Using in vitro models, we investigated the effects of RV infection on human neuronal and primary bronchial epithelial cells. We confirmed that the RV entry receptor ICAM-1is expressed on the surface of PNEs and following infection, RV enters the neurons and replicates with an accompanying cytopathic effect and neuroinflammatory response. To investigate the molecular consequences of the direct infection of neurons, we undertook RNA sequencing of PNEs and observed RV-induced expression of genes responsible for antiviral immune responses, maintenance of cell structure and integrity and pathways that may be associated with neuronal sensitisation believed to underpin the clinical features of cough reflex hypersensitivity. We also observed release of inflammatory mediators from virally infected neurons with upregulation in IL-1β, 1–309, IL-1α, IFN-γ and IL-4 when compared with mock-infected and UV-inactivated control samples. Additionally, RV infects and replicates in cultured PBECs, with accompanying release of IL-1β which was subsequently shown to sensitise neurons to the TRPA1 agonist cinnamaldehyde. Taken together these novel findings support both direct and indirect (bystander) consequences of RV infection and its contribution to cough reflex hypersensitivity.

The expression of ICAM-1 on neuronal cells has been reported on cholinergic airway nerves supporting our hypothesis that airway neurons are directly susceptible to infection [23]. IF for dsRNA, generated during viral replication but not naturally present in human cells, was observed in our infected neuronal cells confirming that the virus not only enters the neurons but replicates in them. Flow cytometry indicated approximately 9% of cells expressed dsRNA and were undergoing active virus replication. Importantly, viral titrations demonstrated that PNEs are permissive for RV reproduction, allowing the production of significant amounts of new infectious viral particles (around 4 to 5-log increase in viral titres compared to the start of the infection) over a period of 24 to 72 h and at a range of low MOIs (0.01, 0.1 and 1), considered physiologically relevant [24].

Infected neurons released inflammatory mediators including IL-1β, interferon-gamma (IFN-γ) and IL-4, which are associated with RV infection of the human airways [25,26,27] and with other neuronal disorders including neuropathic pain and chronic itch [28, 29]. Increased levels of IFN-γ, a cytokine known to induce cough hypersensitivity have been found in the airway of subjects with chronic cough compared to healthy volunteers [26]. We also observed release of the cytokine I-309, also known as Chemokine Ligand 1 (CCL-1), thought to play a role in the development and maintenance of neuropathic pain [30]. Consistent with previous studies [8, 31], our findings demonstrate that neurons infected by respiratory viruses can release inflammatory mediators associated with sensory neuronal dysregulation and disease states, further supporting a role for neuroimmune interactions in viral pathogenesis.

Transcriptomic analysis of infected neurons revealed altered expression of genes related to cell adhesion and structural integrity, consistent with RV-induced cytopathic effects observed in other cell types [32, 33]. We also observed changes in the IFIT3 (Interferon-Induced Protein with Tetratricopeptide Repeats 3) gene involved in the antiviral innate immune response to respiratory viruses including influenza [34]. Interestingly, we observed increased expression of gene signatures linked to neuronal disorders, including distal peripheral sensory neuropathy, supporting the view that chronic cough may have a neuropathic basis [35]. Furthermore, upregulation of cAMP response element-binding protein (CREB1) pathway genes, along with CaMKII-related signalling, are considered important signalling molecules responsible for N-methyl-D-aspartate (NMDA) receptor activation and the development of allodynia, a neuropathic pain state considered similar to cough hypersensitivity [36].

While airway epithelial cells are the primary site for RV replication, we show that RV can also infect and replicate in neuronal cells, triggering neuroinflammatory mediator release. We report that IL-1β, released by both infected PNEs and PBECs, induces heightened sensitivity to the irritant cinnamaldehyde compared to controls suggesting a potential pathway whereby the cough reflex may be sensitised indirectly (in a bystander fashion) following respiratory viral infection.

We acknowledge experimental limitations in that we studied the effect of viral infection on neurons and bronchial epithelial cells in isolation rather than in a more complex array of structural and inflammatory cells that likely interact during viral infection of the human airway. Furthermore, our neuronal model does not entirely replicate the human in vivo circumstance whereby airway neural afferents represent axonal structures projecting from the cell soma (cell bodies) in ganglia located in extra-pulmonary sites. However, it is recognised that viral infection of axonal endings of peripheral neurons is accompanied by anterograde transport of newly synthetised virus to the cell bodies [37, 38]. We believe therefore that our model provides a means to study neuroinflammatory consequences of respiratory viral infection in the human airway.

In conclusion, our data provide evidence for direct and indirect modulation of airway nerves by RV infection. The neuroinflammatory consequences of respiratory viral infection of the human airways has been largely overlooked but does represent a mechanism for accompanying cough reflex hypersensitivity.