Experimental rhinovirus infection induces an antiviral response in circulating B cells which is dysregulated in patients with asthma

Abstract Background Rhinoviruses are the predominant cause of respiratory viral infections and are strongly associated with asthma exacerbations. While humoral immunity plays an important role during virus infections, cellular aspects of this response are less well understood. Here, we investigated the antiviral response of circulating B cells upon experimental rhinovirus infection in healthy individuals and asthma patients. Methods We purified B cells from experimentally infected healthy individuals and patients with asthma and subjected them to total RNA‐sequencing. Rhinovirus‐derived RNA was measured in isolated B cells using a highly sensitive PCR. B cells were stimulated with rhinovirus in vitro to further study gene expression, expression of antiviral proteins and B‐cell differentiation in response rhinovirus stimulation. Protein expression of pro‐inflammatory cytokines in response to rhinovirus was assessed using a proximity extension assay. Results B cells isolated from experimentally infected subjects exhibited an antiviral gene profile linked to IFN‐alpha, carried viral RNA in vivo and were transiently infected by rhinovirus in vitro. B cells rapidly differentiated into plasmablasts upon rhinovirus stimulation. While B cells lacked expression of interferons in response to rhinovirus exposure, co‐stimulation with rhinovirus and IFN‐alpha upregulated pro‐inflammatory cytokine expression suggesting a potential new function of B cells during virus infections. Asthma patients showed extensive upregulation and dysregulation of antiviral gene expression. Conclusion These findings add to the understanding of systemic effects of rhinovirus infections on B‐cell responses in the periphery, show potential dysregulation in patients with asthma and might also have implications during infection with other respiratory viruses.

expression of pro-inflammatory cytokines in response to rhinovirus was assessed using a proximity extension assay. cytokines, including type-I interferons (IFN). While type-I IFNs are essential mediators to induce antiviral immunity, they also play an important role in various B-cell differentiation processes. [9][10][11][12][13] B cells are constantly circulating through the body's tissues and are also present at the mucosal sites of the respiratory tract, where RV infections mainly take place. 14 population in tonsils 16 where RV can be detected. 17 B cell-mediated humoral immune response plays a central role in controlling RV infections 18,19 with neutralizing serum IgG and secretory IgA in the mucosa detected one to 2 weeks after infection, with a role of protection from re-infection with the same strand. 18,19 B cells were also found to internalize RV particles and proliferate in response to RV stimulation in vitro. 20 B cells crucially contribute to eliciting adaptive immune responses by transporting lung-derived viral and bacterial antigens to secondary lymphoid organs in various murine models. [21][22][23] While RV infections usually lead to mild symptoms in healthy adults, they can lead to complications in young children or in adults with underlying chronic respiratory diseases, particularly asthma.

Results
RV-induced wheezing in early life is strongly associated with the development of asthma during later childhood. 24 Furthermore, RV infections are the main cause of asthma exacerbations. 25 Moreover, a less efficient B-cell response was found in patients with asthma due to a bias towards non-virus-neutralizing antibodies. 26,27 Given the frequent incidence of RV infections and the central role of B cells during antiviral responses, it is surprising that the cellular side of this response was not yet addressed in more detail.
Therefore, this study aimed to identify the underlying gene regulatory networks driving the B-cell response during RV infection in vivo; to define the external stimulating factors driving this response; to address whether B cells directly interact with RV in vivo; and to discover potential B-cell functions in addition to well-studied antibody production. Furthermore, since responses to RV were described as less efficient in patients with chronic respiratory diseases, we addressed possible dysregulation of circulating B cells in asthma patients. Here, we report early upregulation of an antiviral gene programme, followed by subsequent upregulation of a pro-inflammatory response in B cells from experimentally infected human individuals.

| ME THODS
A detailed description of the study design, sample preparation, methods used for cell culture, flow cytometric cell sorting, quantitative PCR, detection of RV in human samples, measurement of proteins and methods for RNA-sequencing and data analysis can be found in Appendix S1 section.
In brief, B cells were purified from healthy human subjects that were experimentally infected with RV-A16. Gene expression of sorted B cells was analysed using RNA-sequencing. To confirm results from this first experiment, RNA-sequencing and quantitative PCR were also performed on B cells cultured with RV and IFNα in vitro. RV-induced protein expression of antiviral genes and B-cell subsets were analysed using flow cytometry. Expression of proinflammatory cytokines was measured using a proximity extension assay targeting 92 inflammatory mediators. Gene expression from B cells sorted from experimentally infected asthma patients was compared to gene expression in healthy individuals.

| Intranasal infection with Rhinovirus-A16 elicits antiviral response in peripheral B cells
Immune activation in total peripheral blood cells was observed after infection with respiratory viruses, including influenza, respiratory syncytial virus and rhinovirus (RV). 28 By producing virus-neutralizing antibodies, B cells play a central role in the immune response to RV infections. 15,18,19 However, it is still unclear if or how circulating B cells respond to such local respiratory virus infections. To address this, B cells were purified from peripheral blood mononuclear cells (PBMC) before and after experimental RV infection (day 3 and day 7) from a group of healthy human subjects without asthma or allergies from a previously reported study 29 ( Figure 1A, Figure S1).
Characteristics of the healthy subjects are described in Table S1. The successful inoculation and development of infection of all participating subjects were confirmed previously by RV strain-specific PCR in the nasal lavage and lower airways (see Figure S2 in the Online Appendix S1 section). 29 Figure 1D). Additionally, genes involved in response to unfolded protein were also upregulated on day 7, probably reflecting F I G U R E 1 Intranasal infection with rhinovirus-A16 elicits antiviral response in peripheral B cells of healthy individuals. A, Experimental layout of experimental RV infection experiment. Gene expression of B cells sorted from experimentally RV-infected non-asthmatic human subjects was analysed using RNA-sequencing (B-E). B, Significantly upregulated genes compared to baseline (n = 4) on day 3 (n = 6) and day 7 (n = 6) were analysed for pathway enrichment according to GO Biological Process. C, D, E, Gene expression change shown for virus response genes (C), cytokine genes (D) and response to unfolded protein genes (E). F, G, H, B cells were sorted from in vitro RV-infected PBMC of healthy individuals, and antiviral gene expression (F) or viral RNA expression (G) was analysed using qPCR, n = 7. H, In vitro cultured PBMC were washed 4h after RV infection to remove excess RV virions from cell culture, B cells were sorted at indicated time points and viral RNA expression analysed using qPCR, n = 6. I, RV RNA was detected using a highly sensitive two-step pan-RV PCR. Values are means ± SEM, for in vitro experiments (F-H), RV was used at multiplicity of infection (MOI) of 10 (10 infectious particles per cell), unless noted otherwise increased protein expression ( Figure 1E). Since access to samples from experimentally infected individuals was limited, antiviral response and inflammatory cytokine gene expression was also confirmed in B cells purified from in vitro infected PBMC ( Figure 1F). In line with in vivo results, all genes were upregulated by RV stimulation compared to control. Interestingly, viral RNA level increased by day 3 in a virus-dose dependent manner and was decreased again by day 7 ( Figure 1G)

| Rhinovirus stimulation drives plasmablast differentiation and induces a strong antiviral response in plasmablasts
Next, we aimed to confirm expression of antiviral response genes on protein level as well as timeline of expression in total B cells. As MX1 and IFI44L were also differentially expressed in plasmablasts depending on the isotype, with switched IgM-plasmablasts having higher MX1 ( Figure 3G) and lower IFI44L levels ( Figure 3H) compared to IgM+ plasmablasts. In summary, the antiviral response was most upregulated in the increasing subset of CD27+ CD38+ plasmablasts.

| Upregulation of pro-inflammatory cytokines in B cells is dependent on co-stimulation with RV and IFNα
While we described above that antiviral gene expression was found (CCL4), IL-6, TNF, TNFB (LTA) and VEGFA was confirmed ( Figure 4E).
With exception of MIP-1α/β, all these cytokines were only induced after co-stimulation with IFNα and RV ( Figure 4F). Taken together, these data suggest that inflammatory cytokine genes detected in vivo and in vitro ( Figure 1) were induced in response the co-stimulatory effect of IFNs and direct RV exposure on B cells.

| The peripheral B-cell response to RV infection is elevated in asthma
While RV infections are often asymptomatic or lead to mild symptoms in individuals with no underlying disease, 3-5 they present the most common cause for exacerbations of asthma. 25 Furthermore, skewed and less efficient antiviral immune responses were observed in patients with asthma. 26,27,29,36,38,51,52 To address whether the re-   blood. 28 The number of antiviral genes that were expressed in B cells from asthma patients was higher, and these were more upregulated compared to B cells from healthy subjects. In addition, expression of pro-inflammatory cytokines including IL-6 and TNF was slightly downregulated upon infection ( Figure 5D) while BCR-signalling genes were significantly upregulated ( Figure 5E). Moreover, the average increase in heavy chain expression ( Figure 5F) and expression of plasmablasts markers were more upregulated in patients with asthma ( Figure 5G). Interestingly, expression of type-I IFN receptors was similar between healthy individuals and those with asthma (data not shown). Interaction network analysis using STRING 54 IFI44L  IFIT1  STAT1  IFI6  IFIT3  XAF1  IFI44  IFITM1  MX1  OASL  MX2  OAS3  LYZ  PLSCR1  DDX60  OAS2  EIF2AK2  TRIM22  IFI35  PNPT1  HERC5  IFITM2  IRF9  RV infection ( Figure 5H). In summary, total number of antiviral and antibody-related genes in asthma patients were higher and these were more upregulated compared to infected non-asthmatics individuals suggesting B-cell response is broad and exaggerated at the gene expression level in asthma.

| DISCUSS ION
We report the first comprehensive analysis of the gene regulation in Here, we studied B cells sorted from the peripheral blood of healthy human subjects that were experimentally intranasally infected with RV-A16. We chose RNA-sequencing of highly purified B cells as an unbiased approach to study the suspected antiviral response in circulating B cells. On day 3, circulating B cells showed upregulated antibody and interferon signalling-induced antiviral genes.
While this IFN-driven antiviral response was mostly downregulated on day 7 after infection, a pro-inflammatory cytokine response was upregulated instead. These results are comparable to a study showing early IFN-driven antiviral responses in whole blood, followed by secondary inflammatory responses during influenza infection in humans. 55 We found a relatively high antiviral gene expression variability between individual subjects in which seems to replicate results of an earlier study where gene expression was analysed after experimental infection with RV-A39 in full blood. 28 While a prior study showed a positive correlation of viral load in nasal lavage with neutralizing antibody titre after experimental RV infection, 29 our subject number was too low to determine whether there is a direct relationship between viral load in different tissues and gene expression in peripheral B cells. As we worked with residual samples from a previous study 29 and hence the number of in vivo samples was limited, we confirmed expression of selected antiviral and pro-inflammatory cytokine genes in vitro with similar results. In addition, the upregulated antiviral genes that we detected in purified B cells also overlap to a large extent with genes upregulated in whole blood after RV infection. 28,56 The stimulation driving this gene expression programme seems not to originate from cells of the blood, as IFN expression cannot be detected in whole blood after experimental RV infection. 28 Interestingly, RV is sometimes detected in tonsils 17 where B cells make up about 60% of total cells. 16 Furthermore, B cells were also found in nasal tissue of individuals infected with RV while absent in non-infected controls. 15 Therefore, it seems likely that circulating B cells get stimulated by IFNs released either by phagocytes in airway-associated lymphoid tissue or secreted by airway epithelial cells when they traffic through airway tissue close to infection sites. 14 We showed that circulating B cells isolated from experimentally infected subjects carried RV RNA, suggesting that direct interactions of B cells with RV virions occur in infected humans. In addition to IFN stimulation, this interaction with RV seemed crucial for B cells to express the pro-inflammatory cytokines reported in this study.
Detection of viral RNA on B cells might also suggest that B cells acted as antigen-presenting cells during RV infection, as shown during viral infections in mouse models. 21,22 In vivo interaction of B cells with other viruses, including influenza, 57 respiratory syncytial virus 58 and dengue virus, 59 was also described. In patients having severe dengue virus infection, single-cell sequencing revealed that more than 40% of B cells carried viral RNA. 59 However, it is not possible to assess from our data whether B cells isolated from experimentally infected individuals simply transported viral particles bound to surface receptors as was reported before for virus-like-particles, 21 or whether they internalized virions and were infected. Nonetheless, our in vitro data suggest that at least transient viral RNA production can occur in B cells.
Surprisingly, unlike other antigen-presenting cells, 35-38 B cells did not express IFNs in response to RV and therefore seem to rely on signals from other by standing cells to bring them into an antiviral state. This lack of response seems not limited to RV, as also B cells exposed to respiratory syncytial virus, 58  with asthma were also observed. 63 In addition, a recent study on naturally RV-infected children with asthma also supports the theory that deficient early IFN responses may lead to unchecked virus replication, leading to greater virus loads, which subsequently results in exaggerated IFN responses. 64 Greater virus loads were also observed in the subjects with asthma providing PBMC for the present study (see Figure S2), however, with the much smaller subject numbers studied these differences were not statistically different. Overall, our results suggest that higher viral loads in patients with asthma could lead to an exaggerated antiviral response in the periphery.
A limitation of our study is that gene expression was only measured 3 days after experimental infection, while type-I-IFNs are already secreted within hours. Sampling at earlier, as well as later time points after remission, might help to better understand in which step the antiviral response is dysregulated in subjects with asthma. However, it is practically difficult to obtain these materials in a human in vivo study. In addition, given the relatively low number of subjects, the significant genes should be carefully reviewed and verified in larger cohorts. Nonetheless, the genes that were differentially expressed in B cells after RV infection of healthy individuals in the current study are similar to those found upregulated in earlier studies. 28,65 The peripheral B-cell responses reported here were mostly dependent on type-I-IFN stimulation, as well as exposure to RV-A, which is a positive-sense RNA virus. Therefore, it can be expected that our findings are not restricted to infections with RV-A species only and might represent a general response towards different RV species and comparable viruses infecting the respiratory tract.
Given the frequent incidence of such infections and because of the current SARS-CoV-2 pandemic, antibody responses to respiratory viruses are studied intensively. It could be important to also focus on B-cell functions other than antibody production, including cytokine production and antigen presentation to fully understand the role of B cells responding to these pathogens.

Dr. Willem van de Veen reports grants from Novartis
Forschungsstiftung, and grants from Promedica Stiftung, outside the submitted work. Dr. Sokolowska reports grants from Swiss National Science Foundation, grants from GSK, and grants from Novartis, outside the submitted work. Dr. Glanville has a patent US9937252B2 'Induction of cross-reactive cellular response against rhinovirus antigens' pending. Dr. Gern reports grants from NIH, during the conduct of the study; personal fees from Regeneron, personal fees and