Abstract

Background

   The outbreak of Coronavirus disease 2019 (COVID-19) has become an
   international public health crisis, and the number of cases with dengue
   co-infection has raised concerns. Unfortunately, treatment options are
   currently limited or even unavailable. Thus, the aim of our study was
   to explore the underlying mechanisms and identify potential therapeutic
   targets for co-infection.

Methods

   To further understand the mechanisms underlying co-infection, we used a
   series of bioinformatics analyses to build host factor interaction
   networks and elucidate biological process and molecular function
   categories, pathway activity, tissue-specific enrichment, and potential
   therapeutic agents.

Results

   We explored the pathologic mechanisms of COVID-19 and dengue
   co-infection, including predisposing genes, significant pathways,
   biological functions, and possible drugs for intervention. In total,
   460 shared host factors were collected; among them, CCL4 and AhR
   targets were important. To further analyze biological functions, we
   created a protein-protein interaction (PPI) network and performed
   Molecular Complex Detection (MCODE) analysis. In addition, common
   signaling pathways were acquired, and the toll-like receptor and
   NOD-like receptor signaling pathways exerted a significant effect on
   the interaction. Upregulated genes were identified based on the
   activity score of dysregulated genes, such as IL-1, Hippo, and TNF-α.
   We also conducted tissue-specific enrichment analysis and found ICAM-1
   and CCL2 to be highly expressed in the lung. Finally, candidate drugs
   were screened, including resveratrol, genistein, and dexamethasone.

Conclusions

   This study probes host factor interaction networks for COVID-19 and
   dengue and provides potential drugs for clinical practice. Although the
   findings need to be verified, they contribute to the treatment of
   co-infection and the management of respiratory disease.

   Keywords: COVID-19, dengue, co-infection, host factor interaction
   networks, bioinformatics analyses

Introduction

   Coronavirus disease 2019 (COVID-19), which is due to novel severe acute
   respiratory syndrome coronavirus 2 (SARS-CoV-2), has led to a sudden
   and sharp increase in hospitalizations for pneumonia globally ([39]1).
   In March 2020, the World Health Organization (WHO) publicly announced
   that COVID-19 had become a pandemic. As of June 22, 2021, there have
   been 178,503,429 confirmed cases of COVID-19 worldwide, including
   3,872,457 deaths reported to the WHO, and the numbers continue to
   increase ([40]https://covid19.who.int/).

   Dengue, a mosquito-borne viral infection caused by four dengue virus
   serotypes (DENV 1-4), is found in tropical and subtropical climate
   areas, mainly in urban and semiurban regions ([41]2). According to the
   WHO’s report, the global incidence of dengue has grown dramatically in
   the past few decades
   ([42]https://www.who.int/news-room/fact-sheets/detail/dengue-and-severe
   -dengue). There are an estimated 100–400 million infections each year,
   and approximately half of the world’s population is now at risk of
   contracting the disease.

   There is some cross-reactivity and common pathological processes, such
   as capillary leakage, thrombocytopenia, and coagulopathy, between
   SARS-CoV-2 and DENV ([43]3, [44]4), making it difficult to distinguish
   their common clinical and laboratory characteristics. Overall, patients
   with dengue fever, including positive NS1 and/or IgM serology results,
   need to be differentiated from those with SARS-CoV-2 infection, and
   dengue IgM/IgG testing should be repeated if necessary to identify
   co-infection or serological overlap ([45]5). Two patients in Singapore
   with false-positive results from rapid serological testing for dengue
   were later confirmed to have SARS-CoV-2 infection because of persistent
   fever, and the repeat dengue test (SD Bioline) was negative ([46]6).
   This suggests that misdiagnosis between COVID‐19 and dengue is
   objective.

   Serological overlap makes it difficult to identify symptoms, and
   co-infection complicates each disease. The COVID-19 pandemic and dengue
   fever recurrence epidemic in tropical countries have become global
   health issues. One case of co-infection has been reported in Mayotte, a
   French island in the Indian Ocean ([47]7). Due to the delay in the
   diagnosis of COVID-19 and the transmission of dengue fever by the
   insect vector, measures need to be taken in endemic areas to avoid
   co-infection ([48]8), and the diagnosis of patients with febrile
   disease should be organized to allow for diagnosis of both COVID-19 and
   dengue. In a research from Argentina, prolonged fever time was a
   clinical indication for suspected co-infection in most patients with
   mild COVID-19. Furthermore, in another study, headache was the most
   typical symptom in patients with co-infection ([49]9). Colombia
   ([50]10), Brazil ([51]11), Pakistan ([52]12), and other countries have
   reported that COVID-19 and dengue may cause overlapping epidemics and
   increase the number of critically ill patients, thereby compounding the
   burden on the healthcare system. There are overlapping outbreaks of
   COVID-19 and dengue in dengue-endemic countries. Therefore, identifying
   potential vector breeding sites to protect people from mosquito bites
   and maintaining social distancing to avoid the risk of COVID-19
   transmission are very important ([53]13).

   It is necessary to prepare for dengue outbreaks immediately under the
   premise of controlling the COVID-19 pandemic. The commissioners of
   Lancet Commission on Dengue and Arboviral Diseases announced that an
   approach should be adopted for the prevention and detection of dengue
   and other arboviral diseases during the COVID-19 pandemic in tropical
   and subtropical regions ([54]14). Co-infection of SARS-CoV-2 and dengue
   has become a hot research topic, and the common pathogenesis and
   biological pathways of the two can be used as resection points for
   project research.

   In our study, we compared SARS-CoV-2 and DENV by bioinformatics
   analysis and obtained 460 common core targets and 30 common critical
   pathways through Gene Ontology and KEGG enrichment to identify host
   factor interaction networks between the two viruses. More in-depth and
   more comprehensive analyses of enormous amounts of data on RNA viruses
   such as SARS-CoV-2 and DENV can reveal the molecular mechanisms by
   which viruses infect hosts and can also help in designing therapies.

Materials and Methods

Collection of COVID-19-Associated Genes

   COVID-19-related targets were gathered from [55]GSE147507,
   [56]GSE155249, and [57]GSE157103 by searching the Gene Expression
   Omnibus (GEO) database ([58]https://www.ncbi.nlm.nih.gov/geo). GEO is
   an international public repository that archives and freely distributes
   microarrays, next-generation sequencing, and other forms of
   high-throughput functional genomics data submitted by the research
   community. The [59]GSE147507 dataset involves the transcriptional
   response to SARS-CoV-2 infection by expression profiling with
   high-throughput sequencing and detailed transcriptomic analysis of
   blood, lung, and bronchus samples from COVID-19 cases ([60]15). The
   [61]GSE155249 dataset recorded circuits between infected macrophages
   and T-cells in SARS-CoV-2 pneumonia through high-throughput sequencing
   of alveolar lavage fluid ([62]16). The [63]GSE157103 dataset entails a
   large-scale and multiomics analysis of the severity of COVID-19
   ([64]17). We obtained information by analyzing a set of differentially
   expressed genes (DEGs) from the above three datasets and screened the
   host gene set at FDR (false discovery rate) <0.000001 and | LogFC |≥1.

   In addition, we screened COVID-19-related host factors through PubChem
   ([65]https://pubchem.ncbi.nlm.nih.gov/#query=covid-19), Comparative
   Toxicogenomics Database (CTD) ([66]http://ctdbase.org/), and DisGeNET
   ([67]https://www.disgenet.org/covid/diseases/summary/). PubChem is a
   public repository for archiving biological data of small molecules
   obtained through high-throughput RNA interference screening, aimed at
   identifying critical genes responsible for a biological process or
   disease condition ([68]18). CTD is a public resource for
   literature-based methods to store disparate information for
   toxicogenomics, gene expression profiles, diseases, environmental
   exposures, and pharmaceuticals ([69]19). DisGeNET is a wide-ranging
   platform that integrates data on the genetic basis of human diseases
   and collects data on genotype-phenotype relationships from several
   resources, including Mendelian, complex, environmental and rare
   diseases, and disease-related traits ([70]20). The top 500 hits in each
   database were included. The date of access to the above six databases
   was December 10, 2020.

Collection of Dengue-Associated Genes

   We obtained dengue-related genes from the results of expression
   profiling by high-throughput sequencing, including [71]GSE38246,
   [72]GSE51808, and [73]GSE84331, from GEO. The criteria for screening
   host genes for each data set included FDR <0.000001 and | LogFC | ≥ 1.
   We searched CTD ([74]http://ctdbase.org/), DisGeNET
   ([75]https://www.disgenet.org/), and GeneCards
   ([76]https://www.genecards.org/) at the same time and screened the top
   500 hits in the three databases. Among them, the [77]GSE38246 ([78]21),
   [79]GSE51808 ([80]22), and [81]GSE84331 ([82]23) datasets provide host
   factors required for DENV replication. GeneCards is an authoritative
   compilation of annotative information that provides concise genome,
   proteome, transcriptome, disease, and function data on all known and
   predicted human genes ([83]24). The date of access to the above six
   databases was December 10, 2020.

Protein-Protein Interaction Analysis and Network Construction

   Common host factors between COVID-19- and dengue-related targets were
   identified using the web tools provided by the Van de Peer Lab
   ([84]http://bioinformatics.psb.ugent.be/webtools/Venn/) and input into
   the STRING website (version 11.0, [85]https://string-db.org/) ([86]25)
   to generate a protein-protein interaction (PPI) network; the top 50
   genes were selected. The PPI results were analyzed and visualized using
   Cytoscape-3.8.1 ([87]https://cytoscape.org/) ([88]26, [89]27).
   CytoHubba ([90]http://apps.cytoscape.org/apps/cytohubba) ([91]28) was
   applied to perform network topology analysis. Then, module analysis of
   the PPI network was performed by Molecular Complex Detection (MCODE) in
   Metascape ([92]https://metascape.org/) ([93]29) and exported to
   Cytoscape-3.8.1 for further visualization processing, making the module
   concise and clear.

Gene Ontology and Pathway Enrichment Analysis

   We used R software to perform Gene Ontology and KEGG pathway enrichment
   analyses of genes in common host factor interaction networks. We chose
   a two-sided hypergeometric test and Bonferroni step down for P-value
   correction (P < 0.05). The ShinyGO v0.61 platform
   ([94]http://bioinformatics.sdstate.edu/go/) was used for WIKI pathway
   enrichment analysis.

Inference of Upstream Pathway Activity

   SPEED2 ([95]https://speed2.sys-bio.net/) ([96]30) is a signaling
   pathway enrichment analysis tool that is used to presume upstream
   pathway activity of the host factor interaction networks between
   SARS-CoV-2 and DENV. We first applied a list of gene IDs and then chose
   the Bates test for test enrichment statistics. When clarifying
   transcriptome data, it is important to infer which signaling pathway
   triggers a particular gene expression program rather than to score
   signaling pathway activity.

Tissue-Specific Enrichment Analysis of Top Genes

   Genotype-Tissue Expression (GTEx) was performed by using the GTEx
   Project ([97]https://www.gtexportal.org/home/), which will help
   researchers understand inherited susceptibility to disease by
   identifying regions of the genome that influence whether and how much a
   gene is expressed ([98]31, [99]32). Tissue-specific enrichment
   calculates tissue-specific gene enrichment in an input gene set to
   assign tissue identities to single-cell clusters and differentiated
   embryonic stem cells. We used R software to perform tissue-specific
   enrichment analysis.

Predicting Drugs Through Interaction Networks of Chemicals and Proteins

   STITCH (search tool for interactions of chemicals) integrates
   information about drug-target relationship information for over 68,000
   different chemicals, including 2,200 drugs, and connects them to 1.5
   million genes ([100]http://stitch.embl.de/). We used STITCH to predict
   related drugs that might provide intervention or treatment for these
   diseases.

Results

Identification of Common Host Factors Between COVID-19 and Dengue

   After searching for COVID-19 based on CTD, we retrieved 7,726 related
   host genes. By searching COVID-19 in DisGeNET, 1,843 related genes were
   identified. Similarly, we obtained 651 genes by searching COVID-19 in
   PubChem. The above data were selected from the top 500 genes in
   accordance with relevance scores. From three transcriptomics analyses,
   154 factors were obtained from [101]GSE147507. In addition, 168 host
   factors were obtained from [102]GSE155249, which was the result of the
   study about bronchoalveolar lavage fluid samples. In total, 202
   constituents were acquired from [103]GSE157103, which was the result of
   the study about plasma and leukocyte samples. The selection criteria
   were FDR<0.000001 and |LogFC|≥1. Positive logFC indicates the
   logarithmic degree of upregulation, and negative logFC indicates the
   logarithmic degree of downregulation. Eventually, we included 1,721
   host-related genes by collecting the union of the genes collected in
   the database and the host factors analyzed by transcriptomics.

   On the basis of searches in CTD, DisGeNET, and GeneCards, we identified
   2,679, 496, and 110 targets relevant to DENV, respectively. Among them,
   the top 500 from CTD were included, and the data form the other two
   were all included. Based on transcriptome data analysis of
   [104]GSE38246, [105]GSE51808, and [106]GSE84331, 1, 845, and 276
   related host factors, respectively, were identified.

   We identified 1,721 COVID-19-associated targets and obtained the gene
   names of 1,901 dengue-related genes, with no repetition. After
   collating COVID-19- and dengue-associated targets, these two target
   categories were compared and filtered to identify common elements using
   the web tools provided by the Van de Peer Lab. Ultimately, 460 common
   host factors were obtained, including MMP2, PDF, PFKP, SLC25A3, IGF1,
   CCL4, TLR4, and AhR. The screening process of shared targets between
   COVID-19 and dengue is shown in [107]Figure 1.

Figure 1.

   [108]Figure 1
   [109]Open in a new tab

   The screening process of shared targets between COVID-19 and dengue. In
   total, 460 factors were shared. False discovery rate (FDR) using a
   Benjamini-Hochberg approach (FDR<1×10^-6) and |logFC| ≥ 1. Venn diagram
   tool ([110]http://bioinformatics.psb.ugent.be/webtools/Venn/). CTD:
   Comparative Toxicogenomics Database ([111]http://ctdbase.org/).
   DisGeNET: a platform containing genes associated with human diseases
   ([112]https://www.disgenet.org/). PubChem: a collection of accessible
   chemical information ([113]https://pubchem.ncbi.nlm.nih.gov/).
   GeneCards: a database of human genes ([114]https://www.genecards.org/).
   GEO: Gene Expression Omnibus ([115]https://www.ncbi.nlm.nih.gov/geo/).

PPI Network and Module Analysis

   A PPI network was produced by STRING based on common host factors and
   reimported into Cytoscape-3.8.1 for visual analysis and network
   topology. The order of degree value from high to low was IL-6, AKT1,
   TNF, ALB, TLR4, and STAT3, among others as shown as [116]Figure 2.
   Through MCODE of Metascape, we identified highly interconnected
   portions and screened 11 vital modules on the basis of interaction
   among the 460 targets. The degree of the node can be seen intuitively
   in [117]Figure 3. The greater is the number of points, the higher is
   the degree. In module 1, the hub nodes were GSK3B, TLR4, TBK1, IKBKE,
   and HSP90AA1. In module 2, the hub nodes were BUB1, BUB1B, CENPA, and
   BIRC5. In module 3, the hub nodes were STAT3, IL-4, TGFB1, HGF, and
   PIK3CA. In module 4, the hub nodes were NFKB1, EGFR, and JUN. In module
   5, the hub nodes were AGT, APP, CCL4, CCL5, and CXCL12. In module 6,
   RAD51 was the most important target.

Figure 2.

   [118]Figure 2
   [119]Open in a new tab

   PPI network of common targets between COVID-19 and dengue. In the
   figure, the circle nodes indicate genes, and edges specify interactions
   between nodes.

Figure 3.

   [120]Figure 3
   [121]Open in a new tab

   Module analysis of the 460 shared genes. Each module demonstrates
   different biological process functions. (A) module 1; (B) module 2; (C)
   module 3; (D) module 4; (E) module 5; (F) module 6.

Gene Ontology and KEGG Pathway Enrichment Analyses

   Analysis of the Gene Ontology biology process in the [122]Figure 4
   showed that genes of the host networks are related to the process of
   occurrence and development, including Response to stress, Response to
   organic substance, Response to cytokine, Cellular response to organic
   substance, and Cellular response to chemical stimulus, as well as other
   cell regulation of response to external pathogen stimulation. In
   addition, molecular function analysis in [123]Figure 5 revealed
   signaling receptor binding, identical protein binding, enzyme binding,
   cytokine receptor binding, drug binding, and others. According to KEGG
   pathway enrichment analysis, influenza A, measles, hepatitis B,
   hepatitis C, pathways in cancer, Epstein-Barr virus infection, and the
   Toll-like receptor (TLR) signaling pathway play important roles in
   interaction between COVID-19 and dengue, that shows in [124]Figure 6.
   The TLR Signaling Pathway, Regulation of toll-like receptor signaling
   pathway, Spinal Cord Injury, AGE/RAGE pathway, Oncostatin M Signaling
   Pathway, PI3K-Akt Signaling Pathway, and Photodynamic therapy-induced
   NF-kB survival signaling in WIKI were important biological function
   pathways in the host factor interaction networks, as shown in
   [125]Table 1.

Figure 4.

   [126]Figure 4
   [127]Open in a new tab

   Network of enrichment analysis of GO biological process terms. The size
   of the node represents the number of genes involved in the enrichment
   pathway, and the color from blue to red indicates the P-value from
   small to large.

Figure 5.

   [128]Figure 5
   [129]Open in a new tab

   Network of enrichment analysis of GO molecular function terms. The size
   of the node represents the number of genes involved in the enrichment
   pathway, and the color from blue to red indicates the P-value from
   small to large.

Figure 6.

   [130]Figure 6
   [131]Open in a new tab

   The network of pathway enrichment analysis through KEGG. The size of
   the node represents the number of genes involved in the enrichment
   pathway, and the color from blue to red indicates the P-value from
   small to large.

Table 1.

   Results of pathway enrichment analysis through WIKI.
   Enrichment FDR Genes in list Total genes Functional Category
   5.57E-42 42 102 Wiki : Toll-like Receptor Signaling Pathway
   2.10E-38 44 139 Wiki : Regulation of Toll-like receptor signaling
   pathway
   1.22E-29 35 116 Wiki : Spinal Cord Injury
   1.51E-28 28 66 Wiki : AGE/RAGE pathway
   2.72E-27 27 65 Wiki : Oncostatin M Signaling Pathway
   4.91E-26 49 339 Wiki : PI3K-Akt Signaling Pathway
   1.03E-25 21 35 Wiki : Photodynamic therapy-induced NF-kB survival
   signaling
   4.16E-25 41 234 Wiki : VEGFA-VEGFR2 Signaling Pathway
   6.67E-25 28 87 Wiki : Retinoblastoma (RB) in Cancer
   9.60E-25 23 50 Wiki : Photodynamic therapy-induced AP-1 survival
   signaling
   5.57E-24 26 76 Wiki : Leptin signaling pathway
   1.72E-23 27 88 Wiki : Allograft Rejection
   1.05E-21 21 50 Wiki : Vitamin B12 Metabolism
   1.05E-21 21 50 Wiki : Hepatitis C and Hepatocellular Carcinoma
   1.05E-21 19 37 Wiki : Fibrin Complement Receptor 3 Signaling Pathway
   1.44E-21 23 66 Wiki : Folate Metabolism
   1.61E-21 25 84 Wiki : Apoptosis
   2.12E-21 24 76 Wiki : Prolactin Signaling Pathway
   2.12E-21 27 105 Wiki : Senescence and Autophagy in Cancer
   4.83E-21 16 24 Wiki : IL1 and megakaryocytes in obesity
   8.82E-21 25 90 Wiki:T-Cell antigen Receptor (TCR) Signaling Pathway
   1.54E-20 25 92 Wiki : Corticotropin-releasing hormone signaling pathway
   1.67E-20 28 125 Wiki : Ebola Virus Pathway on Host
   2.80E-20 16 26 Wiki : Cytokines and Inflammatory Response
   3.23E-20 24 85 Wiki : Selenium Micronutrient Network
   1.02E-19 19 46 Wiki : Aryl Hydrocarbon Receptor
   1.02E-19 24 89 Wiki : Pancreatic adenocarcinoma pathway
   1.28E-19 20 54 Wiki : IL-4 Signaling Pathway
   4.30E-19 18 42 Wiki : TNF related weak inducer of apoptosis (TWEAK)
   Signaling Pathway
   4.30E-19 18 42 Wiki : IL-2 Signaling Pathway
   [132]Open in a new tab

Upstream Pathway Activity

   SPEED2 is a signaling pathway annotation enrichment analysis tool with
   gene sets derived from pathway perturbation biological experiments in
   human cell lines. The understanding of incorporated regulatory
   mechanisms can be simplified by identifying modulated pathways upstream
   of differentially expressed genes. Z-values, which we refer to as
   Zrank, were rescaled for each experiment to the interval, providing the
   average signed-rank for this gene across those experiments. If the
   Z-value was higher than zero, the corresponding gene was upregulated;
   if it was lower than zero, it was downregulated. Then, its significance
   was assessed by a P-value, and ranked lists covering all of the
   measured genes were used to assess enrichment or depletion of a
   user-provided gene list. IL-1, Hippo, p53, TNF-α, and TLR occupied a
   pivotal position as shown in as shown in [133]Figure 7.

Figure 7.

   [134]Figure 7
   [135]Open in a new tab

   The rank of signaling pathway activity (adjusted P-value <0.05). The
   color represents the adjusted P-value; the brighter the color, the
   higher the rank.

Tissue-Specific Enrichment Analysis of Host Factor Interaction Networks

   In tissue-specific enrichment analysis, each column represents a
   different tissue, and each line represents a different factor; the
   higher is the value of TPM, the darker is the color, and the density of
   specific distribution of host factors in corresponding tissue is
   higher. The heat map in [136]Figure 8 shows the correlation between
   different samples and factors. For example, the specific distributions
   of host factors in the lung included HSP90AA1, ICAM1, and JUN.

Figure 8.

   [137]Figure 8
   [138]Open in a new tab

   The heat map shows the correlation between different samples and
   factors. Columns represent genes, rows represent tissues, and colors
   represent the level of enrichment, whereby a darker color indicates a
   more significant gene in corresponding tissue.

Predicted Drugs Through Host Target Interactions With Chemicals

   STITCH aims to construct a database that integrates data in the
   literature and various databases of biological pathways, drug-target
   relationships, and binding affinities. We used it to predict
   pharmaceutical preparations according to chemical protein interactions
   and drug-target relationships. The drugs related to most genes mainly
   involve retinoic acid, estradiol, dexamethasone, and estrogen. The list
   of suggested agents for COVID-19 and dengue co-infection can be seen
   intuitively in [139]Table 2. Among them, retinoic acid had the most
   enriched targets, with 471 genes and 84 genes on the list. Resveratrol,
   an anti-inflammatory and antioxidant, showed 180 genes in its
   enrichment target, of which 64 were on the list.

Table 2.

   List of suggested agents for COVID-19 and dengue co-infection.
   Enrichment FDR Genes in list Total genes     Functional Category
   3.90E-59            64           180     resveratrol(CID000445154)
   4.01E-58            68           223     PD98059(CID000004713)
   8.09E-58            59           150     SB203580(CID000176155)
   2.02E-57            67           221     genistein(CID005280961)
   2.02E-57            59           153     p-38(CID100200878)
   4.59E-57            60           163     resveratrol(CID100005056)
   1.75E-56            83           410     dexamethasone(CID000005743)
   1.75E-56            55           129     SB203580(CID100176155)
   4.23E-55            63           201     LY294002(CID000003973)
   6.22E-54            57           157     p-38(CID000200878)
   6.22E-54            62           200     PD98059(CID100004713)
   1.14E-52            84           471     retinoic acid(CID000444795)
   2.00E-52            80           419     retinoic acid(CID100005538)
   1.99E-51            73           340     dexamethasone(CID100003003)
   1.67E-50            53           144     curcumin(CID000969516)
   1.79E-50            58           187     genistein(CID105280961)
   4.16E-50            80           449     estradiol(CID000005757)
   4.56E-49            75           393     estrogen(CID000448537)
   4.76E-49            73           367     estradiol(CID100000450)
   2.68E-48            61           233     cycloheximide(CID000006197)
   3.12E-48            55           175     LY294002(CID100003973)
   6.38E-48            49           127     curcumin(CID100002889)
   1.70E-47            65           285     phorbol ester(CID010282255)
   9.32E-45            58           231     nitric oxide(CID000000945)
   3.35E-44            54           194     nitric oxide(CID100000945)
   1.10E-43            65           325     estrogen(CID100003054)
   1.95E-43            65           328     ethanol(CID000000702)
   1.28E-41            50           174     quercetin(CID005280343)
   3.66E-40            42           113     U0126(CID100005637)
   4.59E-40            54           229     phorbol ester(CID110282255)
   [140]Open in a new tab

Discussion

   Both COVID-19 and dengue are international public health issues with
   adverse effects, and co-infection exacerbates the worldwide burden on
   healthcare systems. Moreover, the similar clinical manifestations and
   serological overlap between COVID-19 and dengue raise the risk of
   incorrect diagnosis. There is also a possibility that overlapping
   immunological cascades may affect not only disease severity but also
   vaccine development ([141]3). In view of this condition, we analyzed
   host factor interaction networks of COVID-19 and dengue co-infection
   from the perspective of biological pathways and attempted to screen
   potential agents. To elucidate mechanisms, we identified common host
   factors and performed further analysis. A PPI network was created, and
   vital modules were analyzed. Subsequently, biological functional
   analysis with Gene Ontology terms, KEGG, and WIKI was completed. In
   addition, signaling pathway activity was assessed, and tissue-specific
   enrichment analysis was carried out. Finally, candidate drugs were
   detected, providing a basis for in-depth research on therapeutic
   strategies.

   This study involved multiple steps. On the one hand, the approach
   illustrated the role of pathogens in disease development processes with
   a progressive logical relationship, echoing the biological effects of
   viruses. The collection source combined the database with transcriptome
   data, and the method combined union and intersection, making the
   dataset more comprehensive and accurate. On the other hand, the
   multiple steps helped in overcoming the limitations of cross-sectional,
   single-scale, and one-dimensional studies. Dynamic evolution in the
   host during processes of infection and immune response were analyzed at
   multiple scales, such as factors, pathways, and tissue, reflecting a
   dynamic, multiscale, and multidimensional research perspective. This
   study involved seven separate sections each with its own distinct
   method and outcome, yet the whole was integrated and complete.

Identification of Common Host Factors Between COVID-19 and Dengue

   After comparing the collected genes, 460 host factors common to
   COVID-19 and dengue were found, including MMP2, PDF, PFKP, SLC25A3,
   IGF1, CCL4, TLR4, and AhR, and further analysis of interaction networks
   was performed.

   COVID-19 pathogenesis is associated with excessive cytokine release,
   such as CCL4 ([142]33). This factor can dominate the chemokine
   signature and shows persistently high concentrations in patients with
   severe COVID-19 ([143]34). Moreover, massive cytokine secretion is
   regarded as a part of the underlying mechanism in dengue. As shown by a
   study from Mathieu Nacher, dendritic cells infected by DENV express
   CCL4, which is associated with vasodilation, endothelial dysfunction,
   and disease severity ([144]35). In SARS-CoV-2-infected cells,
   overproduced plasminogen activator inhibitor-1 (PAI-1) binds to TLR4 on
   macrophages, inducing secretion of proinflammatory cytokines and
   chemokines ([145]36). For patients with dengue, TLR4 on platelets
   binding with DENV NS1 is triggered, causing thrombocytopenia and
   hemorrhage ([146]37).

   The shared factors are thought to be involved in other activities as
   well. In COVID-19 patients, AhR stimulation upregulates expression of
   mucins in alveolar epithelial cells, causing accumulation of alveolar
   mucus and leading to silent hypoxia formation, a unique feature of the
   disease ([147]38). The role of factors in DENV replication was also
   investigated. It has been reported that an increase in DENV titer is
   related to activated AhR; conversely, replication is inhibited by an
   AhR antagonist. In summary, AhR appears to promote DENV replication
   ([148]39).

PPI Network and MCODE Analyses Revealed Key Factors of Host Factor
Interaction Networks

   Targets with high interaction rates in host factor interaction networks
   were obtained from PPI network analysis and might be treatment targets,
   including IL-6, AKT1, TNF, ALB, TLR4, STAT3, and JUN, among others.

   IL-6 is elevated in COVID-19 and dengue and is related to the severity
   of both ([149]40). It can induce the production of other inflammatory
   cytokines by interacting with many different cells and can cause
   E-cadherin expression; this leads to increased endothelial
   permeability, which may result in Acute Respiratory Distress Syndrome
   (ARDS), a fatal comorbidity of COVID-19 ([150]41, [151]42). As reported
   by M Juffrie ([152]43), this process is associated with plasma leakage
   and shock caused by increased permeability in dengue patients. Previous
   research has proposed that IL-6 contributes more to the disease
   pathogenesis of COVID-19 than does DHF, and serum IL-6 levels are
   higher in severe pneumonia patients with COVID-19 ([153]44).

   Virus-induced apoptosis is a significant event that determines
   pathogenicity and virulence, though studies on the function of JUN in
   SARS-CoV-2-induced apoptosis are limited. According to a study on IBV
   (avian coronavirus infectious bronchitis virus), c-JUN seems to protect
   coronavirus-infected cells from apoptosis ([154]45). Interestingly, in
   contrast to c-JUN, the findings from several studies suggest that the
   c-Jun N-terminal kinase (JNK) pathway exerts a crucial effect on
   regulating apoptosis evoked by coronavirus infection ([155]45–[156]47).
   In a study of colony transcriptomes ([157]48), c-JUN was highly
   expressed in DENV2-refractory colonies. In addition, the findings from
   a study about JNK by Avisha Chowdhury suggest that the pathways have a
   broad antiviral function against DENV, depending on apoptosis induction
   and the complement system. Moreover, activated c-JUN participates in
   controlling the rearrangement of membrane structures, which is induced
   by DENV and required for replication ([158]49).

   In total, 11 different plates were obtained with the MCODE module in
   analysis of 460 common host factors. The main function reflected by
   plate B was interferon (IFN) signaling.

   Both SARS-CoV-2 and DENV can promote activation of immune cells,
   leading to the release of proinflammatory cytokines, such as IFN, and
   causing capillary leakage, thrombocytopenia, and coagulopathy,
   pathophysiological similarities between COVID-19 and dengue ([159]3).
   In a study of SARS-CoV-treated mice, interleukins, IFN, and chemokines
   were notably increased within 24 h ([160]50). Among them, IFNs have
   antiviral, antiproliferative, and immunomodulatory functions and are
   regarded as the first line of the body’s antiviral defenses ([161]51).
   Nevertheless, pDCs and monocytes produce more IFN than other cells
   after viral infection, and innate immune response plays a significant
   role in the formation of the cytokine storm ([162]50). In addition,
   IFN‐γ in combination with other inflammatory mediators can facilitate
   plasma leakage, disseminated intravascular coagulation (DIC), and other
   vascular disorders ([163]3), which is the main pathological
   characteristic of dengue.

Enrichment Analysis Showed the Biological Functions of Host Factor
Interaction Networks

   To investigate interaction networks at the molecular level, functional
   enrichment analysis, including functional annotations of significant
   factors and pathway enrichment analysis, was carried out. GO was
   utilized to classify common factors. In this study, the classification
   was executed according to biological process and molecular function
   categories. Some targets were identified as being associated with the
   response to stress, response to an organic substance, response to
   cytokine in biological processes, and response to signaling receptor
   binding, identical protein binding, enzyme binding, and cytokine
   receptor binding in molecular functions. The cytokine response was
   significant.

   SARS-CoV-2 infection gradually induces an excessive inflammatory
   response and dysregulated immune defense, which results in tissue
   damage and causes acute lung injury (ALI) and even ARDS. The mechanism
   may be due to the massive release of cytokines and chemokines, which
   might be a characteristic of COVID-19 ([164]33). The cytokine storm
   triggers plasma leakage and disseminated vascular coagulation, which
   are life-threatening respiratory symptoms in COVID-19 ([165]52).

   The pathogenesis of dengue includes cytokine storms, vascular leakage,
   complement activation, autoimmunity, and antibody-dependent
   enhancement. Through research on patients with dengue hemorrhagic fever
   or dengue shock syndrome, it was observed that cytokines are produced
   and change rapidly during the course of the illness. T cells,
   macrophages, monocytes, and mast cells can produce several soluble
   factors, such as TNFα, IL-6, IL-8, and IL-10, to increase the vascular
   permeability of primary endothelial cells ([166]2).

   Through KEGG and WIKI, we analyzed the biological function of pathways
   in host factor interaction networks, such as influenza A, measles, and
   hepatitis B in KEGG and the TLR signaling pathway, regulation of
   Toll-like receptor signaling pathway, and spinal cord injury in WIKI.

   According to the findings of this study, the TLR pathway has an
   important position, as revealed by KEGG and WIKI. After the viral spike
   protein binds to host cells via ACE2, SARS-CoV-2 RNA is detected by
   TLRs ([167]52). Similarly, TLRs recognize DENV; TLR3 is able to
   establish an antiviral state ([168]53), and TLR 4 can be utilized by
   DENV to activate platelets in vitro ([169]54).

   Based on WIKI, another pathway shared by COVID-19 and dengue is the
   NOD-like receptor signaling pathway. NLRP3 is a member of the pathway,
   and its activation by Viroporin 3a seems to have an important effect on
   the pathogenesis of SARS-COV infection ([170]55). According to relevant
   studies, SARS-CoV-2 infection causes a range of disease manifestations,
   with the most serious being a massive inflammatory response that
   appears to occur via activation of the NLRP3 inflammasome ([171]56).
   Dengue is related to the processing and release of proinflammatory
   cytokines. After infection, the NLRP3-specific inflammasome is
   stimulated by DENV through inflammasome activation. According to the
   present study, the assembly process of the NLRP3 inflammasome complex
   is promoted by the DENV-2 NS2A and NS2B proteins. NLRP3 further exerts
   its effect on IL-1β by regulating its maturation and secretion
   ([172]57), increasing endothelial permeability and causing vascular
   leakage ([173]58).

Pathway Activity and Tissue Specificity of Host Factor Interaction Networks

   In this study, we also assessed the signaling activity of dysregulated
   host factors, which is important to infer disease mechanisms. According
   to the results, upregulated genes include IL-1, Hippo, and TNF α, among
   others.

   IL-1B, a proinflammatory cytokine, plays a vital role in the
   inflammatory response caused by infectious pathogens ([174]59). IL-1B
   is activated in severe COVID-19, inducing vasodilation and permeability
   and providing the conditions for immune cells to arrive at the sites of
   damage. In addition, it can induce complement activation and
   opsonization ([175]60). The production of IL-1B is promoted by
   platelets after infection with DENV, and the synthesis of
   IL-1B-containing microvesicles is induced by DENV, which ultimately
   increases vascular permeability ([176]61). Moreover, IL-1B-31C carriers
   run the risk of dengue shock syndrome (DSS), implying that IL-1B is
   associated with the pathogenesis of DSS ([177]62).

   TNF-α is a vital driver of inflammation ([178]63). According to
   relevant mouse studies in vitro and in vivo, synergistic activity of
   IFN-γ and TNF-α mimics the symptoms of COVID-19 and triggers robust
   cell death ([179]64). Moreover, TNF receptor levels have been reported
   to show a positive correlation with dengue hemorrhagic fever (DHF)
   severity ([180]65). A study of a mouse model of dengue hemorrhage
   implied that TNF-α induces hemorrhage, resulting in endothelial cell
   death ([181]66, [182]67).

   To explore the novel association of tissue with genes in interaction
   networks, tissue-specific enrichment analysis was performed. For
   example, expression of ICAM-1 was highly enriched in the lung.

   Overall, the investigation of endothelial cell adhesion molecules is
   meaningful for understanding the pathophysiology of COVID-19. As
   reported by previous research, expression of adhesion molecules, such
   as ICAM-1 and VCAM-1, is related to systemic inflammation and might
   contribute to coagulation dysfunction ([183]68). In dengue, increased
   expression of ICAM-1 can result in the consequent influx of cells
   promoting inflammation of the endothelium. Relevant studies are in
   favor of the development of new markers for the evaluation, therapeutic
   response, and clinical follow-up of patients with severe dengue
   ([184]69). Thus, ICAM-1 might be important for co-infection.

Identification of Candidate Drugs Against Co-Infection

   Finally, we studied chemical-protein interaction networks with STITCH
   to identify several candidate agents with the potential to treat
   COVID-19 and dengue co-infection, such as resveratrol, curcumin, and
   quercetin. The findings require appropriate validation. For example,
   quercetin has good antiviral, immune modulation, and anti-inflammatory
   properties and is a potential agent for treating co-infection. However,
   quercetin is genetically toxic, mutagenic, and has other insuperable
   side effects ([185]70), but it provides some guidance for drug
   development for the treatment of co-infection.

   Resveratrol is a well-known anti-inflammatory and antioxidant agent
   ([186]71). It can act on several mechanisms of COVID-19, as shown by
   the research of Anderson O. Ferreira. The drug can reduce entry of the
   virus by blocking binding to ACE2 ([187]72). Furthermore, it reduces
   NF-kB, inhibits the TLR4 pathway, and decreases the production and
   expression of some significant inflammatory factors ([188]73). As an
   endothelial barrier protector, it can protect cells from stressful
   conditions to weaken endothelial inflammation ([189]74). Moreover, it
   has a potential antithrombotic effect ([190]75). With regard to dengue,
   the drug can inhibit HMGB1 translocation to increase activation of ISGs
   and subsequently exert an effect by partially suppressing DENV
   replication ([191]71).

Limitations

   However, there are some limitations in this study. First, the
   measurement methods for genomic data differed among studies, which made
   it difficult to achieve complete unity. Second, in view of the
   limitation of disease-gene and drug-target bipartite networks in
   network pharmacology, associated information loss was inevitable
   ([192]76). Furthermore, the analysis methods were dependent on
   computational power and data acquisition ([193]77), which was based on
   current research; hence, it was difficult to predict the complete
   mechanism. Third, the potential of resveratrol, a well-known
   anti-inflammatory and antioxidant agent, to treat co-infection with
   COVID-19 and dengue was based on data mining–based bioinformatics
   tools, and experimental validation is required.

Conclusion

   In this study, mechanisms of co-infection were preliminarily revealed
   in terms of host factor interaction networks. With a series of methods,
   the core factors and pathways were screened, the biological function
   was analyzed, and new ideas for treatment were revealed. However, as
   this study was based on data mining and analysis, there were
   limitations; thus, more validated and rigorous experiments are required
   to verify our prediction.

Data Availability Statement

   The data used to support the findings of this study are available in
   4TU.RearchData (DOI: https://doi.org/10.4121/14769018).

Author Contributions

   WZ conceived and designed the study and plotted the figures based on
   network bioinformatics, using online databases. HW and CL conducted
   data analysis and performed literature searches. QY, TW, and PW carried
   out literature searches and studies of the background of the disease.
   XL, YJ, and SZ reviewed and revised the manuscript. All authors
   contributed to the article and approved the submitted version.

Funding

   This research was funded by grants from the “Double First-Class” and
   High-level University Discipline Collaborative Innovation Team Project
   of Guangzhou University of Chinese Medicine (Grant No.2021XK16),
   Guangdong Provincial Department of Education Innovation Team Project
   (Grant No: 2018KCXTD007), the Key-Area Research and Development Program
   of Guangdong Province (Grant No. 2020B1111100002), the National Natural
   Science Foundation of China (Grant No. 81973814 and No. 81904132), the
   Natural Science Foundation of Guangdong Province (Grant No.
   2017A030310129), the Natural Science Foundation of Guangdong Province
   (Grant No. 2020A1515010589), the Liu Xiaohong Famous Traditional
   Chinese Medicine Inheritance Studio from the Traditional Chinese
   Medicine Bureau of Guangdong Province (Grant No. 201805), the
   Construction Project of Respiratory Department National Clinical
   Medical Research Center (Grant No. 2110200309), 2018 Guangzhou
   University of Chinese Medicine National University Student Innovation
   and Entrepreneurship Training Project (Grant No. 201810572038), 2020
   National College Student Innovation and Entrepreneurship Training
   Project of Guangzhou University of Chinese Medicine (Grant No.
   202010572001), the Student Learning Team Incubation Project of
   Innovation Academy from The First Affiliated Hospital of Guangzhou
   University of Chinese Medicine (Grant No. 2018XXTD003), and the
   Technology Research of COVID-19 Treatment and Prevention and Special
   Project of Traditional Chinese Medicine Application-Research on the
   platform construction for the prevention and treatment of viral
   infectious diseases with traditional Chinese medicine (Grant No.
   2020KJCX-KTYJ-130).

Conflict of Interest

   The authors declare that the research was conducted in the absence of
   any commercial or financial relationships that could be construed as a
   potential conflict of interest.

Publisher’s Note

   All claims expressed in this article are solely those of the authors
   and do not necessarily represent those of their affiliated
   organizations, or those of the publisher, the editors and the
   reviewers. Any product that may be evaluated in this article, or claim
   that may be made by its manufacturer, is not guaranteed or endorsed by
   the publisher.

Acknowledgments