CD8 + T 229 cell subsets. The gating strategy for these analyses is shown in Fig E4. We first evaluated HLA-230 DR + CD38 + expression in EM compartment of CD4 + and CD8 + T cells. CD8 + EM T cell activation 231 was noted in both COVID-19 and MIS-C patients. Whereas in COVID-19 the T cell activation 232 was modest (5-fold, p = 0.0022), MIS-C showed higher activation (13-fold, p < 0.0001) when 233 compared to median HCs. Quantitatively, the amplitude of CD8 + EM T cell activation in HLH was 234 2-fold greater than MIS-C. Although CD8 + EM T cell activation was significantly higher in HLH 235

than MIS-C, we observed a subset of MIS-C patients had similar CD8 + EM T cell activation as of 236 HLH patients (Fig 3, A and C) . Similarly, although the CD4 + EM T compartment was activated in 237

both MIS-C and HLH patients (Fig 3, B and D), CD4 + EM T cell activation was 6 times higher in 238 HLH than MIS-C. Using receiver operating characteristic (ROC) statistics, we calculated optimal 239 threshold value for CD8 + EM T cell activation that can differentiate MIS-C from HCs, and HLH with high sensitivity and specificity ( Fig E5, A provide a more global picture of T cell activation, we also evaluated HLA-DR + CD38 + expression 246 on central memory (CM) and TEMRA populations of CD4 + and CD8 + T cells and also on total 247

CD4 + and CD8 + cells. A similar trend of activation was observed as seen previously in EM 248 compartment ( Fig E6) . 249 T cell activation measured by the expression of HLA-DR + PD-1 + on EM compartment of CD4 + 250 and CD8 + T cells also showed similar trends noted with HLA-DR + and CD38 + expression. The 251 expression of HLA-DR + PD-1 + was significantly elevated both in CD4 + and CD8 + T cells in 252 patients with MIS-C as well as HLH, however, expression of HLA-DR + PD-1 + was much higher 253 in HLH as compared to MIS-C patients (Fig 3, E and F). Since we observed increased activation 254 in CD8 + T EM compartment in MIS-C cohorts, we aimed to further evaluate if increased activation 255 in CD8 EM compartment resulted in relative expansion of CD8 + EM compartment. Therefore, we 256 calculated the ratio of the frequencies of EM and naïve compartment. Interestingly, we found a 257 relative expansion of EM compartment in HLH but not in MIS-C (Fig E7, A) . CD4 + /CD8 + ratio 258 was similar between MIS-C and HLH. When compared to MIS-C, COVID-19 had overall elevated 259 CD4 + /CD8 + ratio, although non-significant (Fig E7, B) . Next, we also evaluated CD4 + and CD8 + 260 J o u r n a l P r e -p r o o f 8   TEMRA populations. Although, TEMRA populations were similar among different patient  261 cohorts, we found an increase of this subset in MIS-C follow-up patients when compared to onset 262 MIS-C patients (Fig E7, C and D) . 263

senescence 265

Patients with MIS-C have been reported to have prolonged presence of SARS-CoV-2 in the GI 266 tract 35 ; thus, we hypothesized that chronic antigenic exposure in MIS-C and associated T cell 267 activation could potentially lead to a post-activation exhaustion state of T cells and show features 268 of proliferation induced senescence. To test these hypotheses, we evaluated the expression of T 269 cell exhaustion and senescence surface markers in CD4 + and CD8 + EM T cells for these patient 270 cohorts. T cell exhaustion was evaluated by co-expression of PD-1 + and Tim3 + on EM CD4 + and 271

CD8 + T cells. Significant increase in T cell exhaustion markers was observed in both MIS-C and 272

HLH patients in both CD4 + and CD8 + EM T cells, however, the frequency of cells expressing 273 exhaustion markers was much higher in HLH when compared to MIS-C (Fig E7, E and F). We 274 also observed a modest increase in exhaustion markers in COVID-19 patients when compared to 275

HCs in both CD4 + and CD8 + T EM compartments. Also, patients with MIS-C displayed 276

significantly higher CD57 + expression on CD8 + EM cells (Fig E7, G) . Interestingly, we did not 277 find any difference in CD57 + expression for COVID-19 and HLH cohort when compared to HCs. 278

Similarly, we found a significant increase in the expression of CD57 + on CD4 + EM T cells in 279 patients with MIS-C (Fig E7, H) individual laboratory parameters that were consistently available for these patients were assessed. 299

A direct comparison of standard parameters of HLH such as ferritin, sIL-2R and sCD163 revealed 300 that patients with HLH had significantly higher plasma levels of ferritin, sIL-2R and sCD163 when 301 compared to patients with MIS-C whereas patients with COVID-19 had lower levels of ferritin, 302 sIL-2R and sCD163 when compared to patients with MIS-C (Fig 4, A-C) . These findings validate 303 the T cell activation and inflammatory signature as seen in our previous observations. 304

Neutrophil to lymphocyte ratio (NLR) has been proposed as a distinguishing parameter in and COVID-19 illness 26, 37 . As reported previously, NLR was higher in MIS-C when compared to 306 COVID-19, with a trend towards significance (median 8.2 vs 3.8, p = 0.06). However, this ratio 307 was significantly lower in patients with HLH compared to COVID-19 (median 0.34 vs 3.8, p = 308 0.005) and MIS-C (median 0.34 vs 8.2, p < 0.0001) (Fig 4, D) . In addition, to determine if the 309 neutrophil counts differ with respect to T cell activation in these patient cohorts, we assessed the 310 ratio of ANC with CD8 + and CD4 + EM T cell activation. We observed that these ratios could easily 311 distinguish these clinical entities as they were higher in COVID-19 when compared to MIS-C and 312 HLH, while HLH had significantly lower ratios when compared to 313 E and F). 314

While previous studies describing immunophenotypic differences in MIS-C have only focused on 316 the acute disease state at onset, follow-up studies are limited 26 . In our study, we assessed T cell 317

based immune markers as well as cytokine/chemokine profile during follow-up. Immune 318 evaluation of patients with MIS-C at follow up revealed significant decrease in both CD4 + and 319

CD8 + EM T cell activation with return to HCs (Fig 3, C-F and Fig E8) . In addition, we observed a 320 decrease in T cell exhaustion as well as senescence markers at follow-up in both CD4 and CD8 321 EM subsets (Fig E7, E-H and Fig E8) . In a subset of MIS-C patients, we also assessed the levels 322 of cytokines and chemokines at follow-up. Almost all of the elevated cytokines had resolved to 323 normal levels in patients with MIS-C (Fig 2, G) . We also assessed some clinical inflammatory 324 markers such as ferritin and CRP at hospital admission and after 7 days of follow-up in some 325 patients with MIS-C where longitudinal data was available and found a significant decrease in 326 their levels on follow-up, suggesting rapid response of systemic inflammation with treatment ( Fig  327  E8 , F). 328

MIS-C and severe COVID-19 have been found associated with rise in acute myocardial markers 330 such as BNP and troponin 26, 38-40 . To investigate whether T cell activation correlates with the 331 cardiac inflammatory markers and disease severity in COVID-19 and MIS-C, we first compared 332 the BNP and troponin levels in COVID-19 and MIS-C. In our cohort, we found that both BNP and 333 troponin are significantly higher in MIS-C as compared to COVID-19 validating the high 334 occurrence of cardiac dysfunction and cardiac injury in MIS-C ( Fig E9, A and B ). We next checked 335

whether there is an association of cardiac dysfunction markers with T cell activation in MIS-C and 336

COVID-19 patients. We found a correlation of troponin levels with CD8 + and CD4 + EM T cell 337 activation ( Fig 5, A for BNP (200 pg/ml; two times the upper limit of normal) and troponin (0.09 ng/ml; two times the 343 upper limit of normal) for calculating the odds ratio of finding high BNP or troponin in patients 344

with high CD8 + EM T cell activation. Our data showed that patients having higher CD8 + EM T 345 cell activation (>15.9%) were 9.1 times (95% CI: 2.7 to 30.1) more likely to have elevated BNP 346 levels (>200 pg/mL) than patients with low T cell activation. Similarly, patients with high CD8 + 347 EM T cell activation were 6.2 (95% CI: 1.8 to 21.3) times more likely to have high troponin (> 348 0.09 ng/mL) levels than the patients who had lower T cell activation. This suggests that T cell 349 activation might directly or indirectly contribute to cardiac pathology with elevation of BNP and 350 troponin levels in these children with MIS-C. 351 We also correlated other inflammatory markers with cardiac dysfunction markers in patients with 352

MIS-C and COVID-19. Ferritin and CRP levels correlated with both BNP and troponin levels ( Fig  353  5 , E-H). Ferritin and CRP levels also correlated with CD8 + EM T cell activation (Fig 6, C) . 354

Thrombocytopenia was frequent in MIS-C. We observed in patients with 355 platelet counts were inversely correlated with CD4 + and CD8 + EM T cell activation (Fig 6, A and  356 B). In addition to T cell activation, other inflammatory markers such as ferritin levels correlated 357 with elevation of liver enzyme alanine transaminase (ALT) and creatinine, but inversely correlated 358 with platelet count in these patients (Fig 6, C) . We also compared these parameters separately in 359

MIS-C and COVID-19 patients. Since steroid treatment might affect some of activation readout 360 and laboratory features over time, we just evaluated MIS-C patients where sampling was done 361 either pre-steroids or within first 48 hours of steroid initiation. We found similar correlations in 362

MIS-C as observed in MIS-C and COVID-19 patients. However, due to the limited number of 363 COVID-19 patients, we observed much lesser correlations between these variables ( Fig E10) . 364 365 Discussion 366 MIS-C is an immune dysregulation state characterized by hyperinflammation with multi-system 367 manifestations including myocarditis, cardiac dysfunction, respiratory failure, acute kidney injury, 368 or gastrointestinal, dermatologic, or neurological involvement 1, 17, 26, 41 . Initially, MIS-C was 369 identified in children, but later similar presentation was also reported in adults (MIS-A) 42-44 . 370 Although several groups have demonstrated an increase in inflammatory markers and T cell 371 activation in 28, 31, 45, 46 , the exact nature and amplitude of hyperinflammation is still 372 poorly defined. Hence a comparison with an established hyper-inflammatory state such as HLH 373 offers additional insight on the immunopathogenesis of MIS-C. A number of similarities exist, but 374 there are also qualitative and quantitative differences in clinical presentation and management of 375 patients with MIS-C and HLH. For example, hyperinflammation in MIS-C has been treated with 376 steroids, IL-1RA (anakinra) and IL-6 blocking antibodies 16, 47 . Similar cytokine blockade and 377 steroids are used for the management of patients with secondary forms of HLH 48 . Although there 378 are similarities between MIS-C and HLH, they differ in some clinical manifestations such as 379 presence of myocarditis leading to cardiac dysfunction, and gastrointestinal manifestation like 380 acute abdomen or inflammatory bowel disease like presentations, which are common in MIS-C 381 but usually not a part of the disease process in HLH 14, 48 . Additionally, pancytopenia and liver 382 function test abnormities are more commonly seen in HLH and infrequent in MIS-C. We 383 hypothesized that despite certain clinical similarities, the amplitude and nature of 384 hyperinflammation might be different in MIS-C when compared to HLH. 385 386 J o u r n a l P r e -p r o o f 11 We found several similarities as well as striking differences between MIS-C and HLH. High T cell 387 activation was found in both MIS-C and HLH, however, degree of T cell activation was lower in 388 MIS-C when compared with HLH. Despite of higher T cell activation in HLH, we found some 389 patients with MIS-C having CD8 + EM T activation comparable to patients with HLH. Although 390 CD4 versus CD8 ratio was similar in MIS-C and HLH, high CD8 + EM/ naïve ratio was observed 391

in HLH, but not in MIS-C, suggesting more dramatic EM T cell expansion in HLH. Though as a 392

group, children with COVID-19 had a modest increase in T cell activation, this finding was not 393 universal in all hospitalized children with COVID-19. In fact, majority of these children had only 394 modest T cell activation when compared to control, suggesting that unlike in MIS-C, T cell 395 activation may not be an important driver of disease morbidity in COVID-19 in children. Though 396

we have shown upregulation of surface markers of exhaustion and senescence in T cells, the fact 397 that they resolve during follow-up could suggest a post-activation state induced transient 398 upregulation of exhaustion markers rather than true exhaustion of T cells 49 . Similarly, transient 399 proliferative stress in EM compartment of T cells could be the cause of upregulation of senescence 400 marker. 401

Despite increased T cell activation in MIS-C, the qualitative nature of T cell activation is still 402 poorly understood. Hence, we performed additional evaluation of plasma cytokines and 403 chemokines in all patient cohorts. We found elevated plasma IFN-γ and associated chemokines 404 (i.e., CXCL9, CXCL10) in both MIS-C and HLH. Interestingly, differences in degree of T cell 405 activation were also seen at the cytokine levels where median values for IFN-γ, CXCL9 and 406 CXCL10 were high in HLH when compared with MIS-C. In addition to these cytokines, elevated 407 IL-6 and TNF might be responsible for amplifying cytokine storm in patients with MIS-C. Innate 408 inflammation can also be a driver of T cell activation 50, 51 . Elevated innate inflammatory cytokines 409 such as IL-18, IL-15, IL-1α, IL-1RA in MIS-C suggest innate inflammatory pathways upregulation 410 could play a role in modulating the T cell activation noted in MIS-C. 411

Steroids and IVIG remain the mainstay of therapy for MIS-C 15, 16, 52 . However, additional biologics 412

were needed for a subset of severely ill patients. Anakinra, tocilizumab and infliximab are among 413 the common biologics used in the steroid-refractory settings 27, 53-55 . Increase in plasma IL-1RA 56 414

and IL-18 57, 58 in patients with MIS-C suggest significant autoinflammatory component of this 415

disease and reinforces the use of IL-1 blockade therapies such as anakinra in MIS-C 59 . Increase in 416

innate inflammatory signature and elevation of IL-6 and TNF in children with MIS-C noted in our 417 study and previous reports would support the use of IL-6 and TNF blockade 28, 45, 46, 55, 60 . As these 418

biologics are broadly used even in adults with severe COVID-19 61-63 , medication shortages have 419 become a concern. In this context, expanding our armamentarium of drugs available to manage 420 MIS-C hyperinflammation would be advantageous. 

In this study, we characterized the similarities and differences in hyperinflammatory states of MIS-482 C and HLH. We found high T cell activation and Th1 type inflammation in both MIS-C and HLH, 483

however, the amplitude of T cell activation and Th1 cytokines was higher in HLH versus MIS-C. 484

In addition to Th1, elevation of Th2 type and angiogenic cytokines and chemokines was unique to 485 MIS-C. We also found that T cell activation as well as other clinical parameters such as ferritin 486

and CRP correlated with cardiac dysfunction markers. Importantly, the hyperinflammation in MIS- Kruskal-Wallis one-way ANOVA followed by Dunn's multiple comparison test for non-normally 779 distributed samples and ordinary one-way ANOVA followed by Tukey's multiple comparison test 780

for normally distributed samples were used for statistical comparison. *p < 0.05, **p < 0.01, ***p 781 < 0.001; ****P < 0.0001; ns, not significant. 782 group. Kruskal-Wallis one-way ANOVA followed by Dunn's multiple comparison test for non-789 normally distributed samples and ordinary one-way ANOVA followed by Tukey's multiple 790 comparison test for normally distributed samples were used for statistical comparison. ***p < 791 0.001; ****P < 0.0001; ns, not significant. 792 793 showing ratio of ANC with CD8 + and CD4 + EM T cell activation. 798 

Population Markers References Central memory (CM) CD4 + T CD4 + CCR7 + CD45RA -(1) Central memory (CM) CD8 + T CD8 + CCR7 + CD45RA -(1) Effector memory (EM) CD4 + T CD4 + CCR7 -CD45RA -(1) Effector memory (EM) CD8 + T CD8 + CCR7 -CD45RA -(1) TEMRA CD8 + T CD8 + CCR7 -CD45RA + (1) TEMRA CD4 + T CD4 + CCR7 -CD45RA + (1) Naïve CD4 + T CD4 + CCR7 + CD45RA + (1) Naïve CD8 + T CD8 + CCR7 + CD45RA + (1) Activated CD4 + EM CD4 + CD45RA -CCR7 -HLA-DR + CD38 + (2-4) Activated CD8 + EM CD8 + CD45RA -CCR7 -HLA-DR + CD38 + (2-4) Senescent CD4 + T CD4 + CD45RA -CCR7 -CD57 + (5-8) Senescent CD8 + T CD8 + CD45RA -CCR7 -CD57 + (5-8) Exhausted CD4 + CD4 + CD45RA -CCR7 -PD-1 + Tim3 + (9) Exhausted CD8 + CD8 + CD45RA -CCR7 -PD-1 + Tim3 + (9) MISC-Pt1

Days from steroids to sample * * * * * 

Why is COVID-19 so mild in children?

511 Incidence of Multisystem Inflammatory Syndrome in Children Among US Persons 512 Infected With SARS-CoV-2

Accumulating evidence suggests 686 anti-TNF therapy needs to be given trial priority in COVID-19 treatment

Trials of 689 anti-tumour necrosis factor therapy for COVID-19 are urgently needed

Anti-692 IL6 treatment of serious COVID-19 disease: A monocentric retrospective experience

Baricitinib restrains 695 the immune dysregulation in patients with severe COVID-19

Baricitinib plus Remdesivir for Hospitalized Adults with Covid-19

Janus kinase inhibitors and 701 major COVID-19 outcomes: time to forget the two faces of Janus! A meta-analysis of 702 randomized controlled trials

Compassionate use of JAK1/2 inhibitor ruxolitinib for severe COVID-19: a prospective 705 observational study

Ruxolitinib in adult patients with secondary haemophagocytic lymphohistiocytosis: an 708 open-label, single-centre, pilot trial

Emapalumab in 710 Children with Primary Hemophagocytic Lymphohistiocytosis

Interleukin-18 induces 713 the production of vascular endothelial growth factor (VEGF) in rheumatoid arthritis 714 synovial fibroblasts via AP-1-dependent pathways

The 716 relationship between severity of coronary artery disease and plasma level of vascular 717 endothelial growth factor

Increased 719 complement activation is a distinctive feature of severe SARS-CoV-2 infection

Complement 722 associated microvascular injury and thrombosis in the pathogenesis of severe COVID-19 723 infection: A report of five cases

Pulmonary Vascular Endothelialitis, Thrombosis, and Angiogenesis in Covid-19. N Engl 726

728 Evidence of thrombotic microangiopathy in children with SARS-CoV-2 across the 729 spectrum of clinical presentations

Proteomic 731 profiling of MIS-C patients indicates heterogeneity relating to interferon gamma 732 dysregulation and vascular endothelial dysfunction

Thinking 734 Beyond HLH: Clinical Features of Patients with Concurrent Presentation of 735

Primarily VEGF-Driven Etiopathogenesis of 738 Tafro Syndrome and Fibroblastic Reticular Cells As a Probable Castleman Cell -739 Qualitative Metasynthesis

Association of vascular 741 endothelial growth factor and renal thrombotic microangiopathy-like lesions in patients 742 with Castleman's disease

Angiotensin II-744 stimulated vascular endothelial growth factor expression in bovine retinal pericytes

Angiotensin II and vascular endothelial 747 growth factor in the vitreous fluid of patients with proliferative diabetic retinopathy

Platelet-derived-growth-factor-750 induced signalling in human platelets: phosphoinositide-3-kinase-dependent inhibition of 751 platelet activation

Endothelial Activation as Potential Mechanisms Behind the Thrombotic Complications 754 of COVID-19 Patients

Standardizing immunophenotyping for the Human Immunology Project

CD38(bright)CD8(+) T Cells Associated with the Development of Acute GVHD Are Activated, Proliferating, and Cytotoxic Trafficking Cells

Peripheral Blood CD38 Bright CD8+ Effector Memory T Cells Predict Acute Graft-versus-Host Disease

T cell activation profiles distinguish hemophagocytic lymphohistiocytosis and early sepsis

CD57+ T lymphocytes and functional immune deficiency

CD57 identifies T cells with functional senescence before terminal differentiation and relative telomere shortening in patients with activated PI3 kinase delta syndrome

CD57 defines a functionally distinct population of mature NK cells in the human CD56dimCD16+ NK-cell subset

Expression of CD57 defines replicative senescence and antigen-induced apoptotic death of CD8+ T cells

Lag-3, Tim-3, and TIGIT: Co-inhibitory Receptors with Specialized Functions in Immune Regulation