People with COPD are more susceptible to viral infection and suffer severe complications with worsened symptoms and frequent exacerbations following infection. This study investigated the transcriptional response of pBECs to RV infection and how this is altered in COPD. We show that the immune response of healthy and COPD cells was characterised by a robust up-regulation of pro-inflammatory and antiviral pathways. However there were clear differences between the COPD and healthy pBECs, including up-regulation of inflammatory genes at baseline and dramatically exaggerated responses to RV infection. We identified 9 genes associated with COPD at baseline and 22 genes altered in COPD but not healthy pBECs in response to RV, not previously reported in COPD, but likely to be important in regulating the exaggerated virus induced inflammation. The increased gene expression in RV infected COPD pBECs also correlated with the corresponding protein release, and enhanced apoptosis; however this enhanced immune response did not reduce viral titre. Furthermore, IFN-β/λ1 pre-treatment resulted in enhanced responses to RV-1B in COPD and healthy pBECs, however COPD pBECs were unresponsive in terms of MDA-5/RIG-I and IFN-β gene induction. Despite this abnormality IFN-β/λ1 pre-treatment still led to significantly reduced viral titre.
The airway epithelium in COPD is exposed chronically to enhanced airway inflammation with increased numbers of neutrophils and lymphocytes that correlate with more severe airflow obstruction, despite the fact that the majority of subjects have ceased smoking, indicating that the airway inflammatory response in COPD becomes self-perpetuating [11, 12]. We found several novel associations with 9 up-regulated genes in COPD pBECs, including the calgranulins S100A8 and S100A9, and proteases ADAM19 and MMP10. S100A8/A9 are small calcium-binding proteins with pro-inflammatory activity and have been reported to increase with steroid resistant neutrophilic inflammation . ADAM19 and MMP10 are involved in tissue repair and remodelling, and single nucleotide polymorphisms in the gene locus containing ADAM19 has been associated with COPD . These observations are consistent with other studies suggesting that the airways of COPD subjects undergo constant cellular repair due to the damages caused by cigarette smoke exposure . Interestingly MMP10 and ADAM19 gene expression has also been shown to be upregulated in pBECs from subjects with asthma .
Upon RV-1B infection, both pBEC groups mounted a robust inflammatory response; however the response was more exaggerated in COPD pBECs. This was in accordance with other studies that showed a more vigorous inflammatory response in COPD pBECs against RV infection . We have identified a number of novel genes whose expression were strongly up-regulated in COPD pBECs in response to RV-1B but not in healthy pBECs. These include important signalling molecules downstream of IL-1 and TLR2, 3 and 4 pathways such as PELI1 , IRAK2  and CH25H . We have previously shown that PELI1 and IRAK2 are up-regulated in the sputum of subjects with neutrophilic asthma . This suggests that IRAK2 and PELI1 play a role promoting neutrophilic airway inflammation, which is triggered by RV infection of the epithelium in COPD. Recently, PELI1 has been shown to be important in regulating the innate immune response of the epithelium to RV, including CXCL-8 production and neutrophil recruitment, without interfering with IFN responses and viral replication . Mouse models have shown that IRAK2 is critical in sustaining late phase inflammatory responses after TLR stimulation, leading to increased production of inflammatory cytokines . These molecules may also be important in T cell related functions, such as T cell tolerance , and promotion of Th17 cell development .
Other important signalling proteins induced by RV in COPD pBECs include PMAIP1, ATF3 and GBP4. PMAIP1, a mitochondria-associated protein, promotes apoptosis . ATF3 senses oxidative stress  and functions to reduce TLR4-mediated NF-κB signalling, therefore serving as negative feedback . GBP4 is an IFN-inducible GTPase that has been shown to disrupt IRF7 activation, thereby inhibiting the induction of type I IFNs . Also potentially important is the up-regulation of Chemokine (C-X3-C motif) ligand 1 (CX3CL1), more commonly known as fractalkine, by RV-1B in COPD pBECs. Soluble CX3CL1 has a chemo-attractant activity for T cells and monocytes; whereas membrane bound protein promotes adhesion of these leukocytes .
Antiviral responses such as IFN-β and -λ1 are critical in limiting viral replication, and were significantly higher in COPD pBECs. This was associated with the higher level of apoptosis after infection, but surprisingly did not affect viral titre in COPD, which was similar to the healthy group. This was in contrast to the study by Schneider et al. that showed enhanced antiviral responses associated with increased replication of RV-39 after infection in COPD . Other studies have examined transcriptional responses to in vivo RV infection of healthy controls using nasal epithelial scrapings , or in vitro RV infection of healthy pBECs . There are many similarities between the changes in gene expression found in reported studies compared to the current study, which includes upregulation of ISGs, antiviral genes, chemokines, and cytokines. However there are also differences in the healthy pBEC responses, including some genes previously reported to be upregulated by RV (e.g. IFNB1, and IL6) that were unchanged in this study. These differences are likely explained by altered experimental conditions, such as differing strains of RV, culture or sampling methods including submerged culture versus air-liquid interface culture versus in vitro infection, time points of RNA sampling, microarray platforms and microarray analysis methods. Nevertheless, this well controlled study adds significant knowledge regarding the differences between healthy and COPD innate immune responses to RV-1B infection under the conditions investigated, which warrant further investigation.
The underlying mechanisms of unchanged viral replication despite induced antiviral responses in COPD pBECs after RV-1B infection is currently unclear. It is possible that the IFNs produced by infection did not efficiently initiate the subsequent inductions of ISGs in COPD pBECs, therefore leading to an unchanged viral titre. However this would not explain the high apoptosis induction in COPD pBECs by RV-1B infection and marked reduction of RV-1B viral replication following IFN pre-treatment. ISGs such as protein kinase R (PKR) have been shown to induce apoptosis via the Fas-associated death domain (FADD) , and induction of which by RV-1B infection in COPD pBECs was higher compared to that in healthy pBECs. Pre-treatment with IFNs also significantly reduced viral replication, further suggesting IFN signalling leading to ISG induction may be functional in COPD. Nevertheless, it is also possible that ISGs were ineffectively up-regulated in COPD pBECs, and apoptosis could also have been induced by other signalling pathways including TNF-α/TNFR1 pathway , and compensated for the reduced ISGs production. This may explain the excess tissue damage and high inflammation that can be caused by high levels of apoptosis in the airways of those with COPD .
While we showed a heightened inflammatory response and antiviral response to RV infection in COPD pBECs in this study, we also demonstrated defects in the antiviral pathway. IFN-β/λ1 pre-treatment led to increased IL-6 and TNF-α mRNA in healthy pBECs, however this was not translated to protein production which was either reduced or unchanged with IFN-β/λ1 pre-treatment. Previous studies have shown exogenous IFN-β decreases RV-1B-induced IL-6 release from healthy and asthmatic pBECs via an unidentified pathway . IFN-β/λ1 pre-treatment also enhanced antiviral responses including MDA-5 and RIG-I, and IFN-β mRNAs in healthy pBECs, leading to marked decrease in viral titre. However, in COPD pBECs IFNs pre-treatment failed to induce MDA-5 and RIG-I and IFN-β mRNA, but did enhance CCL-5, CXCL-10, and IFN-λ1 production, resulting in a decreased viral titre. This indicates that MDA-5-initiated antiviral responses were partially impaired in COPD and led to reduced IFN-β level. It is possible that IFN-β/λ1 pre-treatment significantly up-regulated ISGs such as PKR and MxA, that bound and degraded viral RNAs as RV endocytosed into the host cells. This also suggests differential signalling pathways that regulate type III IFNs other than MDA-5/RIG-I and IFN-β. Indeed, a recent study has identified a cluster of NF-κB binding sites on human IFN-λ1 promoter, and NF-κB was critical for IFN-λ1 induction but not for type I IFNs expression . This is consistent with our results as both NF-κB and IFN-λ1were significantly up-regulated in COPD pBECs.
Apoptosis is another important component of antiviral responses, which was significantly increased in COPD pBECs after RV infection. This may correlate with the increased TNF and IFN levels , which can also up-regulate PMAIP1 gene that promotes the induction of apoptosis . However increased apoptosis did not reduce viral titre. The reason for this observation is unclear; however, it is possible that the quantification methods (TCID50) used in this study may not be as sensitive as direct quantification methods such as plaque assays. Detection of RV-1B by qPCR could only measure the level of total viral RNAs and not differentiate live from dead viruses. Alternatively, high levels of oxidative stress could have contributed to this observation. Superoxide dismutase 2 (SOD2) and ATF3 are important anti-oxidative genes that were up-regulated to a greater extent in COPD pBECs after infection when compared with infected healthy pBECs. Lack of SOD2 in mice can lead to increased oxidative damage to DNA , and lack of ATF3 alters DNA repair mechanisms . This indirectly indicates that the level of oxidative stress is higher in RV infected COPD pBECs compared to that in healthy cells.