key: cord-1023708-qmwatunf
authors: Moolamalla, S.T.R.; Balasubramanian, Rami; Chauhan, Ruchi; Priyakumar, U. Deva; Vinod, P.K.
title: Host metabolic reprogramming in response to SARS-CoV-2 infection: A systems biology approach
date: 2021-07-30
journal: Microb Pathog
DOI: 10.1016/j.micpath.2021.105114
sha: 0bc463716a46847d63a0a76a5b0760982567de69
doc_id: 1023708
cord_uid: qmwatunf

Understanding the pathogenesis of SARS-CoV-2 is essential for developing effective treatment strategies. Viruses hijack the host metabolism to redirect the resources for their replication and survival. The influence of SARS-CoV-2 on host metabolism is yet to be fully understood. In this study, we analyzed the transcriptomic data obtained from different human respiratory cell lines and patient samples (nasopharyngeal swab, peripheral blood mononuclear cells, lung biopsy, bronchoalveolar lavage fluid) to understand metabolic alterations in response to SARS-CoV-2 infection. We explored the expression pattern of metabolic genes in the comprehensive genome-scale network model of human metabolism, Recon3D, to extract key metabolic genes, pathways, and reporter metabolites under each SARS-CoV-2-infected condition. A SARS-CoV-2 core metabolic interactome was constructed for network-based drug repurposing. Our analysis revealed the host-dependent dysregulation of glycolysis, mitochondrial metabolism, amino acid metabolism, nucleotide metabolism, glutathione metabolism, polyamine synthesis, and lipid metabolism. We observed different pro- and antiviral metabolic changes and generated hypotheses on how the host metabolism can be targeted for reducing viral titers and immunomodulation. These findings warrant further exploration with more samples and in vitro studies to test predictions.

The ongoing pandemic of coronavirus disease is a highly infectious respiratory illness in humans caused by a strain called Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) [1, 2] . It was first identified in December 2019 in Wuhan and has spread worldwide, infecting over 188 million people, and causing over 4 million deaths by July 2021. The World Health Organization has declared this ongoing pandemic as a global public health emergency. Most patients have one or more symptoms of fever, cough, shortness of breath, headache, lack of taste/smell, chest pain, and diarrhea, while in severe cases, patients develop severe pneumonia, pulmonary edema, acute respiratory distress syndrome, or multiple organ failure leading to death [3] . The severity is attributed to dysregulation of immune function and host factors like comorbidities; however, their relative contributions are still unclear. It is critical to understand the pathogenesis of SARS-CoV-2 for the design of effective treatment strategies.

Viruses' life cycle involves entering a host, evading the host cell immune response, and viral replication by taking control of host-cellular machinery for protein synthesis. The host-virus protein-protein interaction network controls these processes. The mechanism for infecting hosts may vary depending upon the viral type. SARS-CoV-2 entry into a host cell depends on spike protein (S) binding to the receptor ACE2 and S protein priming by serine protease TMPRSS2 [4] .

The viral entry into host cells is shown to be blocked by the serine protein inhibitor against TMPRSS2. Further, inhibitors based on the crystal structure of the main protease (M pro , 3CL pro ) of SARS-CoV-2 have been proposed as potential treatment strategies [5, 6] . In addition to viral proteins, different host proteins can also serve as drug targets. The host-virus interactome-based drug repurposing strategies have shown great promise in identifying different host protein targets [7, 8] . However, there is minimal knowledge of the mechanism of the SARS-CoV-2 survival strategy.

Due to their complete dependence on host cells to replicate, viruses may have evolved different strategies to reprogram the host metabolism for their replication and survival [9] [10] [11] [12] . Specific host metabolic pathways, including carbohydrate, fatty acid, and nucleotide metabolism, are known to be breached by different viruses upon infection. Each viral species is likely to induce unique metabolic reprogramming of the host cell. The host transcriptomic data in response to SARS-CoV-J o u r n a l P r e -p r o o f 2 is helping to decipher the changes at the level of gene expression. Studies have shown elevated inflammatory cytokine production, low innate antiviral defense, autophagic and mitochondrial dysfunctions in SARS-CoV-2 infected conditions [13] [14] [15] [16] . However, a comprehensive understanding of metabolic reprogramming of the host by SARS-CoV-2 is still lacking.

In this study, we explored the metabolic alterations induced by the SARS-CoV-2 using the transcriptomic data obtained from different human respiratory cell lines and samples collected from patients. We identified metabolic hot-spots and reporter metabolites under various SARS-CoV-2-infected conditions and compared them to obtain robust metabolic changes. A host-virus metabolic interactome of SARS-CoV-2 was reconstructed. We also linked our findings to available proteomics and metabolomics data obtained under SARS-CoV-2 infected conditions. This analysis generates insights into host metabolic response to SARS-CoV-2, which can be targeted for the effective antiviral response.

We generated insights into how the host metabolism is altered in response to SARS-CoV-2 by analyzing the RNA sequencing (RNASeq) raw read count data obtained from 4 human cell lines, including adenocarcinomic alveolar basal epithelial (A549) cells, ACE2 transduced A549 (ACE2) cells, human adenocarcinomic lung epithelial (Calu3) cells, and normal human bronchial epithelial (NHBE) cells infected with SARS-CoV-2 ( Table 1 ) [13] . A549 and ACE2 cell lines were infected with a multiplicity of infection (MOI) equal to 0.2 and 2, whereas Calu3 and NHBE with a MOI of 2. We also used RNASeq raw read count data obtained from lung biopsy [13] and nasopharyngeal swab of healthy and SARS-CoV-2 infected human samples [16] . These datasets are available for download from the gene expression omnibus (GEO) database with accession numbers given in Table 1 . The workflow used to analyze the host metabolic response to SARS-CoV-2 is shown in Figure 1 .

We performed differential gene expression analysis using the DESeq2 (v1.26.1) in R (v3.6.1) for various SARS-CoV-2 infected conditions (cell lines, lung biopsy, and nasopharyngeal swab) to J o u r n a l P r e -p r o o f obtain differentially expressed genes (DEGs) [17] . Further, we obtained DEGs of bronchoalveolar lavage fluid (BALF) and peripheral blood mononuclear cells (PBMC) cells of SARS-CoV-2 patients from Xiong et al., (2020) ( Table 1) . To understand metabolism-specific alterations, we filtered the resulting DEGs based on the genes present in Recon3D, a comprehensive genomescale network model of human metabolism [18] . We compared the DEGs across conditions and extracted key genes based on the number of occurrences to obtain robust gene signatures ("metagenes"). The KEGG pathways associated with DEGs were obtained using Enrichr (adjusted p-value < 0.05) [19] . We also identified transcription factors associated with metabolic DEGs under each condition using the TRUST transcription factor (2019) database in Enrichr (p-value < 0.05) [19] and extracted key transcription factors based on the number of occurrences.

The metabolic DEGs of SARS-CoV-2 were used to extract the reporter metabolites around which the most significant transcriptional changes occur [20] . This approach is based on the hypothesis that genes surrounding a metabolite are co-expressed to maintain homeostasis. Each metabolite is scored based on the gene expression of neighboring genes. We used the Recon3D model to identify the neighboring genes of each metabolite. The Recon3D model contains 10,600 metabolic reactions, 5,835 metabolites, and 1,882 unique metabolic genes spanning nine compartments.

Based on the gene-protein-reaction association rules of the Recon3D model, genes associated with reactions involving a metabolite as a reactant or product were classified as the neighboring genes of that metabolite.

The p-values (pi) obtained from the differential gene expression analysis were transformed into the Z-scores (Zi) using the inverse normal cumulative distribution function (CDF). Metabolites were assigned a Z-score (Zmetabolite) by aggregating the Z-scores of their 'k' neighboring genes (equation 1).

These Zmetabolite scores were corrected for the background distribution using mean (μk) and standard deviation (σk) of aggregated Z-scores obtained by sampling 10,000 sets of k enzymes from the network (equation 2). Corrected Z-scores were then transformed to p-values using CDF.

Metabolites with greater or equal to three neighboring genes (k ≥ 3) and p-values less than 0.05 were identified as reporter metabolites.

Reconstruction of host-SARS-CoV-2 metabolic interactome:

We utilized a recent host-virus protein-protein interactome of SARS-CoV-2 (332 high-confidence interactions between 26 SARS-CoV-2 proteins) to extract the metabolic DEGs targeted by viral proteins [21] . In addition to direct interactions, we also retrieved the indirect metabolic targets present in the immediate neighborhood of all viral protein targets. We used high confidence and curated human protein-protein interaction network (hPPiN) containing 17,063 nodes and 208,760

interactions) to extract the immediate neighbors [22] . The final interactome size was 724 nodes (metabolic and non-metabolic) and 1,018 interactions. The key metabolic processes that are targeted directly and indirectly by viral proteins were identified. Cytoscape version 3.3 was utilized to visualize the networks.

Drug-target network of 2,938 FDA-approved or investigational drugs constructed from databases including DrugBank database (v4.3), Therapeutic Target Database, PharmGKB database, ChEMBL (v20), BindingDB, and the IUPHAR/BPS Guide to Pharmacology was taken from Zhou et al., (2020) [23] [24] [25] [26] [27] [28] [29] . To identify the potential drugs, the proximity of the set of genes 'C' of SARS-CoV-2 metabolic interactome to the set of drug targets 'T' was measured as

where d(c,t) is the average shortest path length of the gene c ∈ C and t ∈ T [30] . The permutation test was performed with randomly selected genes with similar degree distributions of genes in C and T. Z-score was calculated as

J o u r n a l P r e -p r o o f where ̅̅̅ and are the mean and standard deviation obtained by sampling 1,000 times. p-value was calculated based on the permutation test. Drugs with Z-score < -2.0 and p-value < 0.005 were considered significant. From these, we identified the drugs which target one or more upregulated metabolic DEGs. We further prioritized the drugs based on drug categories extracted from the DrugBank database [24] and available clinical trials information (https://clinicaltrials.gov/). Target set enrichment analysis of the identified drugs for KEGG pathways was performed by duplication adjusted hypergeometric test using SignatureSearch R package with q-value cut-off < 0.05 for false discovery rate control [31] .

We first compiled a compendium of all the changes that can serve as a useful resource on SARS-CoV-2 (Table S1 ). This information can be used to extract consensus and condition-specific genes ( Figure S1 ). Genes of cytokine-mediated signaling pathways are present in most infected conditions representing the adverse inflammatory condition ("cytokine storm") as reported by different studies [13, 14] .

We specifically extracted metabolic alterations based on the genes in the genome-scale network model of human metabolism, Recon3D. The metabolic DEGs between normal and infected conditions (p-value < 0.05) is given in Table S2 . In general, A549 and Calu3 have the most metabolic changes compared to other cell lines. The overlap of metabolic DEGs across different cell lines is shown in Figure 2 . We observed that most metabolic DEGs are downregulated in ACE2, A549, and Calu3 cell lines compared to NHBE. There is also an increase in DEGs with an increase in MOI from 0.2 to 2 in ACE2 and A549 cell lines (ACE2_0.2, ACE2_2, and A549_0.2, A549_2). Table 2 shows the top metabolic candidate genes upregulated in most infected conditions. We also identified metabolic pathways and reporter metabolites in each infected cell line.

The MAVS is shown to be downregulated in SARS-CoV-2 infected Caco-2 cells [34] . The innate antiviral immunity is governed by the mitochondrial membrane potential (Δψm) [35] . RLR antiviral signaling also depends on oxidative phosphorylation [36] . Many viruses that manipulate mitochondrial dynamics also alter the NLRP3 inflammasome pathway [37] . Therefore, further analysis is required to test whether mitochondrial dysfunction may lead to sustained inflammasome signaling in SARS-CoV-2.

On the other hand, LDH family (LDHAL6A, LDHAL6B) genes involved in lactate production are significantly upregulated in ACE2_2. The lactate is a natural suppressor of antiviral RLR signaling by binding to the MAVS transmembrane domain [38, 39] . In turn, RLR signaling is shown to suppress glycolysis by inhibiting the enzyme hexokinase HK2 [38] . We observed that HK2 is upregulated in A549_2, ACE2_2, NHBE, and Calu3 cell lines infected with a MOI equal to 2 (Table S2) . HK2 also integrates energy metabolism and cell protection via Akt/mTOR signaling pathway and autophagy [40] . This mutual antagonism between RLR and glycolysis needs to be (Table S1 ). On the other hand, we found that VHL that targets HIF1A to degradation is also upregulated in these cell lines. HIF1A

promotes a shift in mitochondrial metabolism to glycolysis [42] . It is also shown to promote SARS-CoV-2 replication in monocytes [43] . A proteo-transcriptomics analysis of SARS-CoV-2 J o u r n a l P r e -p r o o f infected Huh7 cells has shown dysregulation of Akt/mTOR/HIF1 signaling [44] . The PI3K/Akt/mTOR pathway is also known to control this switch in response to different viral infections [45] [46] [47] . Genes of PI3K controlled inositol phosphate metabolism are upregulated in A549_2, ACE2_2, and Calu3 cell lines infected with a MOI equal to 2 (Figure 3) . However, some genes of this pathway are also downregulated in these cell lines. We hypothesize that PI3K

inhibitors Wortmannin and LY294002, which are known to have anti-cancer activity, maybe repurposed for SARS-CoV-2 antiviral therapy. These inhibitors are shown to inhibit the Middle East Respiratory Syndrome coronavirus (MERS-CoV) replication in vitro [48] .

Genes involved in tryptophan catabolism (KYNU, KMO, IDO1) are upregulated in infected cell lines ( Table 2 and Table S2 ). L-formylkynurenine of tryptophan metabolism is a reporter metabolite of infected conditions (Figure 4) . Metabolomics data from sera of SARS-CoV-2

infected patients show a decrease in tryptophan and an increase in the kynurenine pathway, which correlated with IL-6 levels [49] . IDO1-mediated tryptophan depletion, in the short term, is shown to inhibit viral replication but leads to immunosuppression in the long term [50] .

ASS1 involved in de novo arginine synthesis from aspartate is upregulated in most infected cell lines ( Table 2 ). Inhibition of ASS1 is shown to increase viral replication, and its deficiency affects immune cell activation [51, 52] . These observations suggest that ASS1 is part of the antiviral pathway involved in the host defense. ASNS involved in the synthesis of asparaginase from aspartate is upregulated in A549 and Calu3 cell lines. Aspartic acid, asparaginase, and citrulline are reporter metabolites of infected conditions (Figure 4) . Asparaginase is found to be the novel host factor promoting the replication of human cytomegalovirus (HCMV) [53] . Valine, leucine, and isoleucine (branched-chain amino acids, BCAAs) degradation pathway is also downregulated in infected cell lines (Figure 3) . The degradation of BCAAs generates CoA products that feed into the TCA cycle. l-isoleucine, l-leucine, l-valine, and alpha-ketoisovaleric acid are reporter metabolites of infected conditions (Figure 4) . BCAAs are reported to activate mTOR and interferon signaling by inactivating SOCS3 [54] . BCAA inhibits hepatitis C virus (HCV) replication and promotes infectious particle formation [55] . Serine, glutamine, and leucine are J o u r n a l P r e -p r o o f reporter metabolites in A549, Calu3, and NHBE cell lines (Figure 4) . The metabolites of homocysteine metabolism (l-cysteine and cystathionine) are also observed in A549 and Calu3.

Genes that control polyamine levels are altered in infected cell lines ( Table 2 and S2). These include ODC1 and SAT1, which influence the polyamines' synthesis and catabolism, respectively.

SAT1 is upregulated in all the cell lines, while ODC1 is upregulated in ACE2_0.2, ACE2_2, and A549_2. ODC1 is involved in the generation of putrescine, and SAT1 is involved in catabolizing spermine back to spermidine and putrescine. We found putrescine and N-acetylputrescine as reporter metabolites of infected conditions (Figure 4) . Polyamines are positively charged and bind DNA/RNA to assist in viral replication [56] . They help neutralize the negative charges of DNA and assist in packing DNA into viral particles [57, 58] . Studies show that MERS-CoV is dependent on polyamines for replication [56] . Polyamines are known to control Myc expression, and Myc, in turn, controls the transcription of genes ODC1, SRM, and SMS involved in polyamine synthesis (positive feedback loop) [59] (Figure 5A) . Therefore, the depletion of polyamines can be a strategy to reduce viral infection.

ODC1 inhibitor Difluoromethylornithine, which is in clinical trials for cancer treatment, can be tested for antiviral effects against SARS-CoV-2 [50, 60] . Further, FDA-approved antiviral ribavirin depletes polyamines levels and viral titers by activating SAT1 [61] . Polyamine levels are regulated by a negative feedback loop involving SAT1 ( Figure 5A ). Ribavirin is proposed as a novel coronavirus treatment strategy [62] and is also part of triple-antiviral therapy proposed for alleviating symptoms in mild to moderate SARS-CoV-2 conditions [63] .

Genes of purine and pyrimidine metabolism are mostly downregulated in A549_2, ACE2_2, and Calu3 cell lines (Figure 3) . GUK1 involved in converting guanosine monophosphate to guanosine diphosphate is downregulated in these cell lines. However, GUK1 is proposed as a bottleneck for the viral genome buildup and a drug target for SARS-CoV-2 [64] . CTPS genes involved in cytosine synthesis are downregulated in ACE2 and Calu3, while upregulated in A549_2 and NHBE cell lines. The proteomic study in Caco-2 cells shows that CTPS1 and CTPS2 are J o u r n a l P r e -p r o o f downregulated [34] . CDA, UPP1, UPB1, CMPK2, and TYMP involved in pyrimidine scavenging is upregulated in infected cell lines ( Table S2 ). The compositional difference between human and SARS-CoV-2 RNA shows a deficiency of cytosine in SARS-CoV-2 [65] . CMPK2 is an interferonstimulated gene known to restrict human immunodeficiency virus (HIV) infection [66] . CMPK2

is found adjacent to gene RSAD2, which is well known to inhibit a wide variety of RNA and DNA viruses by synthesizing chain terminators that block DNA/RNA synthesis [67] . NT5E and XDH involved in purine catabolism are upregulated (Table S2) has an antiviral effect on HIV, Nipah virus, and Ebola by blocking the virus's membrane fusion [68] . Recently, CH25H is shown to suppress the SARS-CoV-2 spike protein-mediated membrane fusion [69] . The analog of 25HC (7-hydroxycholesterol) is found to be upregulated in the sera of COVID-19 patients [70] . The distribution of cholesterol influences viral replication, entry, and budding [50] . The cholesterol and fatty acid synthesis are regulated by transcription factor SREBPs activated by a viral infection (HCV) through the PI3K/Akt pathway [71] . LDLR is also upregulated, which RNA viruses use to enter the host [72, 73] . (Figure 4) . HCV and dengue viruses depend on MAPK activated PLA2G4C to target core protein to lipid droplets [74] . Arachidonic acid is shown to suppress the replication of human coronavirus 229E (HCoV-229E) and MERS-CoV [75] .

Genes of the pentose phosphate pathway and folate one-carbon metabolism are mostly downregulated in infected cell lines (Figure 3 and Table S2 ). The cytoplasmic one-carbon metabolism genes DHFR, SHMT1, and MTHFD1 are downregulated in ACE2_0.2, ACE2_2, A549_2, and Calu3, while mitochondrial genes MTHFD2 and MTHFD1L are upregulated in A549_2 and Calu3 cell lines. These pathways play an essential role in maintaining the redox balance by controlling NADPH production and nucleotide synthesis. Further, we observed that the genes of de novo synthesis of glutathione are also mostly downregulated (Figure 3) . These may reflect the oxidative stress condition of the host. Glutathione deficiency is also considered to be a severe manifestation of SARS-CoV-2 [76] . However, SOD2 involved in converting reactive oxygen species (ROS) to peroxides is upregulated in most infected cell lines ( Table 2) . Further, DUOX2 is upregulated in A549_2, ACE2_2, NHBE, and Calu3 cell lines infected with a MOI equal to 2. DUOX2 is usually expressed in epithelial tissues lining the respiratory and intestinal tracts and participates in the host defense against microbial infection at mucosal surfaces by the generation of peroxides [77, 78] . DUOX2 is also an antiviral gene induced in response to cytokines [79, 80] . Superoxide and hydrogen peroxide are reporter metabolites in infected cell lines ( Figure   4 ).

SETD2 and ASH1L that encode histone methyltransferases are upregulated in A549_2, ACE2_0.2, ACE2_2, and Calu3 cell lines. Methylation of STAT1 by SETD2 is shown to be critical for interferon antiviral activity [81] . Further, ASH1L also plays a role in suppressing TLRmediated immune response [82] . The antiviral genes HMOX1 and HMOX2 are also J o u r n a l P r e -p r o o f downregulated [83, 84] (Table S2) . HMOX1 protects cells from programmed cell death by catabolizing free heme, and its induction can have a therapeutic effect [85] . HMOX1 knockout mice showed significantly higher cytokine responses, and its expression is required for the protection against inflammatory insult to the lung [84] .

The metabolic DEGs of lung biopsy, BALF, swab, and PBMC show some overlap with the cell line data but have a comparatively lesser overlap amongst themselves (Figure 2) . It can be noted that the patient samples may be heterogeneous with multiple cell types, including both infected and non-infected cells. The key metabolic genes (NAMPT, SAT1, SOD2, ASS1, CMPK2, DUOX2, PDE4B, ABCA1) in different SARS-CoV-2 infected cell lines are also upregulated in lung biopsy, swab, and BALF ( Table 2) . PDE4B modulates pro-and anti-inflammatory cytokine productions by controlling cAMP levels and is proposed as an effective therapeutic target for many inflammatory diseases [86] . However, most metabolic DEGs are downregulated in swab and lung biopsy samples, while in PBMC, they are upregulated (Figure 2) . Genes of the tricarboxylic acid cycle and oxidative phosphorylation are downregulated in lung biopsy and swab samples and upregulated in PBMC (Figure 6 and Table S2 ). We observed that glycolysis pathway (ENO1, GAPDH, TPI1, PKM, PGM2, HK1, LDHA) and primary transporter of glutamine (SLC1A5) are also upregulated in PBMC (Table S2) . These alterations resemble the condition of aerobic glycolysis termed the "Warburg effect", a hallmark of cancer. However, the observed metabolic changes in PBMC may be linked to the immune response, which is known to trigger a switch to aerobic glycolysis. Lactate is shown to influence cytokine production [87, 88] . Interestingly, Codo et al. (2020) showed that SARS-CoV-2 stimulates glycolysis and that glycolysis promotes SARS-CoV-2 replication in monocytes and macrophages [43] . However, the single-cell transcriptomic data of PBMC from patients did not show viral reads [14, 89] . On the other hand, SARS-CoV-2 is detected in monocytes and lymphocytes from patients, which increased over time from the onset of the disease [90] . Further study is required to understand whether the positive feedback loop between SARS-CoV-2 and glycolysis in immune cells leads to cytokine storm and disease severity ( Figure 5B ). This metabolic adaptation may be mediated by mitochondrial ROS production and HIF1A stabilization, which controls the switch to aerobic glycolysis [43] . (Table S2) . Further, genes involved in BCAA degradation are upregulated. The first step of BCAA degradation (BCAT1) is shown to induce metabolic reprogramming to combat ROS [91] . The metabolites of tyrosine metabolism (homovanillin and 3-methoxytyramine) are also upregulated in PBMC (Figure 7) .

To understand the transcriptional regulation of metabolism, we identified the potential transcription factors based on DEGs' binding site enrichment under each condition. We found that an overlapping set of transcription factors is associated with upregulated and downregulated genes in different infected conditions (Tables 3 and 4) . HIV, and influenza virus [92] , which helps in viral replication and suppressing host cell death.

Transcription factors HIF1A, NFE2L2/NRF2, PPARA, and SERBF1 are also associated with differentially expressed genes. NFE2L2 is involved in oxidative stress, while SERBF1 and PPARA are involved in lipid homeostasis [93, 94] . PPARA stimulation by DNA virus infection suppresses the interferon signaling and impairs immunity against the virus [95] . PPAR controls peroxisome metabolism, which is shown to be essential for different viruses. We observed that the peroxisome function is altered in different cell lines and human samples (Figure 3) .

The host-pathogen interactome of SARS-CoV-2 was used to identify the direct and indirect metabolic targets of viral proteins (Figure 8 and S2) . We found that viral proteins target the host J o u r n a l P r e -p r o o f

The SARS-CoV-2 metabolic interactome was used to identify potential drugs that can have therapeutic effects against the SARS-CoV-2 infection. Drug targets that are in the vicinity of this network were identified using a network proximity measure. We identified 288 drugs (Z-score < -2.0 and p-value < 0.005) that target one or more upregulated metabolic DEGs in infected cell lines ( Table S3) . Targets of these 288 drugs are associated with nucleotide metabolism, amino acid metabolism (arginine, proline, histidine, tyrosine, and phenylalanine metabolism), and lipid metabolism (arachidonic acid metabolism, linoleic acid metabolism, and steroid hormone biosynthesis) ( Table S4) . We further prioritized drugs based on antiviral activity and clinical trial information.

The Venglustat, are shown to block SARS-CoV-2 replication in vitro [98] . In addition to depleting membrane cholesterol, these observations suggest that drugs targeting glycan synthesis are promising candidates for drug repurposing.

We identified Pentoxifylline that targets NT5E in the nucleotide metabolism and PDE4 as a candidate drug against SARS-CoV-2. Eight genes encoding phosphodiesterase are upregulated in at least two infected cell lines (A549, ACE2, and Calu3) ( Table S2) . Pentoxifylline inhibits the replication of several viruses in vitro and has anti-inflammatory, immunomodulatory, and bronchodilatory effects [99] . It is in clinical trials as adjuvant therapy for SARS-CoV-2 J o u r n a l P r e -p r o o f (NCT04433988) [100] . Other phosphodiesterase inhibitors identified include Ibudilast, Sildenafil, and Dipyridamole, which are in clinical trials for SARS-CoV-2 treatment (NCT04429555)[101].

PTGS1 and PTGS2 inhibitors are also candidate drugs against SARS-CoV-2. These include antiinflammatory drugs Naproxen, Niflumic acid, Nimesulide, and Balsalazide, used in treating inflammatory bowel disease and pomalidomide, an immunomodulating antineoplastic agent.

Naproxen is currently in clinical trials for treating critically ill hospitalized patients for SARS-CoV-2 infection (NCT04325633) [102] . We found drugs targeting LTA4H and ALOX5 involved in the biosynthesis of leukotriene as potential candidates. These include Captopril and Ubenimex that target LTA4H, and Zileuton, an anti-inflammatory agent that inhibits ALOX5. Captopril is currently in clinical trials for treating SARS-CoV-2 patients with severe acute respiratory syndrome and pneumonia (NCT04355429) [103] . By targeting the biosynthesis of leukotrienes, these drugs may mitigate the hyperinflammatory response to SARS-CoV-2 [104]. Further, Acetazolamide that inhibits carbonic anhydrases is also identified by this network approach.

Inhibiting carbonic anhydrases is shown to affect the intracellular pH, which can disrupt virusendosome fusion and lysosomal proteases [105] . However, there is a risk of multiple adverse effects using Acetazolamide for treating SARS-CoV-2 [106] .

In summary, we performed a comprehensive comparison of several transcriptomic data and generated hypotheses into how the host metabolic pathway is hijacked in response to SARS-CoV- 

The authors declare no conflict of interest.

All datasets reported in this study are publicly available. : Drugs that target one or more metabolic DEGs (Z-score < -2.0 and p-value < 0.005) identified using the network proximity analysis (xlsx file). J o u r n a l P r e -p r o o f 

A pneumonia outbreak associated with a new coronavirus of probable bat origin

The 2019-new coronavirus epidemic: Evidence for virus evolution

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor

Structure-based design of antiviral drug candidates targeting the SARS-CoV-2 main protease

Designing of improved drugs for COVID-19: Crystal structure of SARS-CoV-2 main protease M(pro)

Network Medicine Framework for Identifying Drug Repurposing Opportunities for COVID-19

Network-based approach to prediction and population-based validation of in silico drug repurposing

Viral hijacking of cellular metabolism

Hijacking the Supplies: Metabolism as a Novel Facet of Virus-Host Interaction

Virus Control of Cell Metabolism for Replication and Evasion of Host Immune Responses. Frontiers in cellular and infection microbiology

Viral activation of cellular metabolism

Imbalanced Host Response to SARS-CoV-2 Drives Development of COVID-19

Transcriptomic characteristics of bronchoalveolar lavage fluid and peripheral blood mononuclear cells in COVID-19 patients. Emerging microbes & infections

Network Analysis and Transcriptome Profiling Identify Autophagic and Mitochondrial Dysfunctions in SARS-CoV-2 Infection

In vivo antiviral host transcriptional response to SARS-CoV-2 by viral load, sex, and age

Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2

Recon3D enables a three-dimensional view of gene variation in human metabolism

Enrichr: a comprehensive gene set enrichment analysis web server 2016 update

Uncovering transcriptional regulation of metabolism by using metabolic network topology

A SARS-CoV-2 protein interaction map reveals targets for drug repurposing

Meta-analysis of host response networks identifies a common core in tuberculosis

A network medicine approach to investigation and population-based validation of disease manifestations and drug repurposing for COVID-19

DrugBank: a comprehensive resource for in silico drug discovery and exploration

TTD: Therapeutic Target Database

PharmGKB: the Pharmacogenomics Knowledge Base

ChEMBL: a largescale bioactivity database for drug discovery

BindingDB: a web-accessible database of experimentally determined protein-ligand binding affinities

The IUPHAR/BPS Guide to PHARMACOLOGY: an expert-driven knowledgebase of drug targets and their ligands

A genome-wide positioning systems network algorithm for in silico drug repurposing

signatureSearch: environment for gene expression signature searching and functional interpretation

Identification and characterization of MAVS, a mitochondrial antiviral signaling protein that activates NF-kappaB and IRF 3

Mitochondrial Interactome: A Focus on Antiviral Signaling Pathways

Proteomics of SARS-CoV-2-infected host cells reveals therapy targets

Mitochondrial-mediated antiviral immunity

RLRmediated antiviral innate immunity requires oxidative phosphorylation activity

The rOX-stars of inflammation: links between the inflammasome and mitochondrial meltdown

Lactate Is a Natural Suppressor of RLR Signaling by Targeting MAVS

Regulation of MAVS Expression and Signaling Function in the Antiviral Innate Immune Response

Hexokinase II integrates energy metabolism and cellular protection: Akting on mitochondria and TORCing to autophagy

Glucose Metabolism Linked to Antiviral Responses

Tee AR. mTORC1 drives HIF-1alpha and VEGF-A signalling via multiple mechanisms involving 4E-BP1, S6K1 and STAT3

Elevated Glucose Levels Favor SARS-CoV-2 Infection and Monocyte Response through a HIF-1alpha/Glycolysis-Dependent Axis. Cell metabolism

Dysregulation in Akt/mTOR/HIF-1 signaling identified by proteo-transcriptomics of SARS-CoV-2 infected cells. Emerging microbes & infections

Activation of the N-Ras-PI3K-Akt-mTOR pathway by hepatitis C virus: control of cell survival and viral replication

The battle over mTOR: an emerging theatre in host-pathogen immunity

An invertebrate Warburg effect: a shrimp virus achieves successful replication by altering the host metabolome via the PI3K-Akt-mTOR pathway

Antiviral potential of ERK/MAPK and PI3K/AKT/mTOR signaling modulation for Middle East respiratory syndrome coronavirus infection as identified by temporal kinome analysis

COVID-19 infection alters kynurenine and fatty acid metabolism, correlating with IL-6 levels and renal status

Interferons: Reprogramming the Metabolic Network against Viral Infection

Argininosuccinate synthetase 1 depletion produces a metabolic state conducive to herpes simplex virus 1 infection

Impaired T cell function in argininosuccinate synthetase deficiency

Asparagine Deprivation Causes a Reversible Inhibition of Human Cytomegalovirus Acute Virus Replication

Malnutrition impairs interferon signaling through mTOR and FoxO pathways in patients with chronic hepatitis C

Valine, the branchedchain amino acid, suppresses hepatitis C virus RNA replication but promotes infectious particle formation

Polyamines and Their Role in Virus Infection. Microbiology and molecular biology reviews : MMBR

Endogenous polyamine function--the RNA perspective

Polyamines in vaccinia virions and polypeptides released from viral cores by acid extraction

Diverse Functions of Polyamines in Virus Infection

Interferon-Induced Spermidine-Spermine Acetyltransferase and Polyamine Depletion Restrict Zika and Chikungunya Viruses

Ribavirin Induces Polyamine Depletion via Nucleotide Depletion to Limit Virus Replication

Novel coronavirus treatment with ribavirin: Groundwork for an evaluation concerning COVID-19

Triple combination of interferon beta-1b, lopinavir-ritonavir, and ribavirin in the treatment of patients admitted to hospital with COVID-19: an open-label, randomised, phase 2 trial

FBA reveals guanylate kinase as a potential target for antiviral therapies against SARS-CoV-2

Cytosine drives evolution of SARS-CoV-2. Environmental microbiology

CMPK2 and BCL-G are associated with type 1 interferon-induced HIV restriction in humans

A naturally occurring antiviral ribonucleotide encoded by the human genome

Interferon-inducible cholesterol-25-hydroxylase broadly inhibits viral entry by production of 25-hydroxycholesterol

Cholesterol 25-hydroxylase suppresses SARS-CoV-2 replication by blocking membrane fusion

Proteomic and Metabolomic Characterization of COVID-19

Hepatitis C virus genotype-3a core protein enhances sterol regulatory element-binding protein-1 activity through the phosphoinositide 3-kinase-Akt-2 pathway

LDL receptor and its family members serve as the cellular receptors for vesicular stomatitis virus

The lowdensity lipoprotein receptor plays a role in the infection of primary human hepatocytes by hepatitis C virus

MAP-kinase regulated cytosolic phospholipase A2 activity is essential for production of infectious hepatitis C virus particles

Characterization of the Lipidomic Profile of Human Coronavirus-Infected Cells: Implications for Lipid Metabolism Remodeling upon Coronavirus Replication

Endogenous Deficiency of Glutathione as the Most Likely Cause of Serious Manifestations and Death in COVID-19 Patients. ACS infectious diseases

Dual oxidase 2 is essential for the toll-like receptor 5-mediated inflammatory response in airway mucosa

Dual oxidases represent novel hydrogen peroxide sources supporting mucosal surface host defense

IFNbeta/TNFalpha synergism induces a non-canonical STAT2/IRF9-dependent pathway triggering a novel DUOX2 NADPH oxidase-mediated airway antiviral response

Lung epithelial NOX/DUOX and respiratory virus infections

Methyltransferase SETD2-Mediated Methylation of STAT1 Is Critical for

Histone methyltransferase Ash1l suppresses interleukin-6 production and inflammatory autoimmune diseases by inducing the ubiquitinediting enzyme A20

Heme Oxygenase 2 Binds Myristate to Regulate Retrovirus Assembly and TLR4 Signaling

Ribavirin induced hemolysis: a novel mechanism of action against chronic hepatitis C virus infection

Mechanisms of cell protection by heme oxygenase-1. Annual review of pharmacology and toxicology

Commentary: Phosphodiesterase 4 inhibitors as potential adjunct treatment targeting the cytokine storm in COVID-19

Lactate Regulates Metabolic and Pro-inflammatory Circuits in Control of T Cell Migration and Effector Functions

In vitro and in vivo Effects of Lactate on Metabolism and Cytokine Production of Human Primary PBMCs and Monocytes

Single-Cell Sequencing of Peripheral Mononuclear Cells Reveals Distinct Immune Response Landscapes of COVID-19 and Influenza Patients

Infection of human lymphomononuclear cells by SARS-CoV-2

Branched-Chain Amino Acid Metabolic Reprogramming Orchestrates Drug Resistance to EGFR Tyrosine Kinase Inhibitors

Hostile takeovers: viral appropriation of the NF-kappaB pathway

Nrf2-Keap1 regulation of cellular defense mechanisms against electrophiles and reactive oxygen species

The SREBP pathway: regulation of cholesterol metabolism by proteolysis of a membrane-bound transcription factor

Metabolic Control of Viral Infection through PPAR-α Regulation of STING Signaling

Sialic acid-Dependent Binding and Viral Entry of SARS-CoV-2

Repurposing of Miglustat to inhibit the coronavirus Severe Acquired Respiratory Syndrome SARS-CoV-2

Glucosylceramide synthase inhibitors prevent replication of SARS-CoV-2 and Influenza virus

Medical science monitor : international medical journal of experimental and clinical research

Efficacy of Pentoxifylline as Add on Therapy in COVID19 Patients

Patients Hospitalized With COVID-19 and at Risk for ARDS, ClinicalTrials.gov Identifier NCT04429555

Efficacy of Addition of Naproxen in the Treatment of Critically Ill Patients Hospitalized for COVID-19 Infection

Efficacy of Captopril in Covid-19 Patients With Severe Acute Respiratory Syndrome (SARS) CoV-2 Pneumonia

Identifier NCT04355429

A Novel Strategy to Mitigate the Hyperinflammatory Response to COVID-19 by Targeting Leukotrienes

The potential effectiveness of acetazolamide in the prevention of acute kidney injury in COVID-19: A hypothesis

COVID-19 Lung Injury and High Altitude Pulmonary Edema: A False Equation with Dangerous Implications. Annals of the

Inhibiting the reproduction of SARS-CoV-2 through perturbations in human lung cell metabolic network