Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Host Response to Respiratory Bacterial Pathogens as Identified by Integrated Analysis of Human Gene Expression Data

  • Steven B. Smith,

    Affiliations Computational Biology, Quantitative Sciences, GlaxoSmithKline, Collegeville, Pennsylvania, United States of America, Institute for Genome Science, University of Maryland, Baltimore, Maryland, United States of America

  • Michal Magid-Slav,

    Affiliation Computational Biology, Quantitative Sciences, GlaxoSmithKline, Collegeville, Pennsylvania, United States of America

  • James R. Brown

    Affiliation Computational Biology, Quantitative Sciences, GlaxoSmithKline, Collegeville, Pennsylvania, United States of America

Host Response to Respiratory Bacterial Pathogens as Identified by Integrated Analysis of Human Gene Expression Data

  • Steven B. Smith, 
  • Michal Magid-Slav, 
  • James R. Brown


Respiratory bacterial pathogens are one of the leading causes of infectious death in the world and a major health concern complicated by the rise of multi-antibiotic resistant strains. Therapeutics that modulate host genes essential for pathogen infectivity could potentially avoid multi-drug resistance and provide a wider scope of treatment options. Here, we perform an integrative analysis of published human gene expression data generated under challenges from the gram-negative and Gram-positive bacteria pathogens, Pseudomonas aeruginosa and Streptococcus pneumoniae, respectively. We applied a previously described differential gene and pathway enrichment analysis pipeline to publicly available host mRNA GEO datasets resulting from exposure to bacterial infection. We found 72 canonical human pathways common between four GEO datasets, representing P. aeruginosa and S. pneumoniae. Although the majority of these pathways are known to be involved with immune response, we found several interesting new interactions such as the SUMO1 pathway that might have a role in bacterial infections. Furthermore, 36 host-bacterial pathways were also shared with our previous results for respiratory virus host gene expression. Based on our pathway analysis we propose several drug-repurposing opportunities supported by the literature.


Bacterial community acquired pneumonia (CAP) is a major morbidity factor world-wide, particularly for the young and elderly [1,2]. In the USA alone, pneumonia is the seventh leading cause of mortality with more than 59,000 deaths in the year 2008 and costing nearly $20 billion in healthcare and productivity costs [1].

Increasing incidents of clinical antibiotic resistance adds further urgency to the development of new therapeutics for CAP and hospital acquired pneumonia (HAP). The most prevalent respiratory bacterial infections are caused by so-called “ESKAPE” species, Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa and Enterobacter spp., which are capable of “escaping” the biocidal action of existing antibiotics [3]. Additional pathogenic species such as Streptococcus pneumoniae show increasing levels of multi-drug resistance further complicating the control of respiratory infections [4]. Meanwhile, the development of new classes of antibiotics has been a major challenge for the pharmaceutical and biotech industries [5].

Any infection involves specific interactions by the pathogen with the host’s cellular proteins and pathways which potentially opens new therapeutic opportunities. While protective immunity is the main host response to resist infection, bacteria are also dependent upon modulation of many host pathways for their proliferation and viability. An example is bacterial suppression of cellular phagocytosis by disruption of toll-like receptor (TLR) cross-talk using specific virulence factors (reviewed in 6,7). However, activation of other host proteins appears to be essential for bacterial pathogenicity. For example, profiling of homogenates made from mouse lungs infected with S. pneumoniae against a kinomics peptide array revealed a wide range of human kinases that were either activated or suppressed [8]. Internationalization of S. aureus by HeLa cells is blocked by inhibition of Src kinase [9]. Human focal adhesion kinase (FAK) plays a role in the invasion of brain microvascular endothelial cells by Group B Streptococcus (GBS), a causative pathogen of neonatal meningitis [10]. The host phosphoinositide-3-kinase (PI3K) lipid signaling pathway is essential for both obligate intracellular bacterial pathogens, such as Legionella pneumophila, Brucella abortus, Mycobacterium tuberculosis and Salmonella enterica [11] as well as opportunistic cellular invaders like P. aeruginosa [12] and GBS [13]. Other host factors exploited by pathogens include integrins, a family of heterodimeric receptors that mediate cellular adhesion, signaling and migration [14].

Targeting host-bacterial interactions as a therapeutic strategy has several potential advantages over current direct-acting anti-bacterial drugs. Arguably, pathogen-host interactions are less vulnerable to the Darwinian selection pressures that drive the rapid evolution of pathogen resistance by antibiotics. Since genetically diverse bacteria often use similar host pathways for cellular adhesion and invasion, therapeutics affecting a broad spectrum of bacterial species could be potentially developed. From the drug development perspective, current pharmaceutical collections have more developable compounds with human protein specificity than anti-microbials [5]. On the other hand, targets need to be selected carefully to avoid significant dampening of host immune responses or other side-effects.

With the growth in host-bacteria interactome datasets, in particular genome-wide gene expression patterns in human cell-lines and clinical subjects under bacterial infection challenge, there is the opportunity to employ computational approaches to discover new antibacterial targets unique to the host [15]. In this study we used an integrative analysis to discover biologically relevant genes and pathways involved in host responses to infections from multiple respiratory bacteria. We adapted a previously published pipeline for the discovery of host-respiratory virus interactions using transcriptomic datasets augmented by other multiple data-sources [16]. Based on stringent quality control filters as well as criteria for clinical importance, we focused on datasets for the Gram-negative bacterium P. aeruginosa and the Gram-positive bacterium S. pneumoniae [1720]. Subsequently, we identified potential repurposed drug targets in those pathways that can be modulated for favorable host responses during infection.

Clinically, co-infection or sequential respiratory tract infections of multiple viruses or bacteria often occur, which can complicate diagnosis and treatment. For example, during influenza viral outbreaks, co-infections by bacterial respiratory pathogens are responsible for elevating risk of death or long-term illness, especially among children [21]. Therefore, we also extended our comparison of host targets across bacterial and viral respiratory infections [16] to identify potential avenues to treat multiple respiratory pathogens.


The National Center for Biotechnology Information’s (NCBI) Gene Expression Omnibus (GEO) database ( accessed September 2011) was queried for human mRNA datasets involving five infectious reparatory bacterial pathogens: Pseudomonas aeruginosa, Streptococcus pneumoniae, Legionella pneumophila, Klebsiella pneumoniae, and Haemophilus influenzae. Subsequent filtering steps (Figure 1) reduced the number of bacteria with suitable datasets to two species, P. aeruginosa and S. pneumoniae (Table 1).

Figure 1. Outline of iterative GEO dataset filtering process.

The analysis pipeline was used to select and quality control GEO datasets linked to human and mammalian mRNA expression under respiratory bacterial challenge. Specific inclusion criteria are described in the Materials and Methods.

Bacterial SpeciesCell TypeArray Platform2GSETimeBacterial strainSample size (treatment/control)
Pseudomonas aeruginosaA549U95Av214693hrN/A3/3
BEAS-2BU133A 2.068024hrN/A3/3
Streptococcus pneumoniaePBMCU133A6269N.S.1N/A13/5
Detroit 562U133 Plus2.085272hrD393/3

Table 1. Profile of included GEO datasets.

Download CSV

GEO datasets were selected based on the following inclusion criteria: 1) the infected cells are normal; 2) only one bacterium is being studied in each “treatment group”; 3) the bacterium infected is wild-type and 4) each “treatment group” and “control group” must have at least 3 samples. All analyzed GEO datasets contain at least one “treatment group” and “control group”. “Treatment” was the experimental variable under study, usually a bacterium type, strain, or time point. “Group” was a collection of individual “samples”, or replicates, each of which originates from their own microarray chip. “Comparison group” was the treatment group compared to a control group. A particular dataset may have more than one comparison group. All criteria for dataset inclusion in the final analysis were chosen prior to the analysis.

Dataset selection, Quality Control (QC) filtering, probe mapping and filtering, and differential gene expression were performed as described previously [16] with the added criterion that the sample could be derived from human serum provided it was not a sepsis-related study, which might involve a different host response of immune system over activation [22]. Our previous methodology is summarized briefly. GEO datasets were included based on four criteria described above and subsequently underwent Quality Control (QC) analysis in Array Studio, which included Median Absolute Deviation (MAD) score, Principle Component Analysis (PCA), pair-wise correlation and kernel density. Probes passing the log2(50) least square mean (LSM) threshold were mapped to their corresponding genes, and differential expression analysis was performed using Array Studio v4.1.1.58 [23]. Differentially expressed genes are those that have a fold change above 1.5 or below -1.5 and a p-value < 0.05.

The differentially expressed gene lists from each comparison group were analyzed for enriched pathways via the Python language package ‘‘ that calculates p-values for each of the 683 pathway maps in the MetaBase v6.14 (Thomson Reuters) using the Fisher Exact test [24]. Pathway significance was defined as a pathway p value < 0.01. To determine pathways enriched across all bacterial species studied, pathways were ranked first by Bacterial Count (BC) then by the pathway’s sum of Normalized Bacterial Expression or NBE (Table S1). A pathway’s BC is defined by the number of bacteria represented by at least one significant comparison group. The NBE for each pathway was calculated using the number of comparisons containing significant pathways within a bacterial species relative to the total number of comparisons within that bacterial species. Ranking the pathways by BC and then NBE resulted in a clearer determination of pathways shared across multiple bacteria, irrespective of time or number of comparison groups.

Repositioned drug candidates for each of the genes in the top pathways were analyzed using Drug Bank as described previously [16]. We looked for genes with a BC of 2 from the top pathways and then searched the literature for potential areas where the candidate therapeutic could be used across different bacterial or viral infections.

We compared the significant gene and pathway overlaps with those from our previous viral study [16]. We used a hypogeometric test to test for significance of top genes and pathways in the present study with top gene and pathway overlap in our previous viral study, respectively. We used the following parameters for the R function phyper(x, m, n, k) [25]:

x= the number of intersecting top genes or pathways between bacterial and viral studies

m=the number of top bacterial genes or pathways

n=the number of total genes (209) or pathways (683) minus the number of top bacterial genes or pathways, respectively.

k=the number of top viral genes or pathways


Dataset Filtering and Analysis

Our analysis approach involved three distinct steps. First, using extensive database searches and stringent quality control (QC) filtering we identified and selected the most suitable human or mammalian gene expression datasets derived from challenges by respiratory bacterial pathogens. Second, rigorous statistical analysis was used to find significant pathways enriched for differentially expressed genes (Table S1). Third, we linked known drugs to targets in these pathways to suggest potential drug repurposing opportunities for respiratory infections.

There were 18 GEO human or mammalian microarray datasets associated with gene expression after exposure to respiratory bacterium Pseudomonas aeruginosa, Streptococcus pneumoniae, Legionella pneumophila, Klebsiella pneumoniae, or Haemophilus influenza (Tables 1 and S2), Three of these 18 datasets were associated with two or more different bacterial species (GSE11051, GSE17221, and GSE6377). We filtered these datasets based on inclusion criterion described in the Material and Methods. Table S2 shows all excluded GSEs and the reasons for rejection.

We identified four candidate GEO datasets for further QC: GSE1469 [19], GSE6269 [20], GSE6802 [17], and GSE8527 [18] (Table 1). Within each GEO dataset, we only considered samples meeting our dataset inclusion criteria for further QC. Specifically, 10 of the 16 samples from GSE1469 were infected with four different mutant P. aeruginosa strains (combinations of exoS, exoT, and exoY gene deletions) while 15 out of the 32 samples from GSE8527 were infected with five different mutant S. pneumoniae strains (all encapsulated strains with lab-generated capsule loci deletions or ∆cps). Thus all of these sample groups were excluded from further analysis. GSE8527 had three S. pneumoniae isolate groups: serotype 2-encapsulated strain D39 (abbreviated: D39), serotype 19F-encapsulated strain G54 (G34) and serotype 4-encapsulated strain TIGR4 (TIGR4). The S. pneumoniae strain G34 was excluded due to small control group sample size while the D39 and TIRG4 groups were independently analyzed. GSE6802 contained respiratory syncytial virus (RSV) and S. aureus samples that were not included in this analysis because there were not associated with respiratory bacterial infections. Similarly, GSE6269 contained S. aureus, Escherichia coli, and Influenza A virus samples that were not included for this analysis. Lastly, only samples originating from the Affymetrix U133A chip in GSE6269 were used for consistency between S. pneumoniae infected and non-infected samples.

Although no single dataset failed QC, sample GSM173246 from GSE6269 failed to meet MAD score, PCA and pair-wise correlation criteria thus was excluded from downstream analysis. Acceptable samples within all four candidate GSE datasets were analyzed for differential expression based on p-value, fold change and LSM thresholds (Table S3).

Bacterial Enriched Host Pathways

Differentially expressed gene lists were used to calculate a pathway enrichment p-value for MetaBase canonical pathways. There were 74 canonical pathways (top pathways) significant in at least one comparison from both P. aeruginosa and S. pneumoniae (BC=2). Two pathways annotated as mouse cell occurrences share redundancy with human-related pathways: “Immune response_Oncostatin M signaling via MAPK in mouse cells” and “Oncostatin M signaling via MAPK in mouse cells” (Table S1). Table 2 lists the 72 top human pathways sorted by the sum of the pathway’s NBE, and whether the pathway was also significant in our previous respiratory virus study [16]. Of the 72 top human pathways, 16 are significant in all comparisons surveyed (i.e., the sum of NBE=2, Table S1). Overall, 29 pathways are classified by Metabase as “Immune response” and four pathways involved bacterial infections or Cystic fibrosis (“Bacterial infections in normal airways”, “Bacterial infections in CF airways”, “Cytokine production by Th17 cells in CF”, and “Mucin expression in CF via IL-6, IL-17”).

Pathway1Major ProcessSum NBECommon with Respiratory Viruses2
Bacterial infections in CF airwaysCF pathways2.00
FGF2-dependent induction of EMTDevelopment2.00
PEDF signalingDevelopment2.00
VEGF signaling via VEGFR2 - generic cascadesDevelopment2.00
Bacterial infections in normal airwaysImmune response2.00
CD40 signalingImmune response2.00X
Gastrin in inflammatory responseImmune response2.00X
HMGB1 release from the cellImmune response2.00
IL-1 signaling pathwayImmune response2.00X
IL-17 signaling pathwaysImmune response2.00X
IL-2 activation and signaling pathwayImmune response2.00X
MIF in innate immunity responseImmune response2.00X
Oncostatin M signaling via MAPK in human cellsImmune response2.00
TLR signaling pathwaysImmune response2.00
TREM1 signaling pathwayImmune response2.00X
Mucin expression in CF via TLRs, EGFR signaling pathwaysCF pathways2.00X
Endoplasmic reticulum stress response pathwayApoptosis and survival1.67X
Chemokines and adhesionCell adhesion1.67X
Cytokine production by Th17 cells in CFCF pathways1.67X
EGFR signaling pathwayDevelopment1.67X
ERBB-family signalingDevelopment1.67
Glucocorticoid receptor signalingDevelopment1.67X
GM-CSF signalingDevelopment1.67X
HGF-dependent inhibition of TGF-beta-induced EMTDevelopment1.67
PDGF signaling via STATs and NF-kBDevelopment1.67X
Regulation of epithelial-to-mesenchymal transition (EMT)Development1.67X
Role of IL-8 in angiogenesisDevelopment1.67
TGF-beta receptor signalingDevelopment1.67
VEGF signaling and activationDevelopment1.67
ATM/ATR regulation of G1/S checkpointDNA damage1.67X
IFN gamma signaling pathwayImmune response1.67X
HMGB1/RAGE signaling pathwayImmune response1.67
HMGB1/TLR signaling pathwayImmune response1.67
IL-6 signaling pathwayImmune response1.67X
Inhibitory action of Lipoxins on pro-inflammatory TNF-alpha signalingImmune response1.67X
MIF-mediated glucocorticoid regulationImmune response1.67X
Signaling pathway mediated by IL-6 and IL-1Immune response1.67X
Mucin expression in CF via IL-6, IL-17 signaling pathwaysCF pathways1.67X
NGF activation of NF-kBApoptosis and survival1.67
Putative SUMO-1 pathwayProteolysis1.67X
Gastrin in cell growth and proliferationDevelopment1.50
TGF-beta-dependent induction of EMT via MAPKDevelopment1.50
Human NKG2D signalingImmune response1.50
IL-15 signalingImmune response1.50X
IL-3 activation and signaling pathwayImmune response1.50X
Murine NKG2D signalingImmune response1.50X
GnRH signalingReproduction1.50X
Role of AP-1 in regulation of cellular metabolismTranscription1.50X
Transcription regulation of amino acid metabolismTranscription1.50
Inhibition of ROS-induced apoptosis by 17beta-estradiolApoptosis and survival1.33
CD28 signalingImmune response1.33
CXCR4 signaling via second messengerImmune response1.33
ICOS pathway in T-helper cellImmune response1.33
T cell receptor signaling pathwayImmune response1.33
Th1 and Th2 cell differentiationImmune response1.33
Role of IAP-proteins in apoptosisApoptosis and survival1.17
ECM remodelingCell adhesion1.17
Cytoskeleton remodelingCytoskeleton remodeling1.17X
FGFR signaling pathwayDevelopment1.17
Growth hormone signaling via PI3K/AKT and MAPK cascadesDevelopment1.17X
Keratinocyte differentiationDevelopment1.17
TGF-beta-dependent induction of EMT via SMADsDevelopment1.17
VEGF-family signalingDevelopment1.17
WNT signaling pathway. Part 2Development1.17X
IL-5 signallingImmune response1.17X
Regulation of eNOS activity in endothelial cellsMuscle contraction1.17X
JNK pathwaySignal transduction1.17
P53 signaling pathwayTranscription1.17X
HIV-1 signaling via CCR5 in macrophages and T lymphocytesImmune response to virus0.83X
NFAT in immune responseImmune response0.83
Role of alpha-6/beta-4 integrins in carcinoma progressionCancer0.83
Ligand-dependent activation of the ESR1/SP pathwayTranscription0.83

Table 2. Significant human pathways affected by respiratory bacteria.

Download CSV

The SUMO-1 pathway was among the list of top pathways not usually associated with canonical innate immune responses. The map for the SUMO-1 pathway with annotation of the network objects’ differential expression with either P. aeruginosa or S. pneumoniae is shown in Figure 2. There are four genes (network objects) that are differentially expressed in both species: CBX4 (Pc2), JUN (c-Jun), NFKBIA (NF-kappa-B1, alpha) and NFKB1 (NF-kappa-B1). All differentially expressed genes except UBA2 (SAE1/2) and Ubiquitin are regulated by SUMO-1 (Table S4 maps gene symbols to network object names).

Figure 2. SUMO-1 Pathway Map.

Black and red circles indicate genes significant in at least one study from Pseudomonas aeruginosa and Streptococcus pneumoniae, respectively. See text for description of pathway.

Potential Drug Repurposing for Respiratory Infections

There were 338 non-redundant differentially expressed genes from the top 72 human pathways. Of these genes, 69 (20%) are differentially expressed in both bacterial species (BC=2) (Table S5). We performed a query on the DrugBank database to identify potential drug candidates for these 69 genes (Table S6). A total of eleven genes were associated with at least one approved drug: JUN, FOS, ICAM1, IFNGR1, IL1Β, ITGB3, PLAUR, PTGS2, SERPINE1, TGFB1, and VEGFA. Seven of these genes associated with at least one anticoagulant, thrombosis, or anti-sepsis therapeutic (Table 3).

Gene symbol1Number of Drugs2Blood-related drug3Indications
PTGS246Acetylsalicylic acidThrombolytics
SERPINE16AnistreplaseThrombolytics; Anticoagulant
Drotrecogin alfaAnti-sepsis

Table 3. Drugs associated with human gene targets commonly associated withStreptococcus pneumoniae and Pseudomonas aeruginosa infections.

Download CSV

We compared the 69 most common pathway genes from the top bacterial pathways with the 178 most common top pathway genes from our previous study on respiratory viruses [16]. There were 38 genes common in at least 50% of the respiratory viral comparisons with a viral count (VC) of five or greater, corresponding to a hypogeometric test p-value of 4.0 x 10-18 (Table S7; Figure 3A). The genes that are differentially expressed in all viruses and bacteria (VC=7; BC=2) are: ATF3, CEBPD, CXCL1, CXCL2, IL8, JUN, NFKBIA, PPP1R15A, STAT1 and TNFAIP3. Comparison of the 67 top human-virus infection pathways with the 72 top human-bacteria pathways revealed a non-significant intersection of 36 common pathways (Table 2; Figure 3B; Table S8).

Figure 3. Overlap between human host gene and pathways identified for respiratory viruses and bacteria infections.

Venn diagrams of top a) pathway genes and b) pathways from previous viral [16] and present bacterial study. Genes have bacterial count (BC) =2 and viral count (VC) ≥5. Pathways have BC=2 and VC=7. Gene overlap is expected to have 4x10-18 probability of occurring by chance under a hypogeometric distribution.


The Host-Bacteria Interactome

In our study, many human host pathways determined to be significantly associated with infections from either S. pneumoniae or P. aeruginosa are also known to be involved in innate immunity which adds support to our analytical approach. For instance, interleukin -1 or IL-1, a significant pathway in all analyzed treatment groups, is important in facilitating CXC chemokine expression and neutrophil recruitment in macrophages and epithelial cells during pneumococcal pneumonia [26]. IL-1Β, a downstream effector of IL-1, is expressed in peripheral blood mononuclear cells (PBMCs) from cystic fibrosis (CF) patients infected by P. aeruginosa [27] via stimulus from the bacterial rhsT virulence gene [28], IL-1Β also plays a crucial role in inflammasome activation [29].

Other interleukin family member pathways we identified include IL-17C which is induced in human bronchial epithelial cells as a consequence of P. aeruginosa and H. influenzae infection [30]. IL-2 is also expressed during bacterial infection and has been suggested as a potential therapeutic target against Mycobacterium tuberculosis (TB) infections [31].

Toll-Like Receptor (TLR) pathways were also significant for S. pneumoniae and P. aeruginosa datasets. TLRs play a key role in activating the innate immune response to bacterial infections via their specific recognition of bacterial and viral nucleic acids as well as bacterial lipopolysaccharide (LPS) and peptidoglycan (reviewed in 31). TLR4 recognizes LPS produced by Gram-negative bacteria while TLR2 is highly responsive to Gram-positive bacteria components. Previous studies have shown TLR4 is over-expressed in response to P. aeruginosa exposure, along with VEGF and IL-23 [32], while TLR2 activation occurs in the presence of S. pneumoniae, S. aureus, E. coli and N. meningitidis [33]. The development of agonists and anti-agonists for TLRs is an intense area of drug development. For example, Stimforte, a TLR4 activator is approved for the treatment of secondary immuno-deficient states caused by chronic bacterial and viral infections. Other TLRs are being investigated as potential targets for sepsis, cancer and multiple immuno-inflammatory diseases.

Several other inflammatory response pathways were also significant in our analysis. The gene TREM1 (Triggering Receptor Expressed on Myeloid cells 1) helps stimulate immune response factors and has been proposed as a biomarker for P. aeruginosa ventilator associated pneumonia (VAP) [34]. VEGF is important in the host defense response against P. aeruginosa [32,35]. VEGF expression is also increased after exposure to S. pneumoniae in the mouse lung along with IL-8, MCP-1, and TNF-α [36] leading to increased vascular permeability [37]. CD40 has been implicated in both P. aeruginosa [38] and S. pneumoniae [39] infection, while PEDF (SERPINF1) is known to inhibit angiogenesis and acts as a neurotrophic factor in neuronal differentiation. This pathway indirectly induces pro-inflammatory genes including IL1Β, IL6, TNF, CCL3 and CCL23 [40,41].

Macrophage Migration Inhibitory Factor (MIF) has been identified as a risk factor for corneal P. aeruginosa infection [42] as well as increased susceptibility to Gram-negative sepsis [43]. A single nucleotide polymorphism (SNP) in the MIF gene is also associated with susceptibility to TB [44]. Other significant pathways from our analysis include Oncostatin M which is secondarily engaged in the innate immune response by cytokines and chemokines [45]. High mobility group box 1 (HMGB1) was recently shown to be elevated in mouse lung exposed to P. aeruginosa, causing hypoxia-induced innate immune response impairment [46]. Gastrin-releasing peptide receptor (GRPR) is linked to TLR4 signaling and antagonistic blockade of GRPR leads to limited protection from lethal sepsis [47]. The EMT (Epithelial to Mesenchymal Transition) pathway was also significant in our analysis adding to recent speculation that viruses and microbes induce EMT via growth and innate immunity signaling pathways [48].

Of the 36 pathways that overlap between respiratory bacteria and viruses, 16 are classified as “immune response” signaling pathways [16]. However, there are additional pathways not so readily associated with immune response. For example, the ER stress response to misfolded proteins has been implicated in chronic disease and autoimmune disease following inflammation due to microbial, viral, and other infections via TLR, NOD and inflammasome signaling [49]. ER stress response has also been linked with several neurodegenerative diseases such as amyotrophic lateral sclerosis, Parkinson’s disease, Huntington’s disease and Alzheimer’s disease [49]. Other common bacterial and viral pathways we found that are not typically associated with innate immunity include DNA damage ATM/ATR regulation of G1/S checkpoint pathway, development WNT signaling pathway and the SUMO-1 pathway.

SUMO1 Role in Host Response

Interestingly, we identified the small ubiquitin-like modifier (SUMO1) pathway as a component of the host-bacteria interactome. SUMO1 is one of four SUMO homologs found in mammals which through a process known as sumoylation, covalently links its protein targets to the E1, E2 and E3-conjugating enzyme cascade for further processing [50]. A key regulatory mechanism for many cellular processes, SUMO1 itself is activated by SUMO1/sentrin specific peptidase 1 (SENP1).

Certain pathogens disrupt the SUMO1 pathway in order to evade host defenses [51]. Cysteine proteases encoded by adenoviruses target host proteins including SUMO1, while herpes viruses indirectly modulate the SUMO pathway by targeting downstream sumoylated promyelocytic leukemia proteins [51]. Sumoylation of human cytomegalovirus polymerase subunit UL44 is associated with increased viral replication [52]. Recent studies revealed interplay between SUMO and DNA viruses while bacterial interaction with the SUMO pathway is less understood. Some evidence suggests that a number of other pathogens, such as the intra-cellular bacteria Yersinia pestis and Listeria monocytogenes modulate this pathway during infection [51].

Previously, we identified the Parkin Ubiquitin Proteasomal System (Parkin-UPS) pathway as a potential factor in respiratory viral infection [16]. In humans, SNPs associated with the gene PARK2, encoding for Parkin, are associated with increased susceptibility to certain infectious diseases such as leprosy, typhoid and paratyphoid fever [53,54]. Interestingly, the SUMO pathway also plays a role in the regulation of the PARK2 gene [55]. Collectively, these findings suggest potential linkages of the SUMO and Parkin-UPS pathways with viral and bacterial pathogenesis.

Drug Repurposing

Interestingly, 12 out of the 72 approved drugs from our DrugBank list are associated with coagulation or sepsis: five anticoagulants/antithrombotics, six thrombolytics and one antisepsis agent (Table 3). These drugs have a variety of targets found in our analysis, namely FOS, ITGB3, PTGS2, PLAUR, SERPINE1, and VEGFA. Most of these genes tend to be up-regulated during infection. An exception is ITGB3, which is down-regulated in P. aeruginosa yet up-regulated in S. pneumoniae models (Table S3).

Coagulation has a role in both innate immune response [56,57] and bacterial evasion of host defense [58]. Bacterial pathogens such as Y. pestis and Group A Streptococci can evade fibrin networks by activating host plasminogen suggesting that these pathogens alter the immune-coagulation response [59,60]. Moreover, patients with acute infections have a higher risk of thrombosis-related conditions, including pulmonary emboli, venous thrombosis and myocardial infarction [60,61]. Infection also often leads to a variety of coagulation processes via activation of the intrinsic pathway, CD40, and TLRs, as reviewed in 60. In the present study, both CD40- and TLR- related pathways were found to be significant (Table S1), further suggesting a potential interplay between coagulation and host immune response to infection.

We suggest ITGB3 (TGB3 integrin, beta 3: platelet glycoprotein IIIa, antigen CD61) and ICAM1 (intercellular adhesion molecule 1) as additional drug target candidates because they were differentially expressed in the top pathways in at least one dataset for both bacterial species. Integrins are a family of heterodimeric receptors (α & β subunits) that mediate cell adhesion, cell-signaling and migration [62] which pathogens are known to exploit for cellular adhesion and invasion (i.e. [63,64]).

Several integrin antagonists are either launched drugs or in late stage development. The drugs Tysabri and Firategrast block α4 integrin-mediated adhesion and invasion of white cells while Abciximab, Eptifibatide and Aggrastat are αIIbβ3 receptor (Glycoprotein receptor or GPR IIb/IIIa) antagonists thus function as anti-coagulants by blocking platelet aggregation. A recent study found that adding Abciximab to Streptococcus gordonii infected cells caused S. gordonii and platelets dissociation [65]. Similarly, Eptifibatide was found to inhibit platelet granzyme B-mediated apoptosis in sepsis models [66]. Clinically, Tirofiban was shown to significantly decrease anti-streptokinase antibody-mediated platelet aggregation in streptococcus-infected patients [67]. These findings suggest that GPIIb/IIIa inhibitors such as Eptifibatide, Abciximab, and Tirofiban may have potential applications against sepsis, S. gordonii infections and other streptococcal infections, respectively. We suggest further investigations into their utility as pan species anti-bacterial therapeutics.

Arrest and firm adhesion of leukocytes to the endothelium is dependent on the activation of ß2 family integrins. ICAM1 is a ligand of a number of integrins such as LFA-1 and facilitates the transmigration of leukocytes across vascular endothelia [68]. We suggest a likely role of ICAM1 in bacterial infections as well. ICAM1 deficient mice show resistance to septic shock as induced by bacterial toxins [69]. ICAM1 is up-regulated and used as a site of entry by several viruses, including human rhinovirus (HRV) [70], West Nile virus [71] and RSV [72]. ICAM1 mAbs have been shown to be effective in blocking HRV infection in cellular systems [73]. Our previous analysis also showed ICAM1 up-regulation by five respiratory viruses [16]. Alicaforsen, an ICAM1 anti-sense siRNA, is currently in clinical trials for the treatment of inflammatory bowel disease [74]. We hypothesize that ICAM1 antagonists could also be repurposed against viral and bacterial infections.

In this study, we have shown the potential for computational approaches to identify host genes and pathways important for respiratory bacterial and viral infections. Identification of potential drug repurposing opportunities provides a path towards clinical application of these results. Our study highlights the importance of human host response to bacterial infections and the need for further well-designed preclinical and clinical studies involving a wide range of pathogens.

Supporting Information

Table S1.

Statistical analysis of pathway enrichment and gene contents.


Table S2.

Rejected GEO datasets and reasons for exclusive from analysis.


Table S3.

QC and statistical analysis of gene expression.


Table S4.

Gene symbol to network object mapping table for SUMO-1.


Table S5.

Differentially expressed genes from top pathways (BC=2).


Table S6.

DrugBank results of targets with BC=2.


Table S7.

Viral and bacterial count of gene by pathogen (Intersection with viral study).


Table S8.

Significant pathways shared between the bacterial and viral studies.



We acknowledge GlaxoSmithKline R&D for their support of this study. We also thank Jacques Ravel, University of Maryland, for his support and encouragement.

Author Contributions

Conceived and designed the experiments: SS MM-S JRB. Performed the experiments: SS. Analyzed the data: SS MM-S. Contributed reagents/materials/analysis tools: SS JRB MM-S. Wrote the manuscript: SS JRB MM-S.


  1. 1. Restrepo MI, Faverio P, Anzueto A (2013) Long-term prognosis in community-acquired pneumonia. Curr Opin Infect Dis 26: 151-158. doi: PubMed: 23426328.
  2. 2. Viasus D, Garcia-Vidal C, Carratalà J (2013) Advances in antibiotic therapy for community-acquired pneumonia. Curr Opin Pulm Med 19: 209-215. doi: PubMed: 23422416.
  3. 3. Pendleton JN, Gorman SP, Gilmore BF (2013) Clinical relevance of the ESKAPE pathogens. Expert Rev Anti Infect Ther 11: 297-308. doi: PubMed: 23458769.
  4. 4. Jenkins SG, Brown SD, Farrell DJ (2008) Trends in antibacterial resistance among Streptococcus pneumoniae isolated in the USA: update from PROTEKT US Years 1-4. Ann Clin Microbiol Antimicrob 7: 1. doi: PubMed: 18190701.
  5. 5. Payne DJ, Gwynn MN, Holmes DJ, Pompliano DL (2007) Drugs for bad bugs: confronting the challenges of antibacterial discovery. Nat Rev Drug Discov 6: 29-40. doi: PubMed: 17159923.
  6. 6. Hajishengallis G, Lambris JD (2011) Microbial manipulation of receptor crosstalk in innate immunity. Nat Rev Immunol 11: 187-200. doi: PubMed: 21350579.
  7. 7. Ishii KJ, Koyama S, Nakagawa A, Coban C, Akira S (2008) Host innate immune receptors and beyond: making sense of microbial infections. Cell Host Microbe 3: 352-363. doi: PubMed: 18541212.
  8. 8. Hoogendijk AJ, Diks SH, van der Poll T, Peppelenbosch MP, Wieland CW (2011) Kinase activity profiling of pneumococcal pneumonia. PLOS ONE 6: e18519. doi: PubMed: 21483672.
  9. 9. Fowler T, Johansson S, Wary KK, Höök M (2003) Src kinase has a central role in in vitro cellular internalization of Staphylococcus aureus. Cell Microbiol 5: 417-426. doi: PubMed: 12780779.
  10. 10. Shin S, Paul-Satyaseela M, Lee JS, Romer LH, Kim KS (2006) Focal adhesion kinase is involved in type III group B streptococcal invasion of human brain microvascular endothelial cells. Microb Pathog 41: 168-173. doi: PubMed: 16949788.
  11. 11. Weber SS, Ragaz C, Hilbi H (2009) Pathogen trafficking pathways and host phosphoinositide metabolism. Mol Microbiol 71: 1341-1352. doi: PubMed: 19208094.
  12. 12. Kierbel A, Gassama-Diagne A, Mostov K, Engel JN (2005) The phosphoinositol-3-kinase-protein kinase B/Akt pathway is critical for Pseudomonas aeruginosa strain PAK internalization. Mol Cell Biol 16: 2577-2585. doi: PubMed: 15772151.
  13. 13. Burnham CA, Shokoples SE, Tyrrell GJ (2007) Invasion of HeLa cells by group B streptococcus requires the phosphoinositide-3-kinase signalling pathway and modulates phosphorylation of host-cell Akt and glycogen synthase kinase-3. Microbiology 153: 4240-4252. doi: PubMed: 18048937.
  14. 14. Kline KA, Fälker S, Dahlberg S, Normark S, Henriques-Normark B (2009) Bacterial adhesins in host-microbe interactions. Cell Host Microbe 5: 580-592. doi: PubMed: 19527885.
  15. 15. Brown JR, Magid-Slav M, Sanseau P, Rajpal DK (2011) Computational biology approaches for selecting host-pathogen drug targets. Drug Discov Today 16: 229-236. doi: PubMed: 21277381.
  16. 16. Smith SB, Dampier W, Tozeren A, Brown JR, Magid-Slav M (2012) Identification of common biological pathways and drug targets across multiple respiratory viruses based on human host gene expression analysis. PLOS ONE 7: e33174. doi: PubMed: 22432004.
  17. 17. Mayer AK, Muehmer M, Mages J, Gueinzius K, Hess C et al. (2007) Differential recognition of TLR-dependent microbial ligands in human bronchial epithelial cells. J Immunol 178: 3134-3142. PubMed: 17312161.
  18. 18. Bootsma HJ, Egmont-Petersen M, Hermans PW (2007) Analysis of the in vitro transcriptional response of human pharyngeal epithelial cells to adherent Streptococcus pneumoniae: evidence for a distinct response to encapsulated strains. Infect Immun 75: 5489-5499. doi: PubMed: 17709418.
  19. 19. Ichikawa JK, English SB, Wolfgang MC, Jackson R, Butte AJ, Lory S (2005) Genome-wide analysis of host responses to the Pseudomonas aeruginosa type III secretion system yields synergistic effects. Cell Microbiol 7: 1635-1646. doi: PubMed: 16207250.
  20. 20. Ramilo O, Allman W, Chung W, Mejias A, Ardura M, Glaser C et al. (2007) Gene expression patterns in blood leukocytes discriminate patients with acute infections. Blood 109: 2066-2077. doi: PubMed: 17105821.
  21. 21. Bhat N, Wright JG, Broder KR, Murray EL, Greenberg ME et al. (2005) Influenza-associated deaths among children in the United States, 2003-2004. N Engl J Med 353: 2559-2567. doi: PubMed: 16354892.
  22. 22. Hotchkiss RS, Monneret G, Payen D (2013) Immunosuppression in sepsis: a novel understanding of the disorder and a new therapeutic approach. Lancet Infect Dis 13: 260-268. doi: PubMed: 23427891.
  23. 23. Omicsoft (2011) Array Studio, version [computer program]. NC, USA: Research Triangle Park.
  24. 24. Fisher RA (1922) On the Interpretation of X2 from Contingency Tables, and the Calculation of P. J R Stat Soc 85: 87-94. doi:
  25. 25. Core R Team (2012) R: A language and environment for statistical computing. Version Vienna, Austria.
  26. 26. Marriott HM, Gascoyne KA, Gowda R, Geary I, Nicklin MJ et al. (2012) Interleukin-1beta regulates CXCL8 release and influences disease outcome in response to Streptococcus pneumoniae, defining intercellular cooperation between pulmonary epithelial cells and macrophages. Infect Immun 80: 1140-1149. doi: PubMed: 22158745.
  27. 27. Tang A, Sharma A, Jen R, Hirschfeld AF, Chilvers MA et al. (2012) Inflammasome-mediated IL-1beta production in humans with cystic fibrosis. PLOS ONE 7: e37689. doi: PubMed: 22649552.
  28. 28. Kung VL, Khare S, Stehlik C, Bacon EM, Hughes AJ, Hauser AR (2012) An rhs gene of Pseudomonas aeruginosa encodes a virulence protein that activates the inflammasome. Proc Natl Acad Sci U S A 109: 1275-1280. doi: PubMed: 22232685.
  29. 29. Menu P, Vince JE (2011) The NLRP3 inflammasome in health and disease: the good, the bad and the ugly. Clin Exp Immunol 166: 1-15. doi: PubMed: 21762124.
  30. 30. Pfeifer P, Voss M, Wonnenberg B, Hellberg J, Seiler F, Lepper PM, Bischoff M, Langer F, Schäfers HJ, Menger MD, Bals R, Beisswenger C (2013) IL-17C is a mediator of respiratory epithelial innate immune response. Am J Respir Cell Mol Biol 48: 415-421. doi: PubMed: 23221046.
  31. 31. Zhang Y, Liu J, Wang Y, Xian Q, Shao L et al. (2012) Immunotherapy using IL-2 and GM-CSF is a potential treatment for multidrug-resistant Mycobacterium tuberculosis. Sci China Life Sci 55: 800-806. doi: PubMed: 23015129.
  32. 32. Paolillo R, Romano CC, Sorrentino S, Mazzola N, Mita L et al. (2011) Expression of IL-23, VEGF and TLR2/TLR4 on mononuclear cells after exposure to Pseudomonas aeruginosa. Int J Immunopathol Pharmacol 24: 961-973. PubMed: 22230402.
  33. 33. Beran O, Potmesil R, Holub M (2011) Differences in Toll-like receptor expression and cytokine production after stimulation with heat-killed gram-positive and gram-negative bacteria. Folia Microbiol (Praha) 56: 138-142. doi: PubMed: 2167426521399944.
  34. 34. Siranović M, Kovac J, Gopcević S, Kelecić M, Kovac N et al. (2011) Human soluble TREM-1: lung and serum levels in patients with bacterial ventilator associated pneumonia. Acta Clin Croat 50: 345-349. PubMed: 22384768.
  35. 35. Breen EC, Malloy JL, Tang K, Xia F, Fu Z et al. (2013) Impaired pulmonary defense against Pseudomonas aeruginosa in VEGF gene inactivated mouse lung. J Cell Physiol 228: 371-379. doi: PubMed: 22718316.
  36. 36. Wilkosz S, Edwards LA, Bielsa S, Hyams C, Taylor A et al. (2012) Characterization of a new mouse model of empyema and the mechanisms of pleural invasion by Streptococcus pneumoniae. Am J Respir Cell Mol Biol 46: 180-187. doi: PubMed: 21885676.
  37. 37. van Der Flier M, Coenjaerts F, Kimpen JL, Hoepelman AM, Geelen SP et al. (2000) Streptococcus pneumoniae induces secretion of vascular endothelial growth factor by human neutrophils. Infect Immun 68: 4792-4794. doi: PubMed: 10899891.
  38. 38. Worgall S, Martushova K, Busch A, Lande L, Crystal RG (2002) Apoptosis induced by Pseudomonas aeruginosa in antigen presenting cells is diminished by genetic modification with CD40 ligand. Pediatr Res 52: 636-644. doi: PubMed: 12409507.
  39. 39. Moens L, Wuyts G, Boon L, den Hartog MT, Ceuppens JL et al. (2008) The human polysaccharide- and protein-specific immune response to Streptococcus pneumoniae is dependent on CD4(+) T lymphocytes, CD14(+) monocytes, and the CD40-CD40 ligand interaction. J Allergy Clin Immunol 122: 1231-1233. doi: PubMed: 18842293.
  40. 40. Sanagi T, Yabe T, Yamada H (2005) The regulation of pro-inflammatory gene expression induced by pigment epithelium-derived factor in rat cultured microglial cells. Neurosci Lett 380: 105-110. doi: PubMed: 15854760.
  41. 41. Yabe T, Sanagi T, Schwartz JP, Yamada H (2005) Pigment epithelium-derived factor induces pro-inflammatory genes in neonatal astrocytes through activation of NF-kappa B and CREB. Glia 50: 223-234. doi: PubMed: 15739190.
  42. 42. Reidy T, Rittenberg A, Dwyer M, D’Ortona S, Pier G et al. (2013) Homotrimeric macrophage migration inhibitory factor (MIF) drives inflammatory responses in the corneal epithelium by promoting caveolin-rich platform assembly in response to infection. J Biol Chem 288: 8269-8278. doi: PubMed: 23372160.
  43. 43. Roger T, Delaloye J, Chanson AL, Giddey M, Le RD et al. (2013) Macrophage migration inhibitory factor deficiency is associated with impaired killing of gram-negative bacteria by macrophages and increased susceptibility to Klebsiella pneumoniae sepsis. J Infect Dis 207: 331-339. doi: PubMed: 23125447.
  44. 44. Li Y, Yuan T, Lu W, Chen M, Cheng X et al. (2012) Association of tuberculosis and polymorphisms in the promoter region of macrophage migration inhibitory factor (MIF) in a Southwestern China Han population. Cytokine 60: 64-67. doi: PubMed: 22742858.
  45. 45. Eberl M, Roberts GW, Meuter S, Williams JD, Topley N et al. (2009) A rapid crosstalk of human gammadelta T cells and monocytes drives the acute inflammation in bacterial infections. PLOS Pathog 5: e1000308.
  46. 46. Patel VS, Sitapara RA, Gore A, Phan B, Sharma L et al. (2013) High Mobility Group Box-1 mediates hyperoxia-induced impairment of Pseudomonas aeruginosa clearance and inflammatory lung injury in mice. Am J Respir Cell Mol Biol 48: 280-287. doi: PubMed: 23087050.
  47. 47. Petronilho F, Vuolo F, Galant LS, Constantino L, Tomasi CD et al. (2012) Gastrin-releasing peptide receptor antagonism induces protection from lethal sepsis: involvement of toll-like receptor 4 signaling. Mol Med 18: 1209-1219. PubMed: 22735756.
  48. 48. Hofman P, Vouret-Craviari V (2012) Microbes-induced EMT at the crossroad of inflammation and cancer. Gut Microbes 3: 176-185. doi: PubMed: 22572828.
  49. 49. Hasnain SZ, Lourie R, Das I, Chen AC, McGuckin MA (2012) The interplay between endoplasmic reticulum stress and inflammation. Immunol Cell Biol 90: 260-270. doi: PubMed: 22249202.
  50. 50. Gareau JR, Lima CD (2010) The SUMO pathway: emerging mechanisms that shape specificity, conjugation and recognition. Nat Rev Mol Cell Biol 11: 861-871. doi: PubMed: 21102611.
  51. 51. Wimmer P, Schreiner S, Dobner T (2012) Human pathogens and the host cell SUMOylation system. J Virol 86: 642-654. doi: PubMed: 22072786.
  52. 52. Sinigalia E, Alvisi G, Segré CV, Mercorelli B, Muratore G et al. (2012) The human cytomegalovirus DNA polymerase processivity factor UL44 is modified by SUMO in a DNA-dependent manner. PLOS ONE 7: e49630. doi: PubMed: 23166733.
  53. 53. Ali S, Vollaard AM, Widjaja S, Surjadi C, van de Vosse E et al. (2006) PARK2/PACRG polymorphisms and susceptibility to typhoid and paratyphoid fever. Clin Exp Immunol 144: 425-431. doi: PubMed: 16734611.
  54. 54. Mira MT, Alcaïs A, Nguyen VT, Moraes MO, Di FC et al. (2004) Susceptibility to leprosy is associated with PARK2 and PACRG. Nature 427: 636-640. doi: PubMed: 14737177.
  55. 55. Um JW, Chung KC (2006) Functional modulation of parkin through physical interaction with SUMO-1. J Neurosci Res 84: 1543-1554. doi: PubMed: 16955485.
  56. 56. Wang Z, Wilhelmsson C, Hyrsl P, Loof TG, Dobes P et al. (2010) Pathogen entrapment by transglutaminase--a conserved early innate immune mechanism. PLOS Pathog 6.
  57. 57. McAdow M, Missiakas DM, Schneewind O (2012) Staphylococcus aureus secretes coagulase and von Willebrand factor binding protein to modify the coagulation cascade and establish host infections. J Innate Immun 4: 141-148. doi: PubMed: 22222316.
  58. 58. Loeb L (1903) The influence of certain bacteria on the coagulation of the blood. J Med Res 10: 407-419. PubMed: 19971581.
  59. 59. Sun H, Ringdahl U, Homeister JW, Fay WP, Engleberg NC et al. (2004) Plasminogen is a critical host pathogenicity factor for group A streptococcal infection. Science 305: 1283-1286. doi: PubMed: 15333838.
  60. 60. Esmon CT (2013) Molecular circuits in thrombosis and inflammation. Thromb Haemost 109: 416-420. doi: PubMed: 23407677.
  61. 61. Esmon CT, Xu J, Lupu F (2011) Innate immunity and coagulation. J Thromb Haemost 9 Suppl 1: 182-188. doi: PubMed: 21781254.
  62. 62. Wolfenson H, Lavelin I, Geiger B (2013) Dynamic regulation of the structure and functions of integrin adhesions. Dev Cell 24: 447-458. doi: PubMed: 23484852.
  63. 63. Scibelli A, Roperto S, Manna L, Pavone LM, Tafuri S et al. (2007) Engagement of integrins as a cellular route of invasion by bacterial pathogens. Vet J 173: 482-491. doi: PubMed: 16546423.
  64. 64. Eto DS, Jones TA, Sundsbak JL, Mulvey MA (2007) Integrin-mediated host cell invasion by type 1-piliated uropathogenic Escherichia coli. PLOS Pathog 3: e100. doi: PubMed: 17630833.
  65. 65. Petersen HJ, Keane C, Jenkinson HF, Vickerman MM, Jesionowski A et al. (2010) Human platelets recognize a novel surface protein, PadA, on Streptococcus gordonii through a unique interaction involving fibrinogen receptor GPIIbIIIa. Infect Immun 78: 413-422. doi: PubMed: 19884334.
  66. 66. Sharron M, Hoptay CE, Wiles AA, Garvin LM, Geha M et al. (2012) Platelets induce apoptosis during sepsis in a contact-dependent manner that is inhibited by GPIIb/IIIa blockade. PLOS ONE 7: e41549. doi: PubMed: 22844498.
  67. 67. Courval M, Palisaitis DA, Diodati JG, Lesperance B, Pharand C (2004) Inhibition of streptokinase-induced, antibody-mediated platelet aggregation with tirofiban after exposure to streptokinase or streptococcal infection. Pharmacotherapy 24: 558-563. doi: PubMed: 15162889.
  68. 68. Reglero-Real N, Marcos-Ramiro B, Millán J (2012) Endothelial membrane reorganization during leukocyte extravasation. Cell Mol Life Sci 69: 3079-3099. doi: PubMed: 22573182.
  69. 69. Xu H, Gonzalo JA, St Pierre Y, Williams IR, Kupper TS et al. (1994) Leukocytosis and resistance to septic shock in intercellular adhesion molecule 1-deficient mice. J Exp Med 180: 95-109. doi: PubMed: 7911822.
  70. 70. Dai J, Wang P, Bai F, Town T, Fikrig E (2008) Icam-1 participates in the entry of west nile virus into the central nervous system. J Virol 82: 4164-4168. doi: PubMed: 18256150.
  71. 71. Xing L, Casasnovas JM, Cheng RH (2003) Structural analysis of human rhinovirus complexed with ICAM-1 reveals the dynamics of receptor-mediated virus uncoating. J Virol 77: 6101-6107. doi: PubMed: 12743267.
  72. 72. Arnold R, König W (2005) Respiratory syncytial virus infection of human lung endothelial cells enhances selectively intercellular adhesion molecule-1 expression. J Immunol 174: 7359-7367. PubMed: 15905583.
  73. 73. Newton P, O’Shea D, Wells E, Moakes K, Dunmore R et al. (2013) Development of a homogeneous high-throughput screening assay for biological inhibitors of human rhinovirus infection. J Biomol Screen 18: 237-246. doi: PubMed: 23207740.
  74. 74. Ghosh N, Chaki R, Mandal SC (2012) Inhibition of selective adhesion molecules in treatment of inflammatory bowel disease. Int Rev Immunol 31: 410-427. doi: PubMed: 23083349.