Altered Transcript Levels of Cytokines in COVID-19 Patients

Majid Samsami Shahid Beheshti University of Medical Sciences School of Medicine Alireza Fatemi Shahid Beheshti University of Medical Sciences School of Medicine Reza Jalili Khoshnoud Shahid Beheshti University of Medical Sciences School of Medicine Karim Kohansal Shahid Beheshti University of Medical Sciences School of Medicine Arezou Sayad Shahid Beheshti University of Medical Sciences School of Medicine Shabnam Soghala Islamic Azad University Shahram Arsang-Jang Zanjan University of Medical Sciences Mohammad Taheri (  mohammad_823@yahoo.com ) Shahid Beheshti University of Medical Sciences School of Medicine https://orcid.org/0000-0001-8381-0591 Soudeh Ghafouri-Fard Shahid Beheshti University of Medical Sciences School of Medicine


Introduction
"Severe acute respiratory syndrome coronavirus 2" (SARS-CoV-2) has been shown to cause "coronavirus disease 2019" . This disorder has led to catastrophic events all over the world. Several lines of evidence support the role of "cytokine storm" in the pathogenesis of severe forms of the disorder (1,2). This term is described as an induction of a cascade of auto-augmenting cytokine secretion resulting from uncontrolled host immune reaction to certain immune stimulators (1). A number of recent investigations have demonstrated elevated levels of pro-in ammatory cytokines in the critically ill COVID-19 patients compared with moderately ill ones (3,4). Induction of immune responses during the course of COVID-19 has been shown to result in the overproduction of enormous quantities of in ammatory proteins, eventually leading to organ damage (5). Up-regulation of pro-in ammatory cytokines such as IL-6, IL-1β, IL-18, and TNF-α in addition to their downstream molecules have been reported in these patients (6,7). Moreover, dysregulation of a number of other in ammatory molecules such as IL-2, IL-7, IL-10 and IFN-γ have been associated with COVID-19 (8). Totally, many studies have shown the important impact of IL-6, IFN-γ and TNF-α in the pathobiology of this condition and related organ damage (5). Thus, identi cation of these in ammatory reactions has signi cance in the treatment of patients with COVID-19 infection (9). Based on the diversity in the regulatory mechanisms of immune reactions in different ethnic groups (10), evaluation of cytokine production in the context of COVID-19 infection in each population would help in the design of appropriated therapeutic options. Therefore, in the present investigation, we analyzed expression levels of some cytokines including those participating in adaptive immune responses (IL-2 and IL-4) and pro-in ammatory cytokines (IFN-G, IL-1, IL-6, IL-17 and TNF-α) in a population of Iranian patients with COVID-19 in comparison with sex-/ agematched healthy controls to unravel the importance of cytokine dysregulation in the pathogenesis of this disorder.

Study participants
The present investigation was performed on patients admitted to Nikan Hospital, Tehran during March 2020 till April 2020. Patients with clinical symptoms of COVID-19 were further assessed through RT-PCR assay on the obtained nasopharyngeal samples. Only those with con rmed molecular diagnosis were enrolled in the study. Control samples were obtained from healthy individuals without no clinical symptom or exposure to the affected individuals. The study protocol was approved by ethics committee of Shahid Beheshti University of Medical Sciences (IR.SBMU.RETECH.REC.1399.046). Informed consent was obtained from all patients. A complete paraclinical investigation and chest CT scan were done for all patients.

Expression assays
As the initial step, 3-5 mL of peripheral blood was gathered from all patients and controls. Then, total RNA was recovered from these specimens using the GeneAll RNA extraction Kit (Seoul, South Korea). Subsequently, total RNA was converted to cDNA using the OneStep RT-PCR Series Kit (BioFact™, Seoul, South Korea). Expression levels of cytokine coding genes were quanti ed in samples obtained from COVID-19 patienst and controls using the RealQ Plus 2x Master Mix (Amplicon, Denmark) using the transcript levels of HPRT1 gene as normalizer. Table 1 displays the features of primers, probes and ampli ed segments. Table 1 Information of primers, probes and amplicons.

Statistical methods
Statistical comparisons were performed using R programming language. Transcript quantities of nine cytokine coding genes were measured from Ct values, considering HPRT1 as the reference gene. The following formula was used: where g indicates any of cytokine coding genes. Then the calculated values were log2 transformed and used for succeeding analyses. Two comparisons were done and the signi cance of difference in mean values between two subgroups was computed using the t-test. Comparisons included: 1. Comparison between expression levels of genes between total patients and healthy controls, and 2. between patients who required admission in intensive care unit (ICU) and who did not. Correlations between expression levels were appraised through the calculation of Spearman correlation coe cients.
ROC curves were plotted using Bayesian Generalized Linear Model, Generalized Linear Model, and Linear Discriminant Analysis with 10-fold cross validation. The linear Discriminant Analysis (LDA) provided the most e cient estimates and in the best setting, the AUC was 1. Youden's J statistic was employed to nd the optimum threshold. DA was then chosen according to the former steps to investigate e ciency of each gene for separating groups. Spearman correlation coe cients were employed to assess the association between patients' information and transcript levels. To compute correlation between categorical and continuous variables point biserial correlation coe cient was used. For all statistical tests, the level of signi cance was set at P value < 0.05.

Results
A total of 91 COVID-19 patients (female/ male ratio: 38/ 53) and 96 healthy subjects (female/male ratio: 39/ 57) entered the study. The mean age (standard deviation) of the patients was 57.17 (16.90) years. Among the admitted patients, 37 patients (40.6%) were admitted in the intensive care unit (ICU). Table 2 shows the paraclinical parameters of the COVID-19 group. Expression assays Figure 1depicts the relative expression levels of cytokine coding genes in COVID-19 patients and healthy subjects.
Expression levels of IFN-G, IL-2, IL-4, IL-6, IL-17, TGF-B, IL-8 and IL-1B were signi cantly higher in COVID-19 patients compared with controls and in both female and male patients compared with sex-matched controls. However, expression of none of these cytokines was different between ICU-admitted patients and other patients except for IL-6 whose expression was lower in the former group compared with the latter (ratio of means=0.33, P value=4.82E-02). Expression of TNF-A was not different between COVID-19 patients and healthy subjects (Table 3).  2 and 3 show the ROC curves depicted by three machine learning models and the best tted one, respectively.
Expression of none of cytokine coding genes was correlated with the assessed clinical/demographic data including age, gender, ICU admission, or CRP/ESR levels ( Figure 2).
Lastly, we measured the correlation between the transcript levels of cytokine coding genes among ICU-admitted COVID-19 patients, Non-ICU-admitted COVID-19 patients and healthy controls ( Figure 5). In the rst group, the most signi cant correlation was found between TNF-A and TGF-B (r=0.91, P value= 9.37E-15). In non-ICU admitted patients and healthy controls, the most signi cant correlations were demonstrated between TNF-A and IFN-G (r=0.91, P value= 1.30E-21 and r=0.92, P value= 7.90E-41, respectively).

Discussion
We examined expression levels of nine cytokine-coding genes among ICU-admitted COVID-19 patients, non-ICU-admitted ones and healthy subjects. We detected over-expression of IFN-G, IL-2, IL-4, IL-6, IL-17, TGF-B, IL-8 and IL-1B in COVID-19 patients compared with healthy subjects and in both female and male patients compared with sex-matched controls. However, expression of none of these cytokines was different between ICU-admitted patients and other patients except for IL-6 whose expression was lower in the former group compared with the latter. Expression of TNF-A was not different between COVID-19 patients and healthy subjects.
Levels of several cytokines have been assessed in different subgroups of COVID-19 patients. For instance, Chen et al. have reported remarkable overexpression of IL-2R and IL-6 in the critically ill COVID-19 patients compared with severe and mild groups (11). However, we detected a trend toward underexpression of IL-6 in the ICU-admitted patients compared with the other group of patients. This nding is not reliable as sex-based comparisons did not verify the difference in the expression of this cytokine between ICU-admitted and non-ICY admitted subjects. Moreover, this nding is in contrast with our previous study demonstrating higher median levels of IL-6 protein in the serum of ICU-admitted patients (12). However, it is worth mentioning that the current study varies with our previous study in the terms of applied technique and source of expression assessment. Anyway, we recommend assessment of expression of IL-6 in larger cohorts of Iranian patients to unravel the possible difference in its expression between Iranian patients and patients from other populations.
Chen et al. did not detect difference in the serum levels of TNF-α, IL-1 and IL-8 among the critical, severe and moderate groups COVID-19 patients (11). This nding is in accordance with our nding regarding similar levels of these cytokine between ICU-admitted and non-ICU-admitted COVID-19 patients.
The observed over-expression of IFN-G in COVID-19 patients is in line with the recently reported augmented nucleoprotein-induced IFN-γ release in these patients (13). Hu et al. have demonstrated lower probability of lung brosis at discharge in patients who had higher baseline levels of IFN-γ (14). Although we did not assess the presence of lung brosis in the admitted patients, we demonstrated similar levels of IFN-G between ICU-admitted and non-ICU-admitted patients. Over-expression of IL-17 in COVID-19 patients has also been reported in other populations (15). In addition, MERS-CoV has been shown to stimulate expression of this cytokine in humans (16). Consistently, Th17 cells have been reported to participate in the cytokine storm stimulated by SARS-CoV-2 (17).
Up-regulation of IL-17, IL-2 and IL-4 levels have also been reported in COVID-19 patients with lung lesions (15).
Similar levels of TNF-A between three study subgroups in the current investigation raises the possibility of ethnic-based differences in the immunological responses in the context of COVID-19 infection, since this cytokine has been repeatedly reported to be increased in these patients and has been suggested as a target of immune-modulatory options (18). SARS-CoV-2 has also been suggested to activate of IL-1β, which consecutively induces other pro-in ammatory cytokines, such as IL-6 and TNF-α (18). However, while we detected over-expression of IL-1B and IL-6 in COVID-19 patients, we could not demonstrate any signi cant difference in the expression of TNF-A.
Then, we measured diagnostic power of cytokine transcripts in diagnosing COVID-19 patients from healthy controls. The AUC value was highest for IL-2 and IL-1B. After combining the transcript levels of all cytokines, AUC, sensitivity and speci city values reached 1.0, 1.0 and 0.99, respectively. Therefore, cytokine levels can be used for distinguishing disease status. For differentiation between ICU-admitted patients and other patients, IL-4 with AUC value of 0.68, had the best diagnostic power among cytokine coding genes. Therefore, these molecules cannot differentiate subgroups of COVID-19 patients. Besides, expression of none of cytokine coding genes was correlated with the assessed clinical/demographic data including age, gender, ICU admission, or CRP/ESR levels.
Lastly, we evaluated the correlation between the transcript levels of cytokine coding genes among three study subgroups. Patterns of correlation between expression levels of genes were more similar between non-ICU admitted patients and healthy controls, implying the altered regulatory mechanisms of cytokines expression in severely affected patients.
In brief, we demonstrated altered levels of several cytokine coding genes in Iranian patients with COVID-19 infection. Our study provides further evidence for contribution of "cytokine storm" in the pathogenesis of moderate/severe forms of COVID-19.

Declarations
Ethics approval and consent to Participant All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent forms were obtained from all study participants. Informed consent forms were obtained from all study participants. The study protocol was approved by the ethical committee of Shahid Beheshti University of Medical Sciences (IR.SBMU.MSP.REC.1399.046). All methods were performed in accordance with the relevant guidelines and regulations.