Background
Non-influenza viral respiratory illnesses cost the US an estimated $39.5 billion annually from both healthcare costs and loss of productivity [
1]. Rhinovirus (RV) is the most common cause of viral upper respiratory illness resulting in rhinitis, sinusitis, pharyngitis, or otitis media, and can lead to the development of bacterial superinfections [
2‐
7]. While most individuals only experience mild symptoms during a RV infection, children, the elderly, the immunosuppressed, and those with asthma, COPD, or cystic fibrosis are predisposed to lower respiratory tract illnesses including wheezing, asthma exacerbations, and respiratory distress that can often result in hospitalization [
8‐
14]. RV has also been implicated in the initiation of asthma, as the presence of wheeze during acute RV infection in the first few years of life is strongly correlated with the development of asthma later in life [
15‐
19]. Three RV species (A, B and C) have been identified, and recent studies have established a link between RV species and illness severity [
14,
20]. The RV-C, discovered in 2006, appear to instigate more severe upper and lower respiratory tract symptoms in infants and children under 5 years of age [
14,
20]. In an effort to explain this phenomenon, Nakagome et al. inoculated airway epithelial cells with multiple members of each RV species, and observed the highest levels of viral replication, cytotoxicity, and release of inflammatory cytokines following inoculation with either RV-A or C [
21].
Virus interactions with the host epithelium are an important determinant of illness severity, and the initial step in this interplay is binding to the target cell via a cell entry receptor. Consider, for example, that cellular retargeting of avian influenza virus in human tissue from ciliated to secretory cells is necessary to produce severe and clinically apparent disease, and also that a broadening of the human coronavirus tropism to alveolar pneumocytes contributed to the Severe Acute Respiratory Syndrome and Middle East Respiratory Syndrome pandemics [
22‐
30]. For RVs, the cellular targets of the RV-A and B have been studied with varying results [
31‐
35]. RV-B14 replication and expression of the major group RV receptor ICAM-1 have been identified in non-ciliated cells derived from the palatine tonsils, and high levels of RV-A16 (A16) replication were noted in cultures with mucus cell metaplasia, which can occur in asthma or COPD [
31,
32,
34]. Significant ICAM-1 expression and A16 binding were also observed on basal epithelial cells [
33,
36]. More recently, Jakiela et al. contradicted previous studies by identifying ciliated cells as the primary target for A16, and further indicated mucus cell metaplasia as a protective factor against A16 infection [
34,
35]. Regarding the RV-C, Bochkov et al. demonstrated the presence of viral RNA in sinus mucosal tissue organ cultures, and subsequently identified cadherin-related family member 3 (CDHR3) as the first and only known RV-C entry factor. However, the specific host cell for the RV-C remains unknown [
37,
38].
We therefore performed a series of experiments to identify the cellular target of the RV-C. Our primary hypotheses were that the RV-C target ciliated airway epithelial cells (AECs), and that infectivity was restricted to cells expressing CDHR3.
Methods
Viruses
RV-C15 (C15) and RV-C15-GFP suspensions were prepared from the pC15-Rz and pC15-GFP plasmids by reverse genetics as previously described [
21,
37‐
39].
Cell culture
Primary human bronchial epithelial cells (HBECs) were extracted from trachea or main stem bronchus trimmings of healthy donor lungs that were transplanted and provided by the University of Wisconsin Department of Surgery, Division of Transplantation and grown by the air–liquid interface (ALI) culture method as described previously [
40,
41]. For some experiments, we used a second culture medium (PneumaCult™, Stem Cell Technologies, Vancouver, BC, Canada) after preliminary experiments demonstrated improved differentiation and susceptibility to RV-C infection. CDHR3 expression was quantified by PCR as previously described [
38].
ALI infections
Basal medium was aspirated from 30–50 day-old ALI cultures that were then washed apically 5x and incubated with 100 μl of apical BEGM containing 1 x 107 (2 x 106 for gene expression and western blot analyses) plaque forming unit equivalents (PFUe) of C15 (3.5 h, 34 °C). The inoculum was then aspirated, basal ALI media replenished, and cultures were incubated for an additional 14.5 h.
Immunofluorescence
Following overnight incubation with C15, ALI cultures were washed 3x apically and basally with PBS and fixed with ice cold 4% PFA (15 min, RT). Membrane inserts were cut in half and removed from transwell plates and placed into a 24-well plate for staining. Cells were permeabilized in 0.3% (v/v) Triton-X100 (10 min, RT), blocked in 0.5% nonfat dry milk in PBST (1 h, RT), and incubated in primary antibody (1:200 in blocking buffer, 1 h, RT), secondary antibody (1:250 in PBST, 1 h, RT) and Syto-13 (0.5uM, 10 min, RT, Life Technologies, Grand Island, NY). Inserts were washed in between staining with PBST, and mounted with ProLong® Gold Antifade Reagent (Life Technologies, Grand Island, NY). Cells were imaged on an Olympus 1X71 fluorescent microscope with a Q imaging Retiga2000R camera, a Nikon Eclipse Ti fluorescent microscope, or Nikon C1 laser scanning confocal microscope (Chiyoda, Tokyo, Japan) with a 60x oil immersion objective. Analysis of digitized images was performed with FIJI/Image J version 1.49 h (NIH, Bethesda, MD).
Immunohistochemistry
Differentiated cell cultures were fixed with 10% normal-buffer formalin, and embedded in paraffin (University of Wisconsin Histology Lab, Madison, WI). Five μm sections were adhered to slides which were deparaffinized and then rehydrated. For antigen retrieval, slides were incubated with proteinase K (40 μg/mL in PBS, 10 min, 37 °C). Peroxidases were blocked (5 min, RT) with Peroxidazed 1 (Biocare Medical, Concord, CA). Slides were blocked (3% FBS, 2% goat serum, 0.2% Tween-20, 1.25% Human BD Fc Block™, 1 h, RT), incubated (1:200 in blocking buffer, 2 h, RT) with anti-C15-VP2 mouse monoclonal antibody (kindly provided by MedImmune Inc., Gaithersberg MD), Mach 4 Universal Probe and then Polymer (15 min, RT each, Biocare Medical, Concord, CA), Betazoid DAB (5 min, RT, Biocare Medical, Concord, CA), and counterstained with CAT hematoxylin or eosin (30s, RT, Biocare Medical, Concord, CA). Images from labeled slides were acquired and analyzed using an Olympus BX60 light microscope with DP Controller and Manager software (Shinjuku-ku, Tokyo, Japan).
Flow cytometry
Basal medium was removed from each well, followed by three washes in calcium-and-magnesium-free-PBS (CMF-PBS) apically, and basally. Cells were trypsinized (200 μl apical, 800 μl basal, 8 min, 37 °C) and suspended vigorously with FBS (200 μl, apical), followed by centrifugation (700 x g, 5 min) and decanting. Samples were treated with 0.1% (v/v) Ghost Dye™ Red 780 (Tonbo Biosciences, San Diego, CA, 20 min, on ice), MeOH (15 min, −20 °C), 0.3% Triton-X100 (10 min, RT) in CMF-PBS, prior to blocking (1 h, RT) in 10% FBS, 0.05% Tween-20, and 1.25% Human BD Fc Block™ (BD Biosciences, San Jose, CA). The samples were then incubated with a first set of primary (1:200, 1 h, RT, in blocking buffer), and secondary (1:1000, 1 h, RT) antibodies, and the second set of primary (1:200, 30 min, RT) and secondary (1:1000, 30 min, RT) antibodies (in blocking buffer). Samples were washed in between all antibody steps (3x, 700 x g, 5 min). Primary antibodies were mouse anti-C15-VP2 (MedImmune, Gaithersburg, MD), mouse anti-FLJ23834 (anti-CDHR3), rabbit anti-acetylated-alpha-tubulin, rabbit anti-Muc5AC, mouse IgG1 isotype, and mouse IgG2b isotype (AbCam, Cambridge, MA). Secondary antibodies (Alexa Fluor 350, Alexa Fluor 568, Alexa Fluor 647) and wheat germ agglutinin (Alexa Fluor 350-conjugated) were from Life Technologies (Grand Island, NY). Data from labelled cells were acquired on a Fortessa (BD Biosciences) that was calibrated using Rainbow Fluorescent Particles (RFP-30-5A, Spherotech, Lake Forest, IL) and analyzed with FlowJo software version 10 (Tree Star, San Carlos, CA). For analysis, we normalized our median fluorescence intensity of CDHR3 (MFICDHR3) data to the double-negative (nonciliated, CDHR3-) population in each experiment to obtain the relative MFICDHR3 (rMFICDHR3).
Western blot
ALI cells were lysed with 2X SDS buffer and proteins were denatured by boiling at 95 °C for 5 min. Then, 15 μL of protein samples were loaded onto mini-Protean TGX gels and protein bands were transferred to PVDF membrane and blocked with 3% non-fat dry milk in TBST. Primary and secondary antibodies were as follows: anti-CDHR3 polyclonal antibody (1:1000, Sigma HPA011218) and anti-rabbit IgG-peroxidase (Sigma A6154, 1:5000) and the substract was SuperSignal West Femto Maximum Sensitivity chemiluminescent substrate (Thermo Scientific, 34095).
Statistics
Data were analyzed using SigmaPlot version 11.0 (Systat Software, Inc., San Jose, CA). One-way Repeated Measures ANOVAs were used to compare three or more groups, and square-root-transformed data was used to analyze data from PneumaCult™-differentiated cultures.
Discussion
Since their discovery in 2006, the RV-C have been identified as important contributors to severe respiratory illness in infants and exacerbations of chronic airway disease, yet little is known about how they interact with airway cells. Here, we demonstrated that among epithelial cells, CDHR3 protein expression in ALI cells is developmentally regulated and largely confined to ciliated cells. Several techniques were used to show that RV-C replicate in ciliated AECs, but not secretory cells. Finally, during the differentiation process, RV-C binding and especially replication were closely correlated to temporal patterns of CDHR3 induction. These findings show evidence of a close relationship between ciliated epithelial cells, CDHR3 expression, and susceptibility to RV-C infection.
Ciliated AECs are targets for multiple RNA viruses (e.g. avian influenza virus, coronavirus, RSV, parainfluenza virus, and RV-A), and infections thereof have been correlated to viral receptor expression [
22,
27,
28,
35,
44‐
49]. CDHR3 genotype was initially linked to the risk of childhood asthma with acute severe exacerbations, and subsequent studies showed that transfection of HeLa cells with CDHR3 rendered the cells permissive for RV-C binding and replication [
50]. Furthermore, the asthma-risk genotype of CDHR3 conferred increased RV-C binding and replication [
38]. Our data extend observations into primary AEC, and demonstrate that CDHR3 expression was high and largely restricted to ciliated cells. Notably, RV-C does not infect secretory cells, which can be overrepresented in chronically inflamed airways [
35,
51]. These findings were consistent for cells differentiated in two different culture media (BEGM and PCM). We found that BEGM-cultures have larger proportions of ciliated cells, but only a small fraction of these cells could be infected with RV-C. Cultures differentiated in PCM had fewer ciliated cells, together with an increased representation of other differentiated cell types (goblet, clara, basal, etc.), more closely mimicking airway biology [
43]. A larger fraction of ciliated cells cultured in PCM could be infected with RV-C, however, the majority of ciliated cells were not infected in either culture system, despite an excess of virus. This finding may indicate the presence of unidentified cofactors or specific cellular properties required for RV-C binding or entry. Many viruses either utilize co-receptors or have cofactors that facilitate cellular binding. Poliovirus, for example, utilizes bacterial polysaccharides to increase the kinetics of binding to its host cells in the gut [
52,
53]. Alternately, some cells may be protected from RV-C infection by overlying mucus, or perhaps CDHR3 exposure to RV-C is reduced by sequestration to lateral membranes (a common location for intracellular cadherins) [
54].
RV-C infection led to changes in cell composition, including a reduction in CDHR3 expression and increased percentage of ciliated cells. The reduction in CDHR3 staining could be due to viral shutoff of host protein synthesis, CDHR3 down-regulation following internalization of the virus/receptor complex, or lysis of infected cells. The apparent increase in ciliated cells following C15-inoculation may occur through infection-induced differentiation from basal cells, or perhaps through transdifferentiation of goblet or clara cells to ciliated cells, a concept proposed in previous studies of airway epithelia and infection [
55].
While our data indicate that CDHR3 expression is necessary for RV-C susceptibility in primary airway epithelial cells, other factors may promote or inhibit RV binding or entry. For example, A16 has been shown to primarily target ciliated cells, even though ICAM-1 is also expressed on some nonciliated cells [
35]. Notably, virus-sized (40 nm) nanoparticles adhere to motile cilia in vitro, demonstrating a potential nonspecific mechanism for viral adhesion to ciliated cells [
56]. These findings raise the possibility that RV may adhere to cilia to facilitate cell entry, and this is consistent with our observations that cilia of infected cells stained positive for C15 (Fig.
3a).
The speckled pattern of C15 infection of AECs is consistent with what has been observed for A16 [
35]. The shedding of infected cells, perhaps before the virus is able to spread to surrounding cells, could contribute to the patchy distribution of infection. The resulting gaps in the epithelium would likely impair barrier function and perhaps facilitate bacterial superinfection [
2,
4,
5,
7]. These effects have been noted following influenza A infections in vitro [
57]. It will be of interest in future studies to determine whether CDHR3 genotype or disease states influence patterns of CDHR3 expression and RV-C infection.
Strengths of these studies included the use of viruses that were cloned from clinical isolates, and the use of primary cultures of airway epithelial cells differentiated at ALI. There are also limitations that should be considered in interpreting these findings. First, the only available anti-CDHR3 monoclonal antibody binds to an intracellular epitope, which prevented a discrete quantification of CDHR3 displayed on the cell surface, where it is accessible to virus. Secondly, our experiments involved cultured cells, and additional studies are needed to map the distribution of RV-C infection in vivo.
Acknowledgments
We thank the Halberg, Sherer, and UWCCC Flow labs for technical assistance.