Differential Induction and Signaling of Type I and III Interferons During Infection of Human Nasal Epithelial Cells With the Four Human Parainfluenza Virus Types

ABSTRACT

Human parainfluenza viruses (HPIVs) commonly cause croup and lower respiratory tract infections in young children. Despite typically self‐limiting with mild symptoms, the innate immune responses to HPIVs remain poorly understood, especially across all four types in human nasal epithelial cells (hNECs), the primary infection site. This study aims to investigate and compare viral replication kinetics and host immune responses in hNECs infected with HPIV types 1 to 4.The hNECs were infected in vitro, and viral replication kinetics, mucociliary function, cell tropism, and innate immune responses were assessed over 72 h. The viral RNA and progeny of HPIV types 1 to 4 were detected by 8 h postinfection (hpi). All HPIV types predominantly targeted ciliated cells of which a significant proportion displayed a unique apical distribution. HPIV‐3 generated the highest level of progeny virus and could infect a fraction of nasal goblet cells. RIG‐I and MDA5 expression was delayed during HPIV‐2 and HPIV‐3 infection (24 hpi), whereas HPIV‐1 and HPIV‐4 induced gradual upregulation from 8 hpi. Minimal upregulation of IFN‐α1 mRNA was induced across all HPIV types. Expression of type I IFN‐β was generally subdued during the earlier stage of infection (24 to 48 hpi) with the four HPIV types. Notably, IFN‐β mRNA expression was significantly elevated for HPIV‐1 infection at 48–72 hpi, whereas only a small increase was observed at 72 hpi for the other HPIV types. Although expression of type III IFN‐λ1 was similarly delayed for all four HPIV types during the earlier stage of infection, a greater upregulation of IFN‐λ1 was detected at 48–72 hpi for all HPIV types (especially HPIV‐1) as compared to type I IFN expression. For HPIV‐1 and HPIV‐4, mRNA expression of CXCL10 and MX1 appeared as early as 8 hpi and gradually increased until 72 hpi. However, for HPIV‐2 and HPIV‐3, mRNA expression of CXCL10 and MX1 was delayed at 8–16 hpi, but subsequently increased significantly at 24–72 hpi. In comparison, IFITM1 mRNA was expressed at relatively weaker levels for the four HPIV types. In conclusion, all four HPIV types exhibit high infectivity and replicative capacity in hNECs while initially evading interferon responses to varying extents. Their innate immune modulation appears to be type‐specific, potentially influencing viral virulence, pathogenesis, and progression of airway disease in high‐risk patients.

Abbreviations

hNECs

human nasal epithelial cells

hNESPCs

human nasal epithelial stem/progenitor cells

hpi

hours postinfection

HPIV

human parainfluenza virus

IFN

interferon

ISGs

interferon‐stimulated genes

PFU

plaque‐forming unit

1. Introduction

Human parainfluenza viruses (HPIVs) are RNA viruses within the Paramyxoviridae family. HPIVs are grouped into four distinct types based on genetic and antigenic variations. HPIV types 1 and 3 are classified in the genus Respirovirus, while HPIV types 2 and 4 belong to the genus Rubulavirus. HPIV infections are associated with distinctive clinical presentations, causing a spectrum of respiratory illness [1]. HPIV‐1 and HPIV‐2 are the leading causes of croup, otherwise known as laryngotracheitis or laryngotracheobronchitis. HPIV‐3 is the most prevalent type often associated with lower respiratory tract infection (LRTI), including bronchiolitis and pneumonia [2]. While illness related to HPIV‐4 infection was previously associated with upper respiratory tract infection (URTI) symptoms, studies on larger sample sizes documented clinical manifestations and severity of HPIV‐4 similar to HPIV‐3 infection [3, 4]. Together, HPIVs are the second most common cause of LRTI in younger children after respiratory syncytial virus (RSV) [5].

Population‐level epidemiological studies have generally reported a high prevalence of HPIV infections among children younger than 5 years [6, 7, 8]. Interestingly, studies of age‐specific HPIV infections in young children reported peaked rates in children less than 2 years of age, with HPIV‐3 infection typically seen in up to 50% of young infants within the first year of life [3, 9]. Besides children and infants, most HPIVs manifest in a variety of respiratory infections, from mild URTI in healthy adults to more severe LRTI, particularly in the elderly and the immunocompromised [10, 11]. These clinical studies have documented HPIV type‐specific infection patterns in terms of viral prevalence patterns in different age groups, clinical presentations, disease severity, and localization of infection in the airway.

Previous studies have investigated host response induced by different types of HPIV [12, 13, 14, 15]. However, these studies reported responses induced in cell line models, which may not be clinically and physiologically relevant to infection of HPIVs in their natural host environment. Differentiated human nasal epithelial cells (hNECs) are highly representative and closely mimic the physiological host response of the upper airway with pseudostratified layer and specialized airway cell types, thus providing a reproducible human host model for the study of natural HPIV infections. However, certain studies of HPIV infection using human airway models only partially recapitulated the host‐virus interactions, with limitations such as small sample size [16, 17, 18]. Currently, holistic understanding of the viral replication kinetics and immune responses during HPIV infections at the early stage of infection is incomplete. Furthermore, many studies focused on a single HPIV type, and detailed investigations into HPIV type‐specific viral pathogenesis and virus‐induced responses are lacking. Given that all HPIV types are known to result in hospitalization in young children, cross‐comparison of different HPIV types is important to elucidate their differential pathogenesis and disease progression in the airway, particularly the upper respiratory mucosa, which serves as the primary portal of entry for respiratory viruses. Therefore, the objectives of our study are to investigate the viral replication kinetics, to characterize and compare the virus‐induced host immune responses between the HPIV types in our established model of human‐derived pseudostratified nasal epithelial cells.

2. Materials and Methods

2.1. Derivation of hNESPCs and in Vitro Differentiation of Human hNECs

Approval to conduct this study was obtained from the National Healthcare Group Domain‐Specific Board of Singapore (DSRB code D/11/228) and the Institutional Review Board of the National University of Singapore (IRB code 13‐509) [6]. In this study, human nasal epithelial stem/progenitor cells (hNESPCs) were derived from tissue biopsies of five subjects who underwent septal plastic surgery at the National University Hospital, Singapore. The medical backgrounds of the donors are summarized in Supporting Information: Table S1. The subjects were not on oral or topical corticosteroid medications 1 month prior to surgery and had no viral infection at the time of surgery. Detailed procedures for the hNEC cultures are provided in the Supplementary Materials.

2.2. HPIV Infection of Fully Differentiated hNECs

Clinical strains of HPIV types 1 to 4 (HPIV‐1/Singapore/DMS7/2016, HPIV‐2/Singapore/DMS21/2015, HPIV‐3/Singapore/DMS38/2015, HPIV‐4A/Singapore/DMS65/2018, and HPIV‐4B/Singapore/DMS60/2015) were obtained from the National Public Health Laboratory in Singapore. HPIVs were propagated by inoculation of monolayers of LLC‐MK2 cells (85%–95% confluent) and incubation at 37°C in 5% CO₂ until cytopathic effect was observed at about 7 days postinoculation. Virus stock was titrated by plaque assay and stored at −80°C. The hNECs from one Transwell were trypsinized, and cell number was counted for the calculation of multiplicity of infection (MOI) before infection. HPIV stocks were diluted using PneumaCult medium, and 100 µL of each strain of the HPIV types 1, 2, 3, 4A, and 4B was inoculated onto hNECs (at MOI of 1) in the apical chamber of the Transwell in 24‐well plates. The hNECs were incubated at 33°C for 1 h, and viral inoculum was removed. PneumaCult medium without virus was inoculated as mock‐infected hNECs control. The apical secretion and hNEC lysates of both HPIV‐infected and mock‐infected hNECs were collected at 8, 16, 24, 48, and 72 h postinfection (hpi) for the relevant assays.

2.3. Viral Quantification Using Plaque Assay

100 µL of 1× Dulbecco's phosphate‐buffered saline (dPBS) was added into the apical chamber of infected and mock‐infected hNECs and incubated for 10 min at 33°C to collect apical secretion and progeny viruses at different time‐points. The samples were stored at −80°C until titration by virus plaque assay. LLC‐MK2 cells (85%–95% confluent) in 24‐well plates were incubated with 200 µL of serial dilutions from 10 ^{− 1} to 10 ^{− 6} of virus from infected hNECs at 37°C for 2 h. Inoculum was removed, replaced with 1 mL of Avicel (FMC Biopolymer) overlay to each well, and incubated at 37°C for 6 days. Avicel overlay was then removed, and cells were fixed with 4% formaldehyde in 1 × PBS for 1 h. The fixed monolayer cells infected with HPIV‐3 were stained with 1% crystal violet, while cells infected with HPIV types 1, 2, 4A, and 4B were stained with 0.25% neutral red in water. TPCK‐trypsin at different final concentrations for each HPIV type (4 µg/mL for HPIV‐1, 2 µg/mL for HPIV‐2 and HPIV‐4B, 1 µg/mL for HPIV‐3 and 3 µg/mL for HPIV‐4A) was optimized, and added to viral inoculum and Avicel to obtain optimal visible and countable plaques [19, 20, 21]. Plaque‐forming unit (PFU) values were calculated as follows:

$$\text{Number}\text{of}\text{plaques}\times \text{Dilution}\text{factor}=\text{Number}\text{of}\text{PFU}\text{per}\text{mL}.$$

2.4. Measurement of Tight Junction Integrity and Ciliary Beat Frequency (CBF) of hNECs

Transepithelial electrical resistance (TEER) was measured at different time‐points using EVOM volt/ohm meter (WPI, Sarasota, FL, USA). CBF was auto‐analyzed using the Sisson‐Ammons Video Analysis system (SAVA, Omaha, NE, USA) as previously described [22]. Detailed procedures are provided in the Supplementary Materials.

2.5. Cytospin Preparation and Immunofluorescence (IF) Staining

At each time‐point, hNECs were dissociated from the Transwell by Trypsin/EDTA solution (Gibco) at 37°C. Dissociated cells were fixed in 4% paraformaldehyde at room temperature for 10 min, washed twice with 1× dPBS, and centrifuged. Cytospin slides (2 × 10⁴ cells per slide) were prepared at 500 rpm for 5 min, and stored at −20°C until staining. To identify the main cell type infected by HPIV, co‐staining of mock‐ and HPIV‐infected hNECs on cytospin slides was performed using antibodies targeting viral protein of each HPIV type and cellular markers of ciliated cells (acetylated α‐tubulin), goblet cells (MUC5AC), and basal cells (p63) of the nasal epithelium. To quantify the percentage of HPIV‐infected cell type, the number of HPIV‐infected ciliated cells, goblet cells, or basal cells over the total number of each cell type was calculated (five fields of each type of HPIV infection from three donors were evaluated at 400× magnification). The percentage of HPIV‐infected ciliated cells showing unique apical distribution pattern (“apical+”) over the total infected ciliated cells in hNECs was also calculated (five fields of each type of HPIV infection from three donors were evaluated at 400× magnification). IF staining of transwell membranes was performed for semi‐quantitative analysis of acetylated α‐tubulin‐positive area in arbitrary units (five fields per individual, n = 3, evaluated at 400× magnification by ImageJ). The detailed IF staining procedures and antibodies used are provided in the Supplementary Materials.

2.6. RNA Extraction and Real‐Time Quantitative Polymerase Chain Reaction (PCR) (qPCR)

Total cell RNA was extracted from hNECs using the RNeasy Mini Kit (Qiagen, Hilden, Germany) following the manufacturer's protocol. RNA concentration and purity were measured using Nanodrop2000 UV‐Vis Spectrophometer (Thermo Fisher Scientific, Waltham, MA, USA), and 1000 ng of total cell RNA (containing viral RNA) were used for cDNA synthesis using qScript cDNA SuperMix (Quanta BioSciences, Beverly, MA, USA). SYBR Green qPCR was performed as previously described, and the ribosomal protein L13a (RPL13A) housekeeping gene was used for normalization [22]. All qPCR primers used are listed in Supporting Information: Table S2.

2.7. RT‐PCR Amplification and Real‐Time Quantitative PCR for HPIV Viral Load

Total cell RNA (containing viral RNA) was extracted using the RNeasy Mini Kit, and 1000 ng was used for cDNA synthesis using Moloney Murine Leukemia Virus (M‐MLV) reverse transcriptase (RT) (Promega, Fitchburg, WI, USA). RNA (1000 ng) with 1 µL of random primers was heated at 70°C for 5 min and placed on ice immediately. Reverse transcription using M‐MLV RT reaction mix was performed at 37°C for 1 h. Real‐time quantitative PCR was carried out using primers specific to different HPIV types [23, 24] (listed in Supporting Information: Table S2). The thermal cycling conditions were 95°C for 10 min followed by 45 cycles, each consisting of 95°C for 10 s, 60°C for 30 s, 72°C for 15 s, and a final hold at 65°C for 1 min.

2.8. Statistical Analysis

The qPCR results were analyzed using GraphPad Prism 8 software (San Diego, California, USA). Since the gene expression levels by qPCR (2 ^{− ΔCt}) were not normally distributed (Gaussian distribution) due to the variability among the different individual‐derived hNECs that were analyzed by GraphPad, the median and interquartile range were used for statistical analysis. The significance level was calculated using one‐way ANOVA, and the nonparametric, grouped, Dunn's multiple comparisons test, unless stated otherwise. Data are presented as fold‐change, and p‐values of < 0.05 were considered significant.

3. Results

3.1. Early Viral RNA Expression and Infectious Virus Progeny Production During Infection of hNECs With HPIV Types 1 to 4

We first examined the viral replication kinetics of HPIV types 1 to 4 during acute infection of hNECs. Viral RNAs of HPIV 1 to 4 were detectable as early as 8 hpi in all five subject‐derived hNECs, subsequently increased exponentially, and peaked at 48 and 72 hpi (Figure 1A–E). Plaque assay demonstrated that infectious virus progeny production was detected as early as 8 hpi, and the titer transiently tapered at 16 hpi before increasing and peaking at 72 hpi for all HPIV types (Figure 1F–J). The live virus titers of HPIV‐1 and HPIV‐3 increased more rapidly, followed by HPIV‐4A, reaching ~ 10⁶ PFU/mL, 2 × 10⁶ PFU/mL, and 7.5 × 10⁴ PFU/mL by 24 hpi, respectively. HPIV‐2 and HPIV‐4B titers reached 5 × 10⁵ PFU/mL and 4.5 × 10⁵ PFU/mL by 48 hpi, respectively. Among HPIV types, the titer of live progeny viruses of HPIV‐3 was the highest (2.75 × 10⁸ PFU/mL) by 72 hpi.

Figure 1.

Exponential increase in viral RNA and live virus titers in hNECs infected with HPIV types 1 to 4. HPIV RNA titers (expressed in delta ∆C_T values) were plotted for (A) HPIV‐1, (B) HPIV‐2, (C) HPIV‐3, (D) HPIV‐4A, and (E) HPIV‐4B. Using plaque assay, infectious virus progeny titers were quantified (PFU per mL) for (F) HPIV‐1, (G) HPIV‐2, (H) HPIV‐3, (I) HPIV‐4A, and (J) HPIV‐4B (n = 5). The corresponding p‐values were determined by using one‐way ANOVA, nonparametric, grouped, Dunn multiple comparison test, and p‐values were calculated by comparison to mock‐infected cells. The data are represented as medians with interquartile values.

3.2. Infection with the Four HPIV Types Mainly Targets Ciliated Cells of Differentiated hNECs in Vitro

Given that the viral load of HPIV peaked at 72 hpi, double IF staining was performed for mock‐ and HPIV‐infected hNECs at 72 hpi. Co‐IF staining of HPIV viral proteins with markers of the three major hNEC types, that is, ciliated cells (acetylated alpha‐tubulin), goblet cells (MUC5AC), and basal cells (p63), revealed that all four HPIV types primarily target ciliated cells. Higher proportions of ciliated cells were infected by HPIV‐1 (51%), HPIV‐3 (65%), and HPIV‐4A (75%) as compared to HPIV‐2 (14%) and HPIV‐4B (12%) (Figure 2A,B). It is worth noting that the differences in the percentage of infected ciliated cells between HPIV2 and HPIV4A infections, and between HPIV4A and HPIV4B infections are statistically significant (p < 0.05). On the other hand, co‐staining of MUC5AC was observed only for HPIV‐3‐infected hNECs (< 20%), while no positive staining of HPIV was observed in p63+ basal cells (Figure 2A,C). Interestingly, we observed a unique pattern of HPIV‐positive staining as shown by clustering of fluorescent signals at the apical side of some infected ciliated cells for all HPIV types (Figure 2D). Out of the total infected ciliated cells, over 70% of these cells displayed the unique apical distribution pattern (“apical+”) for infection with HPIV types 1, 3, 4A, and 4B, while only 43% of HPIV‐2‐infected ciliated cells exhibited apical distribution (Figure 2E). Despite the finding that all HPIV types predominantly target ciliated cells, there was a significant loss of acetylated α‐tubulin‐positive area for hNECs infected by HPIV‐types 1 and 3, while there was minimal loss of ciliated cells for hNECs infected with HPIV types 2 and 4B as determined by IF staining of intact hNECs on Transwell membrane as compared to mock‐infected hNECs (Figure 3A,B). However, there was no significant reduction in mucociliary function of the hNECs during infection with the four HPIV types, as quantified by ciliary beating frequency (CBF) (Figure 3C–G). Cross‐section staining of hNECs on Transwell membrane showed that epithelial integrity of hNECs was maintained throughout infection with the four HPIV types (Figure 3H). We also measured the TEER of the mock‐ and infected hNECs. The HPIV‐infected nasal epithelium remained relatively intact as the TEER consistently remained above 1000 Ω/cm² throughout the 72‐h infection period, even though the TEER was slightly reduced for infected hNECs compared to mock‐infected cells (Figure 3I–M). Our results suggest that infection with any HPIV type causes minimal damage to the epithelial integrity and mucociliary function of hNECs, although we observed loss of ciliated cells during infection with HPIV types 1 and 3.

Figure 2.

Cellular tropism of hNECs during infection with HPIV types 1 to 4. (A) Representative images of co‐staining of ciliated cells (acetylated α‐tubulin‐positive, green), goblet cells (MUC5AC‐positive, green), and basal cells (p63‐positive, green) with structural protein of HPIV types 1 to 4 (red) at 72 hpi. Co‐staining of a fraction of goblet cells with HPIV protein was detected for HPIV‐3 infection only, but not for the other HPIV types. Co‐staining of basal cells with HPIV protein was not detected. Scale bar = 20 µm. (B) Graph of the percentage of HPIV‐infected ciliated cells over total ciliated cells in hNECs. (C) Graph of the percentage of HPIV‐infected goblet cells over total goblet cells in hNECs. (D) Representative immunofluorescence images showing clustered positive signals at the apical side of the HPIV‐infected ciliated cells for hNEC infection with all HPIV types. Scale bar = 10 μm. (E) Graph of the percentage of HPIV‐infected ciliated cells with unique apical distribution pattern (“apical+”) over total infected ciliated cells in hNECs.

Figure 3.

Epithelial integrity and mucociliary function of hNECs during infection with HPIV types 1 to 4. (A) Representative images of co‐staining of mock‐ and HPIV‐infected hNECs cultured on transwell membrane with acetylated α‐tubulin (green) and HPIV viral protein (red) at 72 hpi. Scale bar = 20 µm. (B) Area of α‐tubulin‐positive staining was quantified in arbitrary unit (5 fields per individual, n = 3, evaluated at 400× magnification). (C–G) No significant change in ciliary beating frequency (CBF) was detected at different time‐points after infection with HPIV types 1 to 4 as compared to mock‐infected hNECs (n = 3). (H) Representative cross‐section images of co‐staining of mock‐ and HPIV‐infected hNECs grown on transwell membrane with acetylated α‐tubulin (green) and MUC5AC (red) of at 72 hpi. Scale bar = 20 µm. (I–M) No significant change in transepithelial electrical resistance (TEER) was detected at different time‐points after infection with HPIV types 1 to 4 as compared to the mock‐infected hNECs (n = 3). The p‐value was calculated by one‐way ANOVA and nonparametric Kruskal–Wallis test. The data are represented as medians with interquartile values.

3.3. Comparative Expression of RIG‐I‐Like Receptors During Infection of hNECs With the Four HPIV Types

To examine the activation of RIG‐I‐like receptors (RLRs) during the different HPIV infections, we quantified the mRNA levels of the cytoplasmic pathogen sensors RIG‐I and MDA5. A gradual increase in RIG‐I mRNA expression was observed as early as 8 hpi for HPIV‐1 and HPIV‐4, whereas HPIV‐2 and HPIV‐3 showed no or minimal expression of RIG‐I within first 24 h, indicative of delayed RIG‐I gene expression during the early stage of infection (Figure 4A–E). Similar trends were observed for MDA5 mRNA expression levels (Figure 4F–J).

Figure 4.

Differential patterns of upregulation of cytosolic pattern recognition receptors RIG‐I and MDA5 during hNEC infection by different HPIV types. (A–E) Fold change (FC) of RIG‐I mRNA levels in HPIV‐infected hNECs as compared to mock‐infected hNECs (n = 5). (F–J) Fold change of MDA5 mRNA levels in HPIV‐infected hNECs as compared to mock‐infected hNECs (n = 5). The p‐value was calculated by one‐way ANOVA and nonparametric Kruskal–Wallis test. The data are represented as medians with interquartile values.

3.4. Induction of Type I and III Interferons in Response to hNEC Infection With the Four HPIV Types

We next examined the host antiviral responses elicited by HPIV infections of hNECs. Following the sensing of HPIV infection via the cytosolic pathogen sensors RIG‐I and MDA5, downstream interferon (IFN) expression is induced. Type I and type III IFN systems are critical to human antiviral innate immunity [25] and play crucial roles in defense against various respiratory viruses such as influenza virus, rhinovirus, and SARS‐CoV‐2 [22, 26, 27]. Since most studies on HPIV infection have focused on the activation of type I IFN [12, 28, 29], we explored the holistic responses of type I and III IFN systems during HPIV infection of hNECs. Our results revealed that there was minimal upregulation of type I IFN‐α1 throughout the 72‐h infection period for all HPIV types (Figure 5A–E). The expression of type I IFN‐β was generally subdued during the early stage of infection up to 24 hpi for HPIV‐1, and up to 48 hpi for the other HPIV types (Figure 5F–J). More markedly, HPIV‐1 infection exhibited a significant increase in IFN‐β mRNA levels at 48–72 hpi, whereas only a small increase in IFN‐β mRNA expression was observed at 72 hpi for the other HPIV types. On the other hand, while delayed expression of type III IFN‐λ1 was similarly observed for all four HPIV types during the early stage of infection, a greater upregulation of IFN‐λ1 was detected at 48–72 hpi for all HPIV types (especially HPIV‐1) as compared to type I IFN expression (Figure 5K–O).

Figure 5.

Differential patterns of upregulation of type I and III interferons during hNEC infection by different HPIV types. (A–E) Fold change of type I IFN‐α1 mRNA levels in HPIV‐infected hNECs as compared to mock‐infected hNECs (n = 5). (F–J) Fold change type I IFN‐β mRNA levels in HPIV‐infected hNECs as compared to mock‐infected hNECs (n = 5). (K–O) Fold change of type III IFN‐λ1 mRNA levels in HPIV‐infected hNECs as compared to mock‐infected hNECs (n = 5). Each individual dot represents hNECs from a single donor. The p‐value was calculated by one‐way ANOVA and nonparametric Kruskal–Wallis test. The data are represented as medians with interquartile values.

3.5. Differential Activation of IFN Signaling and Antiviral Interferon‐Stimulated Genes (ISGs) During hNEC Infection With the Four HPIV Types

Since IFNs trigger the activation of a wide array of antiviral genes that collectively establish the host antiviral response in the human airway, we assessed the activity of CXCL10 (a signature chemokine commonly used as a biomarker for respiratory tract viral infections) as well as ISGs MX1 and IFITM1 [30, 31]. IFN‐inducible CXCL10 preferentially signals for Th1 immune response by attracting activated Th1 lymphocytes—it is thus critical to the function of the innate immune system during viral infection [32, 33]. Similar to the pattern of upregulation for RIG‐I and MDA5, the mRNA levels of CXCL10 appeared as early as 8 hpi and gradually increased till 72 hpi for HPIV‐1 and HPIV‐4. However, for HPIV‐2 and HPIV‐3, the mRNA expression of CXCL10 was delayed at 8–16 hpi but subsequently increased at 24–72 hpi (Figure 6A–E). For HPIV‐1 and HPIV‐4, the upregulation of MX1 was gradual throughout the 72‐h infection period. For HPIV‐2 and HPIV‐3, the increase in MX1 mRNA expression was also apparently delayed at 8–16 hpi but subsequently increased significantly at 24–72 hpi (Figure 6F–J). On the other hand, IFITM1 exhibited relatively lower levels of mRNA upregulation in general as compared to mRNA levels of MX1 (Figure 6K–O), suggesting that MX1 may play a more critical role in antiviral responses against all HPIV types.

Figure 6.

Differential patterns of upregulation of IFN‐induced antiviral IFN‐stimulated genes (ISGs) during hNEC infection by different HPIV types. (A–E) Fold change (FC) of CXCL10 mRNA levels in HPIV‐infected hNECs as compared to mock‐infected hNECs (n = 5). (F–J) Fold change of MX1 mRNA levels in HPIV‐infected hNECs as compared to mock‐infected hNECs (n = 5). (K–O) Fold change of IFITM1 mRNA levels in HPIV‐infected hNECs as compared to mock‐infected hNECs (n = 5). The p‐value was calculated by one‐way ANOVA and nonparametric Kruskal–Wallis test. The data are represented as medians with interquartile values.

4. Discussion

In this study, we characterized the replication kinetics and innate immune responses of contemporary clinical isolates representing the four major HPIV subtypes. These strains were selected to reflect currently circulating viruses and to facilitate direct comparison between the four major HPIV subtypes. HPIV‐1 and HPIV‐3 comprise several co‐circulating sublineages—thus, future studies incorporating multiple strains from different sublineages will be valuable to ascertain if there are intrasubtype variations in viral replication and host responses. However, the number of available genome sequences of HPIV‐2 and HPIV‐4 remains limited, which constrained the inclusion of multiple isolates. The HPIV‐4A and HPIV‐4B strains included in this study were therefore chosen as clade‐specific strains. This highlights the need for improved genomic surveillance of these HPIV subtypes to better understand their genetic diversity and evolutionary dynamics. Such efforts will provide a more comprehensive perspective of the biological heterogeneity among HPIVs and its implications for disease pathogenesis and control strategies.

Using the differentiated hNEC model, the novelty of our study is to holistically analyse and highlight distinct features of infection by all four HPIV types in human‐derived nasal epithelium, along with the host cell responses. We found that the titers of infectious virus progeny and viral RNA indicated similar trends of viral replication during infection, and were detectable as early as 8 hpi for all HPIV types. HPIV‐3 exhibited the highest level of live virus progeny production, despite comparable viral RNA titers across all HPIV types. This suggests that HPIV‐3 has the greatest replicative efficiency in the production of live virus during hNEC infection compared to other HPIV types. HPIV‐3 is reported to be the most virulent HPIV and is associated with considerable morbidity and mortality [11, 34]. These observations may be partially explained by the relatively high viral production efficiency of HPIV‐3 in the human airway. Interestingly, while HPIV‐3 is one of the most common causes of lower airway diseases such as bronchiolitis and pneumonia, our study showed that HPIV‐3 is also highly infectious in the nasal epithelial cells of the upper airway. On the other hand, although HPIV‐1 and HPIV‐2 infection of the upper airway commonly results in croup in children, we found that virus progeny production of HPIV‐1 and HPIV‐2 was relatively lower than HPIV‐3. These suggest that viral load alone may not be the main determining factor for causing disease severity of croup. Despite HPIV‐4 being the only HPIV type that is not known to evade the interferon‐induced antiviral effect [35], our study showed that HPIV‐4 replicates efficiently in hNECs and generates infectious virus progeny at levels similar to HPIV‐1 and HPIV‐2. These differences in viral release by the HPIV types in the upper airway imply varying capacities for spreading to the lower respiratory tract, potentially contributing to differences in virulence.

We next compared the effects of infection of hNECs by the four HPIV types in terms of viral tropism and changes in mucociliary function. Our hNEC model previously demonstrated that ciliated cells are the main target for human rhinovirus infection, while H3N2 influenza virus targets both ciliated and goblet cells [27, 36]. Studies have documented that HPIV types 1, 2, and 3 specifically infect ciliated cells in differentiated human tracheobronchial epithelial cells [16, 37, 38]. However, it is not yet known whether HPIVs can infect other human airway cell types, particularly the goblet cells and basal cells, as there was no prior validation using co‐staining of HPIV antigen with the relevant cellular markers. In addition, there is hitherto no study on the cellular tropism of HPIV‐4. Therefore, our study employed IF staining to co‐stain specific HPIV types with the cellular markers for ciliated cells, goblet cells, and basal cells. We report for the first time that all HPIV types primarily target ciliated cells, while HPIV‐3 also targets a fraction of goblet cells (< 20%). HPIV‐1 primarily binds to terminal α2,3‐linked sialic acid (SA) residues, while HPIV‐3 recognizes both α2,3‐ and α2,6‐linked SA receptors in the human airway [39, 40]. In humans, α2,6‐linked SA receptors are mainly found in both ciliated and goblet cells in the upper airway (i.e., nasal turbinates and larynx), and in the upper part of the lower airway (i.e., trachea and bronchus) [41]. In contrast, α2,3‐linked SA receptors are localized to the ciliated epithelium of the bronchioles and alveoli [42]. This distinct spatial distribution likely influences HPIV spread and virulence. Additionally, age‐related differences likely influence susceptibility of individuals to HPIV infections. Widespread α2,6‐linked SA receptor expression in the upper airway is observed in adults, whereas young children predominantly express α2,3‐linked receptors in the respiratory tract [43]. Further research into receptor distribution by age and airway region may clarify how viral tropism affects HPIV type‐specific transmission and pathogenesis.

In addition, we assessed the epithelial integrity and mucociliary function via measurements of TEER and CBF in HPIV‐infected hNECs. HPIV types 1 to 4 cause minimal disruption to the hNEC epithelium and minimal changes to CBF, suggesting preserved epithelial integrity and ciliary function. We found greater cilia shedding for infection with HPIV types 1 and 3, while minimal loss of cilia was observed for HPIV types 2, 4A, and 4B. This aligns with other studies on human airway epithelial (HAE) cells, where infections with HPIV types 1, 2, and 3 reveal intact epithelial layers; moderate loss of ciliated cells for HPIV types 1 and 3, and minimal loss of ciliated cells for HPIV‐2 [16, 37, 44]. Interestingly, a unique pattern of HPIV‐positive staining depicted by clustering of fluorescent signals at the apical side of infected ciliated cells was observed across all four HPIV types. A similar pattern of localization was also observed for HPIV‐3 infection of HAE cells [37]. This apical localization suggests a unique mechanism for viral release via ciliated cells, distinguishing HPIVs from other respiratory viruses such as human rhinovirus and influenza A virus [27, 36]. Interestingly, one study reported the entry and release of the respiratory virus SARS‐CoV‐2 from motile cilia and microvilli of infected hNECs [45]. Similarly, another study also reported apical syncytia formation during SARS‐CoV‐2 infection, as indicated by formation and synchronized release of large clusters of infected cells and syncytia into the apical lumen, contributing to virus dissemination [46]. While our study showed clear localization of HPIV at the apical side of ciliated cells, whether HPIV co‐localizes with cilia—suggesting a potential release mechanism via ciliary trafficking—warrants further investigation.

Besides serving as a physical barrier against respiratory viruses, the nasal epithelium also elicits innate immune responses at the site of infection. RLRs, including RIG‐I and MDA5, are key cytoplasmic pathogen sensors of virus infection, which signal type I and III interferon responses. Our study observed upregulation of RIG‐I and MDA5 during infection by all HPIV types. We found that HPIV‐1, HPIV‐4A, and HPIV‐4B showed gradual upregulation of RIG‐I and MDA5, whereas HPIV‐2 and HPIV‐3 showed no or delayed activation before 48 hpi. Type I (IFN‐β) and III (IFN‐λ1) IFNs were generally delayed and only significantly increased at 48–72 hpi, with type III IFN being the dominant response in hNECs, particularly with HPIV‐1 infection. Interestingly, this delay in activation of IFN was not observed in another respiratory virus infection, that is, influenza A virus. Yan et al. reported a gradual increase in mRNA levels of IFN‐β and IFN‐λ1 from 0 to 8 hpi, with a significant spike at 16 hpi before they peaked at 24 hpi during H3N2 infection of hNECs [27]. The viral accessory proteins of HPIVs likely modulate the induction of host RLR and IFNs. The C proteins expressed by HPIV‐1 and HPIV‐3 likely reduce viral RNA synthesis during replication and limit RLR activation [47, 48]. The stronger suppression or delayed response of both IFN‐β and IFN‐λ1 for HPIV‐3 as compared to HPIV‐1 may be attributed to the single C protein encoded by HPIV‐3 as compared the four C proteins encoded by HPIV‐1. On the other hand, HPIV‐2 also suppressed activation of RLR and IFNs at the early stage of hNEC infection. The V proteins, which are only encoded by HPIV‐2 and HPIV‐4, also restrict the production of viral RNA for HPIV‐2 and reduce RLR and IFN activation [49, 50]. Previous studies using human cell lines showed that HPIV‐4 induced early pro‐inflammatory response, and was incapable of blocking IFN‐induced signaling [14, 29]. Taken together, these observations suggest that each HPIV type uses distinct mechanisms to evade host immunity, warranting further investigation into the roles of C, V, and other viral proteins in host antiviral surveillance using the physiologically relevant pseudostratified hNEC model.

Our study further investigated IFN signaling and ISG regulation during HPIV infection of hNECs. Despite delayed activation of type I and III IFNs, a gradual and robust upregulation of MX1 was observed for HPIV‐1 infection. However, for the other HPIV types, MX1 was only significantly upregulated from 24 to 72 hpi. This ISG induction is driven by the JAK/STAT pathway and involves antiviral signaling mechanisms associated with IFNs. Studies have shown that HPIV type‐specific accessory proteins interfere with IFN signaling differently. HPIV‐1 C protein binds to STAT1 (but not STAT2) to block its nuclear translocation, while HPIV‐3 C proteins inhibit STAT1 phosphorylation [48, 51]. In contrast, HPIV‐2 V proteins target STAT2 for degradation [12, 52], although this targeting may vary depending on host species and cell type [53]. Although the HPIV‐4 V protein reportedly lacks IFN‐antagonistic activity in HeLa cells [29], our study demonstrated that HPIV‐4 delayed activation of IFN at the early phase of infection but effectively induced ISG expression in infected hNECs, similar to other HPIV types. These findings highlight the importance of type‐specific accessory proteins in IFN evasion and their potential roles in HPIV virulence. Models such as the human STAT2 transgenic mouse infected with HPIV‐5, which shows STAT1 degradation and subsequent IFN antagonism, offer valuable tools for investigating these mechanisms in detail [54].

Overall, our study showed that among the four HPIV types, HPIV‐3 replicates and produces infectious virus progeny most efficiently with similar levels of viral RNA loads as compared to other HPIV types. While HPIV‐1 and HPIV‐3 caused a significant loss of ciliated cells during infection, HPIV‐1 induced the highest innate immune response among the four HPIV types. Our in vitro findings demonstrated patterns that are consistent with virus detection rates and clinical features reported in previous clinical studies for all four HPIV types in young children. From 2015 to 2023, HPIV‐3 was the most frequently isolated serotype, followed by HPIV‐1, HPIV‐4, and HPIV‐2 [55, 56, 57]. Han et al. reported that upper respiratory infections were more common among patients infected with HPIV‐2 and HPIV‐3, whereas patients with HPIV‐1 and HPIV‐4 were more frequently diagnosed with lower respiratory infections, including bronchitis, bronchiolitis, abnormal breath sounds, and chest x‐ray abnormalities. Notably, HPIV‐1 was associated with the highest proportion of croup diagnoses compared with other HPIV types [57]. A retrospective cohort study conducted from 2007 to 2018 found that children with HPIV‐1 infection were more likely to present with croup, whereas HPIV‐3 infection was more often associated with pneumonia and longer hospitalization durations [58]. Together with our in vitro results, which showed the strongest immune response for HPIV‐1 followed by HPIV‐3, these findings suggest that differences in host immune activation may contribute to the distinct clinical manifestations observed for each HPIV type.

A limitation of our study is that the hNECs used were derived exclusively from adult patients. This may not fully capture age‐specific responses since HPIV infections (particularly HPIV‐3) are most prevalent in children under five, and afflict up to 50% of infants within their first year [9]. Transcriptomic analysis has reported higher baseline interferon‐associated gene expression in pediatric epithelium as compared to adult, suggesting intrinsic differences in antiviral defense mechanisms and susceptibility of airway epithelium to viral infection in pediatric individuals [59, 60]. Hence, this constitutively heightened basal interferon response induces a preactivated antiviral state which can restrict early SARS‐CoV‐2 replication, thus allowing more effective local innate immune responses, whereas adults show weaker or delayed interferon signaling [46, 61]. Another study on RSV, a highly prevalent respiratory virus of infants, also reported robust virus spread and extensive cell death only in infant bronchial epithelium as compared to adult, suggesting that this intrinsic difference in airway epithelium may play a critical role in pathogenesis and disease progression during virus infection [62]. While our study on HPIV infection of hNECs derived from adult donors showed a delay in interferon response, it may not fully reflect the holistic epithelial immune profile of the pediatric airway. Future studies should compare hNECs derived from pediatric versus adult subjects to explore age‐related differences in airway susceptibility to HPIV infection. Another limitation is the use of only one clinical strain per HPIV type. Broader analysis with multiple strains from all four HPIV types is needed. Ideally, infection models incorporating hNECs co‐cultured with autologous peripheral blood mononuclear cells (PBMCs) provide more comprehensive insights into the coordinated dynamics of innate and adaptive immunity [63].

5. Conclusion

In conclusion, our study revealed that all four types of HPIV are highly infectious and readily replicate in hNECs. Our results indicate delayed IFN activation following HPIV infection, although immune evasion attributed to the variations in type and accessory protein function require detailed investigations. Future studies should explore these virus‐host interactions to better understand the differences in virulence, pathogenesis, and airway disease progression between HPIV types, especially in high‐risk patients.

Author Contributions

De‐Yun Wang and Gavin J. D. Smith conceived the study. Hsiao Hui Ong performed all of the experiments and wrote the manuscript draft. Jing Liu assisted with the in vitro cell culture and experiments. Mark Thong contributed to subject recruitment, obtained informed consent, and helped with the collection and analysis of clinical samples and data. Vincent T. K. Chow provided laboratory resources. Vincent T. K. Chow, Yvonne C. F. Su, and Gavin J. D. Smith critically reviewed and edited the manuscript. All the authors have read and agreed to the published version of the manuscript.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

PIV ms‐Supplementary materials‐2.

Table S1: Patient information. Table S2: SYBR‐Green‐based qPCR primer sequences.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.