Generation of an Induced Pluripotent Stem Cell‐Derived Alveolar Type II In Vitro Model to Study Influenza A Virus Infection and Drug Treatments
Lena Gauthier, Hristina Koceva, Yann Bachelot, Rosanne W. Koutstaal, Puck B. van Kasteren, Marc Thilo Figge, Christian Eggeling, Alexander Sandy Mosig
TL;DR
Researchers created a lab model using stem cells to study how influenza viruses infect lung cells and respond to drugs like oseltamivir.
Contribution
This is the first study to demonstrate influenza A virus infection and drug response in induced pluripotent stem cell-derived alveolar type II cells.
Findings
iAT2 cells grown at air–liquid interface support productive influenza A virus replication.
iAT2 cells show antiviral transcriptional responses and respond to oseltamivir treatment.
The model is scalable and physiologically relevant for influenza research and drug testing.
Abstract
Influenza viruses (IVs) represent a significant global health issue, capable of causing seasonal epidemics and occasional pandemics with substantial morbidity and mortality. The emergence of viral resistance further complicates treatment strategies. In this study, induced pluripotent stem cell‐derived human alveolar type II (iAT2) cells are used to model influenza A virus (IAV) infection and to assess antiviral responses. Cultured at an air–liquid interface (ALI) in transwell systems, iAT2 cells recapitulate key features of the alveolar epithelium and support productive IAV replication. Upon infection, iAT2 cells mounted an antiviral transcriptional response and exhibited sensitivity to oseltamivir treatment, consistent with its established in vivo efficacy. Together, these findings highlight the utility of iAT2 cells as a scalable, physiologically relevant in vitro model for influenza…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5
FIGURE 6
FIGURE 7
FIGURE 8
FIGURE 9- —Stephan Ludwig
- —Deutsche Forschungsgemeinschaft10.13039/501100001659
- —BMFTR Germany
- —State of Thuringia
- —H2020 European Institute of Innovation and Technology10.13039/100010686
- —Leibniz‐Gemeinschaft10.13039/501100001664
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRespiratory viral infections research · Influenza Virus Research Studies · Neonatal Respiratory Health Research
Introduction
1
Influenza, a highly contagious respiratory illness caused by influenza viruses (IVs), poses a significant global health concern. Seasonal epidemic and sporadic pandemic outbreaks contribute to high morbidity and mortality and entail a large socioeconomic burden. The World Health Organization estimates each year a billion cases of seasonal influenza including three to five million severe cases and 250 000 to 650 000 influenza‐associated respiratory deaths globally [1, 2, 3].
IVs are enveloped viruses with a single‐stranded RNA genome [4]. They are subdivided into influenza A, B, C, and D viruses. However, mainly types A and B are relevant for human disease [3, 5, 6]. The pathogenesis of an IV infection depends on the virulence of the infective strain but also on the host response [7, 8]. An effective immune response is required for viral clearing, but some patients react with the overproduction of pro‐inflammatory cytokines, which is often cited as the major cause of acute respiratory failure [9, 10, 11, 12, 13, 14]. Although vaccines are available and remain the primary strategy for preventing influenza, their effectiveness can be compromised due to the rapid evolution of IVs and the seasonal variance of prevalent strains [3, 15, 16, 17, 18, 19]. Therefore, anti‐influenza drugs provide an essential complement to vaccination, in particular for high‐risk populations or in regions where vaccine access is limited [20]. Additionally, adjunctive immunomodulatory therapies are discussed to treat severe influenza, but the traditional one‐size‐fits‐all approach has not been satisfactory so far [21, 22, 23, 24]. However, the advent of individualized medicine offers a promising avenue for better influenza management [24, 25].
Animal models are widely used to study anti‐influenza vaccines and therapeutic treatments due to their ability to mimic lung physiology [26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37]. Yet, ethical concerns and inter‐species differences limit their potential, but in vitro cell culture systems offer a cost‐effective and easy‐to‐handle alternative [26, 38, 39]. Here, immortalized cell lines and primary cells are traditionally used, while induced pluripotent stem cells (iPSCs) have recently arisen as a versatile new opportunity [40, 41]. iPSCs will help to advance our current health system toward personalized medicine as they are generated from patient‐specific cells, allowing for the consideration of patient‐specific factors such as genetic background, immune state, and disease history in in vitro studies [42, 43, 44]. Furthermore, iPSCs can be differentiated into any cell type, including alveolar type II (AT2) cells, which are the primary targets of several respiratory pathogens, including IVs [7, 45, 46, 47, 48, 49]. AT2 cells of the lung line the alveoli at the distal end of the lower respiratory tract. They are positioned at an air–liquid interface (ALI) with their apical surface exposed to the airway lumen and their basal surface in contact with the underlying tissue. The initial contact with respiratory viruses typically occurs via the apical side as pathogens are inhaled [50]. Following infection, IVs exploit the polarized architecture of AT2 cells to ensure its efficient replication and spread. Viral assembly and budding are predominantly directed toward the apical surface, which not only maximizes the release of progeny virions into the airway lumen, promoting transmission via droplets or aerosols, but also allows the virus to evade immune responses that are more active at the basal surface, thus enhancing viral persistence and pathogenicity [51].
Kotton et al. recently highlighted the potential of iPSC‐derived AT2 (iAT2) cells to study the severe acute respiratory syndrome coronavirus type 2 (SARS‐CoV‐2) in transwell inserts with ALI [52]. Notably, the ALI culture in transwell inserts is a powerful tool that mimics the in vivo environment of AT2 cells with an apical air‐exposed side and a basal side exposed to cell culture medium. Furthermore, the ALI helps to maintain the AT2 phenotype over a long time enhancing the accuracy of in vitro studies [52, 53, 54, 55, 56, 57].
In our work, we evaluate the applicability of iAT2 cells as an anti‐influenza drug testing platform. We show that iAT2 cells grown at ALI as 2D cultures in transwells can be infected by the influenza A virus (IAV). We monitor the AT2 phenotype in culture, determine infection conditions for IAV, and analyze the transcriptomic changes upon IAV infection. As a proof of concept, we show that iAT2 cells are responsive to the clinically applied anti‐influenza drug oseltamivir. Overall, our study emphasizes the great potential of iAT2 cells for influenza research that may pave the way toward developing future individualized medical approaches.
Results
2
iAT2 Cells Express Lung Progenitor Markers
2.1
We successfully replicated the protocol for deriving iAT2 cells, which was first published by Jacob et al. [58]. Briefly, iPSCs were differentiated into early lung progenitor cells and further into distal alveolar epithelial cells (Figure 1a). Proper differentiation was monitored by sorting the cells for expression of NK2 homeobox 1 (NKX2.1) and the surface‐active agent (surfactant) protein C (SP‐C). These two proteins are crucial contributors to proper AT2 cell development and function [46, 59]. Flow cytometry analysis revealed the differentiation of 87% NKX2.1^+^ progenitor cells on day 14 or 15 and 96.8% SP‐C^+^ cells on day 30+ (Figure 1b). Self‐renewing iAT2 cells were expanded in 3D cultures as alveolospheres (Figure 1c) and seeded as single cells onto transwell inserts for characterization at any day after day 30 (day X in Figure 1a). Cell growth and epithelial barrier formation in transwells were monitored by transmission light microscopy (Figure 1d). After ALI exposure, iAT2 cells formed confluent monolayers on the insert membranes at 4 and 6 days after seeding, indicating proper barrier function. Further, immunofluorescence analysis at 9 days after seeding revealed strong expression of key alveolar epithelial markers, including the transcription factor NKX2.1, which is essential for lung development and respiratory epithelial identity [46], SP‐C and lysophosphatidylcholine acyltransferase 1 (LPCAT1). SP‐C is a well‐established AT2‐specific marker involved in reducing alveolar surface tension and normal respiratory function [60, 61, 62, 63], while LPCAT1, a lamellar body‐associated enzyme, contributes to pulmonary surfactant homeostasis [64]. Together, these markers confirm successful differentiation and maturation into functional iAT2 cells [63, 65, 66]. Additionally, moderate antigen Kiel 67 (Ki67) expression indicated active cell division reflecting the regenerative potential of the cells (Figure 1e) [67, 68, 69, 70].
Differentiation, cultivation, and characterization of iAT2 cells. (a) Overview of the differentiation, cultivation, and plating procedure of iAT2s as described in the methods. iPSCs: induced pluripotent stem cells; iAT2s: iPSC‐derived alveolar type II cells; ALI: air–liquid interface; NKX2.1: NK2 homeobox 1; SP‐C: surface active agent protein C; for StemDiff, DS/SB, CBRa, CK/DCI, CK/DCI+Y: see methods. (b) Representative flow cytometry plots of the lung progenitor markers NKX2.1 coupled to GFP on day 14 (left) and SP‐C coupled to tdTomato on day 30 (right). (c) Representative transmission light image of alveolospheres within matrigel before passaging on day 35. Scale bar = 30 µm. (d) Representative transmission light images of iAT2s seeded on transwell inserts directly after seeding (day X) and 4 or 6 days after seeding (day X + 4 and X + 6). Scale bars = 50 µm. (e) Immunofluorescence images of iAT2s showing expression of antigen Kiel 67 (Ki67; green, left), SP‐C (magenta, left), lysophosphatidylcholine acyltransferase 1 (LPCAT1; yellow, middle), and NKX2.1 (green, right). Nuclei are displayed in blue. Single z‐planes are shown. Scale bars = 20 µm (left, middle) or 50 µm (right).
iAT2 Cells are Permissive to IAV and Provide a New Platform to Study IV Infections
2.2
We determined the cellular receptiveness to IAV to establish iAT2 cells as an in vitro model for IV infection. During infection, the initial event is the binding of the viral protein hemagglutinin to sialic acids (SAs) at the host cell surface [71]. SAs are acidic sugars typically found as the terminal sugar chain moiety of glycoproteins and glycolipids where they attach to the underlying oligosaccharide via α2,3‐, α2,6‐, α2,8‐, or α2,9‐glycosidic linkages. Human‐pathogenic IVs preferentially bind to α2,6‐linked SAs, while avian IVs mainly recognize α2,3‐linked SAs [72]. Experimentally, SAs are commonly detected using the sugar‐binding lectins Sambucus nigra agglutinin (SNA), specifically binding to α2,6‐linked SAs [73, 74], and Maackia amurensis hemagglutinin (MAH), which targets α2,3‐linked SAs [75, 76]. Prior treatment with recombinant neuraminidase (NA), which catalyzes the hydrolysis of non‐reducing terminal SA residues from carbohydrates and glycoproteins, can serve as an additional control. Fluorescence microscopy of our iAT2 cells stained with SNA or MAH, respectively, revealed strong expression of α2,6‐linked SAs on their apical surface while no α2,3‐linked SAs were detected (Figure 2a). Prior NA treatment resulted in a low SNA signal, further corroborating the presence of terminal SA moieties at iAT2 cells (Figure 2b).
iAT2 cells express α2,6‐linked but not α2,3‐linked sialic acids on their surface. (a,b) Lectin staining of non‐infected iAT2 cells at 9 days after seeding. (a) iAT2s were stained for α2,6‐ or α2,3‐linked sialic acids (SAs). (b) Cells were treated with recombinant neuraminidase (NA) before staining. Images were acquired as z‐stacks. Representative orthogonal views (left and right panels) with insets (middle panels) from one experiment are shown. Pink crosshairs indicate xz‐ and yz‐side view positions displayed in pink rectangles. Yellow rectangles indicate inset positions. Green: α2,6‐ (left and middle panels) or α2,3‐ (right panels) linked SAs; grey: nuclei; scale bars = 50 µm.
Given the clear apical distribution of α2,6‐linked SAs, we further assessed epithelial polarity of iAT2 cells grown at ALI. Actin staining revealed a pronounced, circumferential signal at the apical side, while basal actin stress fibers were absent, consistent with a non‐migratory, epithelial phenotype (Figure 3). E‐cadherin (E‐Cad) localized to the lateral membranes, outlining individual cells and supporting the formation of intercellular contacts. This pattern of cytoskeletal and junctional organization indicates a polarized epithelial architecture, in line with observations from apical‐out distal lung [77] or small airway epithelial [78] organoid models.
iAT2 cells cultured at ALI in transwells show a polarized epithelial architecture. Fluorescence images of iAT2 cells at 9 days after seeding show a pronounced apical actin (magenta) signal and E‐Cadherin (green) localization to lateral membranes. Z‐stack images are shown in orthogonal view. Yellow crosshairs indicate xz‐ and yz‐side view positions displayed in yellow rectangles. Magenta: actin; green: E‐Cadherin; scale bars = 50 µm.
Based on these findings, we next investigated whether the apically localized α2,6‐linked SAs on iAT2 cells are functionally accessible to facilitate IAV infection and, if so, whether IAV exploits the cellular epithelial polarity to direct progeny release predominantly toward the apical surface. To this end, cells were cultured at ALI for 7 days prior to viral exposure. We initially employed a high multiplicity of infection (MOI) of 10 with an initial viral exposure time of 0.5 h, a commonly used adsorption time in cell culture experiments with IAV to allow viral attachment and internalization [79]. At 10 h post‐infection (hpi), apical and basal supernatants were collected for viral titer determination, and cells were fixed for immunofluorescence studies. Titer determinations showed a moderate release of progeny virus toward the apical side implying that iAT2s are generally permissive to IAV infection. Furthermore, only low titers were detected on the basal compartment, indicating directed viral egress primarily via the apical surface (Figure 4a). In addition to evaluating viral titers from supernatants, immunofluorescence microscopy was leveraged to support the iAT2 permissiveness to IAV. Fluorescent staining of the intracellular distribution of the viral nucleoprotein (NP) as well as of the cellular adherens junction component E‐Cad allowed us to study the extent of cell infection and integrity, respectively. At 10 hpi, we mainly observed strong NP signals in the nuclei of infected cells, while only a small subset also displayed weaker NP signals within the cytosol (Figure 4b). Interestingly, the cytosolic NP signal was always more pronounced toward the apical iAT2 side supporting the notion of preferential apical budding. The E‐Cad signal was comparable between infected and uninfected cells indicating no severe effect of the IAV infection on the E‐Cad integrity at 10 hpi.
iAT2 cells are permissive to IAV infection. (a) iAT2s were infected with a high multiplicity of infection (MOI) of 10 or left uninfected (MOI 0), and supernatants were harvested at 10 h post‐infection (hpi). Apical and basal viral titers were determined by plaque assay; data show mean ± SD of 4 individual experiments; PFU: plaque‐forming units. (b) Immunofluorescence staining of viral nucleoprotein (magenta) shows intracellular viral protein distribution at 10 hpi. Images were acquired as z‐stacks, and representative orthogonal views from one experiment are shown. Yellow crosshairs indicate xz‐ and yz‐side view positions displayed in yellow rectangles. Magenta: viral nucleoprotein; green: E‐Cadherin; blue: nuclei; scale bars = 5 µm.
Optimization of IAV Infection Conditions for iAT2 Cells
2.3
IAV replication dynamics in cell culture vary greatly between different cell types, and depending on the application, different infection scenarios must be employed. The two major determinants impacting progeny virus production are infectious dose, defined by MOI, and harvest time given as hpi. To carefully choose appropriate experimental conditions, it is essential to consider the concept of single‐ versus multi‐cycle replication experiments which is well explained by Beyleveld et al. [80]. Since it is estimated that IVs have a replication cycle of approximately 8 h in cell culture [79], supernatants for a single‐cycle infection experiment are commonly harvested at 8 hpi. Thus, 24 or 48 hpi reflects up to six replication cycles. Besides virus input and harvest time, the primary incubation time with virus inoculum influences the initial viral attachment rate when incubated stationary [81]. Moreover, the distinction between apical and basal supernatants is important, since—as highlighted before—in a natural context the cells’ apical surface is exposed to the airway lumen and their basal surface is in contact with the underlying tissue, introducing a polarity with respect to infection cycling.
Since we obtained progeny virus titers upon infection with MOI 10 at 10 hpi, we now applied lower MOIs to study later time points. Upon infection with MOI 5, increasing the initial viral absorption period from 0.5 to 2 h led to higher apical viral titers at 8, 24, and 48 hpi. Viral titers on the basal side were overall low but also tended to increase upon prolonged initial viral exposure (Figure 5a). Upon infection with escalating MOIs (MOI 0.1 to 10 for 8 hpi and MOI 0.001 to 5 for 24/48 hpi), viral titers in apical and basal supernatants increased dose‐dependently. Viral titers on the apical side culminated in the 10^5^ range at 8 and 24 hpi, and in the 10^7^ range at 48 hpi. In contrast, viral titers on the basal side were by 3 to 5 log units lower, indicating preferential viral release from the apical surface (Figure 5b), consistent with our previous observation (Figure 4). Immunofluorescence images showed increasing numbers of infected cells with higher MOIs at 24 hpi, as visualized by viral NP and E‐Cad staining (Figure 5c), supporting both the titer data and the integrity of the epithelial layer during infection. Similar infection dynamics were also observed over an extended time course using a low MOI of 0.1 to maintain epithelial viability. Apical titers steadily increased from 8 to 168 hpi, while basal titers remained low, and E‐Cad staining confirmed monolayer integrity throughout (Figure S1). These findings highlight the robustness of the iAT2 model and its ability to sustain prolonged infection, in contrast to conventional cell lines that typically undergo rapid virus‐induced cytopathic effects.
Establishment of IAV infection conditions for iAT2 cells. (a) iAT2 cells were infected with a multiplicity of infection (MOI) of 5. Increasing initial viral adsorption periods from 0.5 to 2 h and different harvest time points were tested; hpi: hours post‐infection. (b) iAT2 cells were infected with an initial viral adsorption period of 2 h. The impact of increasing MOIs was measured. Apical and basal viral titers in (a) and (b) were determined by plaque assay; data show mean ± SD of 3 (48 hpi in a) or 4 individual experiments (all others); PFU: plaque‐forming units. (c) Immunofluorescence staining of viral nucleoprotein (magenta) at 24 hpi shows intracellular viral protein distribution. Images were acquired as z‐stacks, and representative maximum intensity projections of one experiment are shown. Magenta: viral nucleoprotein; green: E‐Cadherin; blue: nuclei; AB ctrl: control staining without primary antibodies (MOI 5); scale bars = 50 µm.
iAT2 Cells Maintain Epithelial Barrier Integrity During IAV Infection
2.4
To assess the impact of IAV infection on epithelial barrier function, we next measured transepithelial electrical resistance (TEER). iAT2 cells were infected apically with MOI 1 or MOI 10, and TEER was measured prior to infection and at 8, 24, and 48 hpi. In uninfected controls continuously maintained at ALI, TEER values remained stable throughout the experiment. In contrast, TEER decreased transiently at 8 hpi in infected (MOI 1, MOI 10) and mock‐treated cells (MOI 0), followed by a recovery to baseline or higher levels at 24 and 48 hpi. This transient drop was observed for infected and mock‐treated samples but not for ALI controls, suggesting that it may result from apical medium exposure rather than infection, in line with previous reports of lower TEER under submerged conditions compared to ALI cultures [53]. In contrast, EDTA treatment led to a marked and sustained reduction in TEER, confirming effective barrier disruption (Figure 6).
iAT2 cells maintain epithelial barrier integrity during IAV infection. iAT2s were infected with a multiplicity of infection (MOI) of 1 or 10, or mock‐treated with medium (MOI 0). Control cells were kept at an air–liquid interface (ALI) or treated with 6 mm EDTA (pH 7) throughout the experiment. Transepithelial electrical resistance (TEER) was measured prior to infection and at 8, 24, and 48 h post‐infection (hpi). Data show mean ± SD of 4 (8, 24, and 48 hpi) or 2 (0 hpi) individual experiments.
Transcriptomic Profiling Reveals Antiviral and Inflammatory Responses in IAV‐Infected iAT2 Cells
2.5
To characterize the antiviral and inflammatory response of iAT2 cells to IAV infection, we performed bulk RNA sequencing of iAT2 cells infected with MOI 5 or MOI 10 for 24 h. Principal component analysis (PCA) of transcriptomic profiles showed clear separation between infected and control samples along the first principal component, indicating robust infection‐driven transcriptional changes, with moderate variation among replicates (Figure S2). Differentially expressed gene (DEG) analysis identified a set of 70 significantly upregulated and 42 downregulated genes for MOI 5. For MOI 10, we found 84 significantly upregulated and 44 downregulated genes (Figure S2). These transcriptomic alterations indicate activation of multiple immune response pathways upon IAV infection. To explicitly assess epithelial antiviral activation, we compiled a panel of representative markers across key categories (Figures 7 and S3). Several pattern recognition receptors, including DDX58 (RIG‐I), IFIH1 (MDA5), and TLR3, were upregulated alongside a broad set of interferon (IFN)‐stimulated genes (ISGs) such as Mx2, OAS2, IFIT2, IFI6, IFI44L, and RSAD2. Components of the JAK/STAT signaling cascade (STAT1, STAT2, and IRF9) and genes involved in ISGylation and deISGylation (ISG15, USP18) were also elevated. Moderate increases were observed for some proinflammatory cytokines (IL1β, IL18) and chemokines (CXCL2, CXCL11, and CCL20), while housekeeping controls remained stable, supporting pathway‐specific responses rather than global shifts.
Antiviral and inflammatory gene expression signature in iAT2 cells upon IAV infection with MOI 5. iAT2 cells were infected with a multiplicity of infection (MOI) of 5 or left uninfected (MOI 0). At 24 hpi, RNA was extracted and subjected to bulk RNA sequencing. A panel of representative antiviral and inflammatory markers was compiled and grouped into functional categories as indicated with the colored strip at the bottom. Mean expression levels per condition (top) and infection‐induced changes (Δ = MOI 5 – MOI 0, bottom) are visualized in a two‐row heat map. CPM: counts per million; for gene symbols, see Table S1.
Furthermore, gene set enrichment analysis (GSEA) was assessed separately for up‐ and downregulated genes (Figure 8 and Figures S4–S6). Gene ontology (GO, Figure 8a) and Reactome (Figure 8b) pathway analyses confirmed significant enrichment of innate immune and antiviral signaling pathways, including type‐I IFN signaling, defense response to viruses, cytokine‐mediated signaling, and antiviral mechanisms by ISGs. Conversely, GSEA of downregulated genes indicated attenuation of IFNγ‐related signaling together with reduced IL18‐ and IL27‐associated pathways, which typically reflect immune‐cell crosstalk and are less engaged in epithelial monoculture [82]. Similar pathways and marker genes were detected at MOI 10 with higher expression (Figures S3, S4, and S6), indicating dose‐responsive activation and a robust assay dynamic range.
Gene set enrichment analysis of DEGs in iAT2 cells infected with IAV (MOI 5). (a,b) Gene ontology (GO) biological process (BP) analysis (a) and Reactome pathway enrichment (b) of significantly up‐ (left) and downregulated (right) genes in iAT2 cells infected with IAV (MOI 5, 24 hpi) compared to uninfected controls. Analyses were performed using all differentially expressed genes (DEGs; with false discovery rate < 0.05). Top‐ranked pathways are shown, with color‐coding representing gene count per pathway. For pathway identifiers, see Table S2. MOI: multiplicity of infection; IFN: interferon; PW: pathway: ISG: interferon‐regulated gene; neg.: negative; pos.: positive; reg.: regulation; HDL: high‐density lipoprotein; VLDL: very‐low‐density lipoprotein; DDX58: DEAD box Protein 58 (RIG‐I); IFIH1: interferon‐induced with helicase C domain 1 (MDA5); ind.: induction; TLS: translesion synthesis; pol: polymerase; OAS: 2’‐5’‐oligoadenylate synthetase; IL: interleukin; TJ: tight junction; NLRP3: NLR family pyrin domain containing 3; TH1: T‐helper type 1; HSF: heat shock factor; HSP: heat shock protein; SHR: steroid hormone receptor; HSR: heat shock response; PM: plasma membrane.
To benchmark the antiviral response observed in our iAT2 model, we compared our data to a previously published dataset [83]. Bertrams et al. systematically compared the transcriptomic profiles in response to IAV infection of primary AT2 cells, alveolar macrophages (AMs), human lung tissue explants, and the commonly used A549 cell line. They identified an overlapping set of 302 genes, all upregulated in AT2s, AMs, and lung tissue. We reanalyzed their dataset using our own bioinformatics pipeline and compared the resulting gene list to our transcriptomic data. Of the 302 genes reported by Bertrams et al., 233 were also present in our data for MOI 5 and MOI 10, indicating substantial overlap with established antiviral response signatures in primary human lung models (Figures S7 and S8).
iAT2 Cells Provide a Promising New Platform for Anti‐Influenza Drug Testing
2.6
As a proof of concept for iAT2 cells as a putative human infection model for testing of anti‐influenza drugs, we studied the effects of oseltamivir, an inhibitor of viral NA, on progeny virus production at 8 and 24 hpi. iAT2 cells were infected with IAV at MOI 5 or MOI 10 for 8 hpi and with MOI 0.1 or MOI 1 for 24 hpi, and treated with increasing oseltamivir concentrations. At 8 hpi, oseltamivir treatment resulted in a weak but dose‐dependent reduction of apical viral titers. A concentration of 0.1 µm reduced apical viral titers by 23% for both MOI 5 and MOI 10 compared to untreated cells. Further, a dose‐dependent effect of oseltamivir was confirmed. Treatment with 10 µm oseltamivir reduced apical titers by 37% upon infection with MOI 5, whereas a reduction of 58% was observed upon infection with MOI 10 (Figure 9a). At 24 hpi, the dose‐dependent reduction of apical viral titers was more pronounced than after 8 h and culminated in more than 90% inhibition upon treatment with 10 µm oseltamivir (Figure 9b). Viral titers on the basal side remained consistently low at both time points and were not affected by oseltamivir. To exclude a direct, cytotoxic effect of oseltamivir on iAT2s, we measured the cellular metabolic activity via 3‐(4,5‐dimethythiazol‐2‐yl)‐2,5‐diphenyl‐2H‐tetrazolium bromide (MTT) assay after exposure to oseltamivir for 24 h [84]. The absorbance measured for all oseltamivir concentrations was similar to that of healthy control cells which were kept at ALI during 24 h. In contrast, supposedly dead control cells which were incubated with 1% H_2_O_2_ for 24 h showed very low absorbance. Together, these results indicate no cytotoxic effect of oseltamivir on iAT2 cells at the tested concentrations (Figure 9c).
*Oseltamivir dose‐dependently inhibits IAV replication in iAT2 cells. (a) iAT2 cells were infected with IAV at a multiplicity of infection (MOI) of 5 or 10, treated with increasing oseltamivir concentrations, and apical and basal supernatants were harvested at 8 h post‐infection (hpi). (b) iAT2 cells were infected with IAV at MOI 0.1 or 1, treated as in b, and supernatants were harvested at 24 hpi. Viral titers in (a) and (b) were determined by plaque assay; data show mean ± SD of 4 individual experiments; P‐values were calculated using ordinary two‐way ANOVA followed by Dunnett's multiple comparison test with simple effects scheme; *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.0001; PFU: plaque‐forming units. (c) Uninfected iAT2 cells were treated with increasing oseltamivir concentrations for 24 h while control cells were kept at an air–liquid interface (ALI) or treated with H2O2 (1%). Data show mean ± SD of 3 individual experiments; MTT: 3‐(4,5‐dimethythiazol‐2‐yl)‐2,5‐diphenyl‐2H‐tetrazolium bromide.
Discussion
3
Antiviral test systems allowing for the screening of potential compounds and the evaluation of their efficacy and safety play a crucial role in the development of new effective antiviral therapies. Besides animal models, in vitro cell culture systems provide a valuable, cost‐effective, and high‐throughput platform to study IV infections and treatment conditions [38, 39]. However, the effectiveness of these test systems depends on the physiological relevance and translational potential to the human situation. While immortalized cell lines provide an easy‐to‐handle and robust system, they often lack physiological functionality and may respond differently to viral infections than the intact tissue [38, 85, 86]. Thus, patient‐derived primary cells from the nasal, bronchial, or alveolar epithelium are often applied in influenza studies since these cells preserve the properties of the human lung more faithfully [86]. However, primary cells have a limited life span in culture and cannot be expanded indefinitely, limiting the experimental scalability [39, 87, 88]. Therefore, iAT2 cells offer a promising alternative as a human in vitro model as they can be expanded and maintained in culture for extended periods [58, 68].
iAT2‐ALI models have previously been applied to studies of SARS‐CoV‐2 infection, demonstrating the value of this system for modelling alveolar epithelial responses to viral pathogens [89]. In contrast, IAVs have been investigated using iPSC‐derived airway organoids [40, 41] or adult stem cell‐derived alveolar monolayers [90], but typically without ALI culture. Here, we apply an iAT2‐ALI model to IAV infection. Compared to organoid‐based systems, which require microinjection or apical‐out engineering to access the apical surface [91, 92], ALI cultures expose the apical side directly to air, enabling straightforward viral inoculation and sampling. The 2D ALI format also offers a tractable and reproducible platform with greater experimental scalability and more homogeneous cell populations—features well‐suited for high‐throughput antiviral screening and quantitative readouts. In contrast to conventional organoid culture [91, 92], 2D iAT2‐ALI cultures were shown to promote epithelial maturation, polarization, and barrier integrity [53], all of which are critical for modelling apical viral entry, directed budding, and drug responses.
In this study, we demonstrated that iAT2 cells are permissive to IAV infection and support productive viral replication, with progeny virus predominantly released from the apical surface with minimal viral load detected on the basal side. This directional release pattern is reminiscent of the apical shedding observed in previous studies using different cell lines and primary cultures as well as different IAV strains [93, 94, 95]. Furthermore, transcriptomic profiling revealed that infected iAT2 cells mount an innate immune response, characterized by the upregulation of pattern recognition receptors, ISGs, and proinflammatory mediators. Several canonical antiviral effectors, including DDX58 (RIG‐I), IFIH1 (MDA5), TLR3, Mx1 (MxA), as well as OAS and IFIT family members, were induced following IAV infection, consistent with earlier data from primary AT2 cells [96] and from epithelial NCl‐H441 cells co‐cultured with endothelial ISO‐HAS‐1 cells [93]. Notably, direct comparison of our transcriptomic data with a published core antiviral gene signature from primary AT2s, AMs, and lung tissue [83] revealed a significant overlap of 233 out of 302 identified DEGs. However, direct one‐to‐one comparison remains inherently limited due to differences in viral strains, infectious doses (MOIs), sampling time points, and donor variability between the two studies.
In addition to profiling the antiviral response, we assessed epithelial barrier function using TEER measurements and E‐Cad immunostaining. In our model, epithelial integrity was preserved up to 48 hpi. While many conventional in vitro models report early TEER decline and junctional disruption following IAV infection, there is substantial variability in the onset and severity of barrier loss across studies [93, 97, 98, 99, 100, 101]. This likely reflects differences in cell types, viral strains, and experimental conditions. For instance, A549 cells exhibit a low baseline TEER and lack functional tight junctions, which limits their suitability for barrier integrity studies [102, 103, 104, 105]. In contrast, a recent study using a primary alveolar co‐culture model reported stable TEER at 24 hpi, with a drop detected only by 48 hpi and overt tissue damage by 96 hpi [101].
In vivo, alveolar barrier injury typically becomes apparent between days 2 and 7 post‐infection, depending on the viral strain, infectious dose, and host response [106, 107, 108, 109]. Severe infections are often accompanied by widespread tissue damage and pronounced epithelial injury, while in sub‐lethal models, proliferating AT2 cells appear as key drivers of alveolar regeneration during recovery [110, 111]. Comparisons between mild and severe IAV infections suggest that the degree of epithelial damage, particularly to AT1 cells, correlates with disease outcome [112]. Consistently, histopathological studies of fatal human cases reveal extensive alveolar damage and epithelial cell loss, which are closely associated with mortality [113, 114, 115].
Taken together, these observations suggest that our data may reflect an early stage of infection, prior to the onset of overt barrier disruption. Interestingly, upon infection with low doses of IAV, E‐Cad integrity remained intact up to 168 hpi. While this prolonged epithelial stability may not represent the extent of injury observed in severe infections, it provides a stable foundation for future work under moderate infection conditions, thereby expanding the experimental window for drug testing. Furthermore, it resembles a clinically less severe infection, where IV can persist for five days or more [116, 117, 118, 119]. Notably, while epithelial monocultures are suitable for capturing early stages of infection and epithelial‐intrinsic responses, they do not account for the multicellular dynamics that shape barrier injury in vivo. In particular, pro‐inflammatory cytokine production by endothelial cells and macrophage‐mediated tissue damage have been implicated in driving epithelial disruption during IV infection in vivo [120, 121]. To address this, future extensions of our model could incorporate immune components, such as AMs, to investigate how the immune microenvironment influences epithelial stability. In addition to enabling mechanistic studies, the sustained viability of iAT2 cells infected with a low MOI creates an opportunity to test antiviral interventions. This might be relevant for modeling delayed treatment scenarios, as antiviral therapy in clinical practice cannot always be administered immediately after symptom onset. The model may therefore be well‐suited to assess therapeutic efficacy and dosing conditions. In a proof‐of‐concept experiment, we demonstrated that oseltamivir treatment reduced viral titers in a dose‐dependent manner, indicating that iAT2 cells respond to a clinically applied NA inhibitor, and supporting the utility of our model as a platform for evaluating the efficacy of anti‐influenza drug candidates.
While our findings underscore the utility of iAT2 cells for modeling human‐pathogenic IAV infection and antiviral testing, the detected SA linkage pattern warrants further consideration. SNA staining indicated abundant α2,6‐linked SAs on the apical surface of iAT2 cells cultured for 7 days at ALI, whereas MAH staining did not detect α2,3‐linked SAs under the same conditions. In contrast, several histological studies have shown that both a2,6‐ and α2,3‐linked receptors are present in the distal human lung [122, 123, 124], suggesting that our model may underrepresent the full receptor diversity in vivo. Consequently, while the model is well suited for studying infection by human‐pathogenic IAV strains that preferentially recognize α2,6 linkages, its applicability to avian or zoonotic viruses with α2,3‐binding preference may be limited [72]. It remains to be determined whether the observed receptor profile reflects an intrinsic property of iAT2 differentiation at ALI or a maturational state that could be modified by protocol adaptations. Approaches such as prolonged ALI culture, co‐culture with endothelial or AT1‐like cells, or cytokine stimulation could potentially enhance α2,3‐linked SA expression and thereby broaden strain coverage. Notably, in primary human bronchial [125] and tracheal epithelial [126] cultures, sialylation patterns have been shown to change dynamically during differentiation, raising the possibility that alveolar epithelial cells might as well undergo glycan remodeling during maturation.
Beyond receptor composition, additional limitations inherent to iPSC‐derived cell models need to be acknowledged [127, 128, 129, 130], including line‐to‐line and clone‐to‐clone variability, as well as the tendency of iPSC‐derived cells to retain a developmentally immature phenotype. Reprogramming and prolonged culture can yield heterogeneous populations that may complicate validation, scalability, and reproducibility across studies. Nonetheless, transwell systems cultured at ALI can recapitulate key physiological features of the alveolar epithelium, including epithelial polarization and maturation [53, 54]. The differentiation protocol applied here yields a purified iAT2 population [58], and spontaneous differentiation into AT1 cells is not typically observed under static 2D ALI conditions. A separate protocol is required to generate iAT1 cells in 3D cultures [131]. Recent advances, including an autologous iPSC‐derived alveolus‐on‐chip system, demonstrate the feasibility of mixed iAT1/iAT2 models [132]. Building on these advances, future work could implement parallel iAT1 and iAT2 differentiation from the same donor line and establish co‐cultures under ALI conditions, initially on 2D transwell and subsequently in a more complex and dynamic bioengineered 3D alveolus‐on‐chip system [133, 134, 135, 136]. While organ‐on‐chip platforms offer enhanced physiological relevance, 2D transwell‐based ALI cultures provide higher throughput and experimental scalability. Together, these complementary approaches will enable both mechanistic insight and translational antiviral screening.
Conclusion
4
Our study highlights the potential of iAT2 cells as a novel cell culture model for influenza research. The unique features of iAT2s offer valuable opportunities for advancing our understanding of influenza and improving the treatment of this disease. Follow‐up studies should focus on implementing iAT2 cells and other alveolar cell types, such as iAT1 and AMs, derived from iPSCs to create autologous immunocompetent models with improved cellular heterogeneity in a bioinspired arrangement.
Methods
5
Cell Lines and Viruses
5.1
The iPSC dual reporter line BU3 NGST was purchased from Boston University, Center for Regenerative Medicine, and maintained in culture as described previously [46]. Briefly, coming from the BU3 iPSC line [137, 138], BU3 NGST cells expressed two fusion proteins namely NKX2.1 coupled to GFP and SP‐C tagged with tdTomato. The resulting BU3 derivative cell line had the phenotype NKX2.1^GFP^ SP‐C^tdTomato^ (hence, BU3 NGST) [46]. During the differentiation process of iPSCs into iAT2 cells, both reporters served to purify appropriately differentiated cells by fluorescence‐activated cell sorting. Madin–Darby canine kidney (MDCK) cells (NBL‐2; American Type Culture Collection, CCL‐34) were cultured in Eagle Minimum Essential Medium (EMEM; Sigma‐Aldrich, M4655‐500ML) supplemented with 10% fetal bovine serum (FBS; Sigma‐Aldrich, S0615) at 37°C, 5% CO_2_, 95% relative humidity (rH) and passaged every 3 to 4 days. The IV A/Puerto Rico/8/1934 (H1N1) was propagated on MDCK cells. For single‐use virus stocks, supernatants were aliquoted and stored at −80°C. Stock titers were determined by plaque assay.
Buffer Preparations for iAT2 Culture
5.2
Buffers were prepared freshly every time before usage. Complete serum‐free base medium (cSFM): 375 mL Iscove's Modified Dulbecco's Medium (Thermo Fisher Scientific, 12 440 053), 125 mL Ham's F12 medium (Cellgro, 10‐080‐CV), 5 mL GlutaMAX supplement (Thermo Fisher Scientific, 35 050‐061), 5 mL B27 supplement (50×; Thermo Fisher Scientific,17 504 044), 2.5 mL N2 supplement (100×; Invitrogen, 175 02‐048), 3.3 mL bovine serum albumin (BSA, 7.5%; Thermo Fischer Scientific, 15 260 037), 500 µL Primocin (50 mg mL^−1^; Invivogen, ant‐PM‐1), 500 µL ascorbic acid (50 mg mL^−1^; Sigma‐Aldrich, A4544), 1.5 mL 1‐thioglycerol (13 µL mL^−1^; Sigma‐Aldrich, M6145). 10× cAMP/ IBMX medium (10× CI): 50 mL cSFM, 500 µL 3‐isobutyl‐1‐methylxanthine (0.1 m; Sigma‐Aldrich, I5 879), 21.5 mg 8‐bromoadenosine 3′,5′‐cyclic monophosphate sodium salt (Sigma‐Aldrich, B7880). DS/SB medium: 50 mL cSFM, 50 µL SB 431 542 (10 mm; Tocris, 1614), 50 µL dorsomorphin (2 mm; Tocris, 3093). CBRa medium: 50 mL cSFM, 50 µL CHIR 99 021 (3 mm; Tocris, 4423), 50 µL recombinant human bone morphogenetic protein 4 (10 µg mL^−1^; R&D systems, 314‐BP‐010), 50 µL retinoic acid (100 µm; Sigma‐Aldrich, R2625). CK/DCI medium: 45 mL cSFM, 5 mL 10× CI, 50 µL recombinant human keratinocyte growth factor (10 µg mL^−1^; R&D Systems, 251‐KG‐010), 50 µL CHIR 99021 (3 mm), 25 µL dexamethasone (100 µm; Sigma‐Aldrich, D1756). CK/DCI+Y medium: 49.9 mL CK/DCI, 100 µL Y‐27632 dihydrochloride (5 mm; Tocris, 1254).
Buffer Preparations for Virus Infections, Plaque Assays, and TEER Measurements
5.3
Buffers were prepared periodically and stored at 4°C for several months. Phosphate‐buffered saline (PBS): 8 g NaCl (ITW Reagents, A2942,1000), 0.2 g KCl (Carl Roth, 6781.9), 1.14 g Na_2_HPO_4_ (ITW Reagents, 131 679.1210), 0.2 g KH_2_PO_4_ (Carl Roth, 3904.2), ddH_2_O ad 1 L, pH 7.4. Infection PBS (iPBS): 490.5 mL Dulbecco's Phosphate Buffered Saline (Lonza, 17‐512F), 5 mL Penicillin/Streptomycin (Pen/Strep, 100×; Lonza, DE17‐602E), 3.5 mL BSA (30%; Carl Roth, 9401.3), 0.5 mL MgCl_2_ (1 m; Sigma‐Aldrich, M2393‐100G), 0.45 mL CaCl_2_ (1 m; Carl Roth, CN93.1). Infection medium (iDMEM): 495.5 mL Dulbecco's Modified Eagle Medium (DMEM; Anprotec, AC‐LM‐0012), 0.5 mL MgCl_2_ (1 m), 0.45 mL CaCl_2_ (1 m), 3.5 mL BSA (30%), L‐1‐tosylamido‐2‐phenylethyl chloromethyl ketone‐treated trypsin (TPCK trypsin; Thermo Fisher Scientific, 20 233) was added directly before use to a final concentration of 0.167 µg mL^−1^. Plaque medium (PM): 551 mL ddH_2_O, 100 mL Minimum Essential Medium (10×; Gibco, 21430‐020), 10 mL Pen/Strep (100×), 28 mL NaHCO_3_ (7.5%; Anprotec, AC‐DS‐0017), 10 mL diethylaminoethyl dextran (1%; Pharmacia Fine Chemicals, 17‐0350‐01), 7 mL BSA (30%); TPCK trypsin was added directly before use to a final concentration of 0.25 µg mL^−1^. 3% agar: 3 g Oxoid agar (Oxoid, LP0028B), 100 mL ddH_2_O (stored at room temperature (RT)). EDTA (0.1 m), pH 7: 2.9225 g EDTA (Carl Roth, CN06.1), ddH_2_O (100 mL), adjusted to pH 7 with NaOH.
Differentiation of iPSCs and iAT2 Culture
5.4
BU3 NGST cells were differentiated into iAT2 cells following the protocol by Jacob et al. [58]. Briefly, iPSCs were differentiated into iAT2s through three stages by exposing the cells to different media. The definitive endoderm was obtained after 3 days using the STEMdiff Definitive Endoderm Kit (Stemcell Technologies, #05110). The anterior foregut was reached by exposing the cells to DS/SB for another 3 days. For further differentiation into lung progenitors, cells were exposed to CBRa for 8 or 9 days. On day 14 or 15, a purified lung progenitor population was obtained by sorting for NKX2.1‐positive cells. Purified cells were plated in Corning Matrigel Growth Factor‐Reduced Basement Membrane Matrix (Corning, 356 231) in CK/DCI, to allow growth into 3D alveolospheres. Mature iAT2s were identified with a second sorting step for NKX2.1 and SP‐C at day 30+. NKX2.1/SP‐C‐double positive cells were further expanded as matrigel‐embedded alveolospheres with 200 cells µL^−1^ in 500 µL CK/DCI in 24‐well plates. Cultures were maintained at 37°C, 5% CO_2_ for 14 days with medium exchange every 48 h. All experiments were performed using a single donor line. To achieve the required scale, we pooled cells from different differentiations and batches. To mitigate batch‐related effects, the cell differentiation state was verified before pooling.
Seeding of iAT2 Cells
5.5
Transwell inserts (Greiner Bio‐one, 662 641) were placed in 24‐well plates and coated with 100 µL of Corning Matrigel Human Embryonic Stem Cell‐qualified matrix (Corning, 354 277) according to the manufacturer's instructions. Alveolosphere‐containing matrigel droplets were dissolved using dispase (2 mg mL^−1^; Thermo Fisher Scientific, 17105‐041) and alveolospheres were dissociated into single cells using trypsin (0.05%; Gibco, 25‐300‐062). 2 × 10^5^ to 4 × 0^5^ iAT2 cells were seeded onto the prepared transwell inserts (apical wells) in 100 µL CK/DCI+Y, with 500 µL CK/DCI+Y in lower wells (basal wells) and cultivated at 37°C, 5% CO_2_. After 48 h, ALI was established by removing the medium from apical wells. Medium on the basal side was refreshed 24 h later by adding freshly prepared 600 µL CK/DCI. The cells were monitored for confluence during ALI and maintained in culture for an additional 7 days with medium exchange every 48 h until IAV infection.
Lectin Staining and NA Treatment
5.6
iAT2 cells were cultured and seeded as described above. Nine days after seeding, cells were washed once with PBS in apical and basal wells, fixed with paraformaldehyde (3.7%; Carl Roth, 7398.1) in both wells for 15 min at RT, and washed thrice with PBS before PBS was removed from apical and basal wells. In all further steps, solutions were only added to apical wells. Cells were treated with 0.02 U recombinant NA from Arthrobacter ureafaciens (Merck, 10269611001) in 100 µL PBS at pH 5.5 or, as a control, were kept in PBS with neutral pH. After 1 h at 37°C, 5% CO_2_, 95% rH, cells were washed thrice with PBS, blocked with BSA (3% in PBS) for 30 min at RT, and washed again with PBS. In order to stain α2,6‐ or α2,3‐linked SAs, cells were incubated with SNA‐biotin (20 µg mL^−1^; Linaris, B‐1305) or MAH‐biotin (20 µg mL^−1^; Linaris, B‐1265) in BSA (3% in PBS) for 1 h at RT. Cells were washed thrice in PBS and stained with streptavidin‐AF488 (3.3 µg mL^−1^ for SNA or 10 µg mL^−1^ for MAH; Invitrogen, S‐11223), and 4′,6‐diamidino‐2‐phenylindole (DAPI, 1:5000; Invitrogen, D1306) diluted in BSA (3% in PBS) for 30 min at RT. Cells were washed thrice with PBS before PBS was removed from all wells. Membranes were cut off the transwell inserts using a scalpel, mounted on #1 cover slips (Menzel–Gläser, 9. 161 024) using Dako fluorescence mounting medium (Dako, S3023), and stored at 4°C until imaging.
IAV Infection and Oseltamivir Treatment
5.7
iAT2 cells were cultured and seeded as described above. Nine days after seeding, iAT2s were washed once with 100 µL PBS in apical wells and infected with IAV at different MOI ranging from 0 to 10, where the viral stock volumes (*V_IAV stock_ *) were calculated according to Equation (1).
Inocula of respective MOIs were prepared in iPBS. Fifty microliters of inoculum were added to apical wells, and cells were incubated at 37°C, 5% CO_2_, 95% rH for 0.5, 1, or 2 h. After the adsorption period, inocula and media from apical and basal wells were removed, and cells were washed thrice with PBS. Hundred microliters of iDMEM freshly supplemented with TPCK trypsin (0.167 µg mL^−1^) and oseltamivir phosphate (0–10 µm; Sigma‐Aldrich, SML 1606‐100MG) were added to apical wells. CK/DCI (600 µL) freshly supplemented with TPCK trypsin (0.167 µg mL^−1^) and oseltamivir phosphate (0–10 µm) were added to basal wells. Cells were incubated as before until harvest at 8, 24, or 48 hpi. Supernatants from apical and basal wells were collected and frozen at −80°C. Cells were washed thrice with PBS, fixed as above, and stored with PBS in apical and basal wells at 4°C until staining.
IAV Infection Kinetic
5.8
iAT2 cells were seeded and infected as described above with MOI 0.1 and an initial absorption period of 2 h, or left uninfected. Apical and basal supernatants were harvested, and cells were fixed as above at 4, 8, 12, 24, 36, 48, 72, and 168 hpi with an additional medium exchange step in basal wells at 72 hpi. To keep the basal volume constant for all samples, 600 µL CK/DCI freshly supplemented with TPCK trypsin (0.167 µg mL^−1^) were added to the collected basal supernatants at 4 to 72 hpi before freezing. At 72 hpi, basal supernatants of 168 hpi samples were also harvested, basal wells were refilled with 600 µL fresh CK/DCI supplemented with TPCK trypsin (0.167 µg mL^−1^), and the incubation was continued as before. At 168 hpi, the collected basal supernatants were combined with those previously harvested at 72 hpi.
TEER Measurements
5.9
All measurements were done as technical quadruplicates. iAT2 cells were seeded as described above. Additionally, blank inserts with medium but without cells were included. During the last medium exchange prior to IAV infection 1 mL CK/DCI medium was added to the basal wells. At day 9 after seeding, iAT2s were infected as described above with MOI 1 or MOI 10, treated with iPBS (MOI 0), or kept at ALI (blank, ALI, and EDTA samples). After 2 h and washing, 1 mL fresh CK/DCI + TPCK trypsin (0.167 µg mL^−1^) was added to all basal wells. In apical wells, 300 µL iDMEM + TPCK trypsin (0.167 µg mL^−1^) were added in the case of blank, MOI 0, MOI 1, and MOI 10 samples, while negative controls were incubated with 300 µL EDTA (6 mm, pH 7, diluted in iDMEM + TPCK trypsin (0.167 µg mL^−1^)). Positive controls were continuously kept at ALI. Prior to infection (0 hpi), as well as at 8, 24, and 48 hpi, TEER was measured using a Millicell ERS 3.0 Digital Voltohmmeter (Merck, MERS03000) according to the manufacturer's guide. Prior to each measurement, 300 µL iDMEM + TPCK trypsin (0.167 µg mL^−1^) were added to ALI samples and removed again directly after the measurements. After the measurement at 24 hpi, the medium in the apical and basal wells was exchanged with fresh medium. For analysis, the average background resistance (R¯blank) was calculated across all blank samples. R¯blank was then subtracted from all individual measurements (*R_measured_ *) to obtain true tissue resistance values (*R_tissue_ *) (Equation (2)). Since the resistance is inversely proportional to the area of the tissue, the tissue resistance was multiplied with the membrane area (*A_membrane_ * = 0.3 cm^2^) to result in TEER values (Equation (3)).
MTT Assay
5.10
All measurements were done as technical triplicates. iAT2 cells were cultured and seeded as described above. During the next steps, solutions were only added to apical wells while basal wells were kept in culture medium. Nine days after seeding, iAT2s were washed once with 100 µL PBS, treated with 100 µL oseltamivir phosphate (0–10 µm in iDMEM) or H_2_O_2_ (1% in iDMEM; Merck, 1.08597.1000), or ALI was re‐established, and incubated at 37°C, 5% CO_2_, 95% rH for 24 h. After adding 100 µL iDMEM to ALI control cells at 24 h, 25 µL MTT (5 mg mL^−1^; Sigma‐Aldrich, M5655‐1G) were added to all wells, and cells were incubated as above for another 2 h. Medium was removed from basal and apical wells, 50 µL dimethyl sulfoxide (DMSO; Carl Roth, A994.2) was added to apical wells, and cells were incubated on a microplate shaker. After 5 min, all supernatants were transferred to a 96‐well plate, 50 µL DMSO was added in triplicates for background measurement, and the absorbance (A) at 562 nm was measured with a FLUOstar Omega microplate reader (BMG labtech). For analysis, the average background absorbance (A¯ * DMSO *) was calculated and subtracted from all triplicate measurements of a sample (*A_M1_ * – *A_M3_ *), before the average absorbance for that sample (A¯ * sample *) was determined (Equation (4)).
Plaque Assay
5.11
2 × 10^6^ cells well^−1^ were seeded in EMEM supplemented with 5% FBS and 5% Panexin basic (Pan Biotech, P04‐96950) in a total volume of 2 mL in 6‐well plates. The cells were cultivated at 37°C, 5% CO_2_, 95% rH for 1 day until full confluence. Before the plaque assay, harvested volumes of all virus‐containing supernatants were determined using a 100 or 1000 µL pipette. 1:10 serial dilutions of all samples up to 1:10^7^ were prepared in iPBS under thorough vortexing. From now on, two 6‐well plates were processed simultaneously. MDCK cells were washed once with PBS, virus‐containing serial dilutions were vortexed, and 0.5 mL was immediately added per well. Cells were incubated as above for 30 min, during which 3% agar was melted in a microwave, and 17.5 mL pre‐warmed PM were supplemented with TPCK trypsin (0.25 µg mL^−1^). After 30 min incubation, 7.5 mL melted agar was added to the PM‐trypsin aliquot. Virus‐containing serial dilutions were removed from MDCKs, and the cells were directly covered with 2 mL of the PM‐agar mix. The agar was allowed to solidify for 5 min at RT, before plates were placed upside‐down at 37°C, 5% CO_2_, 95% rH. After 3 days, the agar was stained with 800 µL well^−1^ neutral red (20 µg mL^−1^; Merck, N4638‐5G) and incubated as above for 5 h. After neutral red removal, plaques visible as clear spots over the red‐stained agar were counted. Viral titers were calculated according to Equation (5) with an expected volume of 600 µL for basal and 100 µL for apical supernatants.
RNA Extraction and Bulk RNA Sequencing
5.12
Samples were prepared in triplicates. Total RNA was isolated using the QIAGEN RNeasy Mini Kit (QIAGEN, 74104). All steps were performed at room temperature. Cells were rinsed with cold PBS and lysed on the insert with 350 µL RLT buffer. Lysates were homogenized by pipetting and mixed with 70% ethanol. The homogenized solutions were loaded onto RNeasy spin columns. Columns were spun for 30 s at 8000 × g and washed with 700 µL RW1. Columns were spun again and washed twice with 500 µL RPE buffer following the manufacturer's instructions. RNA was eluted in 50 µL RNase‐free water and stored at −80°C. Total RNA samples were shipped on dry ice to Novogene, which carried out library preparation and sequencing using their standard protocols. Details of the provider's workflow are available from Novogene upon request.
Transcriptomic Data Preprocessing and Analyses
5.13
Raw counts were converted to counts per million (CPM) per sample and log transformed. Lowly expressed genes were filtered out using a standard detectability rule: genes with CPM > 1 in at least 3 of 6 samples (i.e., present in ≥50% of samples) were retained. PCA was performed to visualize the sample‐to‐sample structure. DEG analysis was performed with the DESeq2 framework [139] using its Python implementation (pydeseq2) [140]. Analyses were run on the filtered raw count matrix, comparing the infected condition to the control. P‐values were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR). For visualization, negative log_10_ adjusted p‐values were plotted against log_2_ fold changes (FC), and genes were highlighted as “up” or “down” using |log_2_FC| > 1 and FDR (p‐value adjusted) < 0.05. GSEA was assessed separately for up‐ and downregulated genes identified in the infected condition (MOI 5 or MOI 10) in comparison to the control (MOI 0). Enrichment was performed with Enrichr [141] via gseapy (Python) [142] using default parameters and Benjamini–Hochberg correction for multiple testing. The following databases were used: GO Biological Process 2021 [143], GO Cellular Component 2021 [143], GO Molecular Function 2021 [143], KEGG 2021 Human [144], and Reactome 2022 [145]. Pathways with FDR < 0.05 were considered significantly enriched.
For targeted antiviral marker analysis, we defined a literature‐based panel (interferons, pattern recognition receptors, ISGs, JAK/STAT/IRFs, NF‐κB mediators, cytokines, chemokines, restraint/stress factors, plus housekeeping genes). For each gene, we reported the mean log_2_(CPM + 1) per condition and plotted the difference Δ = Infected − Control, shown as heat maps. As an external benchmark, we applied the same approach to the influenza‐induced antiviral gene set reported by Bertrams et al. (GEO: GSE206606) [83], which consisted of genes upregulated upon IV infection in primary human AT2 cells.
Fluorescence Staining
5.14
PBS from apical and basal wells was removed before staining. In all further steps, solutions were only added to apical wells. Cells were permeabilized with triton X‐100 (0.25%; Sigma‐Aldrich, X100‐100ML) for 30 min at RT. For characterization of iAT2s, cells were washed with PBS and blocked with normal donkey serum (NDS, 3% in PBS; Biozol, LIN‐END9010‐10) for 1 h at RT. Samples were incubated at 4°C overnight with primary antibodies diluted in NDS (3% in PBS): mouse‐anti‐Ki67 (1:50; BD Biosciences, 556 003), rabbit‐anti‐LPCAT1 (1:100; Abcam, ab214034), rabbit‐anti‐NKX2.1 (1:100; Abcam, ab76013), and rabbit‐anti‐SP‐C (1:50; Invitrogen, #PA5‐71 680). Cells were washed with PBS and stained with secondary antibodies diluted in NDS (3% in PBS) for 1 h at RT: donkey‐anti‐rabbit Cy3 (1:200; Jackson ImmunoResearch, #711‐165‐152) or donkey‐anti‐mouse AF488 (1:200; Jackson ImmunoResearch, #715‐545‐150), and DAPI (1:200). For determination of the intracellular viral protein load and actin staining, cells were washed thrice with PBS after permeabilization and blocked with BSA (3% in PBS) for 1 h at RT. Incubation with primary antibodies diluted in BSA (3% in PBS) was done at 4°C overnight: mouse‐anti‐NP (1:300; BioRad, MCA400) and rabbit‐anti‐E‐Cad (1:150; Cell Signaling Technology, #3195). Cells were washed thrice with PBS and stained for 1 h at RT with secondary antibodies diluted in BSA (3% in PBS): goat‐anti‐mouse Cy5 (1:300; Dianova, #115‐175‐146), goat‐anti‐rabbit AF488 (1:200; Dianova, #111‐545‐144) or goat‐anti‐rabbit Cy3 (1:300; Dianova, #111‐165‐003), DAPI (1:5000), and, in the case of actin staining, with Flash Phalloidin Green 488 (1:100; BioLegend, 424201). Finally, cells were washed thrice with PBS, mounted on #1 cover slips as above, and stored at 4°C until imaging.
Image Acquisition and Processing
5.15
Transmission light images were acquired with a Primovert microscope (Zeiss, 491206‐0003‐000) equipped with an AxioCam ERc5s (Zeiss, 426540‐9901‐000) using a 4×/NA 0.1 (Zeiss, 415500‐1619‐000) or 10×/NA 0.25 (Zeiss, 415 500‐1605‐001) objective. Fluorescence images were acquired with an Axio Observer.Z1 microscope (Zeiss, 431 007‐9901‐000) equipped with ApoTome.2 (Zeiss, 423 667‐8224‐000), Axiocam 503 mono (Zeiss, 26559‐0000‐000), and HXP 120 V (Leistungselektronik Jena GmbH, LQ‐HXP120‐CAN‐z‐v) using a 20×/NA 0.8 (Zeiss, 420 650‐9902‐000), 40×/NA 0.75 (Zeiss, 420 360‐9900‐000), or 63×/NA 1.4 oil (Zeiss, 420 782‐9900‐799) objective with Immersol 518 F (Zeiss, 444 960‐0000‐000), and filter sets 38 HE (Zeiss, 489 038‐9901‐000), 49 (Zeiss, 488 049‐9901‐000), and 50 (Zeiss, 488 050‐9901‐000). Images were deconvolved using Zen blue (Zeiss, version 2.6) with phase errors correction and deconvolution strength 5. Deconvolved z‐stacks were exported, and further processing was done in ImageJ (NIH, version 1.54 f). For orthogonal displays, fixed minimal (min) and maximal (max) pixel values were set for the AF488 channel, while min and max pixel values for the DAPI channel were determined as follows: The saturation of the respective maximum intensity projection was set to 0.99, and min and max pixel values were extracted. The extracted values were multiplied with 0.8 to retrieve the final min and max pixel values. For maximum intensity projections, fixed min and max pixel values were set for the NP‐Cy5 channel while saturation was set to 0.6 or 0.99 for E‐Cad‐Cy3/AF488 and DAPI channels, respectively. For the control staining without primary antibodies (AB ctrl), min and max pixel values for the Cy3 channel were set according to the lowest and highest values received for the remaining images with 60% saturation.
Statistical Analysis
5.16
For oseltamivir treatments, absolute viral titers were calculated with Excel (Microsoft, version 2016) according to Equation (5), and relative titers were calculated in relation to the apical 0 µm sample of each individual experiment. Every treatment experiment was repeated four times (n = 4), and relative viral titers were transferred into GraphPad Prism (GraphPad Software, version 8.4.3) for graphing and statistical analyses. Significance was tested by ordinary two‐way analysis of variance (ANOVA) followed by Dunnett's multiple comparison test with a simple effects scheme to compare apical and basal titers with their respective 0 µm controls. Data was displayed as mean ± SD, and significance was defined as *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.
Author Contributions
H.K. differentiated, maintained, and characterized iAT2 cells. L.G. performed sialic acid staining, IAV infections, and oseltamivir experiments. Y.B. performed transcriptomics analyses. A.S.M., M.T.F., and C.E. conceived the project, obtained the funding, and supervised the research. H.K. and L.G. wrote the manuscript. R.W.K., P.B.v.K., C.E., M.T.F., and A.S.M. revised the manuscript.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supporting File: adhm70675‐sup‐0001‐SuppMat.pdf.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1J. Paget , P. Spreeuwenberg , V. Charu , et al., “Global Mortality Associated With Seasonal Influenza Epidemics: New Burden Estimates and Predictors From the G La MOR Project,” Journal of Global Health 9 (2019): 020421.31673337 10.7189/jogh.09.020421 PMC 6815659 · doi ↗ · pubmed ↗
- 2A. D. Iuliano , K. M. Roguski , H. H. Chang , et al., “Estimates of Global Seasonal Influenza‐Associated Respiratory Mortality: A Modelling Study,” The Lancet 391 (2018): 1285–1300.10.1016/S 0140-6736(17)33293-2PMC 593524329248255 · doi ↗ · pubmed ↗
- 3WHO , Vol. 2024, World Health Organization (WHO), WHO webpage (2023), https://www.who.int/news‐room/fact‐sheets/detail/influenza‐(seasonal).
- 4F. Krammer , G. J. D. Smith , R. A. M. Fouchier , et al., “Influenza,” Nature Reviews Disease Primers 4 (2018): 3.10.1038/s 41572-018-0002-y PMC 709746729955068 · doi ↗ · pubmed ↗
- 5ECDC , Vol. 2024, European Centre for Disease Prevention and Control (ECDC), ECDC webpage (2022), https://www.ecdc.europa.eu/en/seasonal‐influenza/prevention‐and‐control/antivirals.
- 6CDC , Vol. 2024, Centers for Disease Control and Prevention (CDC), CDC webpage (2023), https://www.cdc.gov/flu/about/viruses‐types.html.
- 7T. Flerlage , D. F. Boyd , V. Meliopoulos , P. G. Thomas , and S. Schultz‐Cherry , “Influenza Virus and SARS‐Co V‐2: Pathogenesis and Host Responses in the Respiratory Tract,” Nature Reviews Microbiology 19 (2021): 425–441.33824495 10.1038/s 41579-021-00542-7PMC 8023351 · doi ↗ · pubmed ↗
- 8Y. Liang , “Pathogenicity and Virulence of Influenza,” Virulence 14 (2023): 2223057.37339323 10.1080/21505594.2023.2223057 PMC 10283447 · doi ↗ · pubmed ↗