Detailed Methods are provided in the online supplement.

Organoid culture

Human airway organoids used in this study were purchased from the Foundation Hubrecht Organoid Biobank (www.hubrechtorganoidbiobank.org, HUB code HUB-07-A2-051, normal lung) and were cultured following previously described protocols [10, 14]. The cells were derived from the left superior lobe (lower airway) of a 53-year-old female donor with no underlying lung disease.

Apical-out polarity switch and mucociliary differentiation of healthy airway organoids

The polarity of the organoids was switched as described by Co et al. [15]. After two days of culturing the organoids in suspension, mucociliary differentiation of organoids was initiated as described by Zhou et al. [16] with the following adaptations: Expansion medium was changed to PneumaCult™-ALI medium supplemented with 2 IU/mL Heparin, 1 µM Hydrocortisone and 10 µM DAPT (Sigma, Cat#D5942). The organoids were then cultured in suspension in a differentiation medium for 16 days.

Bacterial strains and growth conditions

All experiments were conducted with the Pseudomonas aeruginosa PA14 strain, which was generously provided by Dr. Dirk Bumann, Biozentrum Basel, Switzerland. Bacterial cultures were grown as previously published [17].

For experiments with heat-inactivated PA14, bacteria were incubated at 70 °C for 20 min. To confirm the absence of viable bacteria, an aliquot of heat-inactivated PA was plated onto LB plates. In experiments with bacterial cell-free culture supernatant, an overnight culture of PA was centrifuged at 2.000 rcf for 10 min at 4 °C and supernatant was harvested, sterile filtered (0.2 µM), snap-frozen and stored at -80 °C until further usage.

Treatment with pharmacological compounds

Where indicated, organoids were stimulated with either 10 µM Forskolin (Tocris, Cat#1099/10), 1 µM Roflumilast (MedChem Express, Cat#HY-100639) or 50 µM Pyocyanin (Sigma, Cat# R9532-500UL). In a separate experiment, organoids were exposed to 1 mM EHNA (Tocris, Cat#1261).

Colony forming unit (CFU) assay

The quantification of viable intracellular bacteria was performed as described before [17]. Total protein was measured using the Micro BCA™ Protein Assay Kit (Thermo Scientific, Cat#23235) according to the manufacturer’s protocol.

Infection of organoids with PA14 and Gentamicin protection assay

Organoids were infected with 25 × 10^6 bacterial cells/mL Following three hours of active infection, a gentamicin-protection assay was applied to prevent bacterial overgrowth in the cell culture medium. Gentamicin treatment involved washing organoid cultures three times with PBS containing 25 µg/mL gentamicin (Life Technologies, Cat#15710064). For experiments involving gene expression analysis, organoids were incubated in fresh PneumaCult™-ALI medium supplemented with 8 µg/mL gentamicin to effectively eliminate extracellular bacteria, enabling continued culturing of organoids harboring intracellular bacteria [18]. A parallel set of uninfected controls underwent identical washing and incubation procedures. When cells were treated with inactivated bacteria, a concentration of 10^8/mL was used to account for the lack of bacterial growth compared to the viable bacteria during the three-hour active infection phase. Where indicated, organoids were treated with 1:50 (vol/vol) of bacterial cell-free culture supernatant during infection.

Ciliary beating frequency (CBF) measurements and calculation

Beating cilia were recorded at room temperature (20 °C ± 2 °C) through the ocular lens of an inverted light microscope using 10 × magnification and a camera with high optical and time resolution (1920 × 1080pixels at 240fps, Apple, California) positioned on the eyepiece of the microscope with an adaptor as first described by Chen et al. [19]. One-second movies were collected at 240 frames per second (fps) and saved in.mov format. The software Shutter Encoder was used to convert videos to uncompressed YUV format and videos were imported in the Fiji ImageJ package using the ffmpeg plugin [20]. Kymographs of ciliary movement were created for each video by using the line tool as first described elsewhere [21].

Ten kymographs were randomly selected from each video, depicting the ciliary movement of distinct ciliated cells. The calculation of ciliary beat frequency (CBF) involved manual measurement of the pixels between successive peaks in each kymograph wave. CBF was determined using the following equation:

$$ \left( }} \right) \, = \, \frac}\left( }} \right)}}}} $$

Two independent observers assessed CBF and the measuring method was validated by comparing the results to frequencies acquired using two other techniques. First, with equipment for routine PCD (primary ciliary dyskinesia) diagnosis at the ORL Department. High-speed imaging was performed at 130-300fps depending on the region of interest clipping at 37 °C. Kymographs were extracted with ZEISS ZEN2.6 and CBF was determined as described above. Second, photometric CBF measurement based on the deflection of light caused by the ciliary beats was assessed at 37 °C. The diameter of the photosensitive field on the sample was 5 μm and the light changes photograph-multiplied, digitalized at a sample frequency of 400 Hz, and transformed into a time–amplitude signal. A Fast Fourier transformation analysis was performed every 1.6 s (windaq, Dataq Instruments, Ohio, USA) and CBF values were detected at five to ten different sites. Supplementary Video 1 was recorded at 37 °C for representative visualization of ciliary activity and was not used for CBF measurements.

mRNA expression analysis of selected genes by RT-qPCR

The RT-qPCR procedure followed established protocols described by Grubwieser et al., 2023 [17]. TaqMan PCR primers were designed with the PrimerQuest™ Tool (Integrated DNA Technologies, Inc., Iowa, USA) and the primer sequences are listed in Supplementary Table S1. The expression levels of individual mRNAs were calculated relative to the housekeeping gene OAZ1 using the ΔΔCT method.

Immunofluorescence confocal microscopy of organoid and PA14 co-culture

For immunofluorescence imaging, organoids were infected with fluorescent PA14 and fixed with 4% Paraformaldehyde for 30 min. Next, organoids were permeabilized with 0.5% Saponin (Sigma-Aldrich, Cat#SAE0073) for 30 min, stained with Phalloidin-iFluor 647 (Abcam, Cat#ab176759, 1:1000 dilution) or Alexa Fluor® 647 Anti-alpha Tubulin (acetyl K40) antibody (Abcam, Cat#ab218591, 1:1000 dilution) for 2 h at room temperature (20 °C ± 2 °C) and Fluoroshield Mounting Medium With DAPI (Abcam, Cat#104,139) was added before imaging. Slides were imaged immediately after sample preparation using a VS120-S6 fluorescence microscope (Olympus) or the Operetta CLS System (PerkinElmer, Waltham, MA, USA). Images were captured with a 20-x and 40-x objective using 387/440 nm (DAPI) and 650/684 nm (Alexa-flour-647) lasers and filters, under identical exposure times for every sample.

Bioinformatic analysis of RNA sequencing data

FASTQ files from 16 samples from 2 different batches were processed with the nf-core RNA-seq pipeline version 3.10.1 [https://doi.org/10.5281/zenodo.1400710]. In brief, reads were trimmed with TrimGalore v0.6.10 [https://doi.org/10.5281/zenodo.5127899] and were afterwards aligned to the GRCh38 reference genome with GENCODE v38 annotation, using STAR 2.7.10b. Gene expression was quantified using Salmon v1.10.0 [22]. Thereafter, the gene count table was imported into R (v.4.2.3) for downstream analysis. Differential gene expression was performed between infected and control conditions using Bioconductor packages DESeq2 v1.38.3 [23]. To control for batch effect the biological replicate information was added to the DESeq2 design matrix. False-discovery-rates (FDR) were calculated using IHW v1.26 [24]. The filtering of the gene list was performed using threshold values: adjusted P adj.-value < 0.1 and |log2FC|≥ 0.5). The result of the differential expression analysis was used as an input for Gene Set Enrichment Analysis (GSEA) and the over-representation test (ORA). GSEA was performed with the clusterProfiler package v.4.6.2 using the biological processes of the Gene Ontology (GO-BP) database and the Kyoto Encyclopedia of Genes and Genomes (KEGG) as input gene lists. Volcano plots were visualized using the R package EnhancedVolcano v1.16.0. Heatmap plots were generated using ComplexHeatmap v2.14.0. The z-score was calculated by subtracting the mean from the data points and dividing the result by the standard deviation.

RNA extraction and bulk RNA-sequencing

Organoids were centrifuged at 300 g for 10 min at 4 °C and washed twice on ice with cold PBS. Cell pellets were snap-frozen and stored at − 80 °C until further usage. RNA from approximately 1 × 106 cells was isolated using the RNeasy Plus Micro Kit (Qiagen, Cat#74,034) and the QIAshredder (Qiagen, Cat#79,656) according to the manufacturer’s protocol. Genomic DNA was removed with the RNase-free DNase Set (Qiagen, Cat# 79,254) as per manufacturer’s instructions. RNA was eluted in 1X low Tris–EDTA (TE) buffer, containing 10 mM Tris–HCl and 0.1 mM EDTA. The amount and purity of RNA were measured with a Nanodrop Spectrophotometer. In case of poor quality samples were cleaned using the RNeasy MinElute Cleanup Kit (Qiagen, Cat#74,204) according to the manufacturer’s protocol. RNA Integrity number (RIN) was assessed with the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). RNA with a RIN > 7.5 was submitted to QuantSeq 3′-mRNA library preparation (Lexogen, Vienna, Austria) and sequenced with an Ion Proton Sequencer using Ion Proton Hi-Q chemistry (Ion Torrent/Fisher Scientific, Austria). Sequencing data obtained from this experiment are also being used in another manuscript https://doi.org/10.3389/fimmu.2024.1508727, https://zenodo.org/records/13382769) and have been deposited on Zenodo https://doi.org/10.5281/zenodo.12648167.

Electron microscopy

Non-infected and infected organoids were transferred to Eppendorf® vials, passively sedimented and culture media exchanged with Karnovsky’s Glutarladehyde-Formaldehyde fixative. After fixation in 1% aqueous osmium tetroxide and ethanol exposure at ascending concentrations, organoids were embedded in Epon epoxy resin. Additional detail on the method is described elsewhere [25].

Statistical analysis

Statistical analyses of CBF, CFU and qPCR data were performed using GraphPad Prism v.10.1.2 (San Diego, CA, USA). For comparisons of two groups, unpaired Student’s t-test was used. Comparison of more than two groups was performed using one-way ANOVA followed by Tukey’s post-hoc test for multiple comparisons. Gene expression (multiple groups, multiple time points) was analyzed using two-way ANOVA followed by Sidak’s multiple comparisons test. Differences between groups were considered statistically significant when p < 0.05. Correlation between CFU and CBF of infected organoids treated with Forskolin or Roflumilast were calculated using Spearman’s rank correlation test with a 95% confidence interval. The RNA sequencing data was analyzed using R version 4.3.2 as detailed in the supplementary methods.