Measles virus exits human airway epithelia via infectious center sloughing

Measles virus (MeV) is the most contagious human virus, but we do not fully understand why. Unlike most respiratory viruses, MeV does not infect the airway epithelium immediately. MeV traverses the epithelium within immune cells that carry it to lymphatic organs where amplification occurs. Infected immune cells then synchronously deliver large amounts of virus to the airways. However, our understanding of MeV replication in airway epithelia is limited. To model it, we use well-differentiated primary cultures of human airway epithelial cells (HAE) from lung donors. In HAE, MeV spreads directly cell-to-cell forming infectious centers that grow for ∼3-5 days, are stable for a few days, and then disappear. Transepithelial electrical resistance remains intact during the entire course of HAE infection, thus we hypothesized that MeV infectious centers may slough off while preserving epithelial function. After documenting by confocal microscopy that infectious centers progressively detach from HAE, we recovered apical washes and separated cell-associated from cell-free virus by centrifugation. Virus titers were about 10 times higher in the cell-associated fraction than in the supernatant. In sloughed infectious centers, ciliary beating persisted and apoptotic markers were not readily detected, suggesting that they retain functional metabolism. Cell-associated MeV infected primary human monocyte-derived macrophages, modeling the first stage of infection in a new host. Single-cell RNA sequencing identified wound healing, cell growth, and cell differentiation as biological processes relevant for infectious center sloughing. 5-ethynyl-2’-deoxyuridine (EdU) staining located proliferating cells underneath infectious centers. Thus, cells located below infectious centers divide and differentiate to repair the extruded infected epithelial patch. As an extension of these studies, we postulate that expulsion of infectious centers through coughing and sneezing could contribute to MeV’s strikingly high reproductive number by allowing the virus to survive longer in the environment and by delivering a high infectious dose to the next host. AUTHOR SUMMARY Measles virus (MeV) is a respiratory pathogen that infects millions worldwide each year. Although sometimes mischaracterized as an innocuous childhood disease, measles remains a leading cause of death for children under five. MeV is the most contagious human virus and requires vaccination rates above 90% to maintain herd immunity. Global decreases in vaccination rates over the past ten years contributed to recent, widespread MeV outbreaks. We uncover here a novel mechanism by which MeV exits the human airways that may explain why it is much more contagious than other viruses. We document that infected cells containing cell-associated virus slough en masse from the airway epithelial sheet. These expelled infectious centers are metabolically active and can transmit infection to primary human monocyte-derived macrophages more efficiently than cell-free virus particles. Thus, cell-associated MeV can transmit host-to-host, a new paradigm for efficient respiratory virus transmission.


117
To investigate the relevance of infectious center sloughing for virus transmission, we sought to 118 quantify virus load in infected HAE cultures. We collected apical washes, cell lysates, and basolateral 119 media from infected HAE every 3-4 days for 21 days post-infection. Apical washes were gently 120 centrifuged in order to separate cell-free virus in the supernatant from cell-associated virus in the 121 pellet (Fig 2A). We then measured virus titer in cell lysates, basolateral media, and cell-free and cell-122 associated virus from apical media (Fig 2B). High titers were observed in the cell lysates starting at 7 123 days post-inoculation, consistent with microscopy observations. In apical washes, virus titers were 124 very low through day 10 post-inoculation. Starting from day 14, cell-associated virus titers were at 125 least 10-fold higher than cell-free virus titers. These results indicate that most infectious MeV remains 126 cell associated and exits the epithelial sheet via cell sloughing.

128
Infectious centers were collected in the apical washing to assess the viability after sloughing.

129
Immunostaining and confocal microscopy imaging revealed intact nuclei and the F-actin cytoskeleton

143
We next asked if sloughed infectious centers infect macrophages, one of the cell types that ferry 144 virus from the lumen of the airways to the lymphatic organs. To generate macrophages, we isolated 145 monocytes from donated human blood and treated them with the appropriate cytokines to stimulate 146 their differentiation into M2 macrophages (Fig 4A). We then co-cultured these M2 macrophages with 147 extruded infectious centers collected from an apical wash of MeV-infected HAE 14 days post-148 inoculation. As a comparison, we used cell-free virus collected in parallel. Two days later, 149 macrophages were examined for signs of infection using microscopy (Fig 4B). Cell-associated virus 150 (green arrow) spread MeV to nearby macrophages (red arrows); cell-free MeV also infected 151 macrophages, but its lower titers limited the effective MOI. When the number of infected macrophages 152 were quantified by visual counting and normalized to the input titer determined post-hoc, we observed 153 similar levels of infectivity between cell-associated and cell-free MeV (Fig 4C). These experiments 154 suggest that when normalized to input PFU, infectious centers are as effective as cell-free virus in 155 delivering MeV to macrophages. However, since most virus remains cell-associated, sloughed (scRNA-seq) on infected HAE cultures at 3, 7, and 14 days post-inoculation, and as control, mock-161 infected HAE at days 3 and 14 (Fig 5A). Each condition included cultures from 10 pooled matched 162 human donors; similar numbers of cells were sequenced and subjected to equally powered 163 bioinformatic analyses. In total, RNAs from 30,743 cells were sequenced via 10x Genomics scRNA-164 seq.

170
The four additional clusters were defined by a combination of cell type and phenotypic markers: 171 interferon-high, low unique molecular identifier (UMI), mitotic basal, and mitotic surface.

172
HAE are typically mitotically quiescent. However, we identified two small, but distinct clusters of 173 dividing cells, mitotic basal and mitotic surface. These clusters are primarily composed of both GFP+ 174 and GFP-cells from the day 14 timepoint in infected cultures and are almost absent in mock-infected 175 cells (Figs 5C and S2A). Consistent with microscopic evidence showing that basal cells are non-176 permissive to MeV infection, GFP+ basal cells were uncommon (Fig 5D). Of note, the cell type 177 specificity of interferon-high cells could not be determined, but these cells were predominately GFP+ 178 (Fig 5C, D).

179
We also compared the levels of viral RNAs (vRNAs) for each cell type in infected (GFP+ and unbiased signal pathway analysis. As a comparison between the GFP+, GFP-, and Mock groups, lists 187 of differentially expressed genes were generated with a threshold adjusted p-value of 0.05. For GFP+ 188 cells, we identified 91 upregulated genes and 83 downregulated genes; for GFP-cells, 66 upregulated observations (Fig 3). In addition, pathways associated with wound healing, cell growth, and cell 195 differentiation were upregulated in GFP-cells as compared to GFP+ cells (Fig 6 A, B). Cell 196 proliferation genes were upregulated in GFP-cells as compared to GFP+ cells throughout the course 197 of infection (Fig 6D). Altogether, these results indicate that GFP+ cells inhibit apoptotic pathways as 198 innate immune responses develop. In contrast, GFP-cells begin to differentiate. This suggested to us 199 that basal cells situated underneath infectious centers may divide, possibly in preparation for taking 200 over the epithelium-sealing function of the sloughing infected patch.

202
After confirming that baseline transepithelial electrical resistance remained constant following 203 MeV-GFP infection of HAE (Fig 7A), we asked whether cells situated underneath infectious centers 204 proliferate. To identify dividing cells, we used the DNA synthesis marker 5-ethynyl-2'-deoxyuridine 205 (EdU). Indeed, EdU+ cells were localized with infectious centers (Fig 7B, C). We then quantified the 206 kinetics of cell division induction below infectious centers. At day 3 post-inoculation, few EdU+ cells increase of mitotic basal cells over time (Fig 7E). A control EdU+ cell count that excluded infectious 210 centers confirmed the quiescent state of cells not located below infectious centers (Fig 7F, G). These 211 data show that basal cell proliferation is associated with infectious center formation in HAE. Such 212 proliferation may protect the integrity of the epithelium as infectious centers passively slough or 213 actively dislodge the infectious center from the epithelial sheet.

323
The pellet, containing cell-associated virus, was resuspended in 100 µl of USG medium.

589
Images were collected with confocal microscopy and are representative of 3 human donors.