|Home | About | Journals | Submit | Contact Us | Français|
Y.A.B. designed and performed experiments, analyzed data and was the principal author of the paper; A.C.P. performed sequence alignments, statistical receptor footprint analysis and contributed to writing; W.-M.L. performed partial sequencing of clinical isolates, constructed pW10-2R, designed antiviral compound experiments and provided purified HRVs; J.A.R. determined complete genome sequence of HRV-C15; S.P.A. assisted with virus inhibition experiments; X.S. designed and assisted with in situ hybridization experiments; T.R.P. assisted with establishment of the sinus organ culture; N.N.J. and S.B.L. analyzed data and contributed to writing; J.E.G. designed the project, analyzed data and contributed to writing.
A recently recognized human rhinovirus species C (HRV-C) is associated with up to half of HRV infections in young children. We for the first time propagated two HRV-C isolates ex vivo in organ culture of nasal epithelial cells, sequenced a new C15 isolate, and developed the first reverse genetics system for HRV-C. Using contact points for the known HRV receptors, intercellular adhesion molecule 1 (ICAM-1) and low density lipoprotein receptor (LDLR), inter- and intraspecies footprint analysis predicted a unique cell attachment site for HRV-Cs. Antibodies directed to binding sites for HRV-A and -B failed to inhibit HRV-C attachment, consistent with the alternative receptor footprint. HRV-A and -B infected HeLa and WisL cells, but HRV-C did not. However, HRV-C RNA synthesized in vitro and transfected into both cell types resulted in cytopathic effect and recovery of functional virus, indicating that the viral attachment mechanism is a primary distinguishing feature of HRV-C.
Human rhinoviruses (HRV) are positive-strand RNA viruses in the family Picornaviridae. They share a common genome organization, and isolates are typically classified into one of three species (HRV-A, B and C) according to phylogenetic sequence criteria1. HRV is the most frequent cause of the common cold, but recent data have shown that HRV infection has pathogenic potential beyond rhinosinusitis. It is now recognized that 50-85% of asthma exacerbations are due to HRV infections2–4, and that wheezing illnesses in infancy caused by HRV are associated with a high risk of developing childhood asthma5. HRV infections of all types are significant contributors to morbidity associated with exacerbations of other chronic lung diseases such as cystic fibrosis and chronic obstructive pulmonary disease6,7.
Recently, partial or full genome sequencing has been utilized to place clinical isolates within the 99 types of HRV-A and -B. However, cases began to appear, often associated with children hospitalized with lower respiratory tract disease, where the identified HRV sequences were not consistent with those of known HRV types8–14. Full genome sequencing has shown the distinct nature of these HRVs, which are now designated as the HRV-C1,15,16. The pathogenesis and antigenic variability associated with HRV-C, the mechanism of viral attachment to host cells, and potential treatments, have been elusive because unlike HRV-A and B, HRV-C has not been successfully propagated in any cell type in vitro17.
To address these fundamental questions about HRV-C biology, it is first necessary to develop a functional culture system. We now report the characterization and complete genome sequencing of an HRV-C isolate (HRV-C15 type) that can be grown in mucosal organ culture. Bioinformatic comparisons of this new genome and other HRV-C sequences showed unique composition profiles in the putative intercellular adhesion molecule-1 (ICAM-1) and low density lipoprotein receptor (LDLR) footprint locales, inconsistent with all known major and minor group virus-receptor interactions. Full-length HRV-C15 RNA synthesized in vitro from a cDNA clone was shown to be infectious when transfected into cultured cell lines. We have used this methodology to initiate descriptions of the virus biology, tissue tropism, the replication cycle, and response to antiviral and anti-receptor compounds.
We introduced several specimens of nasal lavage fluid (NLF) with high copy number of HRV-C RNA into multiple cell lines (WI-38, WisL, BEAS-2B, A549, HeLa) susceptible to HRV-A and HRV-B infection, and primary cultures of lung fibroblasts and bronchial, sinus and adenoidal epithelial cells from several donors. The latter included cell monolayers, differentiated cells (air-liquid interface method), and primary bronchial epithelial (PBE) cells treated with interleukin 13 to induce goblet cell metaplasia. Assigning an increase in the quantitative (q) RT-PCR signal as an indicator of HRV-C replication, none of these attempts were successful, including those where multiple serial blind passages were undertaken.
Reasoning that organ culture might best mimic natural replication circumstances, surgery-derived byproduct tissues (multiple anonymous donors) from sinus mucosa, adenoids, tonsils and nasal polyps, were cultured. After 1–3 days, viable tissues (continuous ciliary activity) negative for common respiratory viral pathogens were inoculated with representative laboratory HRV-A strains that utilize ICAM-1 (A16) or the LDLR (A1) for cell attachment. Sinus mucosal cultures inoculated with HRV-A16 (109 viral [v] RNA copies ml–1) increased viral RNA by up to 100-fold, even at low (108 vRNA copies ml–1) dose (Fig. 1a). Repeated tests with tonsils and nasal polyps (data not shown) and adenoid tissues (Fig. 1b) exhibited lower (≤ 5-fold) levels of virus replication. Furthermore, a NLF specimen containing 2 × 108 vRNA copies ml–1 of a field HRV-A isolate (A78) responded similar to laboratory strains and was passaged serially into three tissue specimens (Fig. 1c).
Using this approach, a HRV-C15 isolate from a specimen of NLF was inoculated (2.6 × 108 vRNA copies ml-1) into sinus organ culture, and high level viral replication was maintained through seven serial passages (Fig. 1d). During each passage, partial inhibition of ciliary beating was observed at 48–72 h, coinciding with the peak of virus replication, but no other specific cytopathic effects were noted. When similar doses (109 vRNA copies ml-1) of HRV-C15 (Fig. 1e, d11) and HRV-A16 (Fig. 1a, high dose) were used for infection, the growth kinetics were equivalent. Serial passage of a second HRV-C isolate (W23)10 was also accomplished using a similar approach (data not shown). Our attempts to adapt these isolates grown in organ culture to multiple cell lines (listed above) were not successful.
HRV-C15 replication in sinus tissues was analyzed by in situ hybridization using whole mount procedures and a digoxigenin-labeled RNA probe specific for the HRV-C15 plus-strand RNA. Positive hybridization was observed in both of the inoculated sinus organ cultures (Fig. 2a, center and right), but was absent from the mock-infected cultures (Fig. 2a, left). Higher resolution images showed a patchy distribution of HRV-C15-positive cells with the highest signal intensities at the edges of tissues (Fig. 2b, right), suggesting that mechanical injury may increase susceptibility to infection, perhaps by exposing sub-apical epithelial layers expressing virus receptor(s).
The two tissue specimens inoculated with HRV-C15 differed in signal intensity and corresponding virus yield. In the tissue (dn17) with the lower yield (107 vRNA copies ml-1, Fig. 2b, center) the RNA signal was limited to small area at the fragment edge, whereas the higher-yield tissue (dn13, 2 × 109 vRNA copies ml-1), gave many foci, with a greater total density of infected cells (Fig. 2b, right). In tissue cross sections the positive signals originated mainly from middle and upper epithelial layers (Fig. 2c). An outermost sloughing of ciliated cells (Fig. 2c, right) was particularly apparent in the higher titer tissues. There was no evidence for infection of cells deeper than the epithelial layer in either sample.
From the organ culture supernatant fluid (passage 4), we sequenced the complete HRV-C15 genome consisting of 7111 bases. The single open reading frame encodes a polyprotein of 2153 amino acids (the longest among the reported HRV-C strains), flanked by the 5′- and 3′-untranslated regions (UTR) of 607 and 45 nucleotides, respectively. The RNA and deduced polyprotein sequences share characteristic HRV-C features such as a putative internal cis-acting replication element (cre) located within VP2 (1B), Met67/Ser68 cleavage site at the VP4/VP2 junction, species-specific insertions and deletions in the VP1 region, and an isoleucine at the termination of the 3D polymerase15,16,18,19. The full sequence itself shares only 56–62% aligned nucleotide identity with other members of the HRV-C species and less than 41% with HRV-A and B strains.
After our recent complete genome sequencing of the known HRVs, we performed an alignment based on crystal structure superimposition and optimal energy RNA configurations, which we term a structure-based sequence alignment1. This dataset was augmented with the HRV-C15 and other recently published HRV-C sequences13,18,19 in order to infer phylogeny and establish species- or genus-specific motifs to predict cell attachment binding sites of HRV-C. A neighbor-joining tree calculated from the complete genomes (representing 74 HRV-A, 25 HRV-B and 12 HRV-C types) placed HRV-C15 in HRV-C cluster on a branch most closely related to strain n4 (19% p-distance), isolated in China in 2006 (Fig. 3); 5′-UTR analysis revealed HRV-“Ca” type18 clustering (Supplementary Fig. 1 online). The available HRV-C sequences, including HRV-C15, show more collective diversity than the known types of HRV-A and HRV-B, suggesting there is still additional sequence “space” for other strains in this species.
Differences in overall capsid sequence together with the failure of HRV-C15 to replicate in standard cell lines suggest that HRV-C viruses have unique growth requirements, possibly including binding to distinct cellular receptors. Notably, great plasticity and rapid evolution of HRVs have been demonstrated by adaptation of major group HRV (A89) to use another receptor (heparan sulfate) for cell entry20. It was therefore of interest to determine whether the HRV-C have sequence consistency in the same sites as receptor footprints mapped for HRV-A and B21–24.
For each alignment position harboring known contacts for ICAM-1 (HRV-B14, HRV-A16 or human coxsackievirus A21) or LDLR (HRV-A2), the group amino acid populations were recorded and then compared (as groups or internally and pairwise, Supplementary Tables 1 and 2 online). When the ICAM-1 binders were compared among themselves, about 40% of identified footprint positions showed strong (r > 0.6) composition conservation (e.g. HRV-B14-Threonine-VP3-180 (14-T-3-180)). Some of these positions were conserved among all HRVs (e.g. 14-P-1-155 and 16-G-1-148). In contrast, other sites displayed a compositional bias between receptor groups (e.g. 14-P-3-178). Fig. 4a depicts the most abundant residues at each footprint position, when the HRV-A and HRV-B were subdivided by ICAM-1 and LDLR receptor groups, and compared to the HRV-C. There was little commonality between HRV-C and either the ICAM-1 or the LDLR contact points. For example, 6 of 7 LDLR footprint positions, and 47 of 50 ICAM-1 footprint locations were rejected by both statistical methods as compositional matches to the HRV-C (Supplementary Table 2 online, columns “h” or “j”). Overall, the analysis predicts the HRV-C viruses use a different receptor(s), with a distinct compositional footprint.
The distinct HRV-C receptor footprint prediction was then tested in cell culture, using antibodies to block attachment of HRV to ICAM-1 or LDLR. In the absence of these antibodies, binding of HRV-A1 and HRV-A16 to both HeLa and PBE cells were 2–3 logs higher compared to HRV-C15 (Fig. 4b,c). Preincubation of HeLa cells with an ICAM-1-specific antibody reduced binding of major group HRV-A16 but not minor group HRV-A1 (Fig. 4b). In cultures of PBE cells, only HRV-A1 was inhibited by pre-incubation with an LDLR-specific antibody (Fig. 4c).When the experiment was carried out using matched tissue snippets from sinus organ cultures, binding of HRV-C15 was similar to that of the other two viruses. Receptor-specific antibodies inhibited attachment of the respective major (80% inhibition) and minor (58% inhibition) viruses, but there was no effect on HRV-C15 binding (Fig. 4d).
HRVs have been divided into two groups (A and B) based on their sensitivity to different antiviral compounds known as capsid binding agents25. These compounds (e.g., WIN56291, WIN52084) intercalate into a hydrophobic pocket within VP1 protein and prevent viral attachment or uncoating by deforming the receptor binding site or inhibiting a requisite conformational change26,27,28; thus the primary sequence dictating shape and charge of the pocket determine the response to specific antivirals. Preincubation of virus inoculums with WIN56291 inhibited propagation of HRV-A16 (group B) and HRV-C15, but not HRV-B14 (group A), in sinus organ culture (Fig. 4e).
The results of the viral binding studies suggest that HRV-C15 does not replicate in HeLa and PBE cells, which are readily infected with HRV-A and -B, because these cells do not express a cell surface receptor for the virus. Alternatively, these cells could lack an intracellular protein that is required for HRV-C replication. To test this hypothesis, a plasmid (pC15) was developed for cell transfection (as opposed to viral infection) by cloning a full-length cDNA copy of the HRV-C15 genome into the plasmid vector pMJ329 directly downstream of the T7 RNA polymerase promoter (Supplementary Fig. 2 online). The cloned HRV-C15 sequence was identical to the complete genome sequence of passage four HRV-C15 (Fig. 1d, P4) and contained a 3’end poly(A29) tail.
RNA transcripts synthesized in vitro from pC15 induced cytopathic effects when transfected into both HeLa and WisL (fetal lung fibroblast) cell monolayers, and these effects were similar to those induced by parallel transfection of HRV-A16 RNA from pR16.1130 (Fig. 5a). Transfection of about 107 WisL cells (T75 flask) yielded 2–4 × 1010 vRNA copies of virus progeny after purification and concentration steps, and this high output was comparable to that of HRV-A16 transfection. After transfection of lower doses of RNA, only HRV-A16 RNA caused CPE progression over time, confirming that the newly synthesized HRV-C15 virus cannot spread from transfected cells to neighboring (not transfected) cells (Supplementary Fig. 3 online). HRV-C15 virus recovered 24 h after transfection was inoculated (109 vRNA copies ml-1) into mucosal organ cultures, and growth kinetics of progeny virus (Fig. 5b) was similar to those of the parental HRV-C15 (Fig. 1e). Examination of concentrated cell lysates by electron microscopy revealed intact HRV virions (about 30 nm in diameter) as well as the expected empty capsids (Fig. 5c).
HRV-C viruses do not grow in standard cell culture used for virus isolation, and have only been detected using molecular assays. Since 2006, over 60 types of this species have been implicated in both upper and lower respiratory tract infections, particularly in children and in individuals with chronic respiratory diseases31. We employed human organ culture of sinus mucosa to propagate in vitro and study the basic biologic properties of an HRV-C clinical isolate (C15 type). Virus replication was demonstrated by an increase in viral RNA, which localized to focal areas in epithelial cell layers. The high-titer virus obtained after serial passage in sinus organ culture allowed complete genome sequencing and cloning of the full-length cDNA copy of the HRV-C15 genome into a plasmid vector. RNA transcripts synthesized in vitro from this clone were infectious when transfected into cell lines resistant to HRV-C15 infection providing an efficient system for production of recombinant HRV-C. Our experimental findings and bioinformatic predictions indicate unique HRV-C receptor-binding specificity.
Mucosal organ culture is the only model to date that supports HRV-C infection in vitro. We and others have tried to grow these newly discovered rhinoviruses from numerous NLF samples in multiple cell cultures without success, indicating novel growth requirements10,15,16. Moreover, our attempts to adapt high-titer HRV-C15 virus grown in sinus cultures or recombinant virus from transfected cells to a cell line were not successful. HRV-C replication in sinus tissue appears to be limited to the epithelium, and hybridization signals were associated with both non-ciliated and ciliated cells. The degree of virus amplification (high or low virus yields) showed considerable variability, which could be related to the condition of the epithelium (all specimens were from patients with sinusitis), other underlying diagnoses (e.g. allergy or asthma), or perhaps the status of local innate immune elements. Furthermore, the expression of HRV-C-specific receptor(s) on epithelial cells may depend on factors found in vivo such as microbial products or interactions with other types of cells (e.g. dendritic cells or lymphocytes) present in human airways.
Footprints of virus capsid residues that make contact with the cellular receptors have been determined for both major and minor HRV groups21–24. Sequence comparisons within the mapped receptor footprints show variability and a tendency towards a species-specific compositional bias among ICAM-1 binders32. However, amino acid residues corresponding to the documented ICAM-1 footprint have been shown to clearly classify the two receptor groups33. The LDLR binders are harder to define by sequence. They always display a conserved central lysine (e.g. HRV-A2, VP1, Lys224), but this residue is not unique to LDLR binders and predictive modeling instead suggests that the overall binding specificity primarily relies on a favorable combination of charge complementarities and hydrophobic interactions34. Comparisons of the group amino acid populations of the all known ICAM-1 or LDLR contact residues using two correlation metrics (Spearman and Pearson) revealed that the majority of footprint locations were rejected by both methods as compositional matches to the HRV-C suggesting that an alternate receptor is used by these viruses. In agreement with this analysis, very low amounts of HRV-C15 attached to HeLa and PBE cells, and incubation of mucosal organ cultures with blocking antibodies to major and minor HRV receptors showed no effects on HRV-C15 binding.
The development of sinus organ culture and a reverse genetics system have enabled studies of HRV-C15 growth in vitro and provided new insights into its unique replication cycle. HRV-C15 binds to an unknown receptor(s) that is expressed on epithelial cells in differentiated tissues, but is either absent or underexpressed in many cell lines, and is clearly distinct from receptors utilized by other HRV species. It is possible that the lack of knowledge about HRV-C receptor and biology has impeded the development of effective antivirals for HRV, since previous candidate medications were tested only against HRV-A and -B viruses before moving into clinical trials. Even though not all enterovirus species have VP1 motifs that confer susceptibility to capsid binding agents, our results demonstrate that this approach can be used to block HRV-C binding and replication. Thus, these studies provide evidence that there are at least two feasible approaches to the treatment or prevention of HRV infections; refinement of capsid binding agents that target multiple species, and development of competitive antagonists for the HRV receptors. The availability of organ culture and reverse genetics system should greatly facilitate further studies on HRV-C biology and its receptor identification.
We thank Drs. D. Heatley, G. Hartig and J.S. McMurray (University of Wisconsin-Madison) for providing surgical samples; and R. Brockman-Schneider, R. Vrtis, T. Pappas, G. Crisafi, M. Hill, J. Bork, R. Massey, A. Lashua and E. Domyan for technical assistance. This work was supported by NIH grants U19 AI070503, U19 AI070503-04S1, R01 HL080412, R01 HL091490, and the NIAID-funded University of Maryland School of Medicine Genome Sequencing Center for Infectious Disease.
Virus nomenclature follows current recommendations of the ICTV Picornaviridae Study Group, designating the HRV species (A, B, C), followed by the assigned virus type (e.g. HRV-A16, C15 etc.). We propagated laboratory stocks of HRVA1 strain A (minor receptor group), HRV-A16 and HRV-B14 (major receptor group) in H1-HeLa cells (ATCC CRL1958), and purified and titered as previously described29,37. HRV-A78 and HRV-C15 (previously designated W10) are clinical isolates that we identified by partial sequencing of the 5′-UTR as described10.
All protocols for these studies were reviewed and approved by the UW-Madison Health Sciences Institutional Review Board. We obtained residual airway tissue specimens from individuals with chronic sinusitis undergoing endoscopic surgery. We washed, sectioned (4 × 4 mm) and submerged snippets of epithelium along with underlying tissues individually in 24-well plates with 0.3 ml bronchial epithelial growth medium (BEGM, Lonza) and incubated them at 37 °C (5% CO2). Before use, we tested all specimens for preexisting respiratory viruses, including HRV, adenoviruses, influenza, parainfluenza, enteroviruses, respiratory syncytial virus, metapneumovirus, coronaviruses, and bocavirus (Respiratory Multi-Code PLx Assay, EraGen Biosciences)38.
We washed tissue squares (4–5 per well) with PBS (3 × 0.5 ml) before inoculation (2 × 108 vRNA copies ml–1) with a virus suspension. After a period of collective incubation (4–6 h, 34 °C, 5% CO2) we aspirated the media and redistributed the individual tissue samples into separate wells with fresh BEGM (0.3 ml) for further incubation (72–96 h). At harvest, we exposed the combined media and tissue from each well to one freeze-thaw cycle, clarified them by low speed centrifugation (3000 × g, 5 min), then passaged with fresh tissue snippets, or aliquoted and stored (–80 °C).
We evaluated virus concentration by SYBR Green qPCR reagents (Applied Biosystems) using primers (forward, 5′-CCTCCGGCCCCTGAAT-3′; reverse, 5′-AAACACGGACACCCAAAGTAGT-3′) which are complementary to the 5′-UTR of HRV-A1, A16, B14 and C15. We used primers and probe for human actin, beta (ACTB) (#4326315E, Applied Biosystems) to normalize RNA content in cell and tissue lysates. We isolated total RNA from 100 μl of growth medium or mucosal tissue (RNeasy Mini Kit, Qiagen), reverse-transcribed it (TaqMan, Applied Biosystems), and performed PCR in duplicate (ABI 7000 Real-Time PCR System, Applied Biosystems). We derived the standard curve from purified HRVC15 RNA transcripts synthesized in vitro. We performed melting curve analysis of the amplicons after each PCR to confirm reaction specificity.
We cloned a cDNA fragment (867 bases) of HRV-C15, encoding 5’-UTR (partial), VP4 and partial VP2 genes into plasmid pMJ329 after a T7 RNA polymerase promoter in reverse orientation to create plasmid pW10-2R. T7-induced transcription of linearized pW10-2R produced digoxigenin-labeled (Roche) probes that detect the viral genome RNA. We fixed, processed, and hybridized sinus tissue samples according to described protocols39,40. Additional details are provided in the Supplementary Methods online.
We pre-incubated cultured cells or tissue samples (30 min, 37 °C) with monoclonal ICAM-1-specific (# BBA3) or polyclonal LDLR-specific (# AF2148) antibodies (R&D Systems) diluted in growth media (10 μg ml-1) or media alone (control). Next, we added 2 × 108 vRNA copies of virus per well containing either mucosal organ culture or PBE cell monolayers (6-well plate), or per tube containing 1 × 106 HeLa cells in suspension. We incubated cells or tissues with the virus (1 h, 25 °C), washed three times, and analyzed by qPCR. To test the efficacy of the capsid-binding compound WIN56291, we incubated (30 min, 25 °C) virus samples (2 × 108 vRNA copies) in 0.1 ml BEGM containing the compound (0.5 μg ml–1, 0.1% DMSO), followed by inoculation (5 h, 34 °C) into sinus organ cultures. After aspirating unattached virus, we incubated tissues for 72 h and tested by qPCR.
We isolated total RNA (RNeasy Mini Kit, Qiagen) from 100 μl of organ culture supernatant (HRV-C15 passage 4). After reverse-transcription with random primers (TaqMan, Applied Biosystems), we performed PCR with the Platinum PCR SuperMix HF (Invitrogen) and primers (Supplementary Table 3 online). The strategy for cloning full-length cDNA copy of HRV-C15 is described in Supplementary Fig. 2 online. We confirmed each viral cDNA insert by automated sequencing of both strands.
We purified full-length RNA transcripts synthesized from the linearized (BstBI) pC15 DNA (RiboMax large-scale RNA production system T7, Promega) with an RNeasy Mini Kit (Qiagen) and analyzed by agarose gel electrophoresis. We transfected cells with RNA and Lipofectamine 2000 (Invitrogen) mixture (1 to 5 w/v ratio) using manufacturer's recommendations, and incubated for 24–72 h. See the Supplementary Methods online for detailed protocol.
We used Student's t test to analyze viral replication data (SigmaPlot 11.0, Systat Software). Significance was defined at P < 0.05. Statistical evaluation of HRV receptor footprints is described in the Supplementary Methods online.