Respiratory syncytial virus (RSV) is the leading cause of lower respiratory tract infection in children worldwide. The understanding of neonatal RSV pathogenesis depends on using an animal model that reproduces neonatal RSV disease. Previous studies from us and others demonstrated that the neonatal lamb model resembles human neonatal RSV infection. Here, we provide an extensive and detailed characterization of the histopathology, viral load, cellular infiltration, and cytokine production in lungs and tracheobronchial lymph nodes of lambs inoculated with human RSV strain A2 over the course of infection. In the lung, RSV titers were low at day 3 postinfection, increased significantly by day 6, and decreased to baseline levels at day 14. Infection in the lung was associated with an accumulation of macrophages, CD4+ and CD8+ T cells, and a transcriptional response of genes involved in inflammation, chemotaxis, and interferon response, characterized by increased IFNγ, IL-8, MCP-1, and PD-L1, and decreased IFNβ, IL-10, and TGF-β. Laser capture microdissection studies determined that lung macrophage-enriched populations were the source of MCP-1 but not IL-8. Immunoreactivity to caspase 3 occurred within bronchioles and alveoli of day 6-infected lambs. In lung-draining lymph nodes, RSV induced lymphoid hyperplasia, suggesting an ability of RSV to enhance lymphocytic proliferation and differentiation pathways. This study suggests that, in lambs with moderate clinical disease, RSV enhances the activation of caspase cell death and Th1-skewed inflammatory pathways, and complements previous observations that emphasize the role of inflammation in the pathogenesis of RSV disease.
- immune response
human respiratory syncytial virus (RSV) is the leading cause of severe lower respiratory tract illnesses in premature/newborn infants and young children. The World Health Organization (WHO) estimates that in the United States alone, 70,000 to 126,000 infant hospitalizations for pneumonia or bronchiolitis are attributed to RSV every year (43). Furthermore, global annual RSV infections are estimated by the WHO at 64 million, with 160,000 deaths occurring yearly. Immunocompromised individuals (30) and the elderly (14) are also at risk for severe RSV disease, whose manifestations include rhinitis, bronchitis, otitis media, and pneumonia (31). Despite the importance of RSV as a pathogen, there is no currently licensed vaccine for RSV. Challenges encountered for the successful development of a pediatric RSV vaccine include the relatively immature immune response of infants, the potential interference of maternal antibodies with vaccinations, and previous concerns of vaccine-induced enhanced disease severity (reviewed in Ref. 40). However, some progress has been made with prophylaxis using the humanized anti-RSV monoclonal antibody palivizumab (20). Palivizumab is indicated for the prevention of serious lower respiratory tract disease caused by RSV in high-risk children (premature infants and children less than 2 years of age with chronic lung disease or congenital heart disease) (4, 16).
In most children, RSV causes moderate upper respiratory infections, although a small but significant percentage of otherwise healthy children develop severe lower respiratory tract RSV infections that can result in hospitalizations. The mechanisms underlying the spectrum of disease severity from mild to severe are largely unknown, but may include genetic susceptibility to RSV and an immune response of an overzealous, Th2-biased, or inadequate nature (reviewed in Ref. 40).
Although rodents have provided valuable insight into RSV infection and disease, they are not natural hosts for any strain of RSV. In contrast, like human infants, lambs develop RSV disease in natural settings. The airway structure and function in neonatal lambs are similar to human infants (reviewed in Ref. 32). For example, the numbers of Clara cells in lambs are similar to those in humans, whereas mice can have up to 50% of Clara cells in bronchioles (28). Also, submucosal glands are present in both ovine and human lungs, but not in rodents, and these produce lactoperoxidase, which is essential for the epithelial oxidative defense system. Unlike rodents, alveolar development occurs preterm in both humans and lambs (5, 22, 33). Lambs are susceptible to infection by both ovine and bovine strains of RSV, and can be experimentally infected with the human RSV A2 strain (hRSV). In fact, infection of neonatal lambs with hRSV results in pathology resembling infection in human infants, as evidenced by mild peribronchiolar infiltrates of lymphocytes and plasma cells, and bronchiolitis characterized by epithelial cell degeneration, sloughing of degenerating cells, and intraluminal infiltrates of neutrophils (26). These symptoms of hRSV disease are not observed in hRSV-infected rodents. Thus, the neonatal lamb model offers multiple advantages over the rodent model for studying the natural progression of hRSV disease and for investigating intervention strategies that could be relevant to humans.
In the present study, we performed an in-depth analysis of the innate and adaptive immune markers and the cellular components of the inflammatory response in the lungs and lung-draining lymph nodes of lambs infected with hRSV shortly after birth over the course of infection. We used laser capture microdissection (LCM) technology to identify the cellular sources of select inflammatory mediators, and we examined the effect of RSV infection on the architecture of the lung-draining lymph nodes. The results of this investigation provide additional insights into the pathogenesis and immune response to RSV in neonates.
MATERIALS AND METHODS
Animal use and experimental procedures were approved by Iowa State University's Animal Care and Use Committee. In some experiments, we used banked lung samples from lambs that were used in another study to characterize hRSV infection in the lamb model (26). Briefly, healthy lambs 2–3 days of age were sedated with an intramuscular injection of xylazine (0.1 mg/kg), and a sterile bronchoscope was used to instill hRSV strain A2 (5 ml at 2.0 × 107 pfu/ml) or sterile cell growth media (5 ml) in the right mainstem bronchus. Lambs were euthanized with pentobarbital sodium intravenously after 3 (n = 3/group), 6 (n = 3/group), or 14 (n = 3 for control group, n = 4 for RSV-infected group) days after infection. Sections of the right cranial lung lobe were removed, snap-frozen in liquid nitrogen, and stored at −80°C for qPCR analysis. Additional sections of the right cranial lung lobe were fixed in 10% neutral-buffered formalin for immunohistochemical analysis, or placed in cryomolds containing optimum cutting temperature compound (OCT; Thermofisher Scientific, Waltham, MA) at −80°C for immunofluorescence analysis. A second group of healthy lambs 2–3 days of age was also inoculated with either RSV (n = 5) or sterile cell growth media (n = 3) for 6 days for flow cytometry analyses and the study of tracheobronchial lymph nodes. In this group of animals, sections of the right cranial lung lobe and the tracheobronchial lymph nodes were removed and used immediately for flow cytometry analyses. A portion of the tracheobronchial lymph nodes was fixed in 10% neutral-buffered formalin for histological analysis.
Cellular infiltration in the lungs was analyzed by immunofluorescence (IF) microscopy. The IF protocol used here is described elsewhere (36). Briefly, lung frozen sections were fixed in acetone, rehydrated in TRIS buffer, blocked with 5% normal goat serum in PBS, and then washed in TRIS buffer. Sections were incubated with primary antibody (mouse anti-bovine CD11b) at 1:400 in TRIS/PBS + 3% BSA overnight at 4°C, with secondary antibody (Alexa 488-goat anti-mouse IgG, Fab′2) at 1:400 for 30 min at 37°C, counterstained with hematoxylin and finally mounted with an aqueous mounting medium.
Lung samples were incubated for 1 h at 37°C with 7 mg/ml Liberase Blendzyme III (Roche Applied Science, Indianapolis, IN) and 30 μg/ml DNase (Sigma) in PBS (without Ca2+ or Mg2+). After a red blood cell (RBC) lysis procedure (eBioscience), lung cells were plated at 1.0 × 106 cells/100 μl in a 96-well plate, washed in FACS buffer, and incubated with primary antibody (mouse anti-bovine CD1, mouse anti-bovine CD4, mouse anti-bovine CD8, mouse anti-bovine WC1, mouse anti-bovine CD25) or isotype controls for 15 min at 4°C. Cells were stained with Alexa 488 Fab′2 goat anti-mouse IgG for 15 min at 4°C and then fixed in cold 1% paraformaldehyde at 4°C.
Cells were subjected to RBC lysis, plated at 1.0 × 106 cells/100 μl in a 96-well plate, and then fixed and permeabilized using BD Cytofix/Cytoperm solution (BD Biosciences, San Jose, CA). Intracellular staining was performed with mouse anti-bovine TNFα, mouse anti-bovine IFNγ primary antibodies, or isotype controls, followed by Alexa 488 Fab′2 goat anti-mouse IgG secondary antibody. Cells were washed with BD Perm/wash and fixed in cold 1% paraformaldehyde at 4°C. Analysis was performed on a BD FACScanto flow cytometer (Becton Dickson, San Jose, CA), and data were analyzed using FlowJo software v8.8.2 (Tree Star, Ashland, OR).
The LCM procedure was performed as previously described (7, 36). Frozen tissue blocks were cut into 6-μm sections at −25°C and incubated in blocking solution (10% Aurion acetylated/linearized BSA-c solution) and then in primary antibody [1:100 monoclonal mouse anti-sheep CD11b in Common Antibody Diluent (BioGenex, San Ramon, CA) containing 1% normal goat serum and 1% normal swine serum for detection of macrophages and 1:100 monoclonal mouse anti-bovine CD208 for detection of epithelial cells]. After addition of the secondary antibody (1:500 Alexa 488 goat anti-mouse IgG in Common Antibody Diluent), slides were dehydrated through a gradient of ethanol and xylene baths. LCM settings were 80 mV for 800 ms on the Palm-MicroBead. Cells were deposited on HS caps (Arcturus/Molecular Devices). The polymer tabs on each HS cap were peeled off and placed in resuspension lysis buffer for RNA isolation.
Total RNA was isolated from lung tissue as previously described (35) using a procedure based on TRIzol reagent (Invitrogen, Carlsbad, CA). The RNA samples were DNase-treated using TURBO DNase (TURBO DNA-free kit; Ambion, Austin, TX). RNA concentrations and purity were measured by absorbance readings at 260 and 280 nm.
RNA was also isolated from lung macrophages and epithelial cells enriched by LCM using protocols we previously published (7, 36) based on the CellsDirect kit (Invitrogen). Polymer tabs placed in a resuspension lysis buffer were disrupted by mechanical agitation. Samples were incubated at 50°C for 10 min and then at 75°C for 5 min. Samples were DNase-treated for 25 min at room temperature and then incubated with an EDTA solution for 10 min at 70°C. Nuclease-free water and a solution of MgSO4 were added to the samples.
One-Step Real-Time Reverse Transcription PCR
Initial RT-qPCR analysis involved running a test plate for each target to identify the best RNA dilution ranges in which PCR inhibition was not observed, and where amplification efficiencies were better than 80%. Test plate set-up parameters and analysis were performed using PREXCEL-Q (6). RNA samples were used at 0.784 ng/μl in fluorogenic one-step real-time qPCR reactions. Each 20-μl reaction contained 6 μl of RNA sample, 775 nM primers (see Table 1 for primer and probe sequences), 150 nM TaqMan hydrolysis probe (5′-6FAM, 3′-TAMRA-quenched or 5′-6FAM, 3′-MGBNFQ), nuclease-free water, 5.5 mM MgSO4, SuperScript III RT/Platinum Taq mix (Invitrogen), and One-step reaction mix with ROX (Invitrogen). The samples were placed in duplicate wells of 96-well plates (Eppendorf, Westbury, NY). Two negative, no-template control wells were also included. The GeneAmp 5700 Sequence Detection System (Applied Biosystems) was used at 15 min 55°C, 2 min 95°C, and 50 cycles of 15 s 95°C, 30 s 58°C. Quantification cycle (Cq) values generated by the GeneAmp 5700 software were further assessed using custom Excel files employing the efficiency-corrected standard curve-based relative quantification (Pfaffl) approach for qPCR (27). Results are expressed as relative gene expression to uninfected samples. Carryover genomic DNA (gDNA) contamination, as examined by no-reverse transcriptase control reactions, was absent or exhibited gDNA-related Cq values of 12 cycles or more away from the sample Cq values of interest.
Immunolon 1 plates were coated overnight at 4°C with primary antibody (monoclonal mouse anti-bovine TNFα or monoclonal mouse anti-bovine IFNγ; AbD Serotec) diluted to 5 μg/ml with 0.5 M carbonate bicarbonate buffer, pH 9.6. Following a wash (PBS with 0.05% Tween 20), wells were blocked for 1 h at 37°C with blocking buffer (PBS with 1% BSA and 0.05% Tween 20). Standards and samples diluted in wash buffer were added to the plate and incubated for 1 h at 37°C. This was followed by detection using 2.5 μg/ml biotinylated mouse anti-bovine TNFα or IFNγ secondary antibody (AbD Serotec), 1:200 streptavidin-HRP conjugate (Invitrogen), and 1-Step turbo TMB-ELISA substrate solution (Thermo Scientific, Rockford, IL). Absorbance measurements were taken at 450 nm. A standard curve was prepared from the data obtained to calculate the concentrations of the samples in ng/ml.
Immunohistochemistry for Caspase 3 and Scoring System
Paraffin-embedded lung sections were heated at 58°C for 1 h, deparaffinized through a gradient of xylene and ethanol baths, and boiled in antigen retrieval solution (BioGenex). Sections were blocked in 10% normal goat serum plus 10% normal swine serum in PBS for 20 min, then incubated with 1:1,000 rabbit anti-human/mouse active caspase 3 (R&D Systems, Minneapolis, MN) diluted in 5% normal swine serum and antibody diluent for 2 h at 4°C. After washes in PBS, sections were incubated with 1:300 biotinylated goat anti-rabbit IgG (Kirkegaard & Perry Laboratories, Gaithersburg, MD) in antibody diluent plus 5% normal sheep serum for 45 min at room temperature. Slides were treated with a peroxidase block (3% H2O2 in PBS) followed by incubations with streptavidin-conjugated horseradish peroxidase (BioGenex) and the chromogen Nova Red (Vector, Burlingame, CA). Slides were counterstained with Harris' hematoxylin, rehydrated through a series of ethanol and xylene baths, and coverslipped with Permount solution (Sigma, St. Louis, MO). Scoring for caspase 3 immunoreactivity (IR) was performed by a pathologist (M. R. Ackermann) in ×40 fields. The scores ranged from 0–4 and were assigned to bronchiolar and alveolar areas. In bronchioles: 0 = no staining, 1 = 1–10%, 2 = 11–30%, 3 = 31–60%, 4 = more than 60% of airway epithelia had IR to caspase 3. In alveoli, staining occurred in consolidated areas, containing lymphocytes, plasma cells, macrophages, and alveolar epithelia: 0 = no staining, 1 = 1–10%, 2 = 11–30%, 3 = 31–60%, 4 = more than 60% of consolidated areas had IR to caspase 3.
The results were analyzed by one-way ANOVA using SAS version 9.1 (SAS Institute, Cary, NC). Mean values of relative mRNA expression levels of control animals and RSV-infected animals were compared at *P < 0.05, **P < 0.01, and ***P < 0.001 and expressed as means ± SE.
RSV Levels and Cellular Infiltration in the Lung
Previously, we reported that the RSV challenge dose used to infect neonatal lambs caused moderate suppurative bronchiolitis and peribronchiolar mononuclear interstitial pneumonia (26). At 3 days postinoculation (p.i.), lambs had histological changes consistent with RSV infection, but no gross lesions were observed. By day 6, 83% of RSV-infected lambs had gross lesions, which were multifocal, variably sized, and often linear to irregular areas of plum-red consolidation. At day 14 p.i., lambs had minimal histological changes characterized by small areas of alveolar consolidation, with no gross lesions or clinical signs of disease. Also, no viral antigen and minimal viral mRNA were detectable at this time point. Thus, RSV infection was considered to be nearly resolved by day 14 p.i. (26).
Similar observations were made in the hRSV-infected lambs used in the present study. RSV viral mRNA was found to significantly increase from day 3 to day 6 p.i., and to significantly decrease from day 6 to day 14 p.i. (Fig. 1A). We evaluated the effect of RSV infection on inflammation by measuring the cellular influx of monocytes/macrophages (defined here as CD11b+ cells), neutrophils (by morphology), dendritic cells/B cells/monocytes (CD1+ cells), T-helper cells (CD4+ cells), cytotoxic T cells (CD8+ cells), ovine γδ T cells (WC1+ cells), and activated lymphocytes expressing IL-2 receptors (CD25+ cells) into the lungs in response to RSV infection. At both days 3 and 14 p.i., no significant differences in the overall levels of CD11b+ cells in the lungs of control and RSV-infected animals were detected (Fig. 1B). In contrast, the lungs of animals infected with RSV for 6 days revealed an approximate 2.5-fold increase in the recruitment of CD11b+ cells compared with control animals (Fig. 1B). In contrast to observations in mouse models (11), but consistent with human RSV disease (19) and as seen in our previous studies in the lamb model (26), neutrophils were present in lumens of bronchioles at 3 and 6 days p.i.; at 14 days p.i., lesions were resolved and lacked neutrophils (data not shown). In animals examined 6 days p.i., significantly increased levels of CD1+ (Fig. 1C), CD4+ (Fig. 1D), and CD8+ (Fig. 1E) cells were measured. No significant changes in the levels of WC1+ (Fig. 1F) or CD25+ (Fig. 1G) cells were observed in the lungs of RSV-infected animals compared with control animals. The observed peak in infiltration of macrophages at day 6 in the lungs (Fig. 1B) is in agreement with light microscopy findings previously published by our laboratory (26). Similar to findings in human infants (19), our results suggest that monocytes/macrophages and T lymphocytes are the primary cellular immune responses to RSV infection present in the neonatal ovine lung.
RSV-Induced Type I IFNβ is Suppressed During Infection
As type I interferons (IFNs) are crucial to the host antiviral response (8), we determined the effects of RSV infection on the expression levels of these mediators. IFNα transcripts could not be detected in the lungs of either control or RSV-infected newborn lambs (data not shown). However, IFNβ was expressed in these animals (Fig. 2A). At day 3 p.i., control and RSV-infected animals expressed similar levels of IFNβ mRNA, although IFNβ mRNA decreased significantly at days 6 and 14 in RSV-infected animals compared with control animals.
An Inflammatory Response With a Skew Towards a Type 1 Immune Response Is Observed in the Lung Following RSV Infection
The cytokine response in the lungs of RSV-infected lambs was investigated. A pronounced upregulation of IFNγ mRNA was observed in lambs at days 3 and 6 p.i. compared with uninfected control animals (Fig. 2B). The levels of IFNγ mRNA paralleled those of RSV at all time points, including a decrease in expression at day 14 p.i. (Fig. 2B). TNFα mRNA was significantly increased at day 3 p.i., but decreased at both days 6 and 14 p.i. (Fig. 2C). TGF-β decreased steadily with infection (Fig. 2D). IL-10 mRNA levels were inhibited by RSV infection at days 6 and 14 p.i. (Fig. 2E). The mRNA levels of IL-12p35, IL-12p40, IL-4, IL-13, and IL-17 were below the limit of detection (data not shown).
Although protein detection in this model organism is difficult, an attempt was made to measure protein levels of select targets by ELISA. IFNγ protein levels correlated with mRNA levels at days 3 and 6 (Fig. 2F). TNFα protein levels correlated with mRNA at day 3 p.i., but not at days 6 and 14, where protein levels in infected animals were not significantly different from those of control animals (Fig. 2G).
Together, these data suggest that in neonatal animals, RSV infection is associated with a proinflammatory response with Th1 characteristics, as exemplified by the upregulation of IFNγ and a downregulation of anti-inflammatory cytokines during the peak of viral pathogenesis.
Chemokine Expression Levels Following RSV Infection
We also determined the effects of RSV on lung chemokines in the ovine model. IL-8 mRNA levels paralleled those of RSV at day 6 p.i. (Fig. 3A). MCP-1 (Fig. 3B) and MIP-1α (Fig. 3C) showed a similar but more moderate trend in expression as IL-8 at day 6 p.i., a time point at which increased cellular infiltration in the lungs was observed (Fig. 1). Thus, the upregulation of the chemokines IL-8, MCP-1, and MIP-1α following RSV infection correlates with the infiltration of monocytes/macrophages and neutrophils into the lung. In contrast, RANTES mRNA was decreased at all time points, with the most significant inhibition seen at day 6 p.i. (Fig. 3D), suggesting that MCP-1 and MIP-1α, but not RANTES, contribute to the recruitment of T cells into the lungs of RSV-infected lambs.
LCM Studies Reveal the Cellular Source of Chemokines
As RSV infection is associated with a cellular influx into the lungs, an attempt was made to identify the cellular source of key chemokines, MCP-1 and IL-8, using laser capture microdissection (LCM) and qPCR. Using a protocol developed in our laboratory, ovine lung macrophages were enriched (36) and chemokine levels were assessed in those cells. MCP-1 mRNA was detected in LCM-enriched macrophages (Fig. 3E). MCP-1 levels in macrophages were particularly enhanced at day 6 p.i., which coincided with the influx of CD11b+ cells at that time point (Fig. 2A). In contrast, IL-8 mRNA was not detected in LCM-enriched macrophages (data not shown).
Effect of RSV Infection on the Expression of Negative Regulators of T Cell Functions
The expression levels of PD-1 and PD-L1, two negative regulators of T cell functions, were measured in the lungs of hRSV-infected lambs. PD-1 mRNA was significantly decreased by RSV infection at day 3 p.i. but showed no significant changes in expression at days 6 and 14 p.i. compared with control animals (Fig. 4A). PD-L1 mRNA expression was significantly enhanced at both days 3 and 6 p.i., but went back to baseline levels at day 14 (Fig. 4B). Our findings of RSV-mediated PD-L1 upregulation in the lung are consistent with experiments performed in human respiratory tract epithelial cells (38).
Caspase 3 Levels Are Increased at Day 6 of Infection
To assess the apoptotic response to hRSV infection, lung caspase 3 expression was detected by immunohistochemistry. Caspase 3 was not detected in the lungs of any control animals (Fig. 5, A, C, and F) or in the lungs of animals infected with RSV for 3 (Fig. 5B) and 14 days (Fig. 5G). However, an infection with RSV for 6 days, where viral load was found to be maximal, was associated with increased caspase 3 immunoreactivity (IR) in alveolar areas (Fig. 5D) and in airway epithelia (Fig. 5E). In alveolar areas of day 6-infected lungs, ∼11–30% of cells had caspase 3 IR in consolidated areas (average score of 2.36). In bronchiolar areas, ∼1–10% of airway epithelia had caspase 3 IR (average score of 1.06). These findings suggest that, in vivo, increased RSV load is associated with a proapoptotic effect in lungs.
RSV Infection Affects the Architecture of the Lymph Nodes
Gross observations of lung-draining lymph nodes revealed an approximate threefold increase in size and weight following RSV infection (data not shown). Histological observations of the lymph nodes of uninfected neonatal animals revealed that, although they possessed cortical and medullary areas, they lacked obvious follicles (Fig. 6A). Such a morphology has previously been reported in other animals, including calves, and has been attributed to the age and physiological state of the animals (21). However, as shown in Fig. 6B, lymph nodes harvested from lambs 6 days p.i. exhibited dramatically different architecture than those from control lambs: follicular areas indicate lymphoid hyperplasia, with an extensive expansion of the cortical area; the medulla contained large numbers of lymphocytes with numerous tingible-bodied macrophages, which is an indicator of lymphocyte apoptosis following lymphocyte proliferation (Fig. 6C). The change in architecture of the lymph nodes in infected animals suggests an ability of RSV to induce lymphoid proliferation resulting in follicle formation.
Immune Response in Tracheobronchial Lymph Nodes Upon RSV Infection
Similar to the response seen in the lungs at day 6 p.i., an induction in IFNγ mRNA (Fig. 7A) and cells secreting IFNγ (Fig. 7B) was observed in the lung-draining lymph nodes after RSV infection. The lymph nodes of RSV-infected animals had significantly enhanced levels of granzyme B mRNA (Fig. 7C) but no significant changes in CD8+ T cells (Fig. 7D) compared with lymph nodes from control animals. Although TNFα mRNA levels did not significantly change with infection (Fig. 7E), the percentage of cells secreting TNFα in lymph nodes was significantly greater after RSV infection (Fig. 7F). RSV-infected animals displayed enhanced IL-6 mRNA (Fig. 7G) but unchanged levels of PD-L1 mRNA (Fig. 7H). Whereas IL-12 transcripts could not be detected in the lungs, IL-12p35 mRNA and IL12p40 mRNA were expressed in the lymph nodes, but were not significantly changed by RSV infection (data not shown). These findings suggest that, similar to the lungs, the cytokine profile in the lung-draining lymph nodes in response to RSV infection is characterized by a strong proinflammatory response, as exemplified by increased IFNγ and IL-6 levels.
As RSV causes a significant respiratory disease in human infants, there is a profound need for a reliable and consistent neonatal animal model to better understand RSV disease pathogenesis. While no animal model is ideal, an outbred species in which gestation, lung development, immune responses, and susceptibility to RSV closely resembles that of humans would be attractive. Previously, our laboratory has demonstrated that RSV infection in lambs resulted in lung microscopic lesions similar to those seen in RSV-infected human infants (26). In the current study, we expanded on these results and demonstrated that 1) RSV is able to induce an immune/inflammatory response in neonatal lambs; 2) the peak of RSV-induced pathogenesis in the lungs (day 6) is associated with an increased infiltration of macrophages, neutrophils, T helper cells, and cytotoxic T cells; 3) RSV infection induced altered transcription of genes in the ovine lung and tracheobronchial lymph nodes involved in inflammation, IFN responses, and chemotaxis; 4) RSV infection at day 6 was associated with increased caspase 3 levels in the lungs; and 5) tracheobronchial lymph nodes of newborn animals exhibited a dramatic architectural change following RSV infection. To our knowledge, this is the first report of lung responses to RSV infection using an in vivo model that closely mimics human RSV disease.
Our finding that RSV infection induces cellular recruitment to the site of infection is in accordance with those of Johnson et al. (19), who showed the presence of monocytes, macrophages, and CD8+ T cells in lung sections from humans who died from RSV disease. However, while that study reported low levels of CD1a+ dendritic cells in the lungs of infants who died from RSV disease (19), we found increased levels of CD1-expressing cells in the lungs of animals with moderate RSV disease. Presently, it is unknown whether differences in CD1 levels relate to disease outcome. We also observed that cells expressing CD25 were found at low levels compared with levels of CD8+ and CD4+ cells in the lungs of control animals, and that RSV infection resulted in a trend towards decreased levels of CD25+ cells in the lung. Interestingly, in recent studies, a significant impairment in the ability of RSV-infected human dendritic cells to induce T cell activation has been observed (13). One potential mechanism for the impaired dendritic cell-mediated activation of T cells is an interference of contact between T cells and dendritic cells at the level of the immunological synapse (10). Despite similarities in cellular recruitment after RSV infection in moderate (this study) and fatal cases of RSV disease (19), there may be potential differences in the functionality of the recruited cells, and these differences may play key roles in disease outcome.
The generation of appropriate cytokine and chemokine responses is essential for resolution of viral infection. Consistent with studies in human cells (37), we observed a decrease in IFNβ levels after RSV infection in the ovine lung, confirming that RSV inhibits type I IFN. RSV infection also resulted in an increased production of TNFα protein and IFNγ mRNA, but low levels of IFNγ protein. In the lung, IFNγ is primarily produced by NK cells and CD8+ T cells. Although we have not measured NK cell presence in the ovine lung, we suspect that the early production of IFNγ at day 3 is due to NK cells, and that at day 6, both NK cells and CD8+ T cells contribute to increased IFNγ transcripts. The induction of IFNγ at day 6 of infection also correlated with the influx of CD8+ T cells in the lung. Surprisingly, the production of IFNγ occurred independently of IL-12, since we were unable to detect IL-12p35 and IL-12p40 transcripts in the ovine lung. In agreement with our results, a study of respiratory secretions from RSV-infected human infants suggested that the production of IFNγ during RSV infection was not associated with IL-12 or IL-18 (9). In the ovine model, we also found an increase in the production of proinflammatory mediators (IFNγ, IL-8, MCP-1, and MIP-1α) and a concomitant decrease in the expression of the anti-inflammatory cytokines IL-10 and TGF-β at 6 days p.i. in the lung. Although some studies suggest that human infants are prone to develop a Th2-type immune response upon infection (29), studies of cytokine profiles from infants with clinically severe RSV infection reported a similar pattern to our ovine model, with increased IFNγ but low IL-10 and IL-4 (2). In fact, the same study concluded that, regardless of clinical severity, type I immunity was the predominant response (2).
PD-1, PD-L1, and PD-L2 are recently described surface molecules of the B-7 family which play an important role in regulating T cell functions (12). Previous reports have suggested a role for PD-1/PD-L1 interactions in the functional inactivation of virus-specific CD8+ T cells during chronic viral infection (17, 23). Human epithelial cells exposed to RSV have been shown to induce the expression of PD-L1 and PD-L2 (38). Although CD8+ T cells and IFNγ were detected in the lungs of RSV-infected lambs, the observed increase in PD-L1 after infection might cause impairment in the cytolytic function of these cells. In such a case, IFNγ would serve as a proinflammatory cytokine and not fully function in the killing of RSV-infected cells. Results from infants suffering from severe RSV infection showed robust CD8+ T cell activation during convalescence, but not during the peak of viral load and disease severity (24), suggesting that although CD8+ T cells are present in the lungs, the RSV-specific CD8+ effector-memory response is dysfunctional. Results from the present study suggest that the ovine model will serve as a good tool to study RSV-mediated CD8+ T cell dysregulation in a disease model similar to that in human infants.
Our findings of increased IL-8 and MCP-1 at day 6 p.i. with hRSV are consistent with studies performed on nasal washes from RSV-infected individuals (34, 39) and on respiratory epithelial cells cultured in vitro in the presence of RSV (25). RANTES, which has been reported to strongly correlate with RSV disease severity in humans (15, 41), was not increased at day 6 of infection in the ovine lung. The reason for this difference may be due to the fact that we measured RANTES expression in the lung, whereas the mentioned studies were performed from nasopharyngeal and tracheobronchial secretions.
In addition to proinflammatory events, the increased influx of cells into the lungs may result in the activation of cell death pathways. In vitro, it has been shown that RSV can use its F protein, or NS1 and NS2 proteins, to enhance (3) or inhibit (1) apoptosis, respectively. Studies performed by Welliver et al. (42) point to a proapoptotic role of RSV, as lung sections from children with fatal RSV infection exhibited increased caspase 3 staining. Similarly, we found that neonatal lambs exposed to RSV for 6 days showed evidence of caspase 3 staining primarily in bronchioles and alveoli. Thus, our findings suggest that RSV infection activates intracellular pathways leading to cell death responses in epithelial and immune cells of the lung.
RSV infection of lambs was associated with distinct changes in the architecture of the lung-draining lymph nodes and a transcriptional response for proinflammatory cytokines, anti-inflammatory cytokines, and regulators of T cell responses. Although the immune response in the lung-draining lymph nodes appeared less robust than the response in the lungs, we determined that similar to the lungs, a trend towards a proinflammatory response characterized by the presence of IFNγ and TNFα was observed in the lymph nodes. In agreement with Janssen et al. (18), our finding of unchanged levels of CD8+ T cells after RSV infection in the lymph nodes of RSV-infected animals points to NK cells as the likely source of increased IFNγ and the IFNγ-regulated gene granzyme B. Future studies will determine the functional phenotype of CD8+ T cells in the lymph nodes of RSV-infected animals. Thus, infection with RSV appears to trigger immune-modulatory events in the lung-draining lymph nodes.
In conclusion, the current study demonstrated that RSV infection of neonatal lambs resulting in moderate clinical disease caused a rapid activation of a proinflammatory response with a Th1-skew. The chemokines that were increased by infection were also proinflammatory in nature and able to recruit CD4+ and CD8+ T cells, which were seen in the cellular infiltrate. However, the functional ability of these cells to kill RSV-infected cells will require further investigations, since RSV-infected lambs had increased levels of PD-L1. LCM-based studies revealed that MCP-1, but not IL-8, was produced by alveolar macrophages. The marked caspase 3 staining in the lungs at the time of maximal viral load suggested a proapoptotic effect of RSV in vivo. Finally, RSV-infected animals, but not control animals, had marked alterations in the architecture of tracheobronchial lymph nodes, suggesting the ability for RSV to modulate lymphoid proliferation and differentiation pathways. Our findings contribute to the understanding of how RSV modulates the immune response in the neonatal host and highlight the utility of the newborn lamb as a relevant model to study RSV-mediated disease, the host response and pathogenesis of infection, as well as to evaluate the effects of anti-viral and anti-inflammatory agents in vivo.
This research was supported in part by National Institutes of Health Grant RAI-062787A.
This work was supported in part by MedImmune, LLC. F. B. Sow is supported by MedImmune, LLC. S. Krishnan, A. C. Patera, and J. Suzich are all employed by MedImmune, LLC.
We thank the histology lab (Toni Christofferson, Jenny Groeltz-Thrush, and Diane Gerjets), the flow cytometry facility of the office of biotechnology (Dr. Shawn Rigby and Christine Deal), Dr. Tanja Lazic, Dr. Rachel Derscheid, and Dr. Brandon Plattner for technical support. We also thank Dr. Douglas Jones, Dr. Jesse Hostetter, and Dr. Albert van Geelen for helpful discussions during work-in-progress meetings.
- Copyright © 2011 the American Physiological Society