Abstract
In a randomized control trial on convalescent plasma therapy (CPT) in severe COVID-19, we characterized the nature, in terms of abundance of forty eight cytokines, and dimensions, in terms of their interrelationships, of the hyper-immune activation-associated ‘cytokine storm’ in patients suffering from acute respiratory distress syndrome. We found reduced plasma level of the chemokine MCP3 to be a key correlate for clinical improvement, irrespective of therapeutic regimen. We also identified an anti-inflammatory role of CPT independent of its neutralizing antibody content, and a linear regression analysis revealed that neutralizing antibodies as well as the anti-inflammatory effect of CPT both contribute to marked immediate reductions in hypoxia, as compared to patients on standard therapy.
The pandemic caused by the novel coronavirus SARS-CoV-2 has caused over 31 million infections and close to 1 million deaths. The disease caused by SARS-CoV-2, COVID-19, usually happens over two distinct phases in symptomatic individuals. In few patients the early milder phase progresses to a more severe disease and acute respiratory distress syndrome (ARDS), leading in some to untoward fatal outcomes1. A hyper-immune activation response is associated with the severe symptoms, characterized by a systemic deluge of inflammatory cytokines or ‘cytokine storm’2,3. Different therapeutic approaches are currently being explored, either by repurposing specific anti-viral agents, viz. remdesivir4, or by using corticosteroids to affect immunomodulation5. In addition, convalescent plasma therapy (CPT) has emerged as a widely tried strategy against COVID-19, being explored in a number of clinical trials6. Convalescent plasma transfusion is an age-old strategy for passive immunization, with the primary intention to supplement non-recovering patients with antibodies against specific pathogens7. As part of an randomized control clinical trial on COVID-19 CPT, first we aimed at an in-depth characterization of the nature and dimensions of the cytokine storm encountered in the COVID-19 patients progressing to ARDS and then explored effect of CPT, if any, on mitigation of this systemic hyperimmune activation phenotype.
We recruited patients suffering from COVID-19 at the ID&BG Hospital, Kolkata, India, who either had mild COVID-19 disease (N=13) or more severe disease showing evidence for ARDS with PaO2/FiO2 ratio below 300 (N= 33) and collected plasma on day 1 before CPT (time-point 1 or T1) and again on day 3/4 post-enrolment (time-point 2 or T2). We randomized the ARDS patients into either standard-of-care (SOC) group or added CPT to their standard care (CPT group), receiving two transfusions on consecutive days, the first being on the day of enrolment (Day 1 or T1). All experiments were done with appropriate ethical approvals from all involved institutions and with informed consent of the patients recruited (Extended Data Fig. 1A).
First, we analyzed a panel of 48 cytokines in plasma at T1 for deeper characterization of the cytokine storm through comparison between patients having milder disease and patients with evidence of ARDS. We identified a panel of 14 molecules that were significantly higher in ARDS patients (Fig.1A, Extended Data Fig. 2). In addition, we found a significant decrease in plasma abundance of TRAIL in ARDS, a cytokine known to be expressed in cytotoxic T cells and NK cells and involved in killing of virus-infected host cells8, which may signify obviation of infected host cell-directed cytotoxicity at this later phase of the disease. Indeed, the viral load (estimated concomitant to plasma sampling) was significantly higher in patients with milder disease, as expected due to earlier sampling in their disease course (Extended Data Fig. 1B).
A. Heatmap representing normalized plasma abundance of 15 major cytokines compared between patients with mild disease or ARDS at enrolment. B. Extent of correlation between the plasma levels of cytokines and concomitant viral loads as measured by average cycle threshold values for real time PCR of two SARS-CoV2 target genes, compared between patients with mild disease and ARDS. C. Plasma level of selected cytokines correlated with age of patients with ARDS. D. Correlation network of major cytokines at T1 compared between patients with mild disease and ARDS.
Plasma abundance of a number of cytokines in the mild disease was negatively correlated with concomitant viral load, presumably representing the cytokine component of an efficient protective immunity in the milder phase (Fig. 1B). Interestingly, in patients with ARDS there was rather a significant positive correlation of viral load with a few specific cytokines (Fig. 1B). Among them IL-8 and G-CSF presumably represent the usual neutrophil recruitment response triggered by residual virus-infected cells. Notably a number of major cytokine storm components, viz. IL-6, IFNγ, IL1RA, MIP1α, seemed not to correlate any way with the concomitant viral load. Of note here, increasing age was significantly correlated with higher plasma abundance of a number of cytokines (Fig. 1C).
Next to gather some insight on the dimensions of this cytokine storm encountered in ARDS, we analyzed the correlative networks among individual members of the whole panel of cytokines compared between mild and ARDS patients (Extended Data Fig. 3), as well as among the cytokines which were found to be significantly dysregulated in ARDS (Fig. 1D). Interestingly, in ARDS we found more robust assimilation of cytokine co-occurrence networks with greater number of correlative edges compared to mild disease. The most notable was a five member cytokine module, comprising of IL-6, MCP3, MIP1α, IL-1RA and IP10, showing robust correlative upregulation.
We then analyzed the aforementioned cytokine panel in all ARDS patients at T2 and explored if the T1 to T2 change was different between SOC and CPT groups. Intriguingly, on analyzing the panel of 15 cytokines depicted in Fig. 1A, we found a notable effect of CPT, as compared to SOC, in reducing the levels of a number of them, viz. IL-6, IP10 and MCSF (Fig. 2A,B; Extended Data Fig. 4). On the other hand, none of the major cytokines driving the cytokine storm was found to be significantly reduced at T2 in patients receiving SOC (Extended Fig. 4). This anti-inflammatory effect was not dependent on either the anti-SARS-CoV-2 spike IgG content or the capacity for neutralization of spike-ACE2 interaction of CP (Fig. 2A). Of note here, we found strong correlation between anti-spike
A. Heatmap representing log2 fold change between T1 and T2 in the plasma abundance of major cytokines associated with ARD S compared between SOC and CPT groups. Anti-Spike IgG content and nAb content of the convalescent plasma transfused are represented for the CPT group. B. Reduction in plasma levels of IL-6, IP10 and M-CSF from T1 to T2, compared between SOC and CPT groups. C. Correlation network of major cytokines at T2 compared between SOC and CPT groups. D. Kinetics of SpO2/FiO2 ratio in patients over five days post-enrolment, compared between SOC and CPT groups. E. Comparison of SFR5D AUC between SOC and CPT groups. F. Correlation of log2 fold change between T1 and T2 for plasma level of MCP3 and SFR5D AUC. G. SFR5D AUC relationships for nAb content of transfused plasma, log2 fold change of plasma levels of IL-6 and IP10 in the CPT group.
IgG content of plasma and its capacity to neutralize spike-ACE2 interaction, as measured in an in vitro assay9, in our cohort of convalescent donors (Extended Data Fig. 5).
This anti-inflammatory effect of CPT was also apparent, as compared to SOC, when we compared the correlative networks at T2 for the whole panel of cytokines (Extended Data Fig. 6), as well as for the 15 significantly dysregulated ones (Fig. 2C). The correlative edges among the cytokines registered at T2 were significantly less in number in case of CPT, representing a marked attenuation of the cytokine storm in response to CPT and a trend toward a cytokine milieu quite similar to one found in the early milder phase of the disease.
To assess immediate shorter term effect on clinical outcome in patients receiving CPT we analyzed the kinetics of SpO2/FiO2 ratio (S/F ratio) for 5 days following enrolment (Fig. 2D), and also processed the data to represent fraction improvement over 5 days with respect to day 1 by calculating the area under curve (AUC) for the S/F Ratio Curve (SFR5D AUC) (Fig. 2E). CPT was found to affect faster mitigation of hypoxia, as compared to patients receiving SOC only (Fig. 2 D,E). We noted gradual abrogation of this differential response third day onwards after 2nd dose of CPT, indicating that a sustained benefit may require additional transfusions of CPT.
Reduction in the cytokine MCP3 at T2, as compared to T1 was significantly associated with the mitigation of hypoxia, irrespective of whether patients received CPT or not, thus identifying a major pathogenetic molecule underlying COVID-19 ARDS (Fig. 2F). In patients receiving CPT, as expected, the improvements in S/F ratio was significantly correlated with neutralizing antibody (nAb) content of the CP they were transfused with. Interestingly, in a linear regression analysis we identified that IL-6 and IP10, the two major ARDS-associated cytokines that reflected the anti-inflammatory effect of CPT as described in Fig. 2A, also played a major role in this immediate mitigation of hypoxia in combination with nAb content (Fig. 2G; Extended Data Fig. 7). Thus beneficial effect of CPT perhaps mechanistically goes beyond just passive immunization of the recipients and thus should further be explored mechanistically. We envisage that this anti-inflammatory effect of CPT may affect mitigation of other longer term systemic effects of the cytokine storm encountered in severe COVID-19, full appreciation of which awaits end-point analyses in our trial as well as further meta-analyses of data from other clinical trials on CPT accomplished or ongoing elsewhere.
Thus, in this study we could achieve an in-depth characterization of the nature and dimensions of the hyper-immune activation-associated ‘cytokine storm’ in patients suffering from acute respiratory distress syndrome, which could identify a number of hitherto unappreciated features of the disease pathogenesis in severe COVID-19. Moreover, we report here a notable anti-inflammatory effect of COVID-19 convalescent plasma, independent of, but acting in synergy with, its neutralizing antibody content, which may prove to be a composite predictor of response to convalescent plasma therapy in COVID-19 and should be explored while analyzing the clinical outcomes of trials ongoing throughout the world.
Data Availability
All data are available with the corresponding author and will be shared on request.
Methods
Ethical approval
The randomized control trial (RCT) on convalescent plasma therapy and all associated studies and the human sampling needed for them were done with informed consent from the patients as per recommendation and ethical approval from the Institutional Review Boards of CSIR-Indian Institute of Chemical Biology, Kolkata, India (IICB/IRB/2020/3P), Institutional Review Boards of Medical College Hospital, Kolkata (MC/KOL/IEC/NON-SPON/710/04/2020), India and Infectious Disease & Beleghata General Hospital (ID & BG Hospital), Kolkata, India (IDBGH/Ethics/2429). The RCT was approved by Central Drugs Standard Control Organisation (CDSCO) under Directorate General of Health Services, Ministry of Health & Family Welfare, Govt. of India (approval no. CT/BP/09/2020) and registered with Clinical Trial Registry of India (CTRI), under Indian Council of Medical Research, India.
Plasma cytokine analysis
Plasma was isolated from peripheral blood of patients collected in EDTA vials and cytokine levels (pg/ml) were measured using the Bio-Plex Pro Human Cytokine Screening Panel 48-Plex Assay (Bio-Rad, Cat No. 12007283) which quantitates 48 cytokines (FGF basic, Eotaxin, G-CSF, GM-CSF, IFN-γ, IL-1β, IL-1ra, IL-1α, IL-2Rα, IL-3, IL-12 (p40), IL-16, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, GRO-α, HGF, IFN-α2, LIF, MCP-3, IL-10, IL-12 (p70), IL-13, IL-15, IL-17A, IP-10, MCP-1 (MCAF), MIG, β-NGF, SCF, SCGF-β, SDF-1α, MIP-1α, MIP-1β, PDGF-BB, RANTES, TNF-α, VEGF, CTACK, MIF, TRAIL, IL-18, M-CSF and
TNF-β). The plasma samples were diluted (1:3) in sample diluent and the assay was performed using manufacturer’s protocol. The plate was run and analyzed using Bio-Plex® 200 System (Bio-Rad).
RNA Isolation from COVID-19 Samples in TRIzol
RNA from COVID-19 samples in TRIzol samples were extracted using chloroform-isopropanol method. 1/5th volume of chloroform (Cat No. 1024452500, Merck) was added to the TRIzol (Cat No. 15596026, Invitrogen) and mixed thoroughly followed by 5 minutes incubation at room temperature. After centrifugation at 12,000 rpm at 4°C for 15 minutes, aqueous phase was transferred to a fresh microfuge tube. To precipitate RNA, 2-Propanol (Cat No. 1096341000, Merck) was added in equal ratio and incubated on ice for 10 minutes followed by centrifugation at 10,000 rpm at 4°C. RNA pellet was washed twice with 75% ethanol (Cat No. 100983, Merck). The pellet was re-suspended in 20 µL of RNase free water (Cat No. AM9920, Thermo Fisher Scientific) followed by quantitation using NanoDrop (Cat No. ND-2000, Thermo Fisher Scientific).
RT-PCR
qRT-PCR for SARS-CoV-2 detection was performed using the STANDARD M nCoV Real-Time Detection kit (Cat No. 11NCO10, SD Biosensor), approved by Indian Council of Medical Research (ICMR), India. Briefly, 5 µL of RNA was added to the reaction mix, as per manufacturer’s protocol. The RT-PCR was run on QuantStudio 6 Flex Real-Time PCR Systems (Applied Biosystems, Thermo Fisher Scientific) using recommended cycling conditions in a 96 well format. A cycle threshold (Ct-values) cut-offs mean value for both RdRp and E gene was considered as per SD biosensor’s manual for interpreting the results. CY5 labeled Internal Control is used as a positive control.
SARS-CoV-2 Surrogate Virus Neutralization Assay
Neutralizing antibodies against SARS-CoV-2 in human plasma samples from peripheral blood of convalescent donors were detected using GeneScript SARS-CoV-2 Surrogate Virus Neutralization kit (Cat no-L00847). The kit consists of recombinant SARS-CoV-2 Spike protein receptor binding domain fragment conjugated with horseradish peroxidase (HRP-RBD) & human ACE2 receptor protein (hACE2). Presence of SARS-CoV-2 neutralizing antibodies in the plasma samples block interaction between HRP-RBD and hACE2, which is detected through colorimetry. Assay was performed according to manufacturer’s protocol. Plasma samples, and provided positive and negative controls were diluted at a ratio of 1:10 with the sample dilution buffer. HRP-RBD solution was added to the diluted plasma samples and controls at a ratio of 1:1 (vol/vol) and incubated at 37° C for 30 minutes. The 100µl of the above diluted samples and controls were added to the wells and further incubated at 37° C for 15 minutes before wash and substrate addition. Absorbance value was recorded following addition of stop solution at 450nm using Bio-RADiMARK microplate reader. Since presence of SARS-CoV-2 neutralizing antibodies in the plasma samples will result in inhibition of interaction between HRP-RBD and plate-bound human ACE2 protein, and subsequent development of color, assay results are interpreted as inhibition rate of assay reaction. Inhibition rate is calculated as follows: Inhibition = {1- (O.D value of sample/O.D value of negative control)} × 100 [Inhibition values ≥20% signify positive detection of neutralizing antibodies]
ELISA for anti SARS-CoV-2 IgG
Levels of Immunoglobulin G (IgG) specific for SARS-CoV-2 in the plasma isolated from peripheral blood of recovered patients were detected using EUROIMMUN Anti-SARS-CoV-2 (IgG) Elisa kit (Cat No-EI 2606-9601 G). This assay provides semiquantitative estimation of IgG levels against SARS-CoV-2 spike protein. Assay was performed according to manufacturer’s protocol. The assay wells are pre-coated with recombinant S1 domain of the SARS-CoV-2 spike protein. Plasma samples were diluted with provided sample dilution buffer at a ratio of 1:101 (vol/vol). The two reaction steps in this ELISA involves the incubation of diluted patient samples in pre-coated wells following which bound antibodies are detected using enzyme labelled anti human IgG. O.D. was measured at 450nm and reference wavelength of 655nmusing Bio-RAD iMARK microplate reader. Presence of anti SARS-CoV-2 IgG antibodies in the plasma was measured using the following formula: Ratio = Extinction of the control or patient samples/Extinction of calibrator (Ratio ≥ 1.1 is interpreted as positive)
Standard-of-care
All patients infected with SARS-CoV2 at diagnosis either at ER or OPD or through Telemedicine while being in home isolation had received either of these two combinations: Hydroxychloroquine 400 mg BD on first day followed by 400 mg OD for four days plus Azithromycin 500mg OD for 5 days or Ivermectin 12 mg OD for 5days plus Doxycyclin 100 mg BD for 10 days. At the clinical trial site (ID & BG Hospital, Kolkata, India) standard-of-care (SOC) in all patients with evidence for ARDS were: O2 therapy as per requirement in all patients, dexamethasone or equivalent corticosteroid in all patients, for patients with D-dimer <1000 Fibrinogen Equivalent Units (FEU) prophylactic anticoagulation and for patients with D-dimer >1000 ng/ml FEU therapeutic anticoagulation using either low molecular weight heparin or unfractionated heparin, appropriate broad-spectrum antibiotic therapy based on lymphocyte count, clinical and microbiological assessment, blood sugar was maintained below 200mg/dl using appropriate anti-diabetic therapy, appropriate anti-hypertensive agents were used as per requirement to maintain systolic blood pressure 100-140 mm of Hg, diastolic blood pressure at 70-90 mm of Hg and mean arterial pressure >65 mm of Hg. Awake proning for 6-8 hours/day was attempted in all patients with ARDS. O2 therapy was designed to maintain SpO2 >95% using different devices with different efficiencies in O2 supplementation (as denoted by FiO2), viz. nasal canula, face mask, face mask with reservoir, and in patients unable to maintain SpO2 above 90% with face mask with reservoir, high flow nasal cannula (HFNC) or in some cases mechanical ventilation (MV). For S/F ratio kinetics, a value of 89.99 was used for data-points where either HFNC or MV was in use. Of note S/F ratio kinetics was analyzed only in patients having data for at least one intervening time-point in addition to day 1 and day 5.
Convalescent plasma therapy
Plasma was collected from convalescent donors (recovered from RT-PCR positive SARS-CoV-2 infection at least 28 days prior to donation) by apheresis at the Department of Blood Transfusion and Immunohematology, Medical College Hospital, Kolkata, India. All donors were tested for their anti-Spike IgG content in addition to routine screening tests to exclude major blood borne pathogens before apheresis. As per our trial protocol (Clinical Trial Registry of India No. CTRI/2020/05/025209), approved by Central Drugs Standard Control Organisation (CDSCO) India, we randomized the ARDS patients into either standard-of-care (SOC) group as controls or added two consecutive doses of ABO-matched 200ml convalescent plasma on two consecutive days to their standard care (CPT group), the first transfusion being on the day of recruitment (Day 1).
Co-occurrence analyses
Co-occurrence among each pair of cytokines was calculated using Pearson correlation (r) and corresponding p-value of the correlation was measured using a t-distribution. Absolute values of the cytokines were used for the calculation of correlation network and threshold was set to r>0.5, p<0.01 for the complete set of cytokines from mild (n=13) and ARDS (n=33) conditions while the threshold was set to r>0.3, p<0.05 for set of significantly different cytokines between mild and ARDS conditions in SOC (n=16) and Plasma (n=17) sets at T2. All calculations were done using ‘Hmisc’ R package and finally converted to a network file using the ‘igraph’ R package. Visualisation of the network was performed using Cytoscape. Each cytokine was color coded and node size was set in proportion to the median fold change as compared to the same cytokine in the mild datasets.
Statistical analyses
Differential abundance of cytokines was evaluated by Mann-Whitney U Test and Wilcoxon Matched Pairs Test for categorical and ordinal data, respectively. In order to test the monotonous relationship between CT value and cytokine abundance of the patients, spearman’s rank correlation was used, followed by p-value calculation using t-distribution. Similar tests were performed between neutralizing antibody content and IgG antibody content for plasma recipient patients. In order to explain the SFR5D AUC as a function of a combination of cytokines and NAb, linear regression was applied on the ranks of each of the parameters and was implemented using R. The statistical significance was defined as a p value <0.05 (*) and p value <0.01 (**) with two-sided tests, unless otherwise mentioned. Statistical tests were also performed in Statistica 64 (StatSoft).
Author contributions
D.G. conceptualized the study. D.G., S.P. and Y.R. designed the study protocol. P.B. and R.D. did the plasma cytokine analysis. JS performed serological studies. A.L. contributed to the analysis and computation. Y.R. and S.R.P. recruited patients, maintained clinical data and supervised clinical management. R.R., R.M., K.C., A.M., M.M.R., B.S.S., D.R., R.C., B.S. contributed to patient management. J.S.V., S.S. and R.P. did the RT-PCR for SARS-CoV2. S.C. contributed in immunological studies. S.P. and P.B. recruited convalescent donors, D.B., C.M. and P.B. performed donor screening, apheresis and biobanking of convalescent plasma. D.G. and S.P. wrote the manuscript. All authors approved the manuscript.
Competing interests
The authors declare no competing interests.
Acknowledgements
D.G. acknowledges funding for the RCT and associated immune monitoring studies from Council of Scientitific Industrial Research (CSIR), Govt. of India (MLP-129); R.P. acknowledge funding from CSIR (MLP-2005) and Fondation Botnar. D.G. holds a Swarnajayanti Fellowship from Department of Science & Technology (DST), Govt. of India, S.P. holds a Ramanujan Fellowship from Science & Engineering Research Board (SERB), Govt. of India, S.C. holds an Intermediate Fellowship from DBT-Welcome Trust India Alliance and R.P. holds a Ramalingaswami Fellowship from Department of Biotechnology (DBT), Govt. of India. P.B. and A.L. are supported by Senior Research Fellowship from CSIR, R.D. and J.S. are supported by Junior Research Fellowships from University Grants Commission, India. Authors express their gratitude to Anurag Agrawal and Kamakshi Sureka for critical reading of the manuscript.
