key: cord-0748006-9rqep8ll
authors: Bonny, Tania S; Patel, Eshan U; Zhu, Xianming; Bloch, Evan M; Grabowski, M Kate; Abraham, Alison G; Littlefield, Kirsten; Shrestha, Ruchee; Benner, Sarah E; Laeyendecker, Oliver; Shoham, Shmuel; Sullivan, David; Quinn, Thomas C; Casadevall, Arturo; Pekosz, Andrew; Redd, Andrew D; Tobian, Aaron A R
title: Cytokine and Chemokine Levels in COVID-19 Convalescent Plasma
date: 2020-11-26
journal: Open Forum Infect Dis
DOI: 10.1093/ofid/ofaa574
sha: 0884c5579532ed0778c77587c06a49e12d1c789a
doc_id: 748006
cord_uid: 9rqep8ll

BACKGROUND: The efficacy of COVID-19 convalescent plasma (CCP) is primarily ascribed as a source of neutralizing anti-SARS-CoV-2 antibodies. However, the composition of other immune components in CCP and their potential roles remain largely unexplored. This study aimed to describe the composition and concentrations of plasma cytokines and chemokines in eligible CCP donors. METHODS: A cross-sectional study was conducted among 20 pre-pandemic healthy blood donors without SARS-CoV-2 infection and 140 eligible CCP donors with confirmed SARS-CoV-2 infection. Electrochemiluminescence detection based multiplexed sandwich immunoassays were used to quantify plasma cytokine and chemokine concentrations (n=35 analytes). A SARS-CoV-2 microneutralization assay was also performed. Differences in the percent detection and distribution of cytokine and chemokine concentrations were examined by categorical groups using Fisher’s exact and Wilcoxon rank-sum tests, respectively. RESULTS: Among CCP donors (n=140), the median time since molecular diagnosis of SARS-CoV-2 was 44 days(interquartile range=38-50) and 9%(n=12) were hospitalized due to COVID-19. Compared to healthy blood donor controls, CCP donors had significantly higher plasma levels of IFN-γ, IL-10, IL-15, IL-21 and MCP-1, but lower levels of IL-1RA, IL-8, IL-16, and VEGF-A(P<0.0014). Significant differences were also observed in plasma levels of IL-8, IL-15 and IP-10 between CCP donors with low(<40) vs. high(≥160) anti-SARS-CoV-2 neutralizing antibody titers(P<0.0014). The median levels of IL-6, IL-8, TNF-α, IL-12/IL23p40, MDC were significantly higher among CCP donors who were hospitalized vs. non-hospitalized(P<0.05). CONCLUSION: Heterogeneity in cytokine and chemokine composition of CCP suggests there is a different inflammatory state among the CCP donors as compared to SARS-CoV-2 naïve, healthy blood donors.

A c c e p t e d M a n u s c r i p t

The ongoing coronavirus disease 2019 (COVID- 19) pandemic is a challenging global health crisis.

While the incidence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, which causes COVID-19, continues to surge, the scientific community is exploring effective prophylactic and therapeutic options. COVID-19 convalescent plasma (CCP) is currently one of the leading therapies and was recently granted hospital Emergency Use Authorization (EUA) by the U.S.

Food and Drug Administration [1] [2] [3] [4] . Although the efficacy of CCP is primarily thought to be ascribed to neutralizing anti-SARS-CoV-2 antibodies, the composition and potential roles of other immune components in the plasma remain largely unexplored. In this study, the composition and concentrations of plasma cytokines and chemokines were evaluated in potential CCP donors with confirmed SARS-CoV-2 infection, and compared to that in healthy blood donor controls.

A cross-sectional study of potential CCP donors (hereinafter referred to as CCP donors) was conducted as previously described [5] [6] [7] . From April to May 2020, eligible CCP donors (n=140) from the Baltimore/Washington DC area who had a confirmed molecular diagnosis of SARS-CoV-2 RNA (PCR+), were at least 18 years old and met the eligibility criteria for community blood donation (e.g. no pregnancy within last six weeks, never been diagnosed with HIV, HBV or HCV) were included.

During this time period, the FDA did not require a specific titer for CCP donors. Plasma was separated from whole blood within 12 hours of collection and stored at -80˚C until further processing.

In addition, a convenience sample of plasma from 20 platelet blood donors (New York Blood Center, NY) collected prior to December 2019 served as healthy controls (pre-COVID-19 pandemic blood donors hereinafter referred to as controls). Self-reported demographic and clinical information was collected and available for the CCP donors. However, except for ABO blood group, no other information was available for the controls.

A c c e p t e d M a n u s c r i p t 5

The design of the work has been approved by the Johns Hopkins University School of Medicine Institutional Review Board. All study participants (both CCP donors and controls) provided written informed consent.

SARS-CoV-2 neutralizing antibody (nAbs) titers against 100 50% tissue culture infectious doses (TCID 50 ) per 100 L were determined using a microneutralization (NT) assay, as previously described [7] . The nAb titer was the highest plasma dilution that prevented cytopathic effect (CPE) in 50% of the wells tested. nAb area under the curve (AUC) values were estimated using the exact number of wells protected from infection at every plasma dilution.

Highly sensitive, multiplexed sandwich immunoassays using MULTI-ARRAY ® 

A]. Analyte concentrations were calculated per manufacturer protocol (MSD DISCOVERY WORKBENCH® analysis software) and were considered -detectable‖ if both runs of each sample had a signal greater than the analyte-and plate-specific lower limit of detection (LLOD) (i.e., 2.5 standard deviations of the plate-specific blank). Analyte concentrations (pg/mL) from both runs of each analyte were averaged for analysis.

Cytokine and chemokine analytes with <80% detectability in the overall sample were analyzed as binary variables (detectable vs. not detectable) and analytes with80% detectability in the overall sample were analyzed continuously, as previously published [9] . Values for analytes with results below the LLOD were imputed using a single stochastic imputation from truncated log-normal distributions fitted to the detectable values of each analyte (i.e., random values were drawn from the extrapolated tail of the distribution below the LLOD). CCP donors were categorized into groups based on the level of nAb (AUC cutoffs of <40, 40-159, ≥160 as low, middle and high titer, respectively).

Differences in analyte detectability were examined between categorical groups using Fisher's exact  2 tests. Differences in the distribution of continuous analytes were examined between categorical groups using Wilcoxon rank-sum tests.

To assess whether the distribution of log 2 -transformed cytokine and chemokine analytes differed by nAb groups (<40, 40-159, or 160 AUC) among CCP donors, least-squares linear regression was performed separately for each analyte (dependent variable). For each analyte, adjusted  coefficients were estimated from a multivariable model that included adjustment for age (continuous), sex, and days from PCR+ diagnosis (continuous). As a sensitivity analysis, a second model was also A c c e p t e d M a n u s c r i p t 7 constructed that additionally included adjustment for history of COVID-19 hospitalization status as a measure of severity of illness. Similar analyses were performed using modified Poisson regression for binary analytes. Adjustment for multiple comparisons across analytes was performed by using a family-wise Bonferroni correction to control the family-wise error rate at an  of 0.05 [10] ; a twosided P value < 0.0014 (0.05/35 analytes) was considered statistically significant. Analyses were performed in R statistical software (version 4.0.2).

The sample of CCP donors (n=140) was 54% male with a mean age of 42 years (Supplementary Table   1 ). CCP donors were predominantly white (n=108), followed by Asian (n=14), Hispanic (n=5), African American (n=4), and mixed or unknown race (n=9). While most CCP donors had minor symptoms and did not require any hospitalization, 12 (9%) donors had previously been hospitalized due to COVID-19. At the time of blood collection, a median of 44 days (IQR=38-50) had elapsed since participants had been diagnosed with SARS-CoV-2 infection by PCR. There were 60 CCP donors in the low nAb group (43%), 35 in the middle nAb group (25%), and 45 in the high nAb group (32%), for whom epidemiologic characteristics are given in Supplementary Table 1 .

For each of the cytokine and chemokine analytes tested, LLOD range (pg/mL) and total number of detectable samples in CCP donors and controls were calculated. For all analytes, the overall mean coefficient of variation between runs was <15% ( Supplementary Tables 2 and 3 ). The correlation between cytokines and chemokines varied substantially and is shown in (Figure 1) .

A c c e p t e d M a n u s c r i p t 8 Among the analytes with 80% overall detectability, the median levels of IL-6, IL-8, TNF-, IL-12/IL23p40 and MDC were significantly higher among CCP donors who were hospitalized vs. nonhospitalized (Wilcoxon rank sum P <0.05, Supplementary Table 5 ).

Among the analytes with 80% detectability, the distribution of plasma IFN-, IL-10, IL-15, IL-21 and MCP-1 were significantly higher in CCP donors as compared to the controls, whereas the distribution of IL-1RA , IL-8, IL-16, and VEGF-A levels were significantly lower in CCP donors than the controls after adjusting for multiple comparison using Bonferroni correction (Wilcoxon rank-sum P<0.0014) (Figure 2 ). Among CCP donors and the analytes with 80% detectability, the distribution of IL-8, IL-15, and IP-10 were significantly higher in the high nAb titer group (≥160 AUC) compared to the low nAb titer group (<40 AUC) (Wilcoxon rank-sum P<0.0014) (Figure 3 ). In multivariable regression analysis, only the distribution of IL-15 was significantly higher in the high nAb titer group than those in the low nAb titer group (AUC <40) after adjusting for age, sex, days since PCR+ diagnosis as well as for multiple comparisons (P < 0.0014) (Supplementary Table 6 ). Similar results were obtained in the model that adjusted for prior hospitalization status.

Among analytes with <80% detectability, the percent detection of IL-13, IL-1, IL-4, and G-CSF was significantly lower among CCP donors than controls after adjusting for multiple comparisons (Fisher's exact P < 0.0014) (Figure 4 ). There were, however, no statistically significant differences by nAb titer group among CCP donors after adjusting for multiple comparisons ( A c c e p t e d M a n u s c r i p t 9

In addition to nAbs, convalescent plasma may contain cytokines (pro-and/or anti-inflammatory), clotting factors, natural antibodies, defensins, pentraxins and other undefined proteins [11] .

Antibodies in convalescent plasma may also mediate their therapeutic effects through a variety of mechanisms (e.g. direct virus neutralization, antibody mediated complement activation, antibodydependent cellular cytotoxicity, and/or phagocytosis) [11, 12] . While studies have evaluated the acute phase plasma of COVID-19 patients [13] [14] [15] , sparse data exist on recovered COVID-19 patients, and the components in convalescent plasma. In addition, little is known regarding whether a particular cytokines or chemokines profile is associated with higher nAb. This study demonstrates that CCP has a different cytokine and chemokine profile than that of the plasma of SARS-CoV-2 naïve controls, and that IL-8, IL-15 and IP-10 were associated with higher nAb among CCP donors.

Distributions of IFN-, IL-10, IL-15, IL-21 and MCP-1 were significantly higher in CCP as compared to control plasma. Although no significant difference in median levels of these cytokines and chemokines has been observed in the hospitalized CCP donors (n=12), this may still reflect residual effects of the heightened immune response that were present during acute phase of the disease in severe as well as moderately ill COVID-19 patients as previously reported [16, 17] . From a functional standpoint, some of these cytokines (IL-10, IL-15 and IL-21) may be involved in B cell survival, differentiation into plasma cells and class switching, while others may impart an inhibitory effect on B cells in certain situations (e.g. IL-8, IFN-) [18] [19] [20] . In addition, it has recently been reported that individuals with inborn errors of type I IFN immunity is associated with life-threatening COVID-19 pneumonia [21] . Of note, some of the analytes tested (e.g. IL-1RA, IL-8, IL-16, and VEGF-A) were significantly lower in CCP donors than those from the control cohort, the biological relevance of which requires further investigation. This is in contrast to a previous report by Chi et. al. where levels of all 48 cytokines and chemokines tested were very similar between a small sample of convalescent (n=4) and control subjects (n=4) [17] .

A c c e p t e d M a n u s c r i p t 10 In a well-controlled immune response, cytokine and chemokine expression is tightly regulated, and loss of this control may have unintended consequences. For instance, IL-10, a key regulator of immune system homeostasis and an anti-inflammatory cytokine, was also associated with sustained chronic infection when expressed aberrantly [22] . Virus-induced aberrant inflammatory responses have been associated with the pathogenesis of many viral diseases including the coronaviruses [8] .

Excessive cytokine production, also known as hypercytokinemia, has been linked to pulmonary inflammation and acute lung injury in SARS, MERS-CoV, and in SARS-CoV-2 infected patients [8, [23] [24] [25] [26] . Longitudinal immune profiling of COVID-19 has revealed an inflammatory cluster of IL-1α, IL-1β, IL-6, IL-10, IL-12 p70, IL-17A, IFNα, thrombopoietin (TPO), IL-33, IL-16, IL-21, IL-23, IFNλ, eotaxin and eotaxin 3 in patients with severe disease [16] . Transfusion of plasma with elevated levels of some of these cytokines, could contribute to some of the biological effects associated with CCP administration in COVID-19 patients. In this regard, the administration of these cytokine and chemokines may have an immunomodulatory effect in less critical patients via amelioration of the severe inflammatory response (e.g. anti-inflammatory IL-10) and/or antibody production (e.g. IL-21 in plasma cell generation).

There are limitations to this study. The study is cross-sectional and not designed to infer causal associations. The number of available controls was relatively low, and limited demographic and clinical data were available for the control group, which may have resulted in unmeasured confounding. Nevertheless, findings of this study suggest heterogeneity in cytokine and chemokine composition and levels in CCP. Indeed, the biological significance of the variations in several cytokine and chemokine levels observed in this study requires investigation.

The markedly different inflammatory states in the CCP donors observed in this study may have particular significance to long-term recovery from COVID-19, and also the therapeutic potential of CCP. A recent study of 55 people recovering from COVID-19 in China showed abnormal lung scans A c c e p t e d M a n u s c r i p t 11 and lung function three months after discharge from the hospital, and elevated serum levels of Ddimer was associated to people with more lasting lung problems [27] . Even in outpatients with milder COVID-19 illness, delay in complete recovery with prolonged symptoms like cough, fatigue, or shortness of breath has been reported in the US [28] .These studies demonstrate that individuals have markedly different inflammatory states long after symptoms have resolved, and these may contribute to the lingering symptoms (e.g. shortness of breath, cough, gastrointestinal problems, headache, or fatigue) reported in the COVID-19 long-haulers. Thus, the levels of CCP cytokines and chemokines provide important insights into overall recovery status of the convalescing patients, as well as its suitability as a therapy for COVID-19 patients with different disease severity. Larger and longer-term CCP studies are also necessary to evaluate their immunomodulatory effects in COVID-19 patients. M a n u s c r i p t M a n u s c r i p t

Deployment of convalescent plasma for the prevention and treatment of COVID-19

FDA. Recommendations for Investigational COVID-19 Convalescent Plasma

Earlier the better: convalescent plasma

Promoting access to COVID-19 convalescent plasma in low-and middle-income countries

SARS-CoV-2 Antibody Avidity Responses in COVID-19 Patients and Convalescent Plasma Donors

Comparative performance of five commercially available serologic assays to detect antibodies to SARS-CoV-2 and identify individuals with high neutralizing titers

Sex, age, and hospitalization drive antibody responses in a COVID-19 convalescent plasma donor population

Altered cytokine levels and immune responses in patients with SARS-CoV-2 infection and related conditions

Host factors associated with serologic inflammatory markers assessed using multiplex assays

Teoria statistica delle classi e calcolo delle probabilit `a. Pubblicazioni del R Istituto Superiore di Scienze Economiche e Commerciali di Firenze

Plasma therapy against infectious pathogens, as of yesterday, today and tomorrow

SARS-CoV-2 antibody signatures robustly predict diverse antiviral functions relevant for convalescent plasma therapy

Longitudinal dynamics of the neutralizing antibody response to SARS-CoV-2 infection

A dynamic COVID-19 immune signature includes associations with poor prognosis

Cytokine levels in critically ill patients with COVID-19 and other conditions

Longitudinal analyses reveal immunological misfiring in severe COVID-19

Serum Cytokine and Chemokine Profile in Relation to the Severity of Coronavirus Disease 2019 in China

Cytokine-mediated regulation of plasma cell generation: IL-21 takes center stage

IL-8) selectively inhibits immunoglobulin E production induced by IL-4 in human B cells

Human B cell proliferation and Ig secretion induced by recombinant CD40 ligand are modulated by soluble cytokines

Inborn errors of type I IFN immunity in patients with lifethreatening COVID-19

Anti-inflammatory and pro-inflammatory roles of TGF-β, IL-10, and IL-22 in immunity and autoimmunity

Analysis of serum cytokines in patients with severe acute respiratory syndrome

Comparative and kinetic analysis of viral shedding and immunological responses in MERS patients representing a broad spectrum of disease severity

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

Longitudinal characteristics of lymphocyte responses and cytokine profiles in the peripheral blood of SARS-CoV-2 infected patients

Follow-up study of the pulmonary function and related physiological characteristics of COVID-19 survivors three months after recovery

Symptom duration and risk factors for delayed return to usual health among outpatients with COVID-19 in a multistate health care systems network-United States

We are grateful to all participants who enrolled in this study and donated plasma. We thank the National

A c c e p t e d M a n u s c r i p t 13 A c c e p t e d M a n u s c r i p t 17

Note: This analysis is restricted to analytes that had 80% overall detectability. 

Note: This analysis is restricted to analytes that had <80% detectability. There were 20 controls and 140 CCP donors. P values were determined by Fisher's exact test. Only p values < 0.05 are presented. * significant p values after adjusting for multiple comparison using Bonferroni correction (P < 0.0014).