key: cord-0766560-kbshjh6q
authors: Olbei, M. L.; Hautefort, I.; Modos, D.; Treveil, A.; Poletti, M.; Gul, L.; Shannon-Lowe, C. D.; Korcsmaros, T.
title: SARS-CoV-2 causes a different cytokine response compared to other cytokine storm-causing respiratory viruses in severely ill patients
date: 2020-11-16
journal: nan
DOI: 10.1101/2020.11.14.20231878
sha: 93cf1283165e8d7e01075211f9490ed3e716156d
doc_id: 766560
cord_uid: kbshjh6q

Hyper-induction of pro-inflammatory cytokines, also known as a cytokine storm or cytokine release syndrome (CRS) is one of the key aspects of the currently ongoing SARS-CoV-2 pandemic. This process occurs when a large number of innate and adaptive immune cells are activated, and start producing pro-inflammatory cytokines, establishing an exacerbated feedback loop of inflammation. It is one of the factors contributing to the mortality observed with COVID-19 for a subgroup of patients. CRS is not unique to SARS-CoV-2 infection; it was prevalent in most of the major human coronavirus and influenza A subtype outbreaks of the past two decades (H5N1, SARS-CoV, MERS-CoV, H7N9). Here, we collected changing cytokine levels upon infection with the aforementioned viral pathogens through a comprehensive literature search. We analysed published patient data to highlight the conserved and unique cytokine responses caused by these viruses. A map of such responses could help specialists identify interventions that successfully alleviated CRS in different diseases and evaluate whether they could be used in COVID-19 cases.

Introduction 26

The current coronavirus (COVID-19) pandemic has very much focused attention upon this and other 27 viral infectious diseases that the host antiviral immune response is unable to resolve (1) (2) (3) (4) (5) . Indeed, 28 major efforts are now concentrating on how severe acute respiratory syndrome β -coronavirus 2 29

(SARS-CoV-2) alters normal antiviral immune responses (6-15). SARS-CoV-2 causes a wide range 30 of clinical symptoms from asymptomatic, through mild (fever, persistent cough, loss of taste and 31 smell), to severe inflammation-driven pneumonia resulting in multiple organ failure and ultimately 32 death (9-20). SARS-CoV2 induces an anti-inflammatory response attacking both the upper and 33 lower respiratory tract as well as the gut. The upper respiratory tract infection could cause its high 34 infectivity, while the lower respiratory tract and extrapulmonary symptoms are responsible for its 35 severity. Although SARS-CoV-2 appears to modify host inflammatory defences, similar 36 modifications are also observed in other severe respiratory infections caused by viruses such as 37 influenza A, β -coronaviruses SARS-CoV and MERS-CoV. These agents all constitute a global health 38 threat with colossal economic consequences (21, 22) . 39

Although these different viruses cause similar clinical symptoms, the pathogenesis may be driven by 40 different triggers, depending upon the virus. Multiple studies have described an exacerbation of the 41 pro-inflammatory host immune response associated with severe forms of the diseases, including 42 cytokine storms, or cytokine release syndrome (CRS) (1) (2) (3) (4) (5) (9) (10) (11) (12) (13) (14) (15) (16) . Although CRS usually resolves 43

following completion of the antiviral response, it persists in severe cases leading to tissue damage, 44 multiple organ failure and death in critically-ill patients if the clinical intervention is not rapid. In 45 such cases, concentrations of both pro-and anti-inflammatory cytokines are significantly increased in 46 blood and other tissues, including the type-I interferons (IFN) (IFN-, -β, -κ, -ε, -τ, -ω and -ζ) ( H7N9 and the three β -coronaviruses SARS-CoV, MERS-CoV and SARS-CoV-2 ( Figure 1 ). We used 91 the names of each virus and the cytokines and chemokines as search terms, e.g. "SARS-CoV-2 + 92 CXCL10" (Figure 1 ). The collected studies were then screened to retain the studies using only 93 patient-derived data, measured in at least 10 patients. A second pass was done adding "patient" to the 94 search terms, e.g. "SARS-CoV-2 + CXCL10 + patient" in cases where the original search term 95 yielded more than 50 hits. We only considered articles valid if they contained patient-derived data 96 directly; cell line or model organism-based results (and reviews) were excluded. From the main text 97 of the resulting articles, we generated a table containing the presence of the queried cytokines and 98 their direction of change in each disease. We closed the curation on 03/06/2020 (See Supplementary  99  Table S2 for the full list of queried cytokines). The number of discarded articles was estimated using 100 custom python and shell scripts, available in the publication repository. 101

We clustered our data using the clustermap function from the python package seaborn with Jaccard 103 distance and complete linkage method (50). We used all cytokine categories as input. The code is 104 available at our GitHub repository 3 . 105 106 4

Results 107

In order to capture as many studies as possible on which to apply the defined inclusion filtering for 108 our work, we started from a list of relevant cytokines found in a textbook source (48) (Figure 1 ). We 109 only used studies that reported the directional change of measured cytokines. Our curation approach 110 allowed us to highlight shared and differing cytokine responses between influenza A and β -111

coronaviruses, contributing to further understanding of why SARS-CoV-2 in particular differs so 112 much from influenza A CRS-causing viruses but also from other β -coronaviruses, also capable to 113

inducing a cytokine storm in severe cases. 114 1 https://www.biorxiv.org/ 2 https://www.medrxiv.org/ 3 https://github.com/NetBiol/Covid19/tree/master/CRS . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.14.20231878 doi: medRxiv preprint 115 Figure 1 : The literature curation workflow applied in this study. Publications were considered 116 valid for inclusion into our data collection i) only if they contained patient-derived data (model 117 organisms and cell lines were excluded) and ii) if the study data were collected from cohorts of at 118 least 10 participants per group, iii) and if it included a change in cytokine levels. Total hits to queries 119 in bioRxiv, medRxiv and PubTator shown separately in the second box from the top. 55 publications 120

were selected that matched our curation criteria listed above. 121 Table S1 ). Only a small group of 126 cytokines was commonly measured for all viruses (CXCL8, IL-6, CXCL10, IL-2, IL-10, IFN-γ, 127

TNF-α). Across the 55 literature references used here ( Figure 1 ), we first assessed how comparable 128 the number of different cytokines measured in these studies was across the five CRS-causing viruses. 129 Figure 2 shows how variable this number is between virus-specific studies (e.g. 15 for H5N1 and 26 130

for SARS-CoV-2). This variation probably reflects i) the increasing interest developed for CRS-131

causing pathologies over recent years (26 recent studies reported cytokine measurement for SARS-132

CoV-2 against only 10 H5N1-related studies), and ii) the increased availability and sensitivity of 133 multiplex detection method. 134 The influenza A viruses trigger an increase in all cytokine levels measured (Figure 2, is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.14.20231878 doi: medRxiv preprint looking into the detailed patterns of cytokine responses of the various CRS-inducing viruses. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this this version posted November 16, 2020. ;

Infection with either of the two influenza subtypes seems to increase the levels of measured type-I  215  IFN-relevant cytokines, resulting in an antiviral immune response, with the appropriate cytokines  216 showing elevated levels in all influenza A studies (Figure 4, Supplementary Table S1 ). 217

The β -coronavirus-mediated responses show a much more variable IFN response: with SARS-CoV, 218 we see that the type-I IFN response is active, including the downstream-activated IL-12 that reflects 219 the involvement of mature dendritic cells. IL-12 also indirectly activates IFN-γ further downstream. 220 IL-10 is not elevated, which potentially prevents the downregulation of the type-I interferon 221

response. 222

In MERS-CoV infections, the type-I IFN response is induced, but not in all cases (58). In some 223 studies, the levels of IL-12 do not increase, in agreement with IFN-γ also staying at low levels. Yet 224 we see the involvement of the (mostly) anti-inflammatory IL-10. However, caution needs to be 225 applied when looking at IL-10 in an inflammation context, as more and more clinical evidence 226

suggests this cytokine displays pro-inflammatory characteristics in vivo (59,60). 227

We showed here that SARS-CoV-2-mediated infections are characterized by a clear dysregulation of 228 type-I IFN response, and consequently the downstream cytokine signature such as IL-4, IL-12, IL-2, 229 IL-10 and the downstream type II IFN response (Figure 4) . 230 231 5

Discussion 232

In this study, we analysed relevant cytokine levels measured in patients infected each with one of five 233 major viral pathogens through a comprehensive literature curation of published patient data. We 234 generated a map of such responses to help specialists identify routes of interventions to successfully 235 alleviate CRS in different diseases, and evaluate whether they could be used in COVID-19 cases. 236

Based on our literature curation, the five investigated viruses cause atypical cytokine responses in 237 severely-ill patients, reported here in Figure 3 . 238

The cytokine response during viral infection is a dynamic process, with multiple changes in cytokine 239 levels during the course of the infection. During SARS and MERS infection, a slow initial innate 240 immune response accompanied by the infection of alveolar macrophages leads to increased severity 241 of these lower respiratory tract diseases (61-65). Furthermore, a long-lasting pro-inflammatory 242 cytokine production results in high mortality due to the development of severe conditions such as 243 acute respiratory distress syndrome (ARDS) or acute lung injury (9.5% fatality rate for SARS and 244 34.4% for MERS compared to 2.3% for COVID-19 (44)). 245

Severe SARS patients show particularly low levels of the anti-inflammatory cytokine IL-10 ( Figures  246  3, 4) (66). During MERS infection patients develop an expected increased production of IL-10, yet 247 the low levels of IFN-γ-inhibiting IL-4 and IL-2, lead to elevated IFN-γ and the induction of type-II 248 IFN response (Figure 3) (58,67,68) . In contrast, during influenza A infection, the antiviral response 249

activates without much delay with the presence of an intact negative feedback loop. Both viruses 250 considered in our curation induce most of the pro-and anti-inflammatory cytokines downstream of 251 type-I interferon response (Figure 3 ). Although influenza A viruses have effectors that dysregulate 252 IFN-I (e.g. NS1, PB1-F2, polymerase proteins), the IFN-I response is nonetheless sustained, and its 253 excessive activation during severe illness can lead to increased mortality. Furthermore, during H7N9 254 and H5N1 severe infections, TGF-β fails to be activated, contributing to increased pathogenicity (69-255 71). SARS-CoV-2 stands out from the other β -coronaviruses and influenza A viruses, with a highly 256 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.14.20231878 doi: medRxiv preprint perturbed response downstream of Type I IFN signalling, as reflected in the poor balance of 257 measured pro-and anti-inflammatory cytokines (Figures 3 and 4) . Of note, IFN-was found to be 258

increased (similar to the other viruses) only in one small (n=4) patient study, which did not match our 259 inclusion criteria. Type-II IFN-γ was also only increased in patients placed in intensive care units 260 (ICUs) , while it was within normal ranges in other studies (10,72,73). 261

Although the cytokine signalling enabling the reduction of the inflammatory environment is active 262 (Figures 3 and 4) , both influenza viruses H5N1 and H7N9 can cause CRS. In severe cases of 263 infection, CRS could result from insufficient production of important cytokines such as TGF-β (71). 264

Furthermore, the presence of impaired and less abundant effector CD4+ and CD8+ T cells was found 265

to be a characteristic feature accompanying CRS in those diseases. Finally, monocytes, that normally 266 would differentiate from a pro-to anti-inflammatory state with enhanced antigen presentation activity 267

as suggesting that further mechanistic investigation of the cytokine storms during SARS-CoV-2 291 infection will be needed. 292

The ongoing accumulation of patient-derived large data sets will inform the research community and 293

clinicians of the intricacy of host/virus interactions (77 is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.14.20231878 doi: medRxiv preprint

Pedersen SF, Ho Y-C. SARS-CoV-2: a storm is raging. J Clin Invest. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.14.20231878 doi: medRxiv preprint

Control Cytokine Release Syndrome in COVID-19

Induction of pro-317 inflammatory cytokines (IL-1 and IL-6) and lung inflammation by Coronavirus-19 (COVI-19 or 318 SARS-CoV-2): anti-inflammatory strategies

The Role of Cytokines including 320

Interleukin-6 in COVID-19 induced Pneumonia and Macrophage Activation Syndrome-Like Disease

Targeting T-cell senescence 323 and cytokine storm with rapamycin to prevent severe progression in COVID-19

Cytokine-targeted therapy in severely ill COVID-19 patients: 326 Options and cautions

Global Phosphorylation Landscape of SARS-CoV-2 Infection

An interferon-gamma-related 340 cytokine storm in SARS patients

An open 342 resource for T cell phenotype changes in COVID-19 identifies IL-10-producing regulatory T cells as 343 characteristic of severe cases. medRxiv

Elevated levels of plasma 346 cytokines in COVID-19 reflect viral load and lung injury

The pathogenesis and treatment of the `Cytokine Storm' in COVID-348 19

Fatal 353 outcome of human influenza A (H5N1) is associated with high viral load and hypercytokinemia

A(H7N9) virus 356 results in early induction of proinflammatory cytokine responses in both human lung epithelial and 357 endothelial cells and shows increased human adaptation compared with avian H5N1 virus

Novel respiratory viruses: what should the clinician be alert for? Clin 360 Med

Pains and Gains from China's Experiences 364 with Emerging Epidemics: From SARS to H7N9

A Novel Coronavirus 366 Emerging in China -Key Questions for Impact Assessment

Overview of the biology of type I interferons

Cytokine regulation 408 in SARS coronavirus infection compared to other respiratory virus infections

Dysregulation of type I interferon responses in COVID-19. Nat 411 Rev Immunol

Host-Viral Infection Maps Reveal 413 Signatures of Severe COVID-19 Patients

Respiratory Viruses. Encyclopedia of Microbiology

Clinical characteristics and 417 severity of influenza infections by virus type, subtype, and lineage: A systematic literature review. 418 Influenza Other Respi Viruses

Comparative review of respiratory 420 diseases caused by coronaviruses and influenza A viruses during epidemic season. Microbes Infect

COVID-19, SARS and MERS: 423 are they closely related?

425 replication competence, and innate immune responses of the coronavirus SARS-CoV-2 in human 426 respiratory tract and conjunctiva: an analysis in ex-vivo and in-vitro cultures

Comparative tropism, 429 replication kinetics, and cell damage profiling of SARS-CoV-2 and SARS-CoV with implications for 430 clinical manifestations, transmissibility, and laboratory studies of COVID-19: an observational study. 431 Lancet Microbe

Differences and 434 similarities between Severe Acute Respiratory Syndrome (SARS)-CoronaVirus (CoV)

Would a rose by another name smell as sweet?

Chemokines and Their Receptors -Madame Curie 438 Bioscience Database -NCBI Bookshelf

PubTator central: automated concept annotation for 440 biomedical full text articles

Clinical symptoms, immune factors, and 444 molecular characteristics of an adult male in Shenzhen, China infected with influenza virus H5N1

New fronts emerge in the influenza cytokine storm. Semin 447 Immunopathol

Lymphopenia predicts disease 449 severity of COVID-19: a descriptive and predictive study

Immunobiology of the IL-15/IL-15Rα complex as 452 an antitumor and antiviral agent

The human immune response to respiratory 454 syncytial virus infection

Biological features of novel avian 456 influenza A (H7N9) virus. Nature

Viral Invasion and Type I Interferon 458 Response Characterize the Immunophenotypes during COVID-19 Infection

Middle East respiratory 461 syndrome coronavirus: another zoonotic betacoronavirus causing SARS-like disease

Pro-Inflammatory Signaling by IL-10 and IL-22: Bad Habit Stirred Up by 464 Interferons? Front Immunol

Proinflammatory effects of IL-10 during human endotoxemia

Identification 469 of a novel coronavirus in patients with severe acute respiratory syndrome

Genomic 474 characterization of a newly discovered coronavirus associated with acute respiratory distress 475 syndrome in humans

Newly discovered coronavirus as the primary cause of severe acute respiratory syndrome

Pathogenic human coronavirus infections: causes and 480 consequences of cytokine storm and immunopathology

Temporal changes in 483 cytokine/chemokine profiles and pulmonary involvement in severe acute respiratory syndrome. 484 Respirology

MERS-CoV infection 486 in humans is associated with a pro-inflammatory Th1 and Th17 cytokine profile

Clinical progression and 489 cytokine profiles of middle east respiratory syndrome coronavirus infection

H5N1 influenza virulence, pathogenicity and transmissibility: what do we 492 know?

Innate immunity to H5N1 influenza viruses in humans. 494 Viruses

Immune responses to avian influenza viruses

A cluster of health 498 care workers with COVID-19 pneumonia caused by SARS-CoV-2

Clinical features of patients infected 501 with 2019 novel coronavirus in Wuhan, China. Lancet

Pathogenic T cells and inflammatory 503 monocytes incite inflammatory storm in severe COVID-19 patients

Reduction and Functional Exhaustion 505 of T Cells in Patients with Coronavirus Disease 2019 (COVID-19). medRxiv

Tocilizumab shortens time on mechanical ventilation and length of hospital stay in patients with 510 severe COVID-19: a retrospective cohort study

ViralLink: An integrated 512 workflow to investigate the effect of SARS-CoV-2 on intracellular signalling and regulatory 513 pathways