Significance of Catecholamine Biosynthetic/Metabolic Pathway in SARS-CoV-2 Infection and COVID-19 Severity

The SARS-CoV-2 infection was previously associated with the expression of the dopamine biosynthetic enzyme L-Dopa decarboxylase (DDC). Specifically, a negative correlation was detected between DDC mRNA and SARS-CoV-2 RNA levels in in vitro infected epithelial cells and the nasopharyngeal tissue of COVID-19 patients with mild/no symptoms. However, DDC, among other genes related to both DDC expression and SARS-CoV-2-infection (ACE2, dACE2, EPO), was upregulated in these patients, possibly attributed to an orchestrated host antiviral response. Herein, by comparing DDC expression in the nasopharyngeal swab samples of severe/critical to mild COVID-19 cases, we showed a 20 mean-fold reduction, highlighting the importance of the expression of this gene as a potential marker of COVID-19 severity. Moreover, we identified an association of SARS-CoV-2 infection with the expression of key catecholamine biosynthesis/metabolism-related genes, in whole blood samples from hospitalized patients and in cultured cells. Specifically, viral infection downregulated the biosynthetic part of the dopamine pathway (reduction in DDC expression up to 7.5 mean-fold), while enhanced the catabolizing part (increase in monoamine oxidases A and B expression up to 15 and 10 mean-fold, respectively) in vivo, irrespectively of the presence of comorbidities. In accordance, dopamine levels in the sera of severe cases were reduced (up to 3.8 mean-fold). Additionally, a moderate positive correlation between DDC and MAOA mRNA levels (r = 0.527, p < 00001) in the blood was identified upon SARS-CoV-2-infection. These observations were consistent to the gene expression data from SARS-CoV-2-infected Vero E6 and A549 epithelial cells. Furthermore, L-Dopa or dopamine treatment of infected cells attenuated the virus-derived cytopathic effect by 55% and 59%, respectively. The SARS-CoV-2 mediated suppression of dopamine biosynthesis in cell culture was, at least in part, attributed to hypoxia-like conditions triggered by viral infection. These findings suggest that L-Dopa/dopamine intake may have a preventive or therapeutic value for COVID-19 patients.


Introduction
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), that emerged in Wuhan in December 2019 is a positive sense, single-stranded, RNA virus, which may cause severe respiratory infections in humans [1,2]. Symptoms of the disease provoked by the virus (coronavirus disease 2019 or COVID-19) may typically include fever, cough, myalgia or fatigue, sputum, headache, hemoptysis, and diarrhea, as well as dysregulation of the nervous system (loss of smell and taste, consciousness impairment, parkinsonism), and in the most severe cases may also lead to respiratory failure [3][4][5]. Virus infection activates the host immune response, causes dysfunction of the lung epithelial-endothelial barrier, and produces lung and airway hypoxia resulting in oxidative stress [6][7][8]. The induction of systemic hyper-inflammation increases the severity of the disease [9][10][11] and elevates the fatality rate even among younger COVID-19 patients [12].
The SARS-CoV-2 virion enters human cells through the binding of the Spike protein (S) to the angiotensin converting enzyme 2 (ACE2) exopeptidase of the cell membrane that acts as the receptor of the virus [13]. The ACE2 levels are elevated by several factors (such as sex and age), and comorbidities such as cardiovascular diseases, obesity, and diabetes [14], which are known risk factors of the worse clinical manifestations observed in COVID-19 patients [15,16]. Moreover, an N-terminally truncated isoform of ACE2, termed deltaACE2 (dACE2), has been identified as a product of an interferon-stimulated gene (ISG) [17,18]. The dACE2 isoform does not bind to the S protein and is upregulated by SARS-CoV-2 infection in COVID-19 patients [19,20] and in human cell lines and organoids [17]. Expression of the dACE2 gene is higher in airway epithelia and bile duct epithelia in the liver as compared to ACE2, which is generally expressed in human kidney, heart, lung, liver, and vasculature [21,22]. The gene dACE2 has been suggested to be partially responsible for the elevated expression of the ACE2 locus reported in viral infections, including SARS-CoV-2 [17,18]. This was verified by our previous data from nasopharyngeal swab samples of SARS-CoV-2 infected subjects with mild or no symptoms [19], consistently with the induction of other ISGs in COVID-19 infection [20,23,24]. Specifically, we observed significantly higher induction in dACE2 compared to ACE2 in SARS-CoV-2 infected subjects [19]. Furthermore, in our study, the expression of dACE2 seemed to have an inverse relationship with the viral load in these patients, while the expression of ACE2 had a positive association [19]. In addition, several groups have demonstrated elevated levels of ACE2 enzymatic activity in the serum of COVID-19 patients [25][26][27].
Microarray meta-analysis of human tissues has unveiled that the most statistically important co-expressed gene with the ACE2 locus, is L-Dopa decarboxylase (DDC) [28]. The DDC protein catalyzes dopamine and serotonin biosynthesis, which possess immunomodulatory, apart from their neuro-modulatory, function [29][30][31][32]. The DDC gene has also been related to the regulation of cell proliferation and apoptosis [33,34]. We have previously shown that it is the expression of dACE2 that positively correlates with the expression of DDC in human epithelial cells of the nasal cavity, while no correlation has been detected between the expression of DDC and ACE2, regardless of SARS-CoV-2 infection [19]. In intestinal organoids infected by SARS-CoV-2, the ACE2 locus has been found to be co-expressed with other genes encoding enzymes/proteins involved in the metabolism and transport of dopamine and trace amines, in addition to DDC [35]. These genes include the monoamine oxidase A (MAOA), the monoamine oxidase B (MAOB), and several transporter-encoding genes of the solute carrier family (SLC7A9, SLC3A1, SLC6A19 and SLC3A2). Moreover, we have reported that DDC expression was enhanced in the nasopharyngeal tissue of COVID-19 patients with mild or no symptoms, whereas we detected a negative correlation between DDC expression and SARS-CoV-2 RNA levels, also verified in infected Vero E6 cells [19]. In later in vitro studies, Limanaqi et al. [36] observed that the Cells 2023, 12, 12 3 of 35 exogenous administration of dopamine/dopamine agonists in epithelial cells negatively affected SARS-CoV-2 replication, in association with a reduction in D1 and D2 dopamine receptors' mRNA and protein levels, and concurrently upregulated Type-I IFNs and ISGs. The effects of dopamine on viral replication and on D2DRs expression were reproduced by Type-I IFN administration, which in parallel upregulated DDC expression [36]. Our group also showed that the expression and activity of DDC, as well as the treatment of cells with dopamine, reduce the replication of other RNA viruses, including hepatitis C (HCV) and dengue (DENV) viruses [37][38][39]. Complementing our results, an interaction of SARS-CoV-2 viral proteins with the catecholamine degradation enzymes MAOA and catechol-O-methyltransferase (COMT) [40][41][42][43][44], has also been observed. Additionally, the presence of COMT has been deemed as necessary for SARS-CoV-2 infection [45].
It is known that the expression of DDC and ACE2 can also be regulated by hypoxia as well, through mechanisms dependent on hypoxia-inducible factors (HIFs) [39,[46][47][48][49][50][51]. Interestingly, based on the observed upregulation of known hypoxia target genes, such as GLUT1 and VEGFA [52,53], we have indicated that the virus infection triggers hypoxia-like conditions in Vero E6 and A549 cells upon SARS-CoV-2 infection [19], and have suggested hypoxia signaling as a mechanism of DDC and dACE2 co-regulation during infection. Moreover, the expression of GLUT1 in relation to that of other genes, in the blood samples of COVID-19 patients, may be useful in predicting disease severity [54]. Additionally, VEGF expression has been shown to be increased also in infected subjects [53,55]. Indeed, hypoxia or silent hypoxia, that are common clinical manifestations of COVID-19, have also been detected in inflammatory sites of the injured lung tissue in these patients [8,56].
The previously observed differentiated regulation of DDC, ACE2, dACE2 and EPO gene expression in the in vitro infection studies and the nasopharyngeal swab samples of SARS-CoV-2 infected subjects with mild or no symptoms, could be related to an antiviral response of the host in an attempt to restrict the propagation of SARS-CoV-2. Herein, to clarify this discrepancy and to address the implication of the catecholamine pathway in SARS-CoV-2 infection and disease severity, we studied the above-mentioned genes, and other genes involved in the catecholamine pathway in samples from patients and cultured cells infected with SARS-CoV-2. Specifically, we included nasopharyngeal swabs from COVID-19 patients with severe/critical manifestations and compared them to samples from patients with mild/no symptoms of the disease. Moreover, we analyzed the expression of genes related to the catecholamine pathway in whole blood samples of severe cases, and in SARS-CoV-2 infected Vero E6 epithelial cells. Serum dopamine levels in COVID-19 patients were also examined. Finally, we determined the influence of the DDC enzymatic substrate (L-Dopa), and its product (dopamine) on the virus-derived cytopathic effect in cell culture. Altogether, the prognostic and therapeutic values of the catecholamine pathway in COVID-19 disease were determined.

Human Samples
Nasopharyngeal swab samples and whole blood samples were obtained from patients in Greek hospitals or from healthy subjects. All categories of samples used in this study were totally independent and there was no intersection between them.
Nasopharyngeal swab samples were obtained upon routine diagnosis from adult patients tested positive for SARS-CoV-2, with mild or no symptoms (Mild CoV-2 group) or hospitalized in the intensive care unit (ICU) with severe/critical clinical symptoms (Severe CoV-2 group). Samples from subjects with a negative SARS-CoV-2 RT-PCR test constituted the control group for swab tissues (Negative group). All nasopharyngeal swab samples were obtained from the biobank of Eginition Hospital (Athens, Greece).
Whole blood samples were collected, with appropriate and standardized procedures, from COVID-19 patients who required admission to the COVID-19 specialized ward (Ward/CoV-2 group) or the ICU (ICU/CoV-2 group) of the Evangelismos General Hospital. Diagnosis of the viral infection was performed with RT-PCR in the nasopharyngeal swab samples. As control groups, samples from healthy donors (Healthy group), and from patients who were admitted to the ICU but tested negative for SARS-CoV-2 infection (ICU/non-CoV-2 group) were used. The ICU/non-CoV-2 group consisted mainly by subjects who were admitted to the ICU due to head injuries or multi-trauma patients (~70% of total cases).
Dopamine quantitation was performed in serum samples of the COVID-19 patients who required admission to the COVID-19 specialized ward (Ward/CoV-2 group), or in the ICU of the Evangelismos General Hospital. The COVID-19 patients who were hospitalized in the ICU were either treated with dexamethasone (ICU/CoV2 + Dexa group) or not (ICU/CoV-2 group). Serum samples from healthy donors (Healthy group), and subjects not infected with SARS-CoV-2 who were admitted in the ICU (ICU/non-CoV-2 group), were used as controls.

Cell and Virus Culture
Vero E6 cells, originally obtained from ATCC#CRL-1586, were cultured in Dulbecco's modified minimal essential medium (DMEM) (Invitrogen, Waltham, MA, USA) with high glycose (25 mM) after supplementation with penicillin (100 U/mL), streptomycin (100 µg/mL) and fetal calf serum (10% v/v), at 37 • C and 5% CO 2 . A COVID-19 patient's oropharyngeal swab sample was used to obtain the SARS-CoV-2 (isolate 30-287, lineage B1) virus [57]. Vero E6 cells at full confluency were inoculated with the isolated lineage and 4 days post-infection, the media were collected, and the virus was passaged once again in Vero E6 cells. Three days post-infection, the cultured cells were examined for a clear cytopathic effect (CPE), and the virus culture was harvested and stored at −80 • C after filtration (through 0.45µm filters). For titration of the propagated virus, Vero E6 cells were seeded in 96-well plates, inoculated with serial dilutions of the supernatant collected above and further cultured for 96 h at 37 • C. Inverted phase contrast microscopy was carried out to monitor the cytopathic effect (CPE), while the Reed and Muench method was used to determine the TCID 50 [58].

Infection Studies
Vero E6 cells were infected with the SARS-CoV-2 virus (multiplicity of infection, MOI = 0.1), in 6-well plates. As a control, mock-infected cells were used. The virus inoculum was replaced with culture medium 1 h post-inoculation. Total RNA was extracted from cells, 24 and 48 h post-infection using NucleoZOL (Macherey-Nagel), according to the manufacturer's instructions.

Total RNA Extraction from Human Samples
Whole blood samples were stored in tubes with RNA stabilizing solution (Applied Biosystems (Waltham, MA, USA), Life Technologies Corporation (Carlsbad, CA, USA)). Total RNA was extracted using the Tempus spin isolation kit (Applied Biosystems) according to the manufacturer's instructions.
Nasopharyngeal swab samples were collected in appropriate tubes containing 2 mL preservation fluid. Total RNA was extracted using the MagNA Pure LC Total Nucleic Acid Isolation Kit and the MagNa Pure LC 2.0 automated nucleic acid purifier (Roche, Basel, Switzerland), according to the manufacturer's instructions.

RNA Quantification by Reverse Transcription-Quantitative PCR (RT-qPCR)
Total RNA was reverse transcribed to cDNA using the reverse transcriptase of the Moloney murine leukemia virus (Promega, Madison, WI, USA) as recommended by the manufacturer, using oligo d(T)18 primers (New England Biolabs, Ipswich, MA, USA) and the recombinant ribonuclease inhibitor (Takara, San Jose, CA, USA). After cDNA synthesis, real-time quantitative PCR (qPCR) was performed for each target gene, using specific primer sets (Table 1), previously validated [18,33,[59][60][61], and the SYBR-Green Luna Universal qPCR Master Mix (New England Biolabs). Two different primer sets were used for the detection of DDC expression, recognizing the majority of DDC mRNA isoforms (variants: 1, 2, 4, 6 and 7) [29,62]. The first pair binds to the first (forward primer) and second (reverse primer) exons and was used for DDC quantification in the nasopharyngeal swab samples. The second primer set, binds to the second (forward primer) and third (reverse primer) exons and was used in the whole blood samples. The mRNA levels of the house-keeping gene 14-3-3-zeta polypeptide (YWHAZ) were used for normalization. The SARS-CoV-2 positive and negative samples were analyzed concurrently. All qPCR reactions were carried out in the VIIA7 platform (Applied Biosystems). The SARS-CoV-2 RNA quantification was assessed by RT-qPCR using the LightMix Modular Sarbecovirus E-gene Kit for the viral envelope protein (E)-encoding gene quantification and the LightMix Modular SARS-CoV-2 (COVID-19) RdRP Kit for the quantification of the RdRp gene, according to the manufacturer's instructions. Myostatin (MSTN) mRNA levels were quantified, using the LightMix ModularDx Kit MSTN Extraction Control kit (Roche), and served as a reference. The SARS-CoV-2 positive and negative samples were analyzed in parallel.
The cell gene expression data were analyzed utilizing the relative quantification method 2 −∆∆Ct [63]. The expression of the house-keeping gene 14-3-3-zeta polypeptide (YWHAZ) was used for the cell genes' expression normalization.

Elisa Assays
Dopamine levels were quantified in the sera of the following groups: Ward/CoV-2, ICU/CoV-2 + Dexa, ICU/CoV-2, Healthy and ICU/non-CoV-2, using the Dopamine ELISA kit (IBL International, Hamburg, Germany), according to the manufacturer's instructions. The detection limit of this assay is 4 pg/mL and the intra-and inter-assay coefficients of variability are 11% and 16.3%, respectively. Hence, this assay provides high specificity, good accuracy, and precision [64].

Cytopathic Effect Quantitation (CPE)
To record the effect of L-Dopa and dopamine on SARS-CoV-2-infection in vitro, Vero E6 cells were seeded in a 24-well plate (1.8 × 10 5 cells/well) and were inoculated with the produced viral stock (Section 2.2) for 1 h. Subsequently, the virus containing medium was replaced with DMEM complete supplemented or not with L-Dopa or dopamine in a non-cytotoxic concentration (50 µM). Seventy-two hours post-infection, the cells (10 wells per condition) were observed using an inverted microscope with phase contrast for the detection of SARS-CoV-2 induced cytopathic effect (CPE). The CPE foci in the L-Dopa or dopamine treated cells were counted and compared to those in the non-treated cells.

SARS-CoV-2-Human Protein-Protein Interactions (PPIs)
The experimentally determined protein-protein interactions (PPIs) between SARS-CoV-2 and human host proteins were retrieved from the publication of the respective high-throughput yeast-2-hybrid experiments in which all pairwise combinations of 28 SARS-CoV-2 proteins against~16,000 human proteins were systematically tested [65] and the available literature [41][42][43]45,66]. The experientially derived interconnectivity of human host factors was mined from the PICKLE (Protein InteraCtion KnowLedgebasE) meta-database [67].

Measurement of Cellular ATP Content
For measurement of the intracellular ATP content, Vero E6 cells were seeded in a 96-well plate and were subsequently treated or not with three consecutive dilutions of L-Dopa or dopamine (25, 50 and 100 µM). At 72-h post-inoculation, the cells were lysed, and the intracellular ATP levels were evaluated using the ViaLight HS BioAssay kit (Lonza, Basel, Switzerland), according to the manufacturer's instructions, in a GloMax 20/20 singletube luminometer (Promega Corporation, Madison, WI, USA) for 1 s. The total amount of protein was used for normalization. The maximum non-cytotoxic concentrations were chosen to treat the infected cells.

Statistical Analysis
Normality tests were conducted to test the distribution of variables (parametric or non-parametric) to select the appropriate statistical analysis tool each time. The statistical analysis of the comparison of gene expression, dopamine levels and age (continuous parameters) between COVID-19 patients and control groups was performed with either ordinary One Way ANOVA for parametric values, or Kruskal-Wallis test for the nonparametric values, followed by the appropriate post-hoc analysis. Multiple comparisons between the different groups were conducted with Tukey's test for ordinary ANOVA, and with Dunn's test for the non-parametric variables. The diagnostic accuracy of gene expression was evaluated with regression analysis (ROC). The area under curve (AUC), and the cut-off value based on Youden's index were calculated for each gene. The student's t-test (unpaired) or the Mann-Whitney U test were performed to compare the gene expression between men and women. Additionally, sex distribution among the studied groups was determined with the chi-squared test. Pearson's or Spearman's correlation coefficient (r) was performed to evaluate the correlation of (a) the expression levels of the various genes studied, (b) the expression level of each studied gene and the viral RNA, (c) the expression level of each studied gene and age (continuous variable), and the clinical measurements of the patients. The comparison of COVID-19 patients' dopamine levels with those of the control groups, in the presence of absence of comorbidities or mechanical ventilation was performed with Two Way ANOVA. GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA, USA) was used to perform all statistical analyses, and p < 0.05 (two-tailed) was considered statistically significant.

Results
We have previously shown that DDC mRNA levels were upregulated in the nasopharyngeal tissue of COVID-19 patients with mild or no symptoms (Mild CoV-2) compared to subjects not infected with SARS-CoV-2 (Negative) [19]. We also demonstrated that the regulation of this gene is SARS-CoV-2 infection-specific, as its mRNA levels were not altered in patients infected with another respiratory virus, compared to the Negative group. On the other hand, DDC expression negatively correlated with the SARS-CoV-2 viral load in the nasopharyngeal tissue of the same patients. Similarly, DDC was downregulated in cell culture upon SARS-CoV-2 infection. Additionally, the DDC gene displayed substantial coexpression with the ACE2 locus [28], and specifically with the mRNA of the IFN-inducible ACE2 truncated isoform dACE2 [17,18], but not with the SARS-CoV-2 receptor-encoding ACE2 mRNA, despite that both ACE2 and dACE2 genes were upregulated in these patients.

L-Dopa Decarboxylase as a COVID-19 Severity Marker in Nasopharyngeal Swab Samples
To elucidate whether the expression of DDC, ACE2, dACE2, or EPO are also altered in the nasopharyngeal tissue of COVID-19 patients with severe symptoms, we examined 21 samples obtained by SARS-CoV-2 infected patients admitted to the ICU and compared them to the Mild CoV-2 and Negative groups. The median age of the Severe CoV-2, Mild CoV-2, and Negative groups was 64 years (IQR = 61-75), 38 years (IQR = 31-52), and 47 years (IQR = 34-68), respectively ( Table 2). DDC mRNA expression: Regarding the expression of DDC in the nasopharyngeal swabs, we observed a statistically significant reduction of 20 mean-fold (Table S1, p < 0.0001) in the Severe CoV-2 group (median = 0.0007, IQR = 0.0004-0.002) compared to the Mild CoV-2 group (median = 0.034, IQR = 0.016-0.047) ( Figure 1A). Surprisingly, the mRNA levels of DDC in the Severe CoV-2 cohort were even lower (2 mean-fold, p < 0.0001) compared to the Negative control group (median = 0.0041, IQR = 0.0026-0.0057). In addition, DDC expression in this tissue was characterized by high diagnostic accuracy, as determined by the ROC curve analysis, with an AUC of 0.872 (95% confidence interval-CI = 0.770-0.974). The optimal cut-off value was 0.001, with a sensitivity of 57.14% and a specificity of 97.37%. In the Severe CoV-2 group, no statistically significant change in the expression of DDC was observed between men and women ( Figure S1A), nor a correlation with age was found ( Figure S2A), in accordance with our previous results in Negative and Mild CoV-2 subjects [19]. Table 2. Demographic data of subjects whose nasopharyngeal swab samples were used. dACE2 and ACE2 mRNA levels: The dACE2 mRNA levels in the nasopharyngeal tissue of Severe CoV-2 subjects (median = 0.020, IQR = 0.013-0.024) were higher up to 2.5 mean-fold (Table S1, p < 0.0001) compared to the Negative group (median = 0.008, IQR = 0.006-0.01), while there was no difference compared to the Mild CoV-2 patients (median = 0.021, IQR = 0.017-0.031) ( Figure 1B). On the contrary, we detected diminished levels of ACE2 mRNA (median = 0.00051, IQR = 0.0003-0.0012) in the Severe CoV-2 group compared to those obtained from the Negative subjects (median = 0.004, IQR = 0.003-0.007), and the Mild CoV-2 patients (median = 0.007, IQR = 0.004-0.011) (p < 0.0001) ( Figure 1C). Moreover, in the Severe CoV-2 group, the ROC curve analysis revealed a high diagnostic accuracy of the expression of both ACE2 (AUC = 0.900, 95% CI = 0.793-1.000) and dACE2 (AUC = 0.974, 95% CI = 0.936-1.000). The optimal cut-off value for ACE2 expression was at 0.001, with a sensitivity and a specificity of 76.19% and 97.37%, respectively. Regarding dACE2 expression, the cut-off point value was 0.012, with 80.95% sensitivity and 97.37% specificity. Finally, ACE2 and dACE2 expression did not correlate with either sex (Figure S1B,C) or age ( Figure S2B,C). Taking into account that dACE2 is considered to be an ISG [17,18], we also examined the expression of ISG56, which is most widely studied gene, compared to dACE2. The mRNA levels of ISG56 in the Severe CoV-2 group (median = 0.649, IQR = 0.399-0.940) were slightly increased (0.2 mean-fold, p = 0.0413) as compared to the ones of the Negative cohort (median = 0.521, IQR = 0.358-0.814), while they were lower (about 2 mean-fold, p = 0.0002) than in the Mild CoV-2 group (median = 1.194, IQR = 0.572-2.106) ( Figure 1D). Furthermore, the ROC curve analysis of the expression of ISG56 led to a much lower discriminating accuracy (AUC = 0.595, 95% CI = 0.436-0.755). At an optimal cut-off value of 0.097, the specificity and sensitivity were 96.88% and 40%, respectively. Finally, the mRNA levels of ISG56 in this tissue did not correlate with the patients' demographic data ( Figures S1D and S2D).

SARS-CoV
EPO mRNA levels: We have previously shown upregulation of EPO expression in the nasopharyngeal swab samples of COVID-19 patients with mild symptoms [19]. Here, we determined EPO mRNA levels in the Severe CoV-2 cohort and compared them to the Negative and Mild CoV-2 groups. We observed that EPO mRNA levels were significantly upregulated (up to 2.7 mean-folds, p < 0.0001) in the Severe CoV-2 group (median = 0.025, IQR = 0.018-0.037) (Table S1) compared to the Negative one (median = 0.01, IQR = 0.007-0.013) ( Figure 1E), whereas no significant difference was found between the Severe CoV-2 and the Mild CoV-2 (median = 0.025, IQR = 0.017-0.034) cohorts. In addition, the expression of EPO in the Severe CoV-2 patients was characterized by a high diagnostic accuracy, as determined by the ROC curve analysis (AUC = 0.911, 95% CI = 0.832-0.990), with a sensitivity of 66.67% and a specificity of 97.37%, at an optimal cut-off point of 0.021. The expression of EPO also did not correlate with either age or the sex of the patients (Figures S1E and S2E). Dunn's test was performed for multiple comparisons. ROC curves were generated, and AUC, 95% CI, p-values, and cut-off points with their specificity and sensitivity were reported. The red dotted   Dunn's test was performed for multiple comparisons. ROC curves were generated, and AUC, 95% CI, p-values, and cut-off points with their specificity and sensitivity were reported. The red dotted line denotes perfect chance (positive likelihood ratio = sensitivity/(1-specificity). * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs. Negative or Mild CoV-2.
Subsequently, we investigated the possibility that the observed changes in the expression of the above-mentioned genes in the nasopharyngeal swab samples of the Severe CoV-2 group, might be due to the fluctuation of cell populations related to inflammation. More specifically, we studied the expression of the epithelial cell marker EPCAM [68], and the nucleated hematopoietic cell marker CD45 [69], which is also used as a marker of immune cell populations. Amongst the three studied groups, we did not observe any significant alterations in the mRNA levels of EPCAM ( Figure 2A). The mRNA levels of CD45 were slightly increased, albeit not significantly, in the Severe CoV-2 cohort compared to the Negative group ( Figure 2B), which is in agreement with already published data [70]. These findings might possibly suggest that the observed alterations are not related to an enrichment of inflammation-related cell populations expressing these genes in the nasopharyngeal tissue of COVID-19 patients. In support of these results, concerning DDC, further analysis of the single cell gene expression data (RNA-sequencing) of Chua et al., derived from nasopharyngeal swabs of COVID-19 patients (available in Magellan COVID-19 Omics Explorer, https://digital.bihealth.org (accessed on 25 August 2022)), revealed that DDC is almost exclusively expressed in epithelial cells (secretory, ciliated and FOXN4) and not in immune cell populations [21].
Subsequently, we investigated the possibility that the observed changes in the expression of the above-mentioned genes in the nasopharyngeal swab samples of the Severe CoV-2 group, might be due to the fluctuation of cell populations related to inflammation. More specifically, we studied the expression of the epithelial cell marker EPCAM [68], and the nucleated hematopoietic cell marker CD45 [69], which is also used as a marker of immune cell populations. Amongst the three studied groups, we did not observe any significant alterations in the mRNA levels of EPCAM ( Figure 2A). The mRNA levels of CD45 were slightly increased, albeit not significantly, in the Severe CoV-2 cohort compared to the Negative group ( Figure 2B), which is in agreement with already published data [70]. These findings might possibly suggest that the observed alterations are not related to an enrichment of inflammation-related cell populations expressing these genes in the nasopharyngeal tissue of COVID-19 patients. In support of these results, concerning DDC, further analysis of the single cell gene expression data (RNA-sequencing) of Chua et al., derived from nasopharyngeal swabs of COVID-19 patients (available in Magellan COVID-19 Omics Explorer, https://digital.bihealth.org (accessed on 25 August 2022)), revealed that DDC is almost exclusively expressed in epithelial cells (secretory, ciliated and FOXN4) and not in immune cell populations [21]. As reports have shown that in severe cases of COVID-19 there is an impaired local intrinsic immunity against SARS-CoV-2 infection [70], we also examined the expression of LYN and CD74 genes (markers of dendritic and Langerhans cells, respectively), which are normally expressed in the antigen-presenting cells) of the nasopharyngeal tissue. Interestingly, the mRNA levels of both CD74 and LYN were excessively lower in the nasopharyngeal swab samples of the Severe CoV-2 group as compared to the Mild CoV-2 and Negative groups ( Figure 2C,D). The findings concerning CD74 expression are in accordance with previous studies reporting that SARS-CoV-2 infection reduced the expression of CD74 in the nose mucosal tissue [70].

Correlation of the mRNA Levels of DDC with dACE2 in the Nasopharyngeal Swabs of Severe CoV-2 Group
The previously observed correlation between the mRNA levels of DDC and dACE2 in the SARS-CoV-2 infected subjects with mild or no symptoms [19], prompted us to examine this relationship in more severe cases of the disease. In the present study, we observed a strong positive correlation between DDC and dACE2 expression in the Severe CoV-2 group (r = 0.704, p = 0.0005, Figure 3A), as opposed to a weaker correlation previously exhibited in the Negative group [19]. On the contrary, no correlation was detected between the mRNA levels of DDC and ACE2, ISG56 or EPO ( Figure 3B-D). Moreover, the expression of dACE2 did not correlate with ISG56, ACE2, or EPO in this tissue ( Figure S3A-C). Furthermore, because we previously detected a strong negative correlation of DDC and dACE2 mRNA levels with the SARS-CoV-2 viral load in the nasopharyngeal tissue of mild COVID-19 cases [19], we also examined the correlation of these genes with SARS-CoV-2 titers in the Severe CoV-2 cohort. Overall, our results showed that among the genes studied, only the expression of EPO negatively correlated with SARS-CoV-2 viral load levels (r = −0.461, p = 0.353) ( Figure S4). Interestingly, we detected significantly lower viral load levels in the Severe CoV-2 cohort (p = 0.0014) compared to the Mild CoV-2 group, possibly due to higher activation of the host innate immunity in the severe cases, restricting the replication of the virus [11] ( Figure S5).

Reduced DDC and dACE2 Expression in Whole Blood Samples of Hospitalized COVID-19 Patients
Next, we investigated whether the expression of DDC was reduced in whole blood samples of severe COVID-19 cases, which would reinforce our data derived from the nasopharyngeal tissue. For this purpose, we used two SARS-CoV-2 positive hospitalized cohorts, one admitted to a specialized ward (n = 35, Ward/CoV-2) and one hospitalized in the ICU (n = 32, ICU/CoV-2). Moreover, a group of SARS-CoV-2 negative patients admitted to the ICU (n = 30, ICU/non-CoV-2) and a group of healthy subjects (n = 35, Healthy) were used as negative controls. The characteristics of the aforementionedgroups are presented in Table 3 and Table S2. We observed that the expression of DDC was significantly decreased in both ICU/CoV-2 (median = 0.571, IQR = 0.295-0.747) and Ward/CoV-2 (median = 0.158, IQR = 0.115-0.279) patients compared to the Healthy subjects (median = 1.567, IQR = 1.244-2.519) ( Figure 4A), with the greatest decrease shown in the Ward/CoV-2 group (7.5 mean-fold, p < 0.0001; Table S3). The discriminating accuracy determined for DDC mRNA levels was high for both ICU/CoV-2 (AUC = 0.940, 95% CI = 0.886-0.995) and Ward/CoV-2 (AUC = 0.982, 95% CI = 0.961-1.000) cohorts. The optimal cut-off point for Ward-CoV-2 was calculated at 0.539, with a sensitivity of 88.57% and a specificity of 97.14%. The respective cut-off value for ICU/CoV-2 was 0.596, with a sensitivity and a specificity of 56.25% and 97.14%, respectively. Additionally, the expression of DDC did not correlate with either the sex or the age of the patients ( Figures S6A and S7A). Because our data (Section 3.1) suggested DDC as a potential disease severity marker in the nasopharyngeal tissue, we aimed to investigate whether the observed alterations of its expression in the blood are SARS-CoV-2 infection-specific or are due to ICU admission. For this purpose, we determined the mRNA levels of DDC in the ICU/non-CoV-2 patients, as a negative control, and compared them to the ICU/CoV-2 and Healthy groups. The mRNA levels of DDC in the ICU/non-CoV-2 cohort were significantly higher than the ICU/CoV-2 group and did not differ from the healthy subjects. This indicates that the reduction in DDC expression is specific to SARS-CoV-2 infection and is not associated to ICU admission. Table 3. Demographic data and in-hospital measured laboratory parameters of subjects whose blood samples were used.  Taking into account the correlation of DDC and dACE2 expression that we observed in the nasopharyngeal tissues of both mild [19] and severe COVID-19 (present study) patients, we also evaluated dACE2 mRNA levels, in the same whole blood samples that were used for DDC expression analysis. The dACE2 mRNA levels were lower in both the ICU/CoV-2 (median = 0.042, IQR = 0.011-0.066) and the Ward/CoV-2 (median = 0.009, IQR = 0.001-0.030) cohorts, compared to the Healthy group (median = 0.130, IQR = 0.088-0.193), in accordance with the results obtained for DDC. The greatest decrease, of approximately 30 meanfold (p < 0.0001, Table S3), was observed in the Ward/CoV-2 group ( Figure 5A). In agreement to our findings on dACE2 expression, a significant reduction in ISG56 mRNA levels in both the Ward/CoV-2 (1.6 mean-fold, p = 0.0075) and ICU/CoV-2 (up to 3.4 meanfold, p < 0.0001) cohorts was detected compared to the healthy subjects ( Figure 5B). The ROC curve analysis of dACE2 expression revealed a high diagnostic accuracy for both ICU/CoV-2 (AUC = 0.900, 95% CI = 0.826 to 0.975) and Ward/CoV-2 (AUC = 0.967, 95% CI = 0.932-1.000) patients. The optimal cut-off values for the two groups were 0.034 (sensitivity = 41.94% and specificity = 94.12%), and 0.024 (sensitivity = 71.42% and specificity = 97.06%), respectively. A lower diagnostic accuracy was detected for ISG56 expression, especially in the Ward/CoV-2 group (AUC = 0.696, 95% CI = 0.571-0.820). The optimal cut-off points for the ICU/CoV-2 and Ward/CoV-2 patients were calculated at 0.135, with a sensitivity of 35.48% and a specificity of 97.06% for the ICU/CoV-2 cohort, and a sensitivity of 32.35% and a specificity of 97.06% for the Ward/CoV-2 one. Furthermore, there was no significant correlation between the expression of dACE2 or ISG56 with sex or age (Figures S6B,C and S7B,C).

SARS-CoV-2
To examine the possibility that the observed alterations in the expression of the abovementioned genes might be due to a possible enrichment of inflammation-related cell populations in the blood samples of SARS-CoV-2 infected subjects exhibiting severe symptoms, the expression levels of the immune cell markers CD45, CD74 and LYN were analyzed. In the blood, CD45 is mainly considered a marker of myeloid and T-cells, while CD74 and LYN constitute a marker of B-cells. A non-statistically significant decrease in CD45 expression was observed in the Ward/CoV-2 group, as compared to the healthy subjects. Previous studies have reported a reduction in the expression of this gene in COVID-19 patients' blood samples [71]. We did not observe any significant alteration in the expressions of CD74 and LYN in the whole blood of hospitalized COVID-19 patients compared to the Healthy group ( Figure S8A-C).  Taking into account the correlation of DDC and dACE2 expression that we observed in the nasopharyngeal tissues of both mild [19] and severe COVID-19 (present study) patients, we also evaluated dACE2 mRNA levels, in the same whole blood samples that were used for DDC expression analysis. The dACE2 mRNA levels were lower in both the ICU/CoV-2 (median = 0.042, IQR = 0.011-0.066) and the Ward/CoV-2 (median = 0.009, IQR = 0.001-0.030) cohorts, compared to the Healthy group (median = 0.130, IQR = 0.088-  Dunn's test was performed for multiple comparisons. ROC curves were generated, and AUC, 95% CI, p-values, and cut-off points with their specificity and sensitivity were reported. The red dotted line denotes perfect chance (positive likelihood ratio = sensitivity/(1-specificity) =1). **** p < 0.0001. specificity = 97.06%), respectively. A lower diagnostic accuracy was detected for ISG56 expression, especially in the Ward/CoV-2 group (AUC = 0.696, 95% CI = 0.571-0.820). The optimal cut-off points for the ICU/CoV-2 and Ward/CoV-2 patients were calculated at 0.135, with a sensitivity of 35.48% and a specificity of 97.06% for the ICU/CoV-2 cohort, and a sensitivity of 32.35% and a specificity of 97.06% for the Ward/CoV-2 one. Furthermore, there was no significant correlation between the expression of dACE2 or ISG56 with sex or age ( Figures S6B,C and S7B,C).   ROC curves were generated and AUC, 95% CI and cut-off points with their specificity and sensitivity were calculated. p-value was ** p < 0.01, **** p < 0.0001.

Decreased Dopamine Levels in the Serum of Hospitalized COVID-19 Patients
Based on our previous cell culture-derived data supporting that the infection of cells with SARS-CoV-2 reduced DDC expression [19] and the lower DDC mRNA levels detected in the blood of hospitalized SARS-CoV-2 infected subjects, we measured the dopamine levels in the serum of the COVID-19 severe/critical cases of the disease and compared them to two control groups. Specifically, we analyzed two SARS-CoV-2 infected cohorts; one group of patients hospitalized in the specialized COVID-19 ward (n = 27, Ward/CoV-2), and a second group that was admitted to the ICU ( Table 4). The latter was further separated into patients receiving dexamethasone treatment (n = 29, ICU/CoV-2 + Dexa) or not (n = 33, ICU/CoV-2). The Healthy group (n = 33, Healthy), and a group of patients hospitalized in the ICU, tested negative for SARS-CoV-2 infection (n = 27, ICU/non-CoV-2), were used as controls. Decreased dopamine levels were observed in all cohorts of SARS-CoV-2 hospitalized patients, by about 3-3.8 mean-fold (p < 0.0001), as compared to the Healthy subjects (median = 197.8 pg/mL, IQR = 138.2-316.2 pg/mL) ( Figure 6A). The most significant reduction was exhibited in the Ward/CoV-2 group (median = 62.22 pg/mL, IQR = 43.44-182.4 pg/mL). Interestingly, no significant alteration in the serum dopamine levels was observed between the Healthy and the ICU/non-CoV-2 groups. Moreover, lower dopamine levels were detected in both the ICU/CoV-2 (median = 76.43 pg/mL, IQR = 63.76-107.4 pg/mL) and the ICU/CoV-2 + Dexa (median = 93.06 pg/mL, IQR = 71.30-127.6 pg/mL) patients, compared to the ICU/non-CoV-2 subjects (median = 146.3 pg/mL, IQR = 107.9-202.4 pg/mL) (p ICU/CoV-2 = 0.002, p ICU/CoV-2+Dexa = 0.029), indicating that our results on dopamine levels are SARS-CoV-2 infection-specific. The patients who were treated with dexamethasone (ICU/CoV-2 + Dexa) exhibited slightly higher levels of dopamine compared to the non-treated ones, which is in accordance with reported data showing that dexamethasone treatment positively affects the circulating dopamine [72]. Furthermore, the ROC curve analysis revealed a high diagnostic accuracy for the dopamine levels in all groups except from the ICU/CoV-2 + Dexa group ( Figure 6B). All the above data combined, indicate that there is a possible negative association between serum dopamine levels and SARS-CoV-2 infection, independently of ICU admission. Therefore, the reduction in the mRNA levels of DDC in the whole blood samples of the hospitalized COVID-19 patients is accompanied by a significant decrease of serum dopamine levels, and both effects appeared to be SARS-CoV-2 infection-specific. Table 4. Demographic data and in-hospital measured laboratory parameters of subjects whose sera were used.  Next, we investigated whether the observed alterations in the serum dopamine levels in the COVID-19 patients could be related to underlying comorbidities. For this purpose, we tested the presence of underlying diseases as potential dopamine regulators in our three ICU cohorts. No significant correlation between dopamine levels and the underlying comorbidities was detected in the three patient groups ( Figure S9A). Hypertension, the most common comorbidity in our patient cohorts, also had no significant impact on serum  Next, we investigated whether the observed alterations in the serum dopamine levels in the COVID-19 patients could be related to underlying comorbidities. For this purpose, we tested the presence of underlying diseases as potential dopamine regulators in our three ICU cohorts. No significant correlation between dopamine levels and the underlying comorbidities was detected in the three patient groups ( Figure S9A). Hypertension, the most common comorbidity in our patient cohorts, also had no significant impact on serum dopamine levels ( Figure S9B). No effect was exerted by the application of mechanical ventilation in COVID-19 patients ( Figure S9C). These data suggest that systemic dopamine is negatively influenced by SARS-CoV-2 infection independently of the presence of comorbidities.

Alteration of Catecholamine Pathway-Related Genes Expression in Whole Blood Samples of SARS-CoV-2-Infected Hospitalized Individuals
Our aforementioned-results on whole blood DDC expression and serum dopamine levels in hospitalized COVID-19 patients indicated that the activity of the catecholamine pathway is affected by SARS-CoV-2 infection. To further investigate the association between SARS-CoV-2 infection and the function of the catecholamine pathway, the expression of other catecholamine pathway-related genes in blood was examined. These genes include MAOA and MAOB, which catalyze the oxidative deamination of catecholamines in the cytosol, the vesicular membrane-bound dopamine-beta-hydroxylase (DBH), which synthesizes norepinephrine, and the vesicular monoamine transporter (VMAT2), which is responsible for the transfer of catecholamines into the storage vesicles. Significantly higher MAOA mRNA levels (up to 15 mean-fold, Table S3) (p < 0.0001) were detected in the whole blood samples of the ICU/CoV-2 group (median = 0.348, IQR = 0.159-0.586) compared to the Healthy group (median = 0.020, IQR = 0,011-0.030), with a high discriminating accuracy (AUC = 0.992, 95% CI = 0.978-1.000) ( Figure 7A). The cut-off value was calculated at 0.103, with a sensitivity of 93.33% and a specificity of 96.88%. In addition, a slight increase of MAOA expression (up to 3.6 mean-fold, Table S3) (p = 0.046) was detected in the Ward/CoV-2 patients (median = 0.066, IQR= 0.012-0.147) compared to the healthy subjects. However, our results on the Ward/CoV-2 cohort were characterized by a lower diagnostic accuracy (AUC = 0.698, 95% CI = 0.564-0.833), and an optimal cut-off value at 0.097, with a sensitivity and a specificity of 40% and 96.88%, respectively. Interestingly, MAOB expression was similarly elevated in both the ICU/CoV-2 (median = 0.060, IQR = 0.019-0.115) and the Ward/CoV-2 groups (median = 0.065, IQR = 0.026-0.119) (up to 10 mean-fold, p < 0.0001), compared to the healthy subjects (median = 0.005, IQR = 0.003-0.009) ( Figure 7B). The ROC curve analysis showed a high diagnostic accuracy for MAOB expression in the ICU/CoV-2 (AUC = 0.922, 95% CI = 0.843-1.000) and the Ward/CoV-2 (AUC = 0.977, 95% CI = 0.951-1.000) cohorts. For the ICU/CoV-2 group, the optimal cut-off point was estimated at 0.021 (73.33% sensitivity and 97.06 specificity). The cut-off for the Ward/CoV-2 group was 0.021 (80% sensitivity and 97.06% specificity). The elevated expression of MAOA and MAOB in the whole blood samples is in agreement with the reduced serum dopamine levels in these patients. On the contrary, the expression of DBH and VMAT2 in the SARS-CoV-2 infected cohorts did not exhibit any significant alteration in comparison to the healthy subjects ( Figure 7C,D). Finally, the expression of these genes did not correlate either with sex or age ( Figures S6D-G and S10).

Correlation between the Expression of DDC with dACE2 and MAOs in Whole Blood Samples of Hospitalized COVID-19 Patients
The two hospitalized SARS-CoV-2 patient cohorts (Ward/CoV-2 and ICU/CoV-2) were merged, and co-expression correlations for all studied genes were examined in their whole blood samples. A strong positive correlation between the expression of DDC and dACE2 was detected in the SARS-CoV-2 infected patients and the healthy subjects (r = 0.804, p < 0.0001 and r = 0.689, p < 0.0001, respectively) ( Figures 8A and S11A), as determined by the Spearman's correlation coefficient. This is in agreement with our previous data from the nasopharyngeal swab samples of COVID-19 patients with mild/no [19] or severe symptoms (Section 3.2). Moreover, a moderate positive correlation between the expression of DDC and MAOA (r = 0.527, p < 0.0001) was identified in the SARS-CoV-2 infected group, while there was no correlation in the Healthy group ( Figures 8B and S11B). No correlation was found between DDC and MAOB mRNA levels in either group (r = 0.069, p = 0.582) (Figures 8C and S11C). Similarly, we observed that the expression of dACE2 had a moderately strong positive correlation with MAOA (r = 0.4631, p = 0.0001) in the SARS-CoV-2 infected group ( Figure 8D), and no correlation in the Healthy group ( Figure  S11D). No significant dACE2 and MAOB co-expression was detected in neither the patients nor the healthy subjects ( Figure S11E To exclude the possibility that our results in the whole blood expression of the aforementioned-genes in the hospitalized COVID-19 patients are affected by the presence of comorbidities, we grouped the SARS-CoV-2 infected subjects according to the presence of underlying diseases. Our results showed that the mRNA levels of the catecholamine path- To exclude the possibility that our results in the whole blood expression of the aforementioned-genes in the hospitalized COVID-19 patients are affected by the presence of comorbidities, we grouped the SARS-CoV-2 infected subjects according to the presence of underlying diseases. Our results showed that the mRNA levels of the catecholamine pathway-related genes (DDC, MAOA, and MAOB) were not affected (Figure S12A-C) by the presence of comorbidities, while dACE2 expression was negatively affected (p = 0.0083) ( Figure S12D). Thus, the observed alterations in the whole blood expression of the catecholamine biosynthesis/metabolism genes in hospitalized COVID-19 patients are due to SARS-CoV-2 infection.

The effect of SARS-CoV-2 Infection on Catecholamine Pathway-Related Gene Expression in Cultured Epithelial Cells
We have previously shown that SARS-CoV-2 infection significantly reduces DDC expression in Vero E6 infected cells at 24 and 48 h post-infection (h.p.i.) [19]. This was also observed in A549 cells, as indicated by data derived from the RNA-sequencing Skyline database (http://rstats.immgen.org/Skyline_COVID-19/skyline.html (accessed on 28 August 2022)) [73]. To support our aforementioned in vivo results on the effect of SARS-CoV-2 infection on the expression of catecholamine-related genes with cell culture data, the expression of MAOA, MAOB, DBH, and VMAT2 in Vero E6 cells infected for 24 and 48 h was studied. As shown in Figure 9, SARS-CoV-2 infection significantly increased the expression of MAOA (up to 2-fold) at both time points. A moderate increase in MAOB expression at 24 h post-infection, followed by a substantial increase (up to 2.5-fold) at 48 h post-infection was detected. Slightly enhanced mRNA levels of DBH were detected at both time points compared to the mock-infected control ( Figure 9). Moreover, the expression of VMAT2 was significantly enhanced especially at 24 h post-infection. To complement these results with data from human epithelial cells, we studied the impact of SARS-CoV-2 on the expression of catecholamine-related genes in A549 cells, by processing data from the Skyline database. Higher mRNA levels of MAOA, MAOB and VMAT2 were observed upon infection in these cells also, while DBH and COMT expression was not significantly altered ( Figure S13). To conclude, SARS-CoV-2 negatively altered the mRNA levels of DDC, whereas MAOA and MAOB positively, possibly to reduce the intracellular dopamine levels. In turn, non-cytotoxic concentrations of dopamine or L-Dopa ( Figure S14) in SARS-CoV-2 infected Vero E6 cells decreased the virus-related cytopathic effect, by 55% and 59%, respectively ( Figure 10). This suggests that there is a reverse relationship between dopamine levels and SARS-CoV-2 replication in vitro, possibly mediated through dopamine or L-Dopa signaling and/or elevation of the intracellular levels of dopamine. Thus, the effects on the expression of the catecholamine pathway-related genes observed upon infection of Vero E6 cells, which possibly result in lower dopamine levels, favor the propagation of SARS-CoV-2 virus. Therefore, these are possibly caused by the virus itself and not as a cell antiviral response. We have previously shown that SARS-CoV-2 infection significantly reduces DDC expression in Vero E6 infected cells at 24 and 48 h post-infection (h.p.i.) [19]. This was also observed in A549 cells, as indicated by data derived from the RNA-sequencing Skyline database (http://rstats.immgen.org/Skyline_COVID-19/skyline. html) [73]. To support our aforementioned in vivo results on the effect of SARS-CoV-2 infection on the expression of catecholamine-related genes with cell culture data, the expression of MAOA, MAOB, DBH, and VMAT2 in Vero E6 cells infected for 24 and 48 h was studied. As shown in Figure 9, SARS-CoV-2 infection significantly increased the expression of MAOA (up to 2fold) at both time points. A moderate increase in MAOB expression at 24 h post-infection, followed by a substantial increase (up to 2.5-fold) at 48 h post-infection was detected. Slightly enhanced mRNA levels of DBH were detected at both time points compared to the mock-infected control ( Figure 9). Moreover, the expression of VMAT2 was significantly enhanced especially at 24 h post-infection. To complement these results with data from human epithelial cells, we studied the impact of SARS-CoV-2 on the expression of catecholamine-related genes in A549 cells, by processing data from the Skyline database. Higher mRNA levels of MAOA, MAOB and VMAT2 were observed upon infection in these cells also, while DBH and COMT expression was not significantly altered ( Figure  S13). To conclude, SARS-CoV-2 negatively altered the mRNA levels of DDC, whereas MAOA and MAOB positively, possibly to reduce the intracellular dopamine levels. In turn, non-cytotoxic concentrations of dopamine or L-Dopa ( Figure S14) in SARS-CoV-2 infected Vero E6 cells decreased the virus-related cytopathic effect, by 55% and 59%, respectively ( Figure 10). This suggests that there is a reverse relationship between dopamine levels and SARS-CoV-2 replication in vitro, possibly mediated through dopamine or L-Dopa signaling and/or elevation of the intracellular levels of dopamine. Thus, the effects on the expression of the catecholamine pathway-related genes observed upon infection of Vero E6 cells, which possibly result in lower dopamine levels, favor the propagation of SARS-CoV-2 virus. Therefore, these are possibly caused by the virus itself and not as a cell antiviral response.

Meta-Analysis of the Direct Protein-Protein Interactions of the Catecholamine Biosynthesis/Metabolism and Hypoxia Pathways with Specific SARS-CoV-2 Proteins
Finally, by performing meta-analysis of published protein-protein interaction (PPIs) data, we further investigated the association of specific SARS-CoV-2 proteins with the main components of the catecholamine biosynthesis/metabolism and hypoxia pathways. Two main sub-networks of experimentally supported PPIs were identified for the proteins whose gene expression was described above, in both COVID-19 patients and infected cultured epithelial cells. The first involves direct PPIs of the SARS-CoV-2 non-structural proteins 2 and 8 (nsp2, nsp8) with the dopamine catabolizing enzyme MAOA ( Figure 14A). In the second sub-network ( Figure 14B), experimental PPIs were identified between the viral non-structural proteins 9 and 13 (nsp9, nsp13) and L-lactate dehydrogenase A (LDHA) and B chain (LDHB), respectively. Moreover, a bridging interconnection between three human host factors ( Figure 14B), HIF that transcriptionally regulates both EPO and LDHA genes, and a three-member clique constructed by two host genes (LDHA, LDHB) and one viral (nsp13), were revealed.

Meta-Analysis of the Direct Protein-Protein Interactions of the Catecholamine Biosynthesis/Metabolism and Hypoxia Pathways with Specific SARS-CoV-2 Proteins
Finally, by performing meta-analysis of published protein-protein interaction (PPIs) data, we further investigated the association of specific SARS-CoV-2 proteins with the main components of the catecholamine biosynthesis/metabolism and hypoxia pathways. Two main sub-networks of experimentally supported PPIs were identified for the proteins whose gene expression was described above, in both COVID-19 patients and infected cultured epithelial cells. The first involves direct PPIs of the SARS-CoV-2 non-structural proteins 2 and 8 (nsp2, nsp8) with the dopamine catabolizing enzyme MAOA ( Figure 14A). In the second sub-network ( Figure 14B), experimental PPIs were identified between the viral non-structural proteins 9 and 13 (nsp9, nsp13) and L-lactate dehydrogenase A (LDHA) and B chain (LDHB), respectively. Moreover, a bridging interconnection between three human host factors ( Figure 14B), HIF that transcriptionally regulates both EPO and LDHA genes, and a three-member clique constructed by two host genes (LDHA, LDHB) and one viral (nsp13), were revealed. Finally, by performing meta-analysis of published protein-protein interaction (PPIs) data, we further investigated the association of specific SARS-CoV-2 proteins with the main components of the catecholamine biosynthesis/metabolism and hypoxia pathways. Two main sub-networks of experimentally supported PPIs were identified for the proteins whose gene expression was described above, in both COVID-19 patients and infected cultured epithelial cells. The first involves direct PPIs of the SARS-CoV-2 non-structural proteins 2 and 8 (nsp2, nsp8) with the dopamine catabolizing enzyme MAOA ( Figure 14A). In the second sub-network ( Figure 14B), experimental PPIs were identified between the viral non-structural proteins 9 and 13 (nsp9, nsp13) and L-lactate dehydrogenase A (LDHA) and B chain (LDHB), respectively. Moreover, a bridging interconnection between three human host factors ( Figure 14B), HIF that transcriptionally regulates both EPO and LDHA genes, and a three-member clique constructed by two host genes (LDHA, LDHB) and one viral (nsp13), were revealed.

Discussion
Our group has previously identified the impact of SARS-CoV-2 infection on the gene expression of the dopamine biosynthetic enzyme L-Dopa decarboxylase (DDC), and of other genes associated with DDC and SARS-CoV-2 infection, such as the ACE2/dACE2 locus, and the HIF-target gene EPO, in the nasopharyngeal swab samples of COVID-19 patients with mild or no symptoms. These genes were selected based on their role in the infection cycle of SARS-CoV-2 (ACE2) [13], as markers of interferon signaling (ISG56, dACE2) [17,18,74], or survival (EPO) [19]-especially under hypoxic conditions-and based on previous data showing the co-expression link between DDC and ACE2 locus [19,28], and their regulation under hypoxic conditions [39,[46][47][48][49]51]. The expression of these genes was found excessively enhanced in the nasopharyngeal tissue of COVID-19 patients, and interestingly, the mRNA levels of DDC and dACE2 had a reverse correlation with the viral load. The DDC expression regulation appeared to be SARS-CoV-2 infection-specific, as no effect was observed in respective samples of flu-infected patients. On the other hand, based on our cell culture-derived results, SARS-CoV-2 infection downregulated the expression of the aforementioned-genes in Vero E6 cells. Consequently, our previous data from human samples and in vitro studies, suggest that the expression of these genes is affected by SARS-CoV-2-infection, and for the first time reveal that the virus influences the expression of DDC, the product of which catalyzes the production of dopamine, the first catecholamine of the respective pathway. The difference observed in gene expression regulation in the nasopharyngeal swab samples of SARS-CoV-2 infected subjects with mild/no symptoms, compared to the in vitro infection studies, could be attributed to the antiviral response of the host attempting to restrict the virus propagation. Moreover, a strong correlation between the expression of DDC and dACE2 was identified in the patient group, reinforcing the possibility of the involvement of DDC in this antiviral response mechanism against the SARS-CoV-2 virus.
In the present study, we aimed to expand our previous findings, by analyzing the mRNA levels of the same genes in nasopharyngeal swab and whole blood samples of COVID-19 patients with severe symptoms of the disease.
Our results from the three swab cohorts (Mild CoV-2, Severe CoV-2, Negative) identified DDC expression as a potential marker of COVID-19 severity. In contrast to the DDC expression data obtained from the nasopharyngeal swabs of COVID-19 patients with mild/no symptoms [19], the severe cases of the disease exhibited lower DDC mRNA levels compared to the SARS-CoV-2-negative subjects. Additionally, in the blood of hospitalized COVID-19 patients (Ward/CoV-2 group and ICU/CoV-2 group), DDC mRNA levels were reduced, while no alteration was observed in the group of ICU non-COVID-19 patients (ICU/non-CoV-2), compared to the group of healthy subjects (Healthy). Notably, the presence of underlying comorbidities, with hypertension being the most frequent, did not seem to have an impact on the effect of SARS-CoV-2 infection on DDC expression in the hospitalized COVID-19 patients, as similar values were observed in patients with and without comorbidities. Moreover, DDC whole blood mRNA levels in both the ward and ICU groups of COVID-19 patients had a high diagnostic accuracy. The expression of ACE2 in the nasopharyngeal swab samples of COVID-19 patients exhibited the same pattern of changes as DDC; its expression was induced in patients with mild symptoms, while reduced in patients with severe symptoms. Nonetheless, the mRNA levels of this gene in both groups were only slightly altered compared to the non-COVID-19 group (Negative). Additionally, we were unable to detect the mRNA of ACE2 in the whole blood samples and thus, we could not examine whether its expression was affected by SARS-CoV-2 infection, or by the presence of underlying diseases. For the remaining genes, the expression of both dACE2 and EPO was induced in the nasopharyngeal swabs of COVID-19 patients, failing, however, to distinguish patients with mild from the ones with severe symptoms. In the whole blood samples of hospitalized COVID-19 patients, dACE2 and EPO mRNA levels were diminished, compared to healthy subjects. The expression of dACE2 seemed to be influenced by underlying comorbidities. Finally, the expression of ISG56 in the nasopha-ryngeal swab samples of COVID-19 patients exhibited a lower induction than that of DDC in patients with mild symptoms, whereas the group of patients with severe symptoms had equal ISG56 mRNA levels with the non-COVID-19 group (Negative). The ISG56 expression had a much lower diagnostic accuracy compared to DDC expression. On the other hand, in the whole blood samples, ISG56 was significantly downregulated in both hospitalized COVID-19 patient cohorts, in accordance with previous reports on the downregulation of IFN-1 signaling by SARS-CoV-2 [75,76]. Therefore, although dACE2 and ISG56 are considered interferon-stimulated genes, in the nasopharyngeal tissue of the severe COVID-19 cases, their mRNA levels were differentially regulated upon infection. This is in agreement with our observations in SARS-CoV-2 infected Vero E6 cells at 24 h.p.i. [19], suggesting that dACE2 expression may be also subjected to other regulation mechanisms in viral infection. Moreover, in the blood of hospitalized COVID-19 patients, the mRNA levels of DDC and of dACE2 negatively correlated with ISG56. This finding could suggest that DDC and dACE2 expression is associated with the excessive activation of the host immune response against SARS-CoV-2 observed in severe cases. More importantly, in the nasopharyngeal tissue, which is the most easily accessible and most commonly obtained sample from COVID-19 patients, the expression of DDC could serve as the most reliable marker of COVID-19 severity among the genes studied. Hence, it might be able to readily distinguish patients requiring hospitalization. Viral infection has been shown, by our previous studies on HCV and DENV, to affect, in addition to DDC, the expression of other genes encoding enzymes of the catecholamine biosynthesis and metabolism, or catecholamine transporters [37,38]. Therefore, in the whole blood samples of hospitalized COVID-19 patients, we examined the gene expression of MAOA and MAOB, the products of which perform the oxidative deamination of biogenic amines, including dopamine, norepinephrine, and serotonin, DBH, which converts dopamine to norepinephrine, and of the catecholamine transporter-gene VMAT2. Moreover, the expression of these genes was analyzed in SARS-CoV-2-infected Vero E6 cells. Our results showed that SARS-CoV-2 infection did not affect the expression of DBH and VMAT2, but elevated, however, the mRNA levels of MAOA and MAOB in infected cells and in the whole blood samples of hospitalized COVID-19 patients. Interestingly, the MAOA enzyme has been shown to directly interact with the SARS-CoV-2 non-structural proteins nsp2 and nsp8 [41], which reinforces the association of the catecholamine degradation pathway with viral infection. Similar to the DDC mRNA expression, the presence of comorbidities did not affect the mRNA levels of MAOA and MAOB in the whole blood samples of hospitalized COVID-19 patients. Therefore, SARS-CoV-2 infection negatively affects the biosynthetic part and positively the catabolizing part of the dopamine biosynthesis/catabolism pathway both in vitro and in the blood tissue of infected patients with severe symptoms of the disease. At the cellular level, this SARS-CoV-2-mediated regulation of the catecholamine biosynthesis/metabolism pathway could diminish the levels of catecholamines in the cytosol. However, the mRNA levels of MAOA positively correlated with DDC in the blood samples of hospitalized COVID-19 patients, raising a question on the role of serum dopamine levels in these patients, as dopamine is a product of DDC and a substrate of MAOA enzymatic activities. Diminished dopamine levels were found in the hospitalized COVID-19 patients compared to the two control groups, the ICU non-COVID-19 and the Healthy ones. This is in agreement with the results obtained in the same tissue on the mRNA levels of DDC and MAOA/-B. Moreover, neither the underlying comorbidities nor mechanical ventilation seemed to influence dopamine levels. Therefore, the increased expression of MAOA/-B in the blood were not due to enhanced levels of dopamine, but possibly contribute to its reduction. Interestingly, the reduction in blood dopamine levels in COVID-19 patients was predicted first by Nataf et al. based on the effects of SARS-CoV-2 on the catecholamine biosynthesis/metabolism pathway in vitro in infected enterocytes [35].
In turn, exogenous administration of L-Dopa and dopamine in SARS-CoV-2-infected epithelial cells negatively influenced the virus-derived cytopathic effect. Dopamine treatment has been shown to negatively affect SARS-CoV-2 replication, by inducing Type-I interferons and ISG genes, while downregulating the gene expression of pro-inflammatory mediators [36]. Another explanation, supported by our previous data on Huh7 cells, could be a negative effect on the virus exerted by L-Dopa and dopamine uptake by the cells, which then convert L-Dopa to dopamine. Therefore, similarly to HCV and Dengue viruses [37,38], the enhancement of the dopamine intracellular levels could also be responsible for the reduced cytopathic effect observed in SARS-CoV-2 infection. These highlight that L-Dopa or dopamine treatment could be a potential antiviral therapeutic approach especially in cases where interferon response is impaired leading to increased viral titers of SARS-CoV-2 [77]. The dopamine concentrations used in the above in vitro studies are significantly higher than those detected in human blood; however, dopamine levels in many organs/tissues differ substantially from those of blood. Interestingly, in the rodent kidney the detected dopamine levels are comparable to those used in our study to treat the kidney-derived Vero E6 cells [78]. Moreover, we showed that the expression of DDC and dACE2 maintain their strong positive correlation in both swab and blood samples, even though their mRNA levels were reversely regulated in the swab samples of COVID-19 severe cases. In the nasopharyngeal swab and blood samples, there are no indications that the gene expression results are attributed to an enrichment of inflammation related cell populations, as no significant differences were observed in the expression of the studied markers of the immune populations among the different groups of patients.
Blood samples from the hospitalized SARS-CoV-2-infected patients, before infection were not available. Therefore, we cannot exclude the possibility that these subjects might have lower dopamine and DDC expression levels compared to the general population, which could render them more susceptible to severe forms of COVID-19. In this case, dopamine levels could serve as a prognostic marker of disease severity. Moreover, dopamine intake to restore normal levels would probably have significant preventive and/or therapeutic value in these patients. Single nucleotide polymorphisms (SNPs) in the coding region of DDC, which have been reported to be associated with neurological disorders and response to treatment, may be related to reduced activity of the enzyme [79,80]. The frequency of two such polymorphisms in DDC (SNP ID: rs188088947 and SNP ID: rs11575542), which produce missense mutations, was compared between the group of COVID-19 patients with mild/no symptoms and the severe/critical. We did not detect any enrichment of these alleles in the group of COVID-19 patients with severe symptoms (data not shown). However, we cannot exclude the possibility of enrichment, in the group of COVID-19 patients with severe symptoms, of other DDC alleles that may affect for example the expression of the gene.
The levels of LDH and FIB proteins in the blood samples of the hospitalized COVID-19 patients were the only in-hospital routinely measured laboratory parameters that had a moderate positive correlation with the expression of DDC, dACE2, and MAOA genes in this tissue. Furthermore, SARS-CoV-2 severity has been previously linked with higher levels of LDH and FIB proteins [81][82][83][84][85][86][87]. Consistently, increased levels of these proteins were detected in the whole blood samples of the hospitalized groups of COVID-19 patients of the present study, compared to healthy subjects. The expression of the LDH gene was also enhanced in the SARS-CoV-2-infected Vero E6 cells at 24 h.p.i. Although LDH protein levels are usually indicative of tissue/cell damage, the expression of this gene is positively regulated by hypoxia, as the encoded product is a key enzyme of the anaerobic glycolysis pathway the activity of which is induced under low oxygen conditions. Interestingly, the mRNA levels of HIF-1α and GLUT1 genes were also found significantly elevated in the whole blood samples of the ICU COVID-19 patients. Gene HIF-1α encodes the subunit of the HIF transcription factor that is stabilized by hypoxia, while GLUT1 is one of the many genes that the HIF factor upregulates. Gene GLUT1 was also induced in SARS-CoV-2-infected Vero E6 cells at 24 h.p.i. The increased levels of the LDH protein and the induction of the expression of HIF-1α and GLUT1 in the blood samples of the hospitalized COVID-19 patients, are in agreement with several studies highlighting that severe clinical manifestations of COVID-19 are linked with induction of anaerobic glycolysis [88][89][90], and generation of hypoxia-like conditions [91,92].
All the above data, combined with our previous results from Vero E6 and A549 epithelial cells [19], suggest that hypoxia signaling constitutes a possible mechanism of co-regulation of the dopamine biosynthesis/catabolism genes (DDC, MAOA) and dACE2 during SARS-CoV-2 infection. Additionally, a weak positive correlation between the expression of the aforementioned-genes and fibrinogen (FIB) protein levels was detected in the same patient samples. Fibrinogen is known for its chemotactic activities and its role in homeostasis and antimicrobial host defense [93]. The SARS-CoV-2 infection has been reported to elevate fibrinogen levels in more severe cases of the disease [94], which in turn activates HIF-1α and thus induces anaerobic glycolysis [95]. This could suggest that our previously reported data from the in vitro studies, showing that the virus triggers hypoxia-like conditions and manipulates the glycolytic flux towards anaerobic glycolysis to favor its replication, may also reflect the in vivo situation. Additionally, the mRNA levels of EPO were significantly reduced in the SARS-CoV-2-infected Vero E6 cells and in the A549 cells cultured under low oxygen tension, even though the expression of EPO is known to be strongly induced by hypoxia. Both treatments resulted in extensive cell death, therefore it is possible that the expression of EPO is positively regulated by hypoxia, only when cell death mechanisms have not been activated. The EPO expression remained elevated in the nasopharyngeal swabs of hospitalized COVID-19 patients, yet it was severely downregulated in the whole blood samples of the respective group, compared to the healthy subjects. The latter is in agreement with previous reports [8,55,56]. In the nasopharyngeal tissue, the expression of EPO seems to be associated with the activation of cellular survival mechanisms [96,97]. On the other hand, in the whole blood samples, the expression pattern of EPO could either reflect the massive death of many cell populations observed in severe cases of SARS-CoV-2 infection [98], or it could be related to a possible virus-produced dysfunction in hypoxia signaling previously described as a putative mechanism of the ''hypoxia-paradox". According to the latter, in severe COVID-19 cases, reduced levels of the EPO protein are observed, despite the presence of low levels of hemoglobin in these patients [55,99]. Interestingly, recombinant human EPO was used successfully as a treatment in more severe cases of COVID-19 disease [100][101][102]. The observed elevated levels of LDH protein and increased expression of HIF-1α and GLUT1 in the ICU COVID-19 patients likely correlate with an enhanced glucose utilization that could be evidence of the hypoxic conditions caused by the virus infection. Moreover, the interaction of LDHA and LDHB with the SARS-CoV-2 nsp9 and nsp13 proteins has been reported in previous studies. The enzymes LDHA and LDHB have been reported to be implicated in glycolysis and in TCA/oxidative phosphorylation, respectively [103]. These data may suggest that the virus regulates the cell energy-production pathways to favor its propagation. Meanwhile, EPO, which is also transcriptionally regulated by hypoxia (HIF), has not been shown to directly interact with viral proteins.
The present study has some limitations. Although the rapid breakdown of dopamine in most aqueous solutions is a disadvantage of ELISA, this method has been used elsewhere for dopamine measurement in biological fluids such as serum and urine [104][105][106] and seems to have similar results to HPLC [64]. Moreover, accumulated cytosolic-located dopamine, because of its administration in high concentrations, can undergo auto-oxidation, producing toxic quinones that can also generate ROS [38,107]. Consequently, we cannot exclude the implication of ROS on the observed dopamine-mediated reduction in virus-derived CPE. Interestingly, DENV, another virus that causes cell lysis, even though is favored by ROS production [108], is negatively affected by exogenously applied dopamine [38]. Finally, another limitation of this study is the relatively small numbers of patients in a few groups, and that age and sex were not well distributed between some cohorts.

Conclusions
This study suggests that DDC expression levels in the nasopharyngeal tissue could serve as the most reliable marker, amongst the genes studied, of SARS-CoV-2 disease severity. Hence, DDC expression levels could be used as a prognostic tool to avoid the oc-currence of critical manifestations. Additionally, our study reveals a generalized regulation of the catecholamine biosynthesis and catabolism pathway upon SARS-CoV-2 infection in the whole blood samples of hospitalized COVID-19 patients and infected cultured epithelial cells, providing novel knowledge on the viral pathogenesis. Furthermore, in the current study, we provide evidence that the ISG dACE2, which is co-expressed with DDC and MAOA, is possibly regulated by other factors, apart from interferon response, upon SARS-CoV-2 infection. Moreover, we demonstrated that the exogenous administration of non-cytotoxic concentrations of dopamine and L-Dopa in SARS-CoV-2-infected Vero E6 cells led to significantly lower levels of virus-derived cytopathic effect. Finally, we confirmed that SARS-CoV-2 infection induces hypoxia-like conditions in the cell, affecting the expression of catecholamine biosynthesis and catabolism genes, a condition that favors SARS-CoV-2 replication. Considering the above, further investigation is needed to address the impact of the products and precursors of the catecholamine biosynthesis pathway in the viral replication to possibly use them as a therapeutic approach.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/cells12010012/s1. Table S1: Mean values of the studied genes' expression in nasopharyngeal tissue of SARS-CoV-2 positive or negative subjects. Table S2: Demographic data of individuals whose blood samples were used. Table S3: Mean values of the studied genes expression in whole blood samples of SARS-CoV-2 positive or negative subjects.

Institutional Review Board Statement:
This study was based on routine diagnosis samples of nasopharyngeal tissue from the biobank of Eginition Hospital (Athens, Greece) as well as on whole blood samples acquired from the Evangelismos General Hospital after approval by the Evangelismos Hospital Research Ethics Committee (#170/24-4-2020). All collected data were anonymized in standardized forms. Written informed consent was obtained from all subjects. For the minors, consent was obtained from parents or guardians. The study was conducted in accordance with the ethical standards noted in the 1964 Declaration of Helsinki and its later amendments. The study, which is part of DopaHypoCov-SARS-CoV-2 2020 Institut Pasteur International Network program, was approved by the Administration Board of the Hellenic Pasteur Institute.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: All relevant data are within the manuscript and its Supporting information files.