An endogenously activated antiviral state restricts SARS-CoV-2 infection in differentiated primary airway epithelial cells

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the cause of the coronavirus disease-19 (COVID-19) pandemic, was identified in late 2019 and caused >5 million deaths by February 2022. To date, targeted antiviral interventions against COVID-19 are limited. The spectrum of SARS-CoV-2 infection ranges from asymptomatic to fatal disease. However, the reasons for varying outcomes to SARS-CoV-2 infection are yet to be elucidated. Here we show that an endogenously activated interferon lambda (IFNλ1) pathway leads to resistance against SARS-CoV-2 infection. Using a well-differentiated primary nasal epithelial cell (WD-PNEC) culture model derived from multiple adult donors, we discovered that susceptibility to SARS-CoV-2 infection, but not respiratory syncytial virus (RSV) infection, varied. One of four donors was resistant to SARS-CoV-2 infection. High baseline IFNλ1 expression levels and associated interferon stimulated genes correlated with resistance to SARS-CoV-2 infection. Inhibition of the JAK/STAT pathway in WD-PNECs with high endogenous IFNλ1 secretion resulted in higher SARS-CoV-2 titres. Conversely, prophylactic IFNλ treatment of WD-PNECs susceptible to infection resulted in reduced viral titres. An endogenously activated IFNλ response, possibly due to genetic differences, may be one explanation for the differences in susceptibility to SARS-CoV-2 infection in humans. Importantly, our work supports the continued exploration of IFNλ as a potential pharmaceutical against SARS-CoV-2 infection.


Introduction
In late 2019 a novel coronavirus, severe acute respiratory syndrome coronavirus 2 (SARS--CoV-2), emerged in Hubei province in China [1,2]. This virus is the causative agent of the coronavirus disease-19 (COVID-19) pandemic, which has resulted in >5 million deaths in just 26 months. SARS-CoV-2 can infect people of any age. However, severe disease

Virus cultivation and titration
The SARS-CoV-2 clinical isolate BT20.1 was isolated from a patient admitted to the Royal Victoria Hospital, Belfast in June 2020. Characterisation of this strain has previously been described [17]. The virus was passaged 4 times on Vero cells, including isolation. The titres of the virus stocks and experimental samples were determined by plaque assay. In brief, Vero cells were seeded at 5x10 4 cells/well in a 24-well cell culture plate. Twenty-four hours later a 1:10 serial dilution of the virus stock or sample was prepared and inoculated onto the Vero cells. Following 1 h incubation at 37˚C a semi-solid agarose overlay was added. Vero cells were further incubated for 3 days then fixed with paraformaldehyde (final concentration 4% v/v) and stained with crystal violet (5% w/v) for plaque visualisation. The origin and characterization of the clinical isolate RSV BT2a were previously described [18]. RSV titres in biological samples were determined using HEp-2 cells, as previously described [19].

Well-differentiated primary nasal epithelial cell (WD-PNEC) culture, infection and treatment
Nasal epithelial cells were obtained from consented healthy adults with no known history of lung disease or previous SARS-CoV-2 infection. Culture and differentiation protocols were previously described [20,21]. In brief, cells were passaged three times in Promocell Airway Epithelial Cell Growth Medium (Promocell) then seeded onto collagen coated Transwell supports (Corning) at 3x10 4 cells per Transwell. After 4-6 days of submersion air-liquid interface (ALI) was initiated by removing the apical medium. Cells were differentiated using Stemcell PneumaCult ALI medium (Stemcell Technologies) supplemented with hydrocortisone and heparin. Basolateral medium was replaced every 2 days. Complete differentiation took a minimum of 21 days. WD-PNEC cultures were only used when hallmarks of excellent differentiation were evident, including extensive apical coverage with beating cilia and obvious mucus production. WD-PNECs were infected with SARS-CoV-2 (MOI = 0.1) by the addition of inoculum on the apical surface and incubation for 1 h at 37˚C. Cultures were treated basolaterally with 0, 1 or 100 ng/mL human IFNλ3 (R&D systems, Biotechne) for 24 h. Ruxolitinib (Selleck Chemicals) (1 μM) was added to the basolateral medium 48 h before infection and was replaced at 1 and 48 hpi. Apical washes were harvested by adding 200 μL DMEM (low glucose; Gibco), incubation for 5 min and gentle aspiration without damaging the cultures.

Ethics statement
The nasal epithelial cells were collected, with written consent, for the purpose of performance evaluation rather than research. Under Queen's University Belfast procedures, the collection of such samples for the purposes of performance evaluation is exempt from the requirement for ethical review. Specifically, in the emergency surrounding the SARS-CoV-2/COVID-19 pandemic, we sought to establish and evaluate the best protocols for isolating SARS-CoV-2 from clinical samples with a view to developing robust tests for SARS-CoV-2 that maintained its genetic and phenotypic integrity. It is well established that isolation and passage of viruses from clinical samples in immortalised monolayer cell cultures may result in the isolation of virus strains that are specifically adapted to growth in the chosen cell lines. To determine whether we could diminish this selective pressure on isolates and thereby maintain their clinical phenotype and genotype, we obtained nasal epithelial cells from donors by nasal brushing using a well-established minimally invasive protocol. Nasal epithelial cells are the primary targets for SARS-CoV-2 infection in humans. This work allowed us to establish protocols that maintained phenotypic and genotypic characteristics of the SARS-CoV-2 clinical isolates, which is an important step in the development of future diagnostics.

Immunofluorescence
Cultures were fixed with 4% paraformaldehyde (w/v) for a minimum of 1 hour. Cells were permeabilised with 0.2% Triton X-100 (v/v) then blocked using 0.4% bovine serum albumin (BSA) (w/v). Antibodies used are specified in figure legend. Cells were counterstained with DAPI-mounting medium (Vectashield). Images were obtained on a Nikon Eclipse TE-2000U or a Leica SP5 confocal microscope. Quantified genes (raw counts) were processed using edgeR [24] for normalisation purposes and transformed into logCPMs (log counts per million). These logCPM values were subsequently used to generate a heatmap using pheatmap package for comparison and visualisation purposes using a selection of interferon related genes extracted from a selection of pathways annotated in WikiPathways and Reactome Pathway Database (WP2113; R-HSA-913531; WP1904; R-HSA-168928; WP4880; WP1835; R-HSA-909733; WP75), ACE2 and TMPRSS2. SNP calling was carried out using ExactSNP program from the subread package using default parameters [25]. To increase the sequencing depth, we merged together all samples from each donor and used the merged file as input for the SNP calling phase. In addition to SNP calling, we also annotated the variants identified using SnpEff software and the GRCh38.99 database [26].

ELISA analysis
Concentrations of IFN-λ1 in basolateral medium samples from WD-PNECs were measured using a human IFN-λ1/IL-29 ELISA (Invitrogen) according to the manufacturer's instructions.

Resistance to SARS-CoV-2 infection in WD-PNECs
WD-PNEC cultures derived from 4 healthy adult donors demonstrated different susceptibility to SARS-CoV-2 infection and viral growth kinetics ( Fig 1A). Following inoculation apical washes were harvested every 24 h post infection and titrated by plaque assay on Vero cells. WD-PNECs from 3 donors demonstrated similar SARS-CoV-2 growth kinetics with peak viral titres of >10 6 PFU/mL at 48 hpi followed by a plateau or decrease at 72 hpi. Surprisingly, infection of donor 122 did not result in productive infection of SARS-CoV-2, with virus titres continually diminishing relative to titres following inoculation. Immunofluorescence reinforced these results, with high levels of SARS-CoV-2 N expression in WD-PNECs from donors 311, 43 and 32, but with very few infected cells evident in those from donor 122 ( Fig 1B). Consistent with other data, SARS-CoV-2 infection was restricted to the apical surface of the cultures [27]. To determine if resistance to infection was specific to SARS-CoV-2, we infected cultures derived from the same donors with respiratory syncytial virus (RSV), an RNA virus that primarily infects ciliated airway epithelium [28]. There was no evidence of restrictions in RSV growth kinetics in any donor. (Fig 1C).

Pre-activation of IFN pathway in resistant WD-PNECs
The timing and magnitude of IFN responses to SARS-CoV-2 infection is thought to play a role in disease severity. Transcriptomic analysis of RNA extracted at 48 and 96 hpi revealed that there was upregulation of genes associated with both the type I and type III IFN pathways in the mock-infected cultures of donor 122 but not in the other cultures. Interestingly, we did not detect any IFNα from any donor or condition. Following SARS-CoV-2 infection the gene transcription signature of donor 122 did not significantly change (Fig 2A). Similar, or greater, levels of expression of ACE2 and TMPRSS2 were observed in this donor, confirming that resistance to infection was not due to absence or low expression of the attachment and entry factors needed for SARS-CoV-2 infection. The three donors susceptible to infection had similar patterns of gene expression, including an increase in expression of IFNλ-associated genes following SARS-CoV-2 infection (Fig 2A). To confirm these results, IFNλ1 concentrations in  were not permissive to SARS-CoV-2 infection, had high levels of IFNλ1 (>500 pg/mL) in the basolateral medium from mock-infected cultures at 48 and 96 hpi (Fig 2B). Transcriptomic analysis of the IFIH1-mediated interferon induction and toll-like receptor (TLR) signalling pathways revealed that donor 122 had a unique gene transcription signature compared to mock or infected cultures from the other donors (S1 Fig).

Pre-activation of IFN pathway and SARS-CoV-2 resistance phenotype are not transient
To determine if the high baseline level of IFNλ1 in donor 122 cultures was an artifact of culture conditions at the time of the original experiment or the time at which the nasal brush was obtained, a second brush was taken 4 months after the first and cultured in the same manner. When fully differentiated the second cultures were infected with SARS-CoV-2, with similar results (Fig 3A). WD-PNECs cultured from the second nasal brush of donor 122 also had high endogenous levels of IFNλ1 (~400 pg/mL), albeit slightly reduced compared to those evident in the cultures from the first brush. In contrast, IFNλ1 levels from mock-infected cultures from 6 other donors demonstrated background levels of secretion ( Fig 3B).

IFN signalling restricts SARS-CoV-2 infection
SARS-CoV-2 is sensitive to exogenous IFN treatment [29]. We hypothesised that the high baseline level of IFNλ1 in basolateral medium of donor 122-derived WD-PNECs limited the spread of SARS-CoV-2 within the culture and prevented productive infection. Therefore, we blocked the major signalling pathway of both type I and type III IFNs with the JAK1/2 inhibitor ruxolitinib prior to infection. WD-PNECs derived from a donor with low IFNλ1 secretion in response to SARS-CoV-2 infection (donor 311) was used as a control. Ruxolitinib treatment resulted in a massive increase in SARS-CoV-2 titres in apical washes from cultures derived from donor 122 (>3 log 10 PFU/mL), reaching a peak titre of 7.6 log 10 PFU/mL at 72 hpi ( Fig  4A). As expected, there was no significant effect of ruxolitinib treatment on donor 311, which also reached peak viral titres at 72 hpi. Viral titres at 72 hpi were comparable for untreated To demonstrate that IFNλ was likely the underlying cause of resistance to SARS-CoV-2 infection we replicated these conditions in the SARS-CoV-2 permissive donor 43. WD-PNECs were pre-treated with IFNλ in the basolateral medium for 24 h prior to infection (Fig 4B). Treatment with both 1 and 100 ng/mL IFNλ greatly reduced viral titres at 48 hpi by >2 log 10 PFU/mL, resulting in titres like the input inoculum (4.18 log 10 PFU/mL).

Single nucleotide polymorphisms (SNP) analysis revealed differences in pathogen sensing and interferon induction pathways unique to the resistant donor
Unbiased SNP analysis from our RNAseq data demonstrated clear differences between donors, with unique variants in several genes in the SARS-CoV-2-resistant donor compared to the permissive donors. Focusing our SNP analysis to selected genes associated with antiviral responses, we identified 388 SNPs unique to donor 122 and a further 33 in donors 311, 43 and 32 but not detected in donor 122 (Fig 5A). Further conditional filtering for quality and existing SNP (according to dbSNP, NIH) identified 10 SNPs potentially involved in the high endogenous IFN secretion and resistance to SARs-CoV-2 demonstrated in donor 122-derived WD-PNECs (Fig 5B). Interestingly, SNPs unique to donor 122 were found in ADAR (rs3738032), MAVS (rs2089960995), TLR1 (rs4833095), TLR2 (rs1816702; rs3804099) and TLR3 (rs3775296; rs3775291). SNPs present in all donors except 122 were found in ADAR (rs1127313; rs1127326) and IFIH1 (rs1990760). Together these results show that resistant versus permissive donors have unique constellations of SNPs that may explain the molecular mechanisms of susceptibility or resistance to SARS-CoV-2 infection.

Discussion
The WD-PNEC model presents an ideal opportunity to study SARS-CoV-2 infection in a clinically relevant system. The upper respiratory epithelium is thought to be the initial site of infection for SARS-CoV-2 [30]. As such, a comprehensive understanding of SARS-CoV-2 infection of nasal epithelium is critical to understand the early events of the infection process, as they are likely influential in subsequent pathogenesis. Why some patients with COVID-19 experience no or mild symptoms consistent with an upper respiratory tract infection, while others experience much more severe symptoms often resulting in acute respiratory distress syndrome (ARDS) remains to be determined. Consistent with earlier reports, we demonstrated that differentiated nasal epithelial cells can be productively infected with SARS-CoV-2 and that infection is primarily restricted to the apical surface. Surprisingly, however, we provide robust evidence of differential susceptibility of airway epithelium to SARS-CoV-2 infection, with airway epithelium from some individuals being highly susceptible and others resistant. Our data suggest that resistance is likely mediated in part by type III IFNs. Thus, we provide a potential explanation as to why some individuals may be able to control initial SARS-CoV-2 infection.
The household secondary attack rate for SARS-CoV-2 is estimated to be 16.6%, higher than that of SARS-CoV (7.5%) and MERS (4.7%) [31]. While this demonstrates that household contacts play a significant role in onward transmission of the virus, it also highlights that not everyone, despite close contact with an infected individual, will be infected with SARS-CoV-2. Studies on pre-existing immunity to SARS-CoV-2 infections have focused on aspects of the adaptive immune system and have demonstrated reactive T cells in SARS-CoV-2-naive patients being important, possibly due to prior infections with an endemic coronavirus [32]. However, differences in susceptibility to infection due to innate immunity have not been reported previously. Interestingly, a recent pre-print demonstrated a link between inter-individual differences in the expression of anti-SARS-CoV-2 alleles of OAS1 in determining IFNmediated cellular resistance to infection and subsequent disease [33]. However, in our system, we observed enhanced endogenous secretion of IFN and downstream antiviral ISGs. We propose that resistance to SARS-CoV-2 infection is, in part, due to endogenous secretion and signalling of relatively high levels type III IFN in the mucosae independent of SARS-CoV-2 infection. This fits with recent work showing high sensitivity of SARS-CoV-2 to prophylactic IFN treatment. Indeed, SARS-CoV-2 sensitivity to type I and III IFN treatment was previously reported. However, contrary to our findings, these authors did not observe an interferon response to SARS-CoV-2 infection of primary human airway epithelial cells [27]. A study by Dee et al. demonstrated that rhinovirus infection induced an IFN response that blocked subsequent SARS-CoV-2 infection. They proposed that given the prevalence of rhinovirus infections within the population, it could impact the number of COVID-19 cases [34]. In our model, the underlying cause of such IFN secretion remains to be elucidated. Although this phenomenon has only been observed in nasal cells from one donor it provides an exciting precedent to explain resistance to SARS-CoV-2 infection. As such, our data provide the rationale for a large-scale study to determine the frequency of endogenously activated IFN responses and whether it may be implicated in differential susceptibility to COVID-19 disease in humans.
WD-PNECs that were resistant to SARS-CoV-2 infection were derived from a donor with no known history of SARS-CoV-2 infection, was unvaccinated against COVID-19 at the time of both nasal brush samplings and was otherwise in good health. We cannot entirely discount the possibility that donor 122 had a prior, concomitant or persistent bacterial or viral infection at the time of nasal brushing, thereby explaining the elevated endogenous IFNλ1 secretion levels in the corresponding WD-PNECs. However, this is unlikely, as there was no evidence of any infection of the donor at the time of either brushing, the epithelial cells were cultured in the presence of broad spectrum antibiotics, and persistent viral infection of airway epithelial cells for a period of 4 months would be unprecedented. However, epigenetic changes have been reported following viral infection. Of particular note is influenza-induced epigenetic methylation of H3K79, which may play an important role in interferon signalling [35]. That the WD-PNECs cultured from the second nasal brushing 4 months after the initial sample were also resistant to SARS-CoV-2 infection, suggests that this high endogenous level of IFNλ1 was not a transient response, consistent with a genetic or epigenetic mechanism. In addition, cultures from donor 122 were susceptible to infection with RSV. While IFNλ treatment of primary nasal epithelial cell cultures can reduce subsequent RSV growth kinetics, we previously reported that a much higher concentration is needed than that detected in the basolateral medium of donor 122 [36]. In contrast, SARS-CoV-2 is evidently much more sensitive to the levels of IFNλ1 observed in cultures derived from donor 122. Treatment with IFNλ resulted in~3 log 10 PFU/mL reduction in SARS-CoV-2 titres. However, donor 122 showed complete abrogation of SARS-CoV-2 replication despite a lower concentration of IFNλ present in the basolateral medium. This could indicate that there are other factors, in addition to IFNλ, that may explain the resistance of donor 122 to SARS-CoV-2 infection. SNP analysis highlighted differences in IFN signalling pathway genes in SARS-CoV-2-resistant donor 122 compared to those who were permissive. The presence of SNPs in Toll-like receptors (TLRs) is known to affect susceptibility to disease [36]. Here we identified several SNPs unique to donor 122 in TLR1, 2 and 3. Particularly noteworthy is a common variant (minor allele frequency of~0.3 globally) in TLR3 (rs3775291; C>T), which may confer resistance to HIV infection [37]. The mechanism of HIV resistance is still unknown. However, heterozygous individuals have higher cytokine mRNA levels compared to C homozygotes, which could restrict viral infection [38]. This phenomenon appears to be virus-specific, as the same mutation actually confers susceptibility to enteroviral infection [39].
Treatment of human CoV infection with recombinant interferon is not a new concept. Type I IFNs were tested against both SARS-CoV and MERS-CoV with promising pre-clinical results but efficacy was less convincing in clinical trials [40,41]. Early clinical trials assessing the efficacy of type I IFN treatment for COVID-19, alone or in combination with other pharmaceuticals, have initially proved promising but further data are needed to draw definitive conclusions. Similarities between type I and III IFN-induced signalling and induction of antiviral and inflammatory responses are well known. However, type III IFN effects are greatest in mucosal tissues, coincident with the tissue location of its receptor on epithelial cells. Induction of type III IFNs results in prolonged expression of ISGs [42,43]. Type III IFNs have been trialled for treatment of hepatitis C virus (HCV), and hepatitis delta virus (HDV) and, more recently, SARS-CoV-2, with positive results [44][45][46]. Type III IFNs, with tissue-restricted receptor expression, are potentially a much more attractive option for treatment of viral infections than the more inflammatory type I IFNs, which have ubiquitous receptor expression. Our data supports further investigation into IFNλ, directed towards the luminal surface of the airways, to reduce the incidence of severe COVID-19. Here we have identified donor-specific resistance to SARS-CoV-2 infection in WD-PNECs concurrent with high endogenous IFN, likely due to genetic/epigenetic differences. These findings point to a previously overlooked mechanism of resistance to COVID-19 disease and provide the rationale for evaluating the role of elevated endogenous type III IFN secretion at the population level.