CX3CR1 Is a Receptor for Human Respiratory Syncytial Virus in Cotton Rats

ABSTRACT

Respiratory syncytial virus (RSV) has been reported to use CX3CR1 in vitro as a receptor on cultured primary human airway epithelial cultures. To evaluate CX3CR1 as the receptor for RSV in vivo, we used the cotton rat animal model because of its high permissiveness for RSV infection. Sequencing the cotton rat CX3CR1 gene revealed 91% amino acid similarity to human CX3CR1. Previous work found that RSV binds to CX3CR1 via its attachment glycoprotein (G protein) to infect primary human airway cultures. To determine whether CX3CR1-G protein interaction is necessary for RSV infection, recombinant RSVs containing mutations in the CX3CR1 binding site of the G protein were tested in cotton rats. In contrast to wild-type virus, viral mutants did not grow in the lungs of cotton rats. When RSV was incubated with an antibody blocking the CX3CR1 binding site of G protein and subsequently inoculated intranasally into cotton rats, no virus was found in the lungs 4 days postinfection. In contrast, growth of RSV was not affected after preincubation with heparan sulfate (the receptor for RSV on immortalized cell lines). A reduction in CX3CR1 expression in the cotton rat lung through the use of peptide-conjugated morpholino oligomers led to a 10-fold reduction in RSV titers at day 4 postinfection. In summary, these results indicate that CX3CR1 functions as a receptor for RSV in cotton rats and, in combination with data from human airway epithelial cell cultures, strongly suggest that CX3CR1 is a primary receptor for naturally acquired RSV infection.

IMPORTANCE The knowledge about a virus receptor is useful to better understand the uptake of a virus into a cell and potentially develop antivirals directed against either the receptor molecule on the cell or the receptor-binding protein of the virus. Among a number of potential receptor proteins, human CX3CR1 has been demonstrated to act as a receptor for respiratory syncytial virus (RSV) on human epithelial cells in tissue culture. Here, we report that the cotton rat CX3CR1, which is similar to the human molecule, acts as a receptor in vivo. This study strengthens the argument that CX3CR1 is a receptor molecule for RSV.

INTRODUCTION

Human respiratory syncytial virus (RSV) is the leading cause of hospitalization of infants in the United States and the most important viral cause of acute lower respiratory tract infection in children worldwide (1, 2). In the elderly, the disease burden due to RSV infection is as significant as that of influenza virus infection (3, 4). Vaccines and therapies against RSV are urgently needed, and currently 22 vaccine candidates and therapeutic monoclonal antibodies are in various stages of clinical testing (5). To combat disease after RSV infection, a better understanding of RSV pathogenesis is needed, and further clarifying the identity of the receptor for RSV in vivo would enhance progress toward understanding the mechanisms of RSV pathogenesis.

In vitro, heparan sulfate proteoglycans (HSPG) have been identified as the primary receptors for RSV on immortalized cell lines, but they are not likely to be the receptors in vivo due to the lack of heparan sulfate expression on the apical surface of the respiratory epithelium (6–8) which includes ciliated cells, the target for RSV infection. Several other receptors have been proposed for RSV in vivo, including nucleolin, annexin II, surfactant protein A, and CX3CR1 (9–12). Of these, CX3CR1 seems to be the most probable candidate because its expression pattern in the airways, on the cilia of epithelial cells and the apical surface of ciliated epithelium, matches the tropism of RSV (13–17), and antibodies blocking the CX3C motif on the RSV G protein interfere with binding to CX3CR1 (18) in vitro. In addition, CX3CR1 recently has been shown to be a receptor for RSV on primary, well-differentiated, polarized human airway epithelial (HAE) cell cultures, which is a more physiologically relevant model system than immortalized cell lines for RSV receptor studies (19).

CX3CR1 is a G-coupled seven-transmembrane chemokine receptor that is expressed on various cell types, including endothelial and smooth muscle cells, as well as natural killer cells, dendritic cells, monocytes, and T cells (20, 21). Its only known ligand is the CX3C chemokine (CX3CL1) and the RSV attachment (G) protein which contains a conserved CX3C motif (12, 18). Previous investigations into the utilization of CX3CR1 as an RSV receptor in vivo provided conflicting results. C. Johnson et al. found no difference in the viral titers of RSV-infected wild-type and CX3CR1^−/− mice at 4 days postinfection using quantitative real-time PCR (qRT-PCR) (22). However, S. Johnson et al. reported a statistically significant difference (about half a log PFU/g lung) between wild-type and CX3CR1^−/− mice at 5 days postinfection (19). In order to investigate further the role of CX3CR1 as a receptor for RSV in vivo, we performed infection experiments in the cotton rat (Sigmodon hispidus), which is a more permissive small animal model for RSV infection than the mouse (23, 24).

RESULTS

Determination of cotton rat CX3CR1 sequence and protein expression.

The aim of this study was to characterize the cotton rat homolog of human CX3CR1 and define its function in RSV infection. To estimate CX3CR1 expression on cotton rat cells, various antibodies specific for human and mouse CX3CR1 were tested. As human and mouse CX3CR1 share a 90% amino acid identity (25), it was assumed that similarity to cotton rat CX3CR1 might be similarly high and at least some antibodies would cross-react. A rat monoclonal antibody was found to be most reactive with cotton rat splenocytes as analyzed by flow cytometry. Approximately 20% of the total cell population stained positive for CX3CR1, and approximately 25% of those cells expressed CX3CR1 at high levels (Fig. 1).

FIG 1.

CX3CR1 expression on cotton rat splenocytes. Flow cytometry analysis of cells isolated from cotton rat spleen, stained with rat antibodies specific for either human CX3CR1-AF488 (red) or rat IgG2b-AF488 isotype control (blue). Approximately 20% of total cells stained positive for CX3CR1 after correction for isotype control, and of the CX3CR1⁺ population, approximately 25% were CX3CR1 high.

As these results indicated the expression of a CX3CR1 homolog in cotton rats, we determined the gene sequence by RT-PCR amplification of the cDNA using primers based on regions of nucleotide sequence identity between human and mouse CX3CR1. The resulting ∼600-nt segment was sequenced, and the sequence was used to design additional forward and reverse primers, which enabled amplification of the 5′ and 3′ ends of the gene using the RNA ligase-mediated (RLM-RACE) and oligo-capping rapid amplification of cDNA ends (RACE) methods. The complete open reading frame was cloned into a pCR2.1 plasmid and sequenced.

The CX3CR1 open reading frame of the cotton rat is 79% homologous to human and 88% homologous to mouse CX3CR1. The CX3CR1 gene of human is 80% homologous to the mouse. At the protein level, the identity of cotton rat and human was 91% and the identity between cotton rat and mouse was 95% (Fig. 2). At the protein level, human is 90% identical to mouse. Importantly, the seven amino acids which have been identified as necessary for CX3CR1 to bind its CX3CL1 ligand and for chemoattraction of immune cells, and which are primarily located within the amino terminus and third extracellular loop, are conserved among humans cotton rats and mice (Fig. 2) (26).

FIG 2.

CX3CR1 amino acid sequence similarity between cotton rat, mouse, and human. A multiple sequence amino acid alignment is shown for cotton rat (Sigmodon hispidus), mouse (Mus musculus, GenBank: NP_034117.3), and human (Homo sapiens, GenBank: ABS29268.1) CX3CR1. CX3CR1 is a seven-transmembrane cellular receptor. The alignment was performed using TM-Coffee, a multiple sequence alignment tool specific for transmembrane proteins. Amino acids are represented in bold; capitalized letters are residues that are recognized to be important for CX3CR1 binding or chemoattractant activity to its ligand, CX3CL1. Highlighting: blue indicates extracellular regions of the protein (EXT), pink indicates transmembrane regions (TM), and yellow indicates intracellular regions (INT). An asterisk (*) signifies a position with a single, fully conserved residue. A colon (:) signifies conservation between groups of strongly similar properties (>0.5 in the Gonnet PAM 250 matrix). A period (.) signifies conservation between groups of weakly similar properties (≤0.5 in the Gonnet PAM 250 matrix).

CX3CR1 expression in cotton rat airway epithelial cell cultures correlates with RSV infection.

It has been demonstrated that human CX3CR1 is expressed on the cilia of human epithelial airway cells where it enables the uptake of RSV (16, 19). To compare CX3CR1 expression on the airway epithelium of cotton rats to that of humans, cotton rat airway epithelial (CRAE) cultures were generated and maintained at an air-liquid interface. Cross-sections of these cultures showed well-differentiated, polarized epithelial cells that included ciliated cells (Figure 3A). A rabbit antibody against CX3CR1 identified CX3CR1 on the cilia projecting from the apical surface of the ciliated cells. In the respiratory tract, these cells would face the airway lumen and be exposed to pathogens such as RSV (Figure 3B). After inoculation of CRAE cultures with RSV, virus antigen was detected by antibody staining only in the ciliated epithelial cells (Figure 3C), consistent with RSV using the same receptor molecule to infect the ciliated cells of cotton rats and humans.

FIG 3.

CX3CR1 expression and RSV infection of primary cotton rat airway epithelial (CRAE) cultures. Cross-section of paraffin-embedded CRAE cell monolayer on Transwell support membrane. (A) H&E staining of uninfected cells with cilia clearly visible. (B) α-CX3CR1 staining of uninfected cells displaying CX3CR1 expression on the apical surface and on the cilia, or (C) α-RSV staining of cells at 24 h postinfection with an MOI of 0.01 showing RSV infection of ciliated cells, antibody binding visualized by development with DAB (in brown) and counterstained with hematoxylin. Structures are indicated by arrows. Controls with only secondary antibodies did not show any stainings. Magnification of 40×.

CX3CR1 expression in the cotton rat respiratory tract correlates with RSV infection.

To determine expression of cotton rat CX3CR1, the lung and nose of cotton rats was stained with a CX3CR1-specific antibody by immunohistochemistry (IHC). In the upper respiratory tract, CX3CR1 expression was detected most prominently and consistently on the epithelial cell lining of the nasal turbinates (Figure 4A). In the lung, CX3CR1 expression was detected on ciliated columnar epithelial cells of the bronchi and bronchioles, whereas it was not strongly expressed in alveoli (Figure 4B). Overall, CX3CR1 expression in the cotton rat is analogous to that of humans, occurring in the same cell types as well as in the same regions of the respiratory tract. After infection with both strains of RSV A/2 and wild-type RSV A2001/2-20, viral antigen was detected by IHC in ciliated columnar epithelial cells in both the lungs and nasal turbinates of infected cotton rats. Therefore, viral infection occurred in the same cells which had stained positive for CX3CR1 (Fig. 5).

FIG 4.

CX3CR1 expression on cotton rat respiratory epithelium. In contrast to the cuboidal epithelial cells which line most of the respiratory tract of the mouse, the columnar cells which constitute most of the respiratory epithelium in the cotton rat are of a similar morphology to those of humans. IHC analysis of (A) nasal turbinate and (B) bronchiolar localization of CX3CR1 expression in an uninfected cotton rat. CX3CR1 is detected primarily at the apical surface and on the cilia of respiratory epithelium. Anti-CX3CR1 MAb visualized by development with DAB (3,3'-diaminobenzidine) (in brown) and counterstained with hematoxylin. Magnification of 40×.

FIG 5.

RSV infection of cotton rat respiratory epithelium. Immunohistochemical analysis of nasal turbinate (A, C, E) and bronchiolar (B, D, F) localization of RSV antigen in tissues collected from naive cotton rats (A, B), cotton rats at 4 days postinfection (intranasal) with 10⁵ TCID₅₀ RSV A/2 (C, D), or cotton rats at 4 days postinfection with 10⁵ TCID₅₀ wild-type RSV A2001/2-20 (E, F). RSV-specific polyclonal goat serum visualized by development with DAB (in brown) and counterstained with hematoxylin. Magnification of 100× (A, C, E) or 200× (B, D, F).

CX3CR1 supports RSV infection in a cultured cell line with reduced heparan sulfate expression.

The previous experiments had established a clear correlation between CX3CR1 expression in vitro and in vivo with RSV infection. In order to establish a function for CX3CR1 as a receptor, CHO cells deficient in xylosyltransferase I (CHO-pgsA-745) were transfected with cotton rat CX3CR1. CHO-pgsA-745 cells express reduced amounts of heparan sulfate proteoglycans and therefore reduced infection with RSV. Transfection of CHO-pgsA-745 cells with a plasmid expressing the cotton rat CX3CR1 protein resulted in an increased number of RSV-infected cells (Fig. 6).

FIG 6.

In vitro expression of CX3CR1 increases infection with RSV. CHOpgsA745 cells and CHOpgsA745 cells transfected with cotton rat CX3CR1 were inoculated with a recombinant RSV expressing green fluorescent protein (GFP). At 24 and 48 h after infection, the number of infected cells was significantly increased in CX3CR1 expressing cells with an MOI of 0.001. Multiple t tests. Statistical significance was determined using the Holm-Sidak method, with alpha equal to 0.05. Each row was analyzed individually, without assuming a consistent SD. ***adjusted P < 0.001; ****adjusted P < 0.0001.

An antibody against the RSV G protein (131-2g), but not soluble heparan sulfate, prevents RSV infection in cotton rats.

Binding and entry of RSV into a host cell is dependent on the RSV G (attachment) and F (fusion) glycoproteins. In vitro, RSV with a deletion of the G protein gene is able to replicate in the absence of the G protein by using the F protein for both attachment and fusion, but viral replication is impaired (27). In mice, the RSV G protein is required for efficient viral replication (28). The G protein contains a central conserved and cysteine noose region flanked by two heavily glycosylated regions (29). Within the cysteine noose is a CX3C motif which has been shown to bind to CX3CR1 and a heparin binding domain adjacent to the CX3C motif (12). Antibody 131-2g binds to a conserved epitope near the CX3C motif of the G protein and inhibits interaction with CX3CR1. On HAE cells, preincubation of RSV with antibody 131-2g leads to a decrease in RSV infection (19), whereas preincubation with soluble heparan sulfate has no effect on viral entry. However, on immortalized cell lines which primarily use heparan sulfate proteoglycans as a receptor (30), preincubation of RSV with soluble heparan sulfate inhibits infection and antibody 131-2g does not.

To assess the effect of antibody 131-2g and soluble heparan sulfate on RSV infection in vivo, RSV was preincubated with the 131-2g (20 μg/ml), soluble heparan sulfate (100 μg/ml), or soluble keratan sulfate (100 μg/ml) before being inoculated intranasally into cotton rats. Four days postinoculation, lungs were collected from the animals and viral titers were determined. Lungs from animals which received RSV preincubated with heparan sulfate produced virus titers equivalent to those from samples from control animals which had received RSV preincubated with keratan sulfate (Fig. 7). Lungs from cotton rats inoculated with RSV preincubated with 131-2g antibody did not contain detectable levels of virus, indicating that the CX3C motif interaction with CX3CR1 is necessary for efficient RSV infection in vivo.

FIG 7.

In vivo replication of RSV preincubated with RSV G protein-specific antibody (131-2g) or soluble heparan sulfate. Preincubation of RSV with the antibody 131-2g, which prevents G protein interaction with host CX3CR1, prevents efficient viral replication in cotton rats. Heparan sulfate proteoglycans are the receptor for RSV on immortalized cell lines but not HAE cultures. Preincubating RSV with soluble heparan sulfate before inoculation does not affect viral growth in cotton rats. RSV alone and keratan sulfate (KS) served as controls. Viral titer in lungs measured on day 4 postinfection by TCID₅₀ assay, limit of detection 102 TCID₅₀. Five animals per group, ANOVA P = 0.0052, Dunnett’s multiple-comparison test, **adjusted P < 0.01 compared to RSV.

Decreased CX3CR1 expression reduces RSV infection in vivo.

In order to further confirm the role of CX3CR1 in RSV infection in the cotton rat, expression of the protein was inhibited in vivo by peptide-conjugated phosphorodiamidate morpholino oligomers (PPMOs) (31). PPMOs are water soluble and delivery-enabled DNA-like antisense oligomers. After entering a cell, PPMOs bind to a specific mRNA sequence by base pairing and sterically blocking translation of the mRNA. Two CX3CR1-blocking PPMOs were designed based on the cotton rat CX3CR1 sequence. The first PPMO was specific to the positions 34 to 10 (-34/-10) region of the 5′ UTR, and the second PPMO targeted the first 25 nucleotides after the AUG translational start site. Cotton rats were treated twice intranasally with 5 mg/kg of either CX3CR1-targeting or control (nonspecific) PPMO at 48 and 24 h before RSV inoculation. On day 4 postinoculation, cotton rats treated with the PPMOs demonstrated a nearly 10-fold reduction in virus titer. Animals treated with (-34/-10) PPMOs had an 84% reduction of viral titer in the lungs (Figure 8A) and an 89% reduction of viral titer in the nasal turbinates (Figure 8B). The AUG PPMO treatment group exhibited a 75% reduction of viral titer in the lungs (Figure 8A) and a 74% reduction of viral titer in the nasal turbinates (Figure 8B).

FIG 8.

Decreased CX3CR1 expression leads to decreased RSV titer in vivo. Cotton rats treated with peptide-conjugated phosphorodiamidate morpholino oligomers (PPMOs) against CX3CR1 have lower viral titers at day 4 postinfection in (A) lungs and (B) nasal turbinates. Intranasal treatment with 5 mg/kg/day of PPMO was administered at 48 and 24 h before intranasal inoculation of 10⁵ TCID₅₀ RSV A/2. PPMOs targeted to CX3CR1 mRNA at the first 25 bases including the start codon (AUG), upstream of the start codon from positions 34 to 10 (-34/-10), or nonspecific control PPMO (control). Viral titer was measured on day 4 postinfection by TCID₅₀ assay, limit of detection 10² TCID₅₀. Four animals per group, one-way unpaired t test *adjusted P < 0.05, **adjusted P < 0.01 compared to control.

The CX3C motif in the RSV G protein is important for infection in vivo.

These data indicated the importance of CX3CR1 as a receptor and of G protein as a receptor-binding protein in cotton rats. To specifically test the role of the CX3C motif in the neck of the G protein cysteine noose (12), cotton rats were inoculated with RSV mutants deleted for the G protein, RSV-ΔG, or RSV containing mutations in the CX3C motif. RSV C186S contains a mutation of the last cysteine of the CX3C motif. RSV C173/176S contains two mutations, C173S and C176S, which although not directly mutations of the CX3C motif, prevent the formation of the cysteine noose by removing their partner cysteines and likely disrupting the structure of the CX3C motif. RSV CX4C has an additional amino acid (alanine) inserted between the two cysteines of the CX3C motif.

Four days after infection with these virus mutants, cotton rat lungs were harvested for the determination of viral load. Whereas RSV expressing the wild-type G protein grew well in cotton rats, none of the RSV mutants did (Fig. 9). These results demonstrate the importance of the CX3C motif for RSV infection in vivo.

FIG 9.

RSV G protein mutants do not replicate in vivo. RSVs with mutations in the G protein were tested in cotton rats for their in vivo viability; wild-type RSV A/2 (RSV), G-deletion (RSV-ΔG), C173S and C176S mutations (no cysteine noose [C173/176S]), mutation of second C of CX3C motif (C186S), and insertion of additional amino acid between the two Cs of CX3C motif (CX4C). Viral titer in lungs measured on day four postinfection by TCID₅₀ assay, limit of detection 102 TCID₅₀. Four animals per group, ANOVA P < 0.0001, Dunnett’s multiple-comparison test, ****adjusted P < 0.0001 compared to RSV.

DISCUSSION

Many cellular molecules have been described as potential receptor candidates for RSV (9–12, 32, 33). Although for all of them some interaction between these molecules and either RSV F or RSV G protein could be demonstrated, most failed a more rigorous analysis of in vitro infection and typically lacked in vivo data. Receptor candidates for RSV should not exclusively be tested on permanent cell lines because heparan sulfate proteoglycans, present on all cell lines, act as a receptor in vitro but not in vivo, leading to irrelevant results. The best in vitro models for testing receptor usage are currently human epithelial airway cultures which should express the natural in vivo receptor, like human epithelial cells in situ. Using this in vitro model, Johnson et al. demonstrated that CX3CR1 functions as a receptor for RSV on well-differentiated primary human airway epithelial cultures (19). Further testing in mice with a gene deletion in CX3CR1 demonstrated a small reduction in viral titers, which differed from a previous study by another group.

A possible explanation for the low or lack of effect of deleting CX3CR1 in mice could be the low replication of RSV typically seen in mice (23). To overcome this restriction, the cotton rat model was chosen to investigate CX3CR1 as an in vivo receptor for RSV. After intranasal inoculation, the cotton rat is more permissive to RSV infection than is the mouse (23). This is possibly due to intrinsic differences in cellular permissivity between the two species, which is a phenomenon that has been observed with measles virus and its receptor, CD150 (34).

The cotton rat model provides high predictive power for both the effectiveness and required dosage of RSV treatments in humans (23). RSV clinical isolates grow well in the cotton rat without the need for adaptation (personal observation and references 35–38). Another advantage of the cotton rat model is the similarity of the lung cell composition between the cotton rat and humans. In contrast to that in adult mice, in cotton rats and humans, the primary cell type in bronchial epithelium is a ciliated columnar cell (39). As a consequence, RSV infection in cotton rats and humans is typically localized to bronchial tissue, whereas in the mouse, infection is primarily in the alveoli (39).

We found that the CX3CR1 sequence similarity between human and cotton rat or mouse is similar, 91% and 90%, respectively. The CX3CR1 of both cotton rat and mouse share the conserved amino acids at residues known to be important for binding of CX3CL1 and chemotactic activity, namely, human E13-D16, D25, E254, and D266 (26). However, it is possible that one or more other amino acids in the human CX3CR1 are important for RSV G protein binding and that the cotton rat protein has this amino acid(s) but the mouse does not.

Similar to the expression of CX3CR1 on HAE cultures (16), cotton rat CX3CR1 was found on the cilia of the ciliated epithelial cells in CRAE cultures and in situ in the respiratory tract, and its expression correlated with RSV infection. Transient expression of the CX3CR1 in a cell line deficient in heparin sulfate proteoglycan and inhibition of RSV infection of CRAE cultures by a neutralizing antibody specific for the CX3C region of the G protein both demonstrated the importance of the CX3C region as a receptor in RSV infection.

CX3CR1 is highly expressed on the respiratory epithelium in culture, but its role in infection cannot yet be tested in vivo by knockout in genetically modified cotton rats. As an alternative, we used CX3CR1-specific morpholinos to interfere with the in vivo expression of the CX3CR1 protein. These experiments provided further evidence for the role of CX3CR1 as a receptor for RSV, similar to the complementary experiment of using viruses with mutations in their G protein which abolish or severely reduce G protein interaction with human CX3CR1. The latter data are in agreement with studies of Ha et al. (40). These authors found a decrease in virus replication after infection of an RSV with a mutation (CX4C) in the CX3CR1 binding site, further supporting the role of cotton rat CX3CR1 as RSV receptor.

Here, using the most permissive small animal model for RSV infection, the cotton rat, we determined that RSV’s CX3CR1 gene has conserved the amino acids known to be required for binding to its ligand, CX3CL1, demonstrated that RSV with mutations in the CX3C site of the G protein cannot replicate in the cotton rat, found CX3CR1 on the cilia of primary airway epithelial cultures derived from cotton rats, and found that specific in vivo inhibition of CX3CR1 expression inhibits RSV infection in vivo. Based on these results, we conclude that CX3CR1 is an important receptor for RSV in the cotton rat, consistent with previous reports (14, 16, 19) implicating CX3CR1 as a receptor for RSV on primary human airway epithelial cells.

MATERIALS AND METHODS

Cotton rats.

Inbred cotton rats (Sigmodon hispidus) of 4 to 8 weeks of age were purchased from Envigo (Indianapolis, IN) and maintained in standard polycarbonate cages in a BSL-2 barrier system. Environmental conditions were held at 22 ± 2°C, 30 to 70% relative humidity, and 12 h light/12 h dark cycle, and food and water were provided ad libitum. Euthanasia was performed via CO₂ inhalation. Inoculations were performed intranasally on isoflurane-anesthetized animals. All animal experiments were approved by the Institutional Animal Care and Use Committee of The Ohio State University.

Intranasal infection of cotton rats.

Animals inoculated with RSVs received a dose of 10⁵ 50% tissue culture infective dose (TCID₅₀) of virus in 100 μl phosphate-buffered saline (PBS) distributed evenly between the nares. Treatment groups consisted of 4 to 5 cotton rats per group. For experiments with heparan sulfate, keratan sulfate, and monoclonal antibody 131-2g (gifted from Mark Peeples at Nationwide Children’s Hospital Research Institute), RSV was incubated with either heparan or keratan sulfate (Sigma-Aldrich) at a concentration of 100 μg/ml or with antibody 131-2g at a concentration of 20 μg/ml in PBS at room temperature for 30 min before inoculation. Tissues (lung and nasal turbinates) were homogenized in PBS and the titer was determined after storage at −80°C. Virus growth in vivo was determined by TCID₅₀ assay of lung or nasal turbinate homogenates. In brief, 10⁵ HEp-2 cells/well of a 48-well plate were plated overnight and infected with 10-fold serial dilutions for 1 h. The virus solution was replaced with minimal essential medium (MEM)/2% fetal calf serum and virus-positive wells were scored for cytopathic effect 4 days later and calculated according to Reed (41).

Viruses and cells.

Stocks of RSV A/2 and wild-type RSV A2001/2-20 (obtained from BEI Resources, NIAID, NIH) were grown on HEp-2 cells in MEM/2% fetal bovine serum (FBS), pelleted, and resuspended in MEM/10% trehalose. Virus was titrated by 50% tissue culture infectious dose (TCID₅₀) assay on HEp-2 cells. RSV A2001/2-20 was isolated from a nasal wash from an infant with RSV bronchiolitis in Nashville, Tennessee on February 20, 2001. The following recombinant RSV mutants were used: RSV-SF is a virus with a deletion of the G protein gene (RSV-ΔG) (27); RSV-C173/176S contains two amino acid substitutions in the G protein, C173S and C176S, resulting in the loss of the cysteine noose structure (normally present due to the disulfide bonds of C173 with C182 and C176 with C186), and the RSV-C186 mutant contains a single amino acid substitution in the G protein at position 186 (which, in addition to being the fourth cysteine in the noose structure, is the second cysteine in the CX3C motif) (28). The RSV-CX4C mutant contains an alanine insertion at position 186 resulting in the loss of the CX3C motif (42).

Construction of recombinant RSV expressing red fluorescent (tdTomato) protein (rrRSV131).

Recombinant RSV expressing red fluorescent (tdTomato) protein (rrRSV131) was constructed by replacing the green fluorescent protein gene in the full-length RSV cDNA clone MP224 (43), cloned into the XmaI site between the leader and the NS1 gene. The tdTomato gene was synthesized (Integrated DNA Technologies), preceded and followed by the N gene start and end sequences. The cDNA was also modified by the addition of restriction sites in the intergenic regions between M/SH (PvuI), SH/G (SacI, ApaI), G/F (SacII), and F/M2 (XhoI). The long 3′ untranslated region of the SH gene was also deleted. The rrRSV131 virus was rescued by transfecting the full-length HC131 cDNA plasmid, along with plasmids encoding the RSV N, P, L, and M2-1 proteins, into HeLa cells as described previously (44). These cultures were simultaneously infected with modified recombinant vaccinia virus expressing T7 RNA polymerase (MVA-T7) which drives transcription of each plasmid (45). The resulting virus, rrRSV131, was propagated in HeLa cells grown in Dulbecco’s modifed Eagle’s medium (DMEM) supplemented with 5% FBS and passaged four times to reach a titer of 10⁷ PFU/ml.

Generation of cotton rat airway epithelial (CRAE) cell cultures.

Cotton rat airway epithelial (CRAE) cell cultures were generated as described with the exception that cotton rat tracheas were used in place of human respiratory tissue (46). Briefly, cells from the trachea of cotton rats were grown on collagen-coated Transwell inserts (Corning, Inc., Corning, NY) until reaching confluence and forming tight junctions, at which time the upper medium was drawn off, and the cultures were maintained at the air-liquid interface to form differentiated, polarized epithelial cell cultures.

Cloning and sequencing of cotton rat CX3CR1.

Cotton rat CX3CR1 was first amplified from cotton rat spleen and thymus cDNA (SuperScript first-strand synthesis kit, Invitrogen) by PCR using primers designed by comparing sequence homology between human, rat, and mouse: fwd, 5′-GTCCTCGCCCTCACCAACAGC-3′, rev. 5′-TTGAGAGTCTCCAGGAAAATCA-3′. The resulting partial gene sequence of cotton rat CX3CR1 was used to design primers for full-length, RNA ligase-mediated rapid amplification of 5′ and 3′ cDNA ends (GeneRacer kit, Invitrogen). The sequence of the forward primer was 5′-C CTCGTGGTGGTGGTCTTCTTCCTCTT-3′, and the sequence of the reverse primer was 5′-TCACTCAAGGCCAGGTTCAGGAGGTAGA-3′. The GeneRacer kit, along with SuperScript III RT (Invitrogen), was used per the manufacturer’s instructions, and the complete coding sequence of CX3CR1 was obtained. CX3CR1 was cloned into pCR2.1-TOPO plasmid (Invitrogen) using the forward primer 5′-TCACCATGCCCACTTTCTTCCCTGG-3′ and the reverse primer 5′-GCCTTTCAGAGCAGAATAGACCCCTC-3′ per the manufacturer’s instructions. Correct placement and orientation of CX3CR1 were confirmed by sequencing the plasmid with the included M13 primers. CX3CR1 was inserted into pcDNA3.1 mammalian lentivirus-transfected cells expression vector (Invitrogen) from pCR2.1 by restriction enzyme digest with BamHI and XhoI and confirmed by sequencing of the plasmid.

Needle pairwise sequence alignment tool (EMBL-EBI, EMBOSS) was used to calculate amino acid and nucleotide similarity/homology between CX3CR1 of cotton rat, human (NCBI version numbers: mRNA EU006531.1, protein ABS29268.1), and mouse (NCBI version numbers: mRNA NM_009987.4, protein NP_034117.3). Multiple sequence alignment was performed using Position-Specific Iterative/TransMembrane-Coffee (PSI/TM-Coffee, version 11.00) (47). Translation of cotton rat CX3CR1 cDNA sequence to a protein sequence was performed using ExPASy Translate tool at the Bioinformatics Resource Portal (Swiss Institute of Bioinformatics).

Flow cytometry.

Cotton rat splenocytes were prepared by mechanical disruption of spleen through a 70-μm cell strainer (BD, Franklin Lakes, NJ). Cells were stained with 10 μg/ml of either rat α-human CX3CR1-AF488 (D070-A48 MBL) or isotype control (M090-A48 MBL). Mean fluorescence was measured by FACSCalibur, and FlowingSoftware (version 2.5.1, Turku, Finland) was used for data analysis.

Immunohistochemistry.

Tissues were collected immediately after CO₂ asphyxiation. Samples stained for RSV were collected at 24 h (CRAE cells) or 4 days (lungs and nasal turbinates) postinfection. Nasal cavities, lungs, and CRAE cell cultures were fixed in freshly prepared 4% paraformaldehyde for 48 to72 h. Cotton rat noses were decalcified in 0.35 M EDTA solution. Tissues were processed, paraffin-embedded, and sectioned for staining. Stainings included routine H&E staining, staining with rabbit α-human CX3CR1 antibody (cross-reactive with rat) diluted 1:1,000 (ab8021 Abcam), or, for RSV-infected samples, staining with goat α-RSV antiserum diluted 1:1,000 (0601, ViroStat Portland, ME). After cells were rinsed with wash buffer (Dako catalog number S3006), biotinylated secondary antibodies were applied for 30 min at room temperature. The detection signal was amplified by using a standard avidin-biotin complex kit (Vector Laboratories, Burlingame, CA), and the reaction was developed by using 3,3′-diaminobenzidine.

Peptide-conjugated phosphorodiamidate morpholino oligomers (PPMOs).

Two translation-blocking PPMOs were designed from the cotton rat sequence of CX3CR1: “CX3CR1(-34/-10),” sequence 5′-CTGGGCAGATAGGAACTGAACTTCT-3′, and “CX3CR1-AUG,” sequence 5′-CCAGTCCAGGGAAGAAAGTGGG[CAT]-3′. The sequences are complementary to the CX3CR1 mRNA. As a control, a nonspecific PPMO was used, “NC705,” sequence 5′-CCTCTTACCTCAGTTACAATTTATA-3′. Phosphorodiamidate morpholino oligomers (PMOs) (Gene Tools, LLC Philomath, OR) were conjugated to a cell-penetrating peptide, to enhance uptake by the cell (48). For inoculation into cotton rats, the PMOs were diluted in PBS and administered intranasally in 100 μl. Each cotton rat received 5 mg/kg body weight at 48 and 24 h before challenge with 10⁵ TCID₅₀ RSV A/2 in a volume of 100 μl PBS.

Statistical analysis.

Statistical analysis was performed using GraphPad Prism version 7.00 (GraphPad Software, La Jolla, CA, USA). Data were analyzed by analysis of variance (ANOVA), unless noted otherwise. Dunnett’s test was used to perform multiple comparisons of experimental groups to the control group. Adjusted P value of <0.05 was considered statistically significant. All data are presented as mean ± standard deviation (SD).

Data availability.

The nucleotide sequence of the cotton rat CX3CR1 coding sequence is available from GenBank under accession number MZ254651.