Flagellin From Pseudomonas aeruginosa Modulates SARS-CoV-2 Infectivity in Cystic Fibrosis Airway Epithelial Cells by Increasing TMPRSS2 Expression

In the coronavirus disease 2019 (COVID-19) health crisis, one major challenge is to identify the susceptibility factors of severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) in order to adapt the recommendations for populations, as well as to reduce the risk of COVID-19 development in the most vulnerable people, especially patients with chronic respiratory diseases such as cystic fibrosis (CF). Airway epithelial cells (AECs) play a critical role in the modulation of both immune responses and COVID-19 severity. SARS-CoV-2 infects the airway through the receptor angiotensin-converting enzyme 2, and a host protease, transmembrane serine protease 2 (TMPRSS2), plays a major role in SARS-CoV-2 infectivity. Here, we show that Pseudomonas aeruginosa increases TMPRSS2 expression, notably in primary AECs with deficiency of the ion channel CF transmembrane conductance regulator (CFTR). Further, we show that the main component of P. aeruginosa flagella, the protein flagellin, increases TMPRSS2 expression in primary AECs and Calu-3 cells, through activation of Toll-like receptor-5 and p38 MAPK. This increase is particularly seen in Calu-3 cells deficient for CFTR and is associated with an intracellular increased level of SARS-CoV-2 infection, however, with no effect on the amount of virus particles released. Considering the urgency of the COVID-19 health crisis, this result may be of clinical significance for CF patients, who are frequently infected with and colonized by P. aeruginosa during the course of CF and might develop COVID-19.


INTRODUCTION
As of October 30, 2021, the coronavirus disease 2019  pandemic, caused by severe acute respiratory syndrome (SARS)-coronavirus (CoV)-2, has infected nearly 245 million people globally and led to >4.9 million deaths (https://covid19. who.int). In this health crisis, one of the major challenges is to identify the susceptibility factors of the infecting virus in order to adapt public health recommendations and to reduce the risk of getting COVID-19, particularly in the case of the most vulnerable people: patients with common chronic respiratory diseases such as asthma and chronic obstructive pulmonary disease, and patients with less common or rare chronic respiratory diseases such as cystic fibrosis (CF). Given their lung impairments, patients with chronic respiratory diseases can reasonably be expected to face an elevated risk of developing severe COVID-19, but the magnitude of this risk remains uncertain (1). Together with clinical follow-up studies conducted to more accurately estimate the disease risk of these patients, basic research on the pathophysiology of SARS-CoV-2 infection should provide critical insights into how COVID-19 affects patients with respiratory diseases.
The aforementioned COVID-19 development in patients is particularly relevant in the case of people with CF (pwCF). CF is caused by variants in the gene CFTR (CF transmembrane conductance regulator), with the most frequent variant being F508del, which leads to aberrant function of airway epithelial cells (AECs). During the course of CF, the lungs of the patients are inflamed and chronically infected by various pathogens, including Pseudomonas aeruginosa, the most prevalent pathogen (2). The most recent multinational report identified 181 cases of pwCF infected by SARS-CoV-2, and recorded 7 deaths (3); among the 181 pwCF, 82% were symptomatic, 47% were hospitalized, and 51% showed airway infection by P. aeruginosa.
Here, we show that the main component of P. aeruginosa flagella, the protein flagellin (Pa-F), upregulates TMPRSS2 expression in AECs, especially in patients' cells deficient for CFTR, through Toll-like receptor-5 (TLR5) and p38 activation. Importantly, this enhanced TMPRSS2 expression is associated with an increase in the level of SARS-CoV-2 infection.

Reverse Transcription-qPCR
Human RNA was isolated using a NucleoSpin RNA/miRNA kit (Macherey Nagel, Duren, Germany). RT was performed using a high-capacity cDNA kit (Applied Biosystems, Foster City, CA, USA). Real-time qPCR was performed by using an ABI QS3 with a Sensifast Probe Lo-Rox Kit (Bio-technofix, Guibeville, France), TaqMan probes for ACE2 (Hs01085333_m1), TMPRSS2 (Hs00237175_m1), FURIN (Hs00965485_g1), and GAPDH (Hs02786624_g1), and a cDNA template. For relative quantification, the expression level of target genes was normalized to the expression of GAPDH relative to the reference group (specified in the figure legends) used as a calibrator and was calculated using the 2 −DDCt method.

SARS-CoV-2 Infection and Viral Quantification
Fully polarized Calu-3 cells grown at the air-liquid interface were infected with SARS-CoV-2 (strain BetaCoV/France/IDF0571/ 2020; accession ID EPI_ISL_411218) at a multiplicity of infection of 1, as previously described (1 h of contact with the virus followed by a change of the medium, and analysis at 24h) (16). Viral quantification through RT-qPCR targeting of ORF1b-nsp14 was performed as described (16).

Statistical Analysis
Differences among groups were assessed for statistical significance by using Prism 9.00 software (GraphPad Software, La Jolla, CA, USA), as indicated in the figure legends. P < 0.05 was considered statistically significant.

Data Availability
RNAseq data used here are from a transcriptomic study (17) which RNAseq raw datafiles are available in the European Nucleotide Archive (ENA) (primary accession number PRJEB9292). http://www.ebi.ac.uk/ena/data/view/PRJEB9292.

ACE2, FURIN, and TMPRSS2 Expression In CF and Non-CF Primary Human AECs Upon P. aeruginosa Infection
We first examined ACE2, FURIN, and TMPRSS2 expression from a previous transcriptomic study performed using primary hAECs; the cells were isolated from control (non-CF) donors and pwCF homozygous for the CFTR F508del variant, and were infected by P. aeruginosa (17). At baseline (time 0 h), similar ACE2 and FURIN mRNA expression levels were observed in non-CF and CF primary hAECs ( Figure 1A), whereas TMPRSS2 expression was significantly higher in CF primary hAECs ( Figures 1A, B). Importantly, P. aeruginosa infection increased TMPRSS2 mRNA expression over time in CF but not in non-CF primary hAECs ( Figures 1A, B), whereas the infection did not affect ACE2 and FURIN expression ( Figure 1A).
Because the most critical proinflammatory factor from P. aeruginosa present in the sputum of pwCF is flagellin (18), we next exposed CF primary hAECs to flagellin. Treatment with flagellin increased the mRNA level of TMPRSS2 without increasing that of ACE2 ( Figure 1C) or FURIN (not illustrated). This effect was observed in both non-CF and CF  ACE2, FURIN, and TMPRSS2 Expression in CFTR-Sufficient and -Deficient Calu-3 Cells Exposed to P. aeruginosa Flagellin To investigate the mechanism underlying the aforementioned increase in TMPRSS2 expression and to eliminate the interindividual variability, we sought to identify AEC lines expressing detectable levels of ACE2 and TMPRSS2 mRNA and protein. Thus, we measured ACE2 and TMPRSS2 expression in the AEC lines Calu-3, Beas-2B, and 16HBE (Figure 2A), which revealed that Calu-3 cells expressed higher mRNA levels of ACE2 and TMPRSS2 relative to the other cell lines, and that ACE2 protein was detected only in Calu-3 cells.
In accord with what was observed in CF primary hAECs, we found that exposure of Calu-3 cells to P. aeruginosa flagellin significantly increased TMPRSS2 mRNA expression ( Figure 2B) in a dose-dependent manner (Supplementary Figure 1) without affecting the transcripts levels of ACE2 ( Figure 2C) and FURIN (not illustrated). This increase in TMPRSS2 expression was more notable in Calu-3-CFTR-KD than in Calu-3-CFTR-WT ( Figure 2B), and the TMPRSS2 upregulation was also detected at the protein level ( Figure 2D). As expected, flagellin induced the synthesis of the proinflammatory cytokines interleukin (IL)-8 and IL-6 (Supplementary Figure 2A, B) both in Calu-3-CFTR-WT and Calu-3-CFTR-KD cells, and this inflammatory response was relatively higher in the Calu-3-CFTR-KD cells, which agrees with previous work showing that CF epithelial cells from human (19) or porcine (20) origin exhibit an enhanced inflammatory response to flagellin.
To ascertain whether the observed effect of flagellin is specific to the bacterial source of the protein, we used ultrapure flagellin isolated from S. Typhimurium serovar Typhimurium (St-F) in our assays, which revealed that St-F induced similar TMPRSS2 expression in Calu-3 cells as did flagellin isolated from P. aeruginosa ( Figure 3A). By contrast, recombinant flagellins (standard or vaccigrade ™ ) from S. Typhimurium did not affect the expression ( Figure 3A).
Flagellin is known to activate TLR5 and the downstream p38 mitogen-activated protein kinase (MAPK) signaling pathway (21); thus, we tested the involvement of this pathway in the observed TMPRSS2 induction. We showed that the upregulation of TMPRSS2 expression depended on TLR5 signaling ( Figure 3B). When a p38 inhibitor was used, flagellin-induced TMPRSS2 expression was diminished in Calu-3-CFTR-KD cells ( Figure 3C) and primary CF hAECs ( Figure 3D). Furthermore, an inhibitor of nuclear factor-kappa B (NF-kB) also reduced the TMPRSS2 induction by flagellin. Accordingly, the results of western blotting confirmed that flagellin stimulated p38 phosphorylation as well as NF-kB activation ( Figure 3E). Moreover, as shown previously (21), we further observed that this NF-kB activation depended on p38 activity ( Figure 3E).
To determine whether TMPRSS2 induction was restricted to the lung epithelium, we examined the flagellin effect in an intestinal cell line, Caco-2/TC7. Our results showed that flagellins from P. aeruginosa and S. Typhimurium, which were able to induce IL-8 production (Supplementary Figure 3A), exerted no effect on TMPRSS2 expression in these cells (Supplementary Figure 3B).

Influence of TMPRSS2 Induction by P. aeruginosa Flagellin on SARS-CoV-2 Infectivity in CFTR-Sufficient and -Deficient Calu-3 Cells
Lastly, we investigated whether TMPRSS2 induction by flagellin influences SARS-CoV-2 infectivity. After infection with SARS-CoV-2, the intracellular nsp14 viral mRNA level was increased, and this level was significantly higher in Calu-3-CFTR-KD than -CFTR-WT cells and was even more notably elevated when the cells were pre-stimulated with flagellin ( Figure 4A). By contrast, the extracellular nsp14 viral mRNA level, measured in the apical supernatant of AECs as a surrogate for viral production, was significantly lower in Calu-3-CFTR-KD cells than in Calu-3-CFTR-WT cells ( Figure 4B). Whereas pre-stimulation with flagellin did not affect viral-particle release in Calu-3-CFTR-WT cells, a lower, albeit not statistically significant, level of nsp14 mRNA was measured at the apical side of Calu-3-CFTR-KD cells pre-stimulated with flagellin. To ensure that viral particles release was not due to an increase in the permeability of the epithelial barrier by flagellin, we measured TEER and observed that flagellin treatment was not associated with a loss of epithelial integrity (Supplementary Figure 4). Lastly, IFN-b and IFN-l measurements in basal supernatants were not induced either by the virus or the flagellin in the different conditions tested (not illustrated).

DISCUSSION
In this study, we showed that exposure of AECs to flagellin from P. aeruginosa induces an increase in TMPRSS2 expression, which is dependent on TLR5 and p38 MAPK activation. Notably, prior exposure of AECs to flagellin results in increased infectivity of SARS-CoV-2 (illustrated in Figure 5).
We found that TMPRSS2 is more expressed in hAECs from pwCF as compared with the level in controls. Although the  hAECs were isolated from a limited number of pwCF, this observation agrees with previous results obtained using excised lungs, where RNA in situ hybridization revealed that TMPRSS2 expression was higher in pwCF than in non-CF patients (7). We also observed that flagellin from P. aeruginosa increases TMPRSS2 expression in primary bronchial epithelial cells from both pwCF and controls. The variability of the flagellin-elicited response in the two groups is likely due to the effect of several donor-related factors, such as sex, age, and CF clinical history in the case of pwCF. An age-related increase in TLR5 expression and sensing has been observed in human monocytes (22). Moreover, the single-nucleotide polymorphism TLR5 c.1174C>T, which is common in the general population (23) and generates a variant that acts as a modifier gene in CF (24), might also contribute to this variability. Unfortunately, it was not possible to have age-matched donors which constitutes a limitation in our study.
To elucidate the specific contribution of CFTR in the level of TMPRSS2 induction by flagellin, we used isogenic Calu-3-CFTR-WT and Calu-3-CFTR-KD cells. We found that TMPRSS2 is highly induced in Calu-3-CFTR-KD cells, and we further confirmed previous observations indicating that CFTRdeficient bronchial cells show an elevated inflammatory response to flagellin, characterized by increased levels of IL-6 and IL-8 (19). Flagellin is known to activate TLR5 and downstream p38 and NF-kB in Calu-3 cells (21). Confirming these results, TMPRSS2 induction by flagellin was found here to depend on p38 and NF-kB, both in Calu-3 cells and primary CF hAECs. Interestingly, a recent phosphoproteomic study in Vero-6 cells revealed that SARS-CoV-2 stimulates the p38 pathway, and that pharmacological inhibition of p38 shows antiviral efficacy (25). Thus, this study and our results here suggest that inhibiting the p38 pathway could represent a potential COVID-19 therapy.
The increase in intracellular viral mRNA levels in CF cells exposed to flagellin, which indicates an elevated level of infection, is likely the result of the upregulated expression of TMPRSS2. Accordingly, TMPRSS2 inhibition by using the serine-protease  infection process does not allow to study at best the resulting inflammatory response. Indeed, it was shown in Calu-3 cells that SARS-CoV-2 induces a significant but delayed IFN-b in comparison to Sendaï virus or synthetic dsRNA (poly I:C) (27). Further investigations conducted using primary cells differentiated at the air-liquid interface will be necessary to specifically characterize the response of CF bronchial epithelial cells. Nevertheless, the model used here is relevant. As noted in the introduction section, secretory cells are infected by SARS-CoV-2; this was demonstrated in previous studies conducted using singlecell RNA-seq, either ex vivo with lung biopsies of patients infected with SARS-CoV-2 (4), or in vitro with a reconstituted epithelium at the air-liquid interface (6,8). Calu-3 cells polarized at the airliquid interface present the characteristics of secretory cells (mucus production) and express naturally (i.e., without exogenous overexpression) the required proteins (ACE2, TMPRSS2) for infection by SARS-CoV-2, which is not the case with the other epithelial cell lines tested in this study.
Several studies have delineated the antiviral capacity endowed by flagellin against other respiratory viruses, including influenza A (28). Furthermore, flagellin was recently suggested to be capable of modulating the innate immune response and thereby eliminating SARS-CoV-2 and resolving COVID-19 (29). Accordingly, the use of recombinant flagellin as an adjuvant in vaccine development has been considered (30). However, we observed here that recombinant flagellin (standard or vaccigrade ™ ) from S. Typhimurium did not affect TMPRSS2 expression. Thus, although these data were obtained in vitro, it is likely based on the aforementioned finding that pathogen-targeting vaccines developed using recombinant flagellin as an adjuvant will not produce negative effects in the case of concomitant exposure to SARS-CoV-2. Moreover, the observed upregulation of TMPRSS2 expression induced by flagellin appears specific to the lung epithelium, because the effect was not replicated in Caco-2/TC7 cells, which are epithelial cells of intestinal origin. During their lifetime, pwCF are also infected with other flagellated bacteria including Burkholderia cenocepacia and Stenotrophomonas maltophilia (31). In particular, flagellin from Burkholderia cenocepacia also known to activate TLR5 (32,33), could also modulate TMPRSS2 expression.
The question of whether CF patients face an increased risk of developing a severe form of COVID-19 is a topical one and a source of discussion (34). Clinical follow-up results obtained to date indicate that pwCF, both adults (3,35) and children (36), do not show an elevated risk of developing severe COVID-19 as compared with the general population. However, pwCF with advanced CF disease (associated with older age, CF-related diabetes, lower lung function, having received an organ transplant) might develop a severe clinical course (3,35).
In a recent French study, we compared the baseline clinical characteristics of 31 pwCF infected by SARS-CoV-2 during the first wave of the pandemic to that of the overall French CF population (n = 6,913; >90% of all French CF cases) (37). The pwCF with COVID-19 were found to be older and more frequently chronically colonized with P. aeruginosa (37). However, considering the small number of patients, these results must be interpreted with caution. Thus, whether the risk of developing severe COVID-19 is increased in pwCF because of their P. aeruginosa infection remains unresolved and will require further meta-analysis performed using international cohorts. Moreover, a recent study including 874 individuals with COVID-19 showed that carriers of CF-causing variants (N=40) may be more likely to develop severe COVID-19 (38). This reinforces the interest in studying the specific role of CFTR in the pathogenesis of COVID-19.
In conclusion, we have shown that exposure of CF AECs to flagellin from P. aeruginosa can enhance SARS-CoV-2 infectivity. Further clinical follow-up studies and in vitro experimental investigations into the mechanisms associated with the specific host response of primary CF cells to SARS-CoV-2 infection should help elucidate this matter and provide insights for future clinical care.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ebi.ac. uk/ena, PRJEB9292.

AUTHOR CONTRIBUTIONS
LG, MR, and OT, designed experiments. MR, JB, CC, VB, OT, AP, JM, MG, JO, conducted the experiments. LG wrote the manuscript. MR, JB, VB, MR-C, HC and OT critically revised the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
LG received a grant from the Facultéde Medecine Sorbonne Université(AAP COVID19).