Distinct Mucoinflammatory Phenotype and the Immunomodulatory Long Noncoding Transcripts Associated with SARS-CoV-2 Airway Infection

SARS-CoV-2 productively infects a 3D airway tissue model and induces inflammatory and mucous responses

To model the acute respiratory infection conditions, a 3D airway tissue culture model of primary human respiratory tract epithelial cells (RTECs) was infected with SARS-CoV-2 primary isolate (USA-WI1/2020, BEI Cat # NR-52384) at 1 multiplicity of infection (MOI) (see Supplemental Methods for details). Viral RNA was assessed in apical washes, in the basal culture supernatant, and in cells; from RNA samples extracted at 0, 1, 4, 24, and 48 h post-infection (hpi). Viral RNA increased by 3.5-fold from 8.9×10⁵ viral genomic equivalents per ml at 1 hpi to 3.1×10⁶ at 24 hpi in apical washes (Figure 1A), there was no change in VL in basal culture supernatant which remained at ∼3.0×10⁵ genomic equivalents per ml at all the time-points (Figure 1B), and the level of vRNA in total cell RNA increased by ∼6 fold at 24 hpi (Figure 1C). Thus, 3D airway cells productively infected with an actively replicating SARS-CoV-2 shed the virions almost exclusively from the apical surface.

• Download figure
• Open in new tab

Figure 1. SARS-CoV-2 infection of human respiratory epithelial cells induces a robust mucoinflammatory phenotype in a 3D airway tissue model.

Respiratory tract epithelial cells (RTECs) differentiated on air-liquid interface were infected with 1 MOI of SARS-CoV-2 clinical isolate (USA-WA1/2020 isolate) and analyzed at 0, 1, 4, 24, and 48 h post-infection (hpi). Viral loads determined in (A.) the apical wash, (B.) the basal culture supernatant, and (C.) in the cell RNA. Relative expression levels of inflammatory factors (D.) IL-6 and (E.) ICAM-1 mRNAs, and (F.) the epithelial secretory factor SCGB1A1 mRNA. Changes in the mRNA expression of airway mucins, (G.) MUC5AC, (H.) MUC5B, (I.) MUC2, and (J.) MUC4 mRNA levels; and the mucin regulatory transcriptional factors, (K.) SPDEF and (L.) FOXA3 as analyzed by qPCR of the total cell RNA. (M.) Representative micrographs of RTECs at 0 and 48 hpi showing MUC5AC (shown in green) immunoreactivity along with the DAPI stained nuclei (shown in blue), scale - 5µ. (N.) Percentage of MUC5AC+ cells within each treatment group. Quantification of (O.) IL-6 and (P.) ICAM-1 levels in media supernatants as assessed by ELISA (n=4/gp from 2 independent experiments; *p<0.05; **p<0.01; ***p<0.001 by ANOVA).

In addition, we found the expression of the viral entry host factor, ACE2, significantly reduced at all time points following infection (see Supplemental Figure 1A). In contrast, the expression of S protein processing enzyme TMPRSS2 remained constant for first 24 hpi but was markedly suppressed at 48 hpi (Supplemental Figure 1B). Thus, SARS-CoV-2 infection directs ACE2 and TMRPSS2 suppression in a RTEC 3D model, as reported earlier (8, 20). We also found altered expressions of several inflammatory factors charged with directing airway tissue remodeling and/or controlling host immune responses and airway mucin expression (21, 22). Among the inflammatory factors, although CXCL-8 expression was unaffected following infection (Supplemental Figure 1C), interleukin-6 (IL-6) was significantly elevated (9-fold) at 1 hpi (Figure 1D), whereas intercellular adhesion molecule-1 (ICAM-1) was elevated (>5-fold) at 24 hpi (Figure 1E) and airway secretoglobulin 1A1 (SCGB1A1), also known as club cell secretory protein (CCSP) was induced by 5-fold from 1 to 24 hpi with then returned to baseline levels at 48 hpi (Figure 1F).

Notably, we also observed robust mucin hyperexpression in our 3D airway tissue model. SARS-CoV-2 infection robustly upregulated secretory mucin MUC5AC (21, 23-26) transcription by more than 50-fold by 1 hpi and maintained hyperexpression at all the time-points (Figure 1G). In addition, expression of the constitutive epithelial mucin, MUC5B, involved in the airway antimicrobial response (27-29) was increased by 48 hpi (Figure 1H), and similar to MUC5AC, cell surface-associated mucins, MUC2 (Figure 1I), and MUC4 (Figure 1J) levels were also increased by 1 hpi then remain elevated at all time-points. Furthermore, known transcriptional regulators of mucin expression (30, 31) were similarly upregulated following infection with >2-fold induction of SPDEF levels (Figure 1K) and >20-fold induction of FOXA3 levels at 1 hpi (Figure 1L).

Next, protein expressions were assessed in transwell cells by immunostaining. In agreement with our quantitative RT-PCR analyses, immunostaining confirmed robust immunopositivity for MUC5AC expression (Figure 1M) with ∼40% MUCAC+ cells at 48 hpi compared to <5% cells at 0 hpi (Figure 1N). Levels of secreted inflammatory factors, IL-6 (Figure 1O) and ICAM-1 (Figure 1P) in basal media supernatants were also elevated at 24 and 48 hpi. Thus, the inflammatory and mucous responses of respiratory epithelial cells are acutely moderated by SARS-CoV-2 infection. Although roles for airway mucins in innate antiviral immunity are well documented, SARS-CoV-2 associated mucin signatures have not been thoroughly evaluated to date (32). Specifically, these mucins are heavily glycosylated and could result in higher viral binding/retention due to interaction with sialic acids (33) and contribute to viral infection and/or transmission.

SARS-CoV-2 infected airway tissue model show increased expression of immunomodulatory lncRNAs

In addition to mucoinflammatory factor encoding mRNAs, another type of RNA, lncRNAs, have now been found to play vital roles in regulating the innate immune responses (12, 13) and recent studies have shown that specific lncRNAs are differentially expressed during host-pathogen interactions following viral infection (18, 19, 34, 35). That said, our group identified two novel epithelial lncRNAs, LASI (lncRNA on antisense strand to ICAM-1) and TOSL (TNFAIP3-opposite strand lncRNA) that are differentially expressed during the mucus hypersecretory response (36) and blocking these lncRNA expression alleviates the mucin hyperexpression. Notably, we find LASI transcripts are highly elevated (10-fold) at 1 hpi (Figure 2A), whereas TOSL expression is increased by 1 hpi and were >6-fold higher at 48 hpi (Figure 2B). However, in contrast to recent clinical reports (19, 37, 38), we find NEAT1 (Nuclear Enriched Abundant Transcript 1) lncRNA levels reduced following infection (Figure 2C), and no change in the expression of MALAT1 (Metastasis Associated Lung Adenocarcinoma Transcript 1) another immunomodulatory lncRNA (39, 40) MALAT1 (Metastasis Associated Lung Adenocarcinoma Transcript 1) expression (Supplemental Figure 1D). In agreement with our qRT-PCR based expression, dual-RNA FISH analysis showed co-expression of SARS-CoV-2 N1 vRNA and LASI transcripts around the nuclear/perinuclear region (Figure 2D) with elevated expressions at 24 hpi compared to 4 hpi as assessed by the histological-score (H-score) quantitation H-score (Figure 2E). Thus, SARS-CoV-2 infection of our 3D airway culture model resulted in a strong mucoinflammatory phenotype with concurrent increases in associated lncRNA expressions. Therefore, our data suggest that these lncRNAs may be serve as biomarkers for measuring SARS-CoV-2 disease progression. However, these need to be confirmed using large number of clinical samples.

• Download figure
• Open in new tab

Figure 2. SARS-CoV-2 infection differentially affects the expression of inflammation associated lncRNAs in airway epithelial cells.

Relative expression levels of select immunomodulatory lncRNAs, (A.) LASI, (B.) TOSL, and (C.) NEAT1 in airway epithelial cells following 0, 1, 4, 24, and 48 h of SARS-CoV-2 infection. (n=4/gp from 2 independent experiments; *p<0.05; **p<0.01; ***p<0.001 by ANOVA). (D.) Colocalization of SARS-CoV-2 vRNA and LASI transcripts in infected cells at 0, 4, and 24 hpi as determined by the dual-FISH analysis and structured illumination microscopy. Representative micrograph of dual-FISH-stained cells showing SARS-CoV-2 nucleocapsid vRNA (in red) and LASI transcripts (in green) along with DAPI-stained nuclei (in blue), scale – 2 µ. (E.) H-score quantitation of vRNA and LASI transcripts per cell at 4 and 24 hpi (n=9-10 cells/gp; *p<0.05; **p<0.01; by Student’s t-test).

COVID-19 patients with higher viral load show increased mucoinflammatory response

Large cohort studies revealed there was no difference in SARS-CoV-2 viral load between symptomatic and asymptomatic patients, whereas disease severity and mortality in symptomatic patients is directly correlated with viral load (38, 41, 42). The mucosal immune response of upper respiratory tract plays a critical role in viral shedding and replication. To explore this correlation, in our study, we divided subjects into 2 groups based on the qRT-PCR cycle threshold (C_T) of SARS-CoV-2 N1 nucleocapsid, with a C_T>30 considered a low viral load (Lo-VL) and a C_T<30 high viral load (Hi-VL). As detailed in Supplemental Table 1, nasopharyngeal swabs collected from a total of 20 subjects (10 Hi-VL and 10 Lo-VL) were examined. Twelve symptomatic subjects were hospitalized and had at least one and/or combination of comorbidities viz., obesity, chronic disorders, hypertension, and diabetes. Hi-VL subjects were found to contain >100-fold higher vRNA levels than Lo-VL subjects (Figure 3A). Interestingly, no significant differences were observed in ACE2 (Figure 3B) and TMPRSS2 (Figure 3C) mRNA levels between groups. In contrast, Hi-VL samples exhibited ∼1.5-fold higher IL-6 (Figure 3D) and ICAM-1 (Figure 3E) inflammatory factors mRNA expressions, as well as 4-fold higher SCGB1A1 expression (Figure 3F). Hi-VL patient samples also exhibited robust secretory mucin levels, with ∼10-fold higher MUC5AC (Figure 3G) and 4-fold higher MUC5B (Figure 3H) expressions. Notably, there was a trend towards reduced MUC2 expression (Figure 3I) in Hi-VL subjects with ∼10-fold higher MUC4 (Figure 3J) expression compared to Lo-VL subjects. Expression of SPDEF was also increased by ∼5-fold in Hi-VL subjects (Figure 3K), but in contrast to our 3D models, we failed to detect any FOXA3 transcripts in patient nasopharyngeal swab samples. These data clearly suggests that SARS-CoV-2 can regulate the expression of several cellular receptors by inducing inflammatory responses.

• Download figure
• Open in new tab

Figure 3. COVID-19 patients with high nasopharyngeal viral load show increased mucoinflammatory phenotype compared to the patients with low viral load.

Total RNA from nasopharyngeal swab samples of COVID-19 patients were analyzed. Based on the SARS-CoV-2 nucleocapsid viral RNA expression, the patients with high viral load (Hi-VL) had average C_T values of 25.9±1.3 whereas those with low viral load (Lo-VL) had average C_T values of 35.5±0.7 (n=10/gp). Relative mRNA expression of: (A.) SARS-CoV-2 viral RNA (vRNA) and viral entry related host factors, (B.) ACE2 and (C.) TMPRSS2 mRNA; Innate inflammatory factors, (D.) IL-6, (E.) ICAM-1, and (F.) SCGB1A1; Airway mucins (G.) MUC5AC, (H.) MUC5B, (I.) MUC2, and (J.) MUC4; and (K.) SPDEF transcriptional factor in the Hi-VL vs Lo-VL patient swab samples (n=10/gp; *p<0.05; **p<0.01; ***p<0.001 by Student’s t-test).

Respiratory mucosa lncRNAs are associated with SARS-CoV-2 viral load

While we found no difference in MALAT1 lncRNA levels between Hi-VL and Lo-VL subjects (Supplemental Figure 2A), the expression of select immunomodulatory lncRNAs, LASI (Figure 4A), TOSL (Figure 4B), and NEAT1 (Figure 4C) were significantly upregulated in Hi-VL patient samples. In addition, dual-RNA FISH analysis of nasal swab epithelial cells from Hi-VL subjects showed enriched SARS-CoV-2 nucleocapsid vRNA and LASI transcripts in nuclear and cytosolic regions (Figure 4D). The H-score quantitation confirmed vRNA and LASI transcripts were 1.5-fold higher in Hi-VL samples compared to Lo-VL samples (Figure 4F), corroborating our qRT-PCR data. Nasal epithelial cell mucin protein expression was also analyzed for the mucin protein expression by immunostaining for MUC5AC (Figure 4E) and MUC5B (Supplemental Figure 2B) expression of dual-FISH labeled cells. MUC5AC immunopositivity in Hi-VL subjects was ∼10-fold higher than in Lo-VL samples as quantified by mean fluorescence intensity (MFI) per cell (Figure 4G) whereas MUC5B expression did not significantly differ (Supplemental Figure 2C). Thus, both mucoinflammatory responses and immunomodulatory lncRNA expression of nasopharyngeal cells are significantly altered in Hi-VL SARS-CoV-2 infected patients as compared to Lo-VL patients. Current studies are ongoing to identify the molecular targets/binding partners of these lncRNAs.

• Download figure
• Open in new tab

Figure 4. COVID-19 patients with high viral load show increased levels of immunomodulatory lncRNAs and the MUC5AC mucin expression.

Relative expression of immunomodulatory lncRNAs, (A.) LASI, (B.) TOSL, and (C.) NEAT1 in the Hi vs Lo patient swab samples. (D.) Colocalization of SARS-CoV-2 vRNA and LASI transcripts in nasal swab samples as detected by dual-FISH analysis. Representative micrographs from low (Lo) and high (Hi) viral load patient samples display detection of SARS-CoV-2 nucleocapsid vRNA (green) and LASI transcripts (red) along with DAPI-stained nuclei (blue). (E.) Nasal swab cells showing immunoreactive MUC5AC expression (shown in white) in the dual-FISH labeled cells, scale – 2 µ. (F.) H-score quantitation of vRNA and LASI transcripts in COVID-19 patients. (G.) Relative quantitation of mean fluorescence intensity (MFI) of MUC5AC expression in Lo-VL and Hi-VL patient samples. (n=10/gp for data in A, B, C, & F; and n=4/gp for data in G; *p<0.05; **p<0.01; ***p<0.001 by Student’s t-test)

We next performed in-silico analysis of FASTQ files obtained from RNA-sequencing of lung biopsies from COVID-19 patients and healthy controls as earlier reported (32, 33). Notably, although only a limited number of samples were analyzed, there was a markedly higher expression of MUC5AC, MUC5B, and SPDEF genes in COVID-19 patients (Supplemental Figure 3A) whereas there were no changes in IL-6, CXCL-8, and ICAM-1 expressions (Supplemental Figure 3B). In terms of lncRNAs, MALAT1 trended lower in COVID-19 patients, LASI lncRNA expression trended higher and was significantly elevated in at least one COVID-19 patient, and there was no observed difference in NEAT1 lncRNA levels between COVID-19 patients and healthy controls (Supplemental Figure 3C). That said, these data do further corroborate the association of mucoinflammatory response and lncRNAs with COVID-19 pathophysiology observed in our 3D models and clearly warrant further analyses of larger data sets and the underlying molecular mechanisms.

In conclusion, the present study employing both a 3D airway tissue model of SARS-CoV-2 infection and COVID-19 patient samples provides the first experimental association of SARS-CoV-2 viral load with mucoinflammatory phenotype and specific airway mucosa lncRNAs despite limited cohort sample size, possible variation in swab sampling, and/or the cell numbers per sample potentially contributing to data incongruity. Importantly, we have recently shown that blocking lncRNA LASI can help suppress mucus hyperexpression (34) and hence, could be a feasible strategy to regulate COVID-19 associated mucoinflammatory response. As initial host-viral interactions in the upper respiratory tract are crucial for modulating subsequent respiratory and systemic immune responses to SARS-CoV-2 and other co-infections and environmental exposures; targeting these lncRNAs could potentially rescue the debilitating outcomes of the COVID-19 sequelae and help improve current disease management strategies.

Statistics

Statistical methods are described in Supplemental Methods. Briefly, grouped results were expressed as mean ± SEM and P values less than or equal to 0.05 were considered significant.

Study approval

The COVID-19 patient samples were collected at University of Miami Biobank facility according to the approved institutional review board protocol. All studies involving SARS-CoV-2 wild type virus were conducted in the University of Nebraska Medical Center (UNMC) BSL-3 high containment core facility according to the institutional biosafety committee approved protocol. The formalin-fixed and frozen samples were analyzed at the Florida International University according to the approved institutional biosafety committee protocols in a designated BSL-2+ facility.