CCL7 and IRF-7 Mediate Hallmark Inflammatory and IFN Responses following Rhinovirus 1B Infection Free

Rhinovirus (RV) infections are the most common factor associated with exacerbations of asthma (1, 2). Exacerbations of asthma induced by viral infections are difficult to treat and are a major cause of asthma morbidity, often associated with a loss of disease control by provocation of airway hyperreactivity (AHR) and lung inflammation (3, 4). As such, much research has been directed toward identifying the underlying mechanisms of RV-induced asthma exacerbations with the hope of developing novel therapeutics.

Normally RV infections induce a rapid production of IFN-α and IFN-β along with the recruitment of neutrophils following detection of RV by epithelial cells and macrophages (5–7). Robust IFN responses are capable of limiting inflammation and reducing the duration, severity, and spread of viral infection (8). Many asthma sufferers generally exhibit elevated eosinophil and neutrophil recruitment upon RV infection and a diminished production of IFN-α and IFN-β. Therefore, it is important to investigate the mechanisms of IFN production and general, antiviral, and inflammatory responses. It is also of pivotal importance to study the chemokines released upon infection, due to their ability to induce leukocyte trafficking and their potential to control the intensity of an immune response (9).

In this study, we infected BALB/c mice with minor group RV strain 1B (RV1B) and conducted an array on RNA extracted from blunt dissected airway tissue. The most significantly upregulated genes following RV infection included CCL7 (also known as monocyte chemotactic protein-3) and IFN regulatory factor (IRF)-7. In this study, we show a functional role of CCL7 in the chemotaxis of neutrophils, macrophages, and eosinophils in RV-induced exacerbation of allergic airways disease (AAD) in vivo and an important role of IRF-7 in promoting type 1 IFN responses to RV infection.

All experiments were approved by the Animal Care and Ethics Committee of the University of Newcastle. Six- to 8-wk-old male BALB/c mice obtained from the specific pathogen-free facility of the University of Newcastle were housed within approved containment facilities in Bioresources at the Hunter Medical Research Institute (Newcastle, NSW, Australia) with ad libitum access to food and water under a 12 h light/dark cycle. All experiments were conducted in accordance with the ARRIVE guidelines.

Twenty-four hours before inoculation with RV1B, mice were i.p. administered with 0.8 mg/kg (20 μg/200μl) anti-CCL7 Ab (R&D Systems) or a species-matched isotype control Ab. In other experiments, mice were intranasally administered with 3.75 nmol of an IRF-7–targeted small interfering RNA (IRF-7 siRNA) intranasally (3.75 nmol/25 μl) or nonsense control siRNA 24 h before innoculation with RV. Both siRNAs were obtained from Dharmacon.

Lyophilized preparation of milled house dust mites (HDM) containing Dermatophagoides pteronyssinus extract was obtained from Greer Laboratories (Lenoir, NC). Mice were sensitized and challenge to HDM followed by infection with RV, as previously reported (10, 11). Briefly, mice were treated intranasally under light isoflurane anesthesia with HDM (50 μg/50μl in sterile saline) daily for 3 d during sensitization. Mice were then intranasally challenged 12 d later, with HDM (5 μg/50 μl) during 4 d to induce AAD. Nonsensitized mice (saline) received sterile endotoxin-free saline only. Allergic (HDM) mice were intranasally infected with RV (50 μl containing 1 × 10⁷ virions) or UV-inactivated RV 24 h after the last HDM challenge to exacerbate preexisting AAD. Mice were euthanized 24 h after the RV infection at the peak of host response. Twenty-four hours postinfection was chosen for our analyses, as innate immune responses induced by RV1B in BALB/c mice peaks 24 h postinfection and RV infection is quickly resolved (12) by 4 d postinfection.

AHR was measured as previously described (13–15). Briefly, mice were anesthetized with ketamine-xylene combination (Illum) and total lung resistance and dynamic compliance were measured invasively (Buxco). Mice were mechanically ventilated and increased lung resistance to nebulized methacholine was expressed as percentage change from control (baseline).

Bronchoalveolar lavage fluid (BALF) was collected by cannulating the trachea of euthanized mice and flushing the airways twice with 1 ml HBSS. BALF was centrifuged at 800 × g for 10 min at 4°C. Pelleted cells were treated with RBC lysis buffer (12 mM NaHCO₃, 0.1 mM EDTA, 155 mM NH₄Cl [pH 7.35]) for 5 min and centrifuged as above. Cell pellets were resuspended in 100 μl HBSS and total number of viable cells was determined by trypan blue exclusion in a hemocytometer. Following a cytospin, slides were stained with May–Grünwald–Giemsa, blinded, and differential cell counts were determined from a count of 200 cells per slide.

An excised lung lobe was fixed in 10% neutral buffered formalin for 24 h and then stored in 70% ethanol. Carbol’s chromotrope-hematoxylin and periodic acid–Schiff (PAS) stains of 5-μm slices of fixed lungs were used for semiquantitative histopathologic scoring (range, 0–10) as previously described (16). Scores are a sum of perivascular edema (0, absent; 1, mild to moderate, present in <25% of the perivascular spaces; 2, moderate to severe, in 25–75% of perivascular spaces; or 3, severe, present in >75% of perivascular spaces), perivascular/peribronchial acute inflammation (0, absent; 1, perivascular edematous inflammation with fewer than five neutrophils per high-power field; 2, moderate perivascular inflammation, involving some peribronchial spaces, with more than five neutrophils per high-power field; or 3, severe, acute perivascular and peribronchial inflammation with numerous neutrophils encircling >50% bronchioles), bronchiole goblet cell metaplasia (0, absent; 1, two or fewere bronchioles with fewer than five goblet cells; or 2, large numbers of goblet cells present), and inflammation in alveolar spaces (0, absent; 1, present in <25% of alveolar spaces; or 2, present in >25% of alveolar spaces). Images representative of all mice in a group were taken from tissue sections using Image-Pro Plus software and enhanced in Adobe Photoshop CS5.1 to reduce brightness (−30) and enhance contrast (+100)

Thoracic contents were extracted from euthanized mice and forceps were used to separate airways from the parenchyma by blunt dissection (17), resulting in the isolation of several generations of airway tissue. TRIzol (Ambion, Carlsbad, CA) was used to extract airway mRNA and concentrations were determined by spectrophotometry (NanoDrop) for reverse transcription with BioScript (Bioline, Alexandria, NSW, Australia) to generate cDNA. Quantitative PCRs (qPCRs) were performed according to manufacturer’s instructions with SYBR Green (Invitrogen, Mulgrave, VIC, Australia) using the primer sequences in Supplemental Table I. Copy numbers of target genes were obtained by referencing CTs to a standard curve of known concentration. All values were normalized to the endogenously expressed control gene, hypoxanthine phosphoribosyltransferase.

Mouse lung cells were mechanically extracted using gentleMACS tubes as per manufacturer’s instructions and stained with anti–CD11c-FITC, anti–CD11b-PerCP (BD Pharmingen), anti–MHC class II-PE, and anti–F4/80-allophycocyanin (eBioscience) to determine myeloid dendritic cell and macrophage numbers or anti–mPDCA-1-allophycocyanin to determine plasmacytoid DC (pDC) numbers (Miltenyi Biotec). We determined numbers of positive cells by flow cytometry (FACSCanto, Becton Dickinson).

A single excised lung lobe from each mouse was snap frozen before homogenization in buffers and protease inhibitors as recommended in manufacturers’ instructions. Levels of CCL7 (eBioscience), CCL11, CCL20, CXCL2 (R&D Systems), and p-ERK1 were determined in clarified lung lysates by ELISA. Quantification of activated NF-κB subunits was performed with a TransAM NF-κB transcription factor assay (Active Motif). All concentrations were normalized to lung lobe weight.

All graphs in figures were created in GraphPad Prism 6.0 (GraphPad Software, La Jolla, CA) are expressed as means ± SEM. A Student t test or two-way ANOVA was used as appropriate with an α value of 0.05 for all analyses. All significantly different results shown are a comparison with the isotype or nonsense control, unless otherwise stated.

CCL7 is the most significantly upregulated gene at the peak of RV infection in vivo (www.ebi.ac.uk/arrayexpress, access no. E-MTAB-2826), which was confirmed by qPCR (Fig. 1A) and ELISA (Fig. 1B). Therefore, we administered a neutralizing Ab against CCL7 and subsequently infected BALB/c mice with RV. Twenty-four hours after infection, the peak of viral load, and host response, mice were sacrificed for analyses. Anti-CCL7 treatment did not affect CXCL2 (also known as MIP2-α) release, demonstrating that the observed effects were specifically due to CCL7 inhibition (Fig. 1B). Anti-CCL7 suppressed AHR induced by RV (Fig. 1C), concurrent with a decrease in neutrophil and macrophage numbers in BALF (Fig. 1D).

Intraperitoneal administration of anti-CCL7 24 h prior to RV infection suppressed levels of the phosphorylated MAPK ERK1 and active NF-κB p65 and p50 (Fig. 2A, 2B). Anti-CCL7 treatment reduced expression of IRF-7 and IRF-5 mRNA, but not IRF-3 mRNA (Fig. 2C), and it dampened IFN-β release (Fig. 2E). Anti-CCL7 did not affect viral titer as determined by qPCR of positive-strand RV1B RNA (Fig. 2D). To explore the relationship between CCL7, IRF-7 and IFN-β, we assessed pDC and macrophage populations in the lung by flow cytometry. Flow cytometric analysis of pDCs revealed that anti-CCL7 did not reduce the number of pDCs recruited by RV (Fig. 2F). However, alveolar macrophage numbers in lung tissue were reduced to baseline by αCCL7 (Fig. 2F).

To explore the role of IRF-7 upregulation during RV infection, we administered IRF-7 siRNA intranasally 24 h prior to RV infection. IRF-7 siRNA successfully suppressed IRF-7 mRNA expression at 24 h postinfection without affecting other IRFs, such as IRF-5 or IRF-3 (Fig. 3A). This resulted in complete suppression of IFN-α and IFN-β production, but, similar to inhibition of CCL7 and subsequent impaired IRF-7 expression, it had no effect on RV titer (Fig. 3B, 3C). IRF-7 siRNA treatment reduced macrophage and neutrophil numbers in the BALF (Fig. 3E), but this did not result in a significant reduction in AHR compared with RV-infected mice that received nonsense siRNA control (Fig. 3D). IRF-7 siRNA treatment did not affect CCL7 or CXCL2 production or reduce levels of p-ERK1 or active NF-κB subunits (Fig. 3F–H).

To investigate the role of CCL7 in RV-induced exacerbations of AAD, we sensitized and challenged mice to HDM and on the day of final challenge, 24 h prior to RV infection, mice received anti-CCL7 i.p. CCL7 was upregulated in mice treated with an isotype control Ab and was successfully neutralized in the lung after anti-CCL7 treatment (Fig. 4A) and resolved 4 d postinfection along with clearance of RV infection (Supplemental Fig. 1). CCL7 inhibition also downregulated CCL20 (also known as MIP3-α), but not CXCL2 (Fig. 4A) or CCL11 (also known as eotaxin-1) (Fig. 4B). BALF neutrophils, macrophages, and eosinophils were reduced with anti-CCL7 (Fig. 4C). However, anti-CCL7 treatment did not ameliorate AHR (Fig. 4D). Furthermore, NF-κB p65 and p50 activity remained elevated in anti-CCL7–treated allergic mice (Fig. 4E). CCL7 neutralization during RV infection of allergic mice did not affect IRF-7, type I IFN mRNA, or RV1B copy numbers (Supplemental Fig. 2).

Next, we investigated the effects of IRF-7 inhibition in RV-induced exacerbation of AAD. We sensitized and challenged mice to HDM and on the day of final challenge, 24 h prior to RV infection, mice received IRF-7 siRNA intranasally. IRF-7 mRNA expression was significantly induced 24 h following RV infection and resolved 4 d postinfection (Supplemental Fig. 1). IRF-7 siRNA effectively reduced IRF-7 mRNA, resulting in abrogated type 1 IFN responses with no effect on viral titer (Fig. 5A–C). Unlike anti-CCL7, IRF-7 siRNA treatment did not suppress inflammation or chemokine production in the context of an RV-induced exacerbation (Fig. 5D–H).

We employed histopathological scoring of 5-μm sections of fixed mouse lungs to evaluate pulmonary inflammation in allergic mice based on the presence of perivascular edema, perivascular and peribronchiolar inflammation, mucus production, and alveolar inflammation. All allergic mice displayed high numbers of periodic acid–Schiff⁺ cells as well as perivascular and peribronchiolar inflammatory infiltrates and, in some cases, alveolar inflammation. Representative images of RV exacerbation groups are shown in Fig. 6A. RV infection of allergic mice did not increase pulmonary histopathological scores compared with allergic mice given UV control. Furthermore, inhibiting IRF-7 or neutralizing CCL7 during RV exacerbation did not affect pulmonary histopathology scores, although all allergic mice were scored significantly higher than saline controls (Fig. 6B, 6C).

Our study has dissected the role of CCL7 and IRF-7 in RV-induced inflammatory and antiviral responses.

CCL7 activates CCR2, which is expressed on neutrophils to mediate chemotaxis (18). It has previously been shown that CCL7 also binds to CCR3 on eosinophils to induce chemotaxis and activation (19). Airway epithelial cells and macrophages produce CCL7, which may orchestrate cell recruitment in the context of oxidative stress (20, 21). Induction of CCL7 during viral infections has been observed in asthmatic patients (22) and in models of RV infection dissecting the role of CCL2 (23). Our study significantly extends these findings by showing that CCL7 inhibition limited macrophage and neutrophil influx into the lung and the development of AHR in nonallergic mice, which was associated with suppression of the MAPK signaling factor ERK1 and NF-κB subunits p50 and p65. In contrast, AHR and NF-κB activation was not significantly reduced by CCL7 inhibition in allergic mice although inflammation was limited. Thus, our data suggest an association between NF-κB activation and AHR, which has been reported previously (10, 11, 24). NF-κB promotes Th2 polarization and subsequent IL-13 production as well as IgE production from B cells, resulting in mast cell degranulation after allergen-specific IgE cross-linkage, both of which directly result in propagation of AHR in the context of allergy (24–28). In the context of RV exacerbations, we have shown previously that by restoring protein phosphatase 2A (PP2A) activity, levels of NF-κB subunits were reduced, which protected against AHR (11). Furthermore, when we inhibited PP2A activity with an siRNA targeted to PP2Acα (the active or catalytic subunit of the polyprotein) in these experiments, both NF-κB activity and AHR were restored (10).

We showed that CCL7-mediated inflammation is not essential for the development of RV-induced AHR in allergic airways disease. Dissociation between AHR and inflammation has been previously reported in other models of allergic airways disease (29) and in subjects with asthma (24, 30).

Interestingly, CCL7 neutralization in RV infection also suppressed IRF-7 and IFN-β expression, which was associated with reduced macrophage but not pDC numbers in the lung. A role of CCL7 in macrophage recruitment is further highlighted by the observation of a correlation between nasal CCL7 production and the recruitment of macrophages in children with virus-induced asthma (31). A previous report showed IFN-α and IFN-β production by macrophages upon norovirus infection, which was abolished in IRF-3 and IRF-7 double knockout cells (32). It is therefore possible that CCL7 increases IRF-7 expression and IFN-β production in the lung by regulation of macrophage inflammation upon RV infection.

To elucidate the role of IRF-7 in RV infection, we employed siRNA-mediated inhibition. IRF-7 is thought to be a master regulator of type 1 IFN production (33–36), and induction of IRF-7 mRNA has been reported previously (37) in asthmatic children during periods of respiratory viral infections. As expected, IRF-7 silencing suppressed IFN-α and IFN-β responses. Interestingly, IRF-7 inhibition also reduced neutrophil and macrophage numbers in a similar fashion to CCL7 antagonism but did not reduce AHR or NF-κB activity. Although our results therefore support an important role of IRF-7 in RV-induced antiviral responses through induction of IFN production, they also reveal an unexpected role of IRF-7 in modulating the inflammatory response upon in vivo RV exposure. Recently, IRF-7 was identified to control pro- to anti-inflammatory (M1-to-M2) macrophage phenotype switch, and IRF-7 expression reduced proinflammatory macrophage activity (38). This may be relevant to RV infection, and the synergistic effects of CCL7 on IRF-7 expression observed in our study may regulate the balance between the proinflammatory activity of lung macrophages and their antiviral capabilities.

Collectively, we dissect the role of CCL7 and IRF-7 in RV-induced antiviral and inflammatory responses and show an important collaborative function for the recruitment and activation of inflammatory cells in the lung and for mounting an IFN-β response to RV.

We thank Dr. Ana Pereira de Siqueira, Heather MacDonald, Jane Grehan, Leon Sokulsky, Matthew Morton, and the staff from Bioresources of the Hunter Medical Research Institute for technical assistance.

S.L.J. has received consultancy fees from Centocor, Sanofi Pasteur, Synairgen, GlaxoSmithKline, Chiesi, Boehringer Ingelheim, Grünenthal, and Novartis; has patents planned, pending, or issued (U.K. patent application no. 02 167 29.4 and international patent application no. PCT/EP2003/007939, U.K. patent application no. GB 0405634.7, and U.K. patent application no. 0518425.4); and has stock/stock options in Synairgen.