Primary Ciliary Dyskinesia Patient-Specific hiPSC-Derived Airway Epithelium in Air-Liquid Interface Culture Recapitulates Disease Specific Phenotypes In Vitro

Primary ciliary dyskinesia (PCD) is a rare heterogenic genetic disorder associated with perturbed biogenesis or function of motile cilia. Motile cilia dysfunction results in diminished mucociliary clearance (MCC) of pathogens in the respiratory tract and chronic airway inflammation and infections successively causing progressive lung damage. Current approaches to treat PCD are symptomatic, only, indicating an urgent need for curative therapeutic options. Here, we developed an in vitro model for PCD based on human induced pluripotent stem cell (hiPSC)-derived airway epithelium in Air-Liquid-Interface cultures. Applying transmission electron microscopy, immunofluorescence staining, ciliary beat frequency, and mucociliary transport measurements, we could demonstrate that ciliated respiratory epithelia cells derived from two PCD patient-specific hiPSC lines carrying mutations in DNAH5 and NME5, respectively, recapitulate the respective diseased phenotype on a molecular, structural and functional level.


Introduction
Motile cilia, membrane-bound organelles, are involved in the developmental processes and function of different organs. They are present on various cells, e.g., in the ventricular system of the brain, on nodal cells during development that are responsible for left-right asymmetry, in the fallopian tubes, and in the epithelium of the respiratory tract [1]. Mutations in genes involved in the assembly and function of motile cilia can lead therefore to complex disorders summarized as motile ciliopathies [1]. Primary ciliary dyskinesia (PCD) is a rare genetic disorder belonging to the group of motile ciliopathies, with current Scientific, Waltham, MA, USA), following the manufacturer's instructions. hiPSC were cultivated on Geltrex TM (Gibco, Billings, MT, USA) in an E8 cell culture medium (in housemade). The medium was exchanged daily. Cells were passaged every 3 to 4 days using Accutase TM (Gibco, Billings, MT, USA) for up to 10 passages.

Magnetic Activated Cell Sorting (MACS) of Lung Progenitors
On day 14 of differentiation, cells were detached from culture vessels as described above. Cells were harvested in basal media with 10 µM Y-27632 and spun at 216× g for 3 min at 4 • C. After straining and counting, up to 15 × 10 6 cells were first incubated in MACS buffer (PBS w/o (Gibco, Billings, MT, USA), 1.0% BSA (Sigma Aldrich, Saint Louis, MO, USA), 2 mM EDTA (Sigma Aldrich, Saint Louis, MO, USA) with FcR blocking (Miltenyi Biotec, Bergisch Gladbach, Germany) for 10 min at 4 • C and mouse anti-human CPM antibody (FUJIFILM Wako; 1:200 dilution) for additional 20 min at 4 • C. Unbound antibody was removed by washing once with MACS buffer before cells were incubated in anti-mouse IgG 2a+b microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) for 20 min at 4 • C. After washing once with MACS buffer, cells were separated using the QuadroMACS Separator (Miltenyi Biotec, Bergisch Gladbach, Germany). NKX2.1 flow cytometry analysis was used to determine the proportion of lung progenitor cells before and after sorting.

Flow Cytometry Analysis
For assessment of successful DE induction, flow cytometry analysis was performed on day 3 of differentiation according to our previously published protocol [26,27]. Shortly, cells were detached with Accutase TM and stained with anti-CXCR4, anti-cKIT, and anti-EpCAM in FACS Buffer (PBS w/o, 1.0% FCS (Pan Biotech, Aidenbach, Germany), 1 mM EDTA) for 30 min on ice. After washing twice with FACS buffer, cells were resuspended in FACS buffer containing 1.7 µg/mL DAPI (Sigma-Aldrich, Saint Louis, MO, USA) for dead cell exclusion for flow cytometry measurement. Lung progenitor cultures at day 14 of differentiation were detached as described above and fixed and permeabilized using the transcription factor staining buffer kit (Miltenyi Biotec, Bergisch Gladbach, Germany). After washing with flow buffer (PBS w/o containing 1.0% BSA), cells were stained with anti-NKX2.1 for 30 min on ice. Cells were washed three times with FACS buffer and resuspended in FACS buffer for flow cytometry measurement.
Antibodies and respective dilutions are listed in Supplementary Table S3. All flow cytometry measurements were performed using the MACSQuant Analyzer 10 (Miltenyi Biotec, Bergisch Gladbach, Germany) and data were analyzed with FlowJo software (Ashland, OR, USA).

Differentiation of Lung Progenitor Cells towards Pseudostratified Airway Epithelium on Air-Liquid Interface Culture
After MACS sorting, purified lung progenitor cells were resuspended in a small airway epithelial cell growth medium (SAECGM; PromoCell, Heidelberg, Germany) supplemented with 1% penicillin/streptomycin (Gibco, Billings, MT, USA), 1 µM A83-01 (Tocris, Bristol, UK), 0.2 µM DMH-1 (Tocris, Bristol, UK) and 5 µM Y-27632 [28]. For initiation of airliquid interface (ALI) culture, 2.65 × 10 5 cells/cm 2 were seeded per transwell (Greiner Bio-One, Frickenhausen, Germany), coated with 804 G conditioned medium [28]. During the expansion phase, the medium was renewed every other day in the apical and basal chambers. Once cells have formed a confluent cell layer (usually after 4 days), the medium in the apical and basal chamber was switched to PneumaCult TM -ALI medium (STEMCELL Tech., Vancouver, BC, Canada) containing 1% penicillin/streptomycin. After 48h, the medium in the apical chamber was removed to conduct the airlift and the medium in the basal chamber was renewed. Cells were differentiated on ALI cultures for 28 days for molecular characterization and >35 days for functionality assessment with medium exchanges every two to three days.

Quantitative RT-PCR
ALI cultures were lysed at day 28 after airlift using TRIzol (Thermo Fisher Scientific, Waltham, MA, USA). RNA isolation was performed using the Nucleospin RNA II Kit (Macherey-Nagel, Nordrhein-Westfalen, Germany), and reverse transcription of 500 ng RNA was conducted using the RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA). cDNA was diluted 1:5 and RT-qPCR was conducted using SsoAdvanced TM Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) with PrimePCR Assays (CCDC40, CK5) or self-designed primers (FOXJ1, P63, NGFR, NKX2.1, MUC5AC, CCSP, betaActin, GAPDH). Primer sequences and assay IDs are listed in Supplementary Table S2. Samples were pipetted as duplicates and run on the CFX Connect TM Real-Time System (Bio-Rad, Hercules, CA, USA). Ct values were averaged and normalized to the housekeeping genes GAPDH and betaActin.

Cytospin
Prior to staining for ciliary protein markers, airway epithelium was dissociated into single cells or small clumps. For this purpose, the membrane of day 28 ALI cultures was removed from the transwell using a scalpel and incubated in Accutase TM supplemented with 200 µg/mL DNase I (Roche) for 15-20 min in a shaking 37 • C water bath. Dissociation was stopped with RPMI medium (Gibco, Billings, MT, USA) and cells were spun at 138× g for 3 min. After resuspending in RPMI medium, cells were centrifuged onto a Superfrost ® microscope slide at 650 rpm for 5 min using Cytospin2 (Shandon, Cambridge, UK). Slides were dried at room temperature and stored at −20 • C until used.

Paraffin Embedding and Sectioning
For fixation, the medium was aspirated from the apical and basal side and replaced with 4% paraformaldehyde (PFA), and was incubated for 15 min at room temperature. In order to remove mucus from the airway epithelium, the apical side of ALI cultures was washed repeatedly with PBS (Gibco, Billings, MT, USA) before fixation. After fixation, transwells were washed with and stored in PBS at 4 • C. For embedding, the membrane was cut out of the transwell using a scalpel, dehydrated (Leica Biosystems, Nußloch, Germany), and embedded in paraffin (Leica Biosystems, Nußloch, Germany). Using a microtome (Leica Biosystems, Nußloch, Germany), cross-sections of 3 µm thickness were generated.

Immunofluorescence Staining
Before staining paraffin-embedded samples, paraffin was removed from cross-sections by first heating the sections to 60 • C for 30 min, followed by incubation in xylol for 2 × 15 min. Rehydration of the sections was conducted by stepwise incubation in ethanol (100%, 90%, 80%, and 70%) for 2 min each and finally, incubation in water. For antigen retrieval, slides were incubated in 10 mM citrate buffer, pH6 (Sigma Aldrich, Saint Louis, MO, USA) for 60 min at 133 • C.
Prior to staining of cytospin samples, slides were taken from the −20 • C storage and fixed immediately with 4% PFA for 2 min at room temperature.
For immunofluorescence staining of cytospin samples and cross-sections, samples were washed twice with Tris-buffered saline (TBS) and were blocked with blocking buffer (TBS containing 5% donkey serum (Biozol, Munich, Germany) and 0.25% Triton X-100 (Sigma-Aldrich, Saint Louis, MO, USA)) for 1 h at room temperature. Primary antibody incubation was performed in a staining buffer (TBS with 1% BSA) at 4 • C overnight. Primary antibodies and respective dilutions are listed in Supplementary Table S3. Slides were washed three times with TBS and secondary antibody incubation was performed in staining buffer for 30 min at room temperature. Secondary antibodies are listed in Supplementary Table S4. Samples were washed three times with TBS and nuclei were stained with TBS containing 1.7 µg/mL DAPI (Sigma-Aldrich, Saint Louis, MO, USA) for 5 min. After washing three more times with TBS, sections were mounted with mounting medium (Dako, Nowy Sącz, Poland) and dried overnight at room temperature. Images were taken with the Axio Observer A7 (Zeiss, Baden-Württemberg, Germany) by optical sectioning using the Apotome. Pictures were processed using ZenBlue 3.0 software (Zeiss, Baden-Württemberg, Germany).

Transepithelial Electrical Resistance (TEER) Analysis
Prior to measurement, ALI cultures were supplied with fresh PneumaCult TM -ALI medium on the basal side and covered with 750 µL of PBS supplemented with 100 µg/mL Primocin (InvivoGen, San Diego, CA, USA) on the apical side of the transwell. Electrical resistance between the apical and basal chamber is measured with an EVOM3 electrical volt-ohm meter (World Precision Instruments, Sarasota, FL, USA) using the STX4 electrode (R Total ). Measurements were carried out in technical duplicates. A transwell without cells was used for blanking (R Blank ). TEER values were calculated as follows:

Transmission Electron Microscopy
ALI cultures were washed with PBS as described above and subsequently fixed and stored in 150 mM HEPES buffer containing 1.5% PFA and 1.5% glutaraldehyde. Membranes were cut off the ALI inserts and fixation and embedding were carried out as described in [29]

PCD Detect Software Analysis
For the assessment of ultrastructural dynein arm defects, PCD Detect Software was used, following the developers' workflow and using the same alignment settings [30]. From each cilia cross-section, 9 cut-outs, each including a microtubule doublet with outer and inner dynein arms, were manually selected. At least 5 cilia cross-sections (equivalent to 45 cut-outs) from different cells were analyzed in each sample. In this study, we present the averaging of all features as color contour maps of electron density (blue: high density; yellow: low density).

Mucociliary Clearance (MCC) Measurement
Only cultures at day 35 of ALI or older were used for MCC measurements to ensure substantial maturity of MCC. To remove accumulated mucus and debris, ALI cultures were submerged with PBS for 10 min, then the PBS was carefully suctioned off. To visualize MCC, fluorescent microspheres (1-micron diameter, Invitrogen, Waltham, MA, USA) were diluted 1:1000 in PBS, and 20 µL of this suspension was added to the surface of each ALI culture. Particle movement was recorded using a Zeiss Axioscope fluorescence microscope equipped with an Orca Flash 4.0 camera (Hamamatsu, Shizuoka, Japan) and a temperaturecontrolled chamber that was preheated to 37 • C. Apart from a few initial test recordings, standardized acquisition settings using a 10× objective were 1024 × 1024 pixels with 1.3 µm/pixel resolution at a framerate of 20 frames per second for a total duration of 10 s. For each condition, at least 3 experiments (different differentiations or clones (hiPSCs), or different isolations (pALI)) were performed. In each experiment, 5 fields of view (FOVs) with visible particle movement (sometimes only Brownian motion) were recorded from 2 insert cultures each. The particle trajectories were extracted from the raw movies using ImageJ/Fiji [31] with the Trackmate plugin [32]. Movies that did not contain analyzable bead signals due to debris or overgrowth were excluded. Trajectories were further analyzed to compute the average trajectory speed Uk and MCC coverage C k (the fraction of the FOV containing MCC) of each FOV k using customized code in Matlab (Mathworks), as reported previously [33]. For statistical analysis, the pooled FOV-averaged trajectory speed and MCC coverage values (the U k and C k for all k = 1 . . . n FOVs, respectively) per condition were compared with the non-parametric Kruskal-Wallis test, followed by Tukey-Kramer multi-comparison test, to test for different median values.

Ciliary Beat Frequency (CBF) Measurement
Only cultures at day 35 of ALI or older were used for CBF measurements. CBF was recorded using a Zeiss Axioscope microscope, allowing for oblique phase contrast recordings, equipped with an Orca Flash 4.0 camera (Hamamatsu, Shizuoka, Japan) and a temperature-controlled chamber that was preheated to 37 • C. Koehler illumination was set up for all recordings. Two different procedures were performed to measure CBF. In non-PCD pALI and non-PCD iALI cultures, CBF was measured at the tissue level, and single-cell CBF was measured in DNAH5 mut and NME5 mut cultures, as the movement of cilia could not be resolved at the tissue level. For tissue-level CBF measurements, ALI cultures were washed as above. High-speed movies of CBF (ca. 140 frames per second for a duration of at least 1.5 s) were recorded using a 40× long-distance phase contrast objective at a spatial resolution of 0.3 µm/pixel in a 512 × 512 pixel frame. For each condition, at least 3 experiments (different differentiations (hiPSCs) or different isolations (pALI)) were performed. From the cultures of each experiment, 6 FOVs with visible CBF were recorded. We measured cilia beat frequencies by applying Fourier spectral analysis, as previously described [34]. The mean CBF was calculated for each window of 32 × 32 pixels, resulting in a maximum of 16 mean CBF values per FOV if all windows contained ciliary beat, and from these values, the median CBF of the entire FOV was determined.
For single-cell CBF measurements, the cell layer was dissociated by incubating the ALI cultures for 20 min in prewarmed Accutase TM in a conical tube. DNase I at a concentration of 200 µg/mL was added to remove any free DNA resulting from the dissociation process. After 20 min warm culture medium was added at a ratio of 3:1 and the cells were centrifuged at 170 to 240× g for 5 min. The resulting pellet was resuspended in 1 mL culture medium to obtain a high cell density and 10 µL of this suspension was placed onto a glass bottom imaging dish (Ibidi, Munich, Germany). To confirm the viability of the cells, the same amount of life-dead stain (Invitrogen, Waltham, MA, USA) was added to the cell suspension droplet, and the solution was covered with a round glass coverslip in order to prevent evaporation. The measurements were performed using a 40× oil phase contrast objective (NA 1.4) under standardized acquisition settings, which were 800 frames per second for a duration of 2 s (NME5 mut ) and up to 10 s (DNAH5 mut ) of recording time and a spatial resolution of 0.16 µm/pixel in a 265 × 265 pixel frame. For each condition, at least 3 experiments (different differentiation (hiPSCs) or different isolations (pALI)) were performed, and in each experiment, at least 9 FOVs, i.e., 9 ciliated cells, were recorded. After each CBF movie recording a brightfield image and fluorescent images of the life-dead staining were recorded to confirm viability. The recordings were analyzed using the kymograph function in ImageJ. Kymographs of individual or groups of cilia were generated and the frequency was calculated by counting the amount of pixel intensity changes over time in the kymograph. Whenever a wave pattern was observed, the number of peaks were counted, else the dark lines were counted and divided by two to obtain a full beat cycle count. In the case of non-harmonic oscillation observed in DNAH5 mut -derived ciliated cells, stationary vibrations were not included in the count. A median was calculated from all counted frequencies per ciliated cell. For statistical analysis, the pooled tissue-level medians and single-cell medians per condition were compared with the non-parametric Kruskal-Wallis test, followed by Tukey-Kramer multi-comparison test, to test for different median values.
In order to direct the hiPSCs towards respiratory airway epithelial cell fate, a stepwise differentiation protocol was used that induced definitive endoderm (DE), anterior foregut endoderm (AFE), and lung progenitor cells ( Figure 1A). The protocol was slightly adapted individually for all cell lines regarding seeding densities and duration of AFE induction.

PCD Specific hiPSC-Derived Epithelial Cells show Impaired Expression of Cilia Proteins
In order to visualize ciliary protein expression, ALI cultures at day 28 post airlift were dissociated into single cells and stained for DNAH5 and NME5 protein, respectively. Immunofluorescence staining against AcTub was used to identify cilia and confirmed the presence of cilia with a normal appearance in all cell lines. In ciliated cells generated from non-PCD diseased primary cell-derived ALI cultures (non-PCD pALIs; Figure 2A,B) as well as non-PCD hiPSC-derived ALIs (non-PCD iALIs; Figure 2C,D), expression of DNAH5 and NME5 was observed that was associated with the cilia. In contrast, DNAH5-mutant ciliated cells (DNAH5mut Cl. 22 and Cl. 24) showed the absence of DNAH5 protein in the cilia. NME5 protein expression remains unchanged in these cells ( Figure 2E-H). In ciliated cells derived from NME5-mutant hiPSCs, cilia show a lack of NME5 protein expression and unchanged DNAH5 expression ( Figure 2I,J).

Transmission Electron Microscopy Analysis Shows Impact of DNAH5 and NME5 Mutations on Ciliary Ultrastructure
To further investigate the impact of each specific mutation on the cilia axoneme ultra-structure, transmission electron microscopy (TEM) of generated iALI cultures was performed. Analysis of cilia of non-PCD diseased hiPSC-derived cells showed a normal 9 + 2 ultrastructure pattern in more than 95% of analyzed cilia axonemes ( Figure 3A,B) as expected. In contrast, NME5-mutants show a 9 + 2 structure only in 56.6% of cilia axonemes and frequent numerical abnormalities or disorganization of cilia ultrastructure ( Figure 3A,B). 5.1% of axonemes exhibited a 9 + 0 ultrastructure pattern, which is consistent with previous findings of Cho and colleagues [36]. Furthermore, we detected additional abnormal ultrastructural patterns. Most abundantly, 23.4% of axonemes exhibited an 8 + 1 pattern. A smaller subset of axonemes showed additional configurations of microtubule transpositions (e.g., 7 + 4, 8 + 4), additional microtubules (9 + 4 pattern), and/or microtubular disorganization. These findings resemble phenotypes associated with mutations in other radial spoke proteins (RSPH1, RSPH3) [37,38].  lar disorganization. These findings resemble phenotypes associated with mutations in other radial spoke proteins (RSPH1, RSPH3) [37,38]. For investigating outer dynein arm (ODA) defects in DNAH5-mutated cilia, TEM cross-sections were analyzed using the PCD Detect software [30]. This software allows overlapping and averaging of microtubule features to generate enhanced signals and improved structural resolution. By averaging the electron density of multiple outer microtubule doublets, the shortening of outer dynein arms in DNAH5-mutated cilia compared to non-PCD controls was detected and visualized in color-coded density plots ( Figure 3C).  For investigating outer dynein arm (ODA) defects in DNAH5-mutated cilia, TEM cross-sections were analyzed using the PCD Detect software [30]. This software allows overlapping and averaging of microtubule features to generate enhanced signals and improved structural resolution. By averaging the electron density of multiple outer microtubule doublets, the shortening of outer dynein arms in DNAH5-mutated cilia compared to non-PCD controls was detected and visualized in color-coded density plots ( Figure 3C).

PCD Specific hiPSC-Derived ALI Cultures show Altered Ciliary Beat Frequencies
To evaluate the maturation and functional consequences of the DNAH5 and NME5 mutations in PCD, we measured ciliary beat frequencies (CBF) using high-speed microscopy on day 35 after airlift or later. In non-PCD cells (pALI and iALI) ciliary beat was visible directly in ALI cultures, enabling us to use tissue-level recordings and automated analysis of CBF ( Figure 4A). In PCD cells (DNAH5mut and NME5mut), the ciliary beat was too sparse or abnormal for direct analysis of ALI cultures, and we, therefore, conducted a manual analysis of CBF in single-cell suspensions ( Figure 4B). Non-PCD iALI cultures showed a relatively low variation of CBFs and a median value of 8 Hz (IQR = 1.3), which compared well to the similarly distributed CBF values in non-PCD pALI with a median value of 11.4 Hz (IQR = 3.0). Compared to non-PCD iALIs, DNAH5mut cultures showed significantly lower values of CBF, with a median of 2 Hz (IQR = 1.2), whereas NME5 cultures exhibited significantly higher and highly variable values of CBF, ranging from 5 to 19 Hz with a median of 15 Hz (IQR = 4.5) ( Figure 4C).

PCD Specific hiPSC-Derived ALI Cultures Show Impaired Mucociliary Clearance (MCC)
Since functional MCC depends on many factors, including ciliary beat coordination and orientation [39,40], CBF alone is not a good predictor of MCC [33]. Indeed, visually

PCD Specific hiPSC-Derived ALI Cultures Show Impaired Mucociliary Clearance (MCC)
Since functional MCC depends on many factors, including ciliary beat coordination and orientation [39,40], CBF alone is not a good predictor of MCC [33]. Indeed, visually comparing the ciliary beat in suspended cells of all conditions revealed metachronal and coordinated beats in non-PCD pALI and iALI cultures (Supplementary Videos S1 and S2), whereas we observed sparse strokes and occasional vibrations in DNAH5-mutant cells (Supplementary Video S3), which is consistent with the literature [23,41,42]. NME5-mutant cells showed an asymmetrical beating pattern (Supplementary Video S4) and sporadically exhibited circular instead of whipping cilia movements (data not shown). This has not been described for NME5 defects, yet, but resembles findings in other radial spoke defects [24,[43][44][45]. Altogether these defects suggest that the cilia beat in the PCD mutants might be ineffective at transporting fluids. We, therefore, proceeded to measure a direct proxy of MCC by recording the transport of suspended fluorescent microspheres on the intact tissue surface at day 35 post-airlift or later ( Figure 5A). As expected, the non-PCD pALI cultures generated the fastest flow that covered most of the tissue surface. Non-PCD iALI cultures created significantly faster and more expansive MCC than DNAH5 and NME5-mutated cultures ( Figure 5B), demonstrating that the functionally healthy versus the diseased phenotype of mutated airway epithelia from PCD patients is exhibited by the hiPSC-derived cultures. comparing the ciliary beat in suspended cells of all conditions revealed metachronal and coordinated beats in non-PCD pALI and iALI cultures (Supplementary Videos S1 and S2), whereas we observed sparse strokes and occasional vibrations in DNAH5-mutant cells (Supplementary Video S3), which is consistent with the literature [23,41,42]. NME5-mutant cells showed an asymmetrical beating pattern (Supplementary Video S4) and sporadically exhibited circular instead of whipping cilia movements (data not shown). This has not been described for NME5 defects, yet, but resembles findings in other radial spoke defects [24,[43][44][45]. Altogether these defects suggest that the cilia beat in the PCD mutants might be ineffective at transporting fluids. We, therefore, proceeded to measure a direct proxy of MCC by recording the transport of suspended fluorescent microspheres on the intact tissue surface at day 35 post-airlift or later ( Figure 5A). As expected, the non-PCD pALI cultures generated the fastest flow that covered most of the tissue surface. Non-PCD iALI cultures created significantly faster and more expansive MCC than DNAH5 and NME5-mutated cultures ( Figure 5B), demonstrating that the functionally healthy versus the diseased phenotype of mutated airway epithelia from PCD patients is exhibited by the hiPSC-derived cultures.

Discussion
To date, mainly primary cells or non-airway immortalized cell lines (e.g., HEK293 cell line) have been used for PCD in vitro modeling or drug screening [46][47][48]. However, both cell sources harbor severe limitations for such applications. Non-airway immortalized cell lines do not allow assessment of the therapeutic effect on a functional level, and primary airway cells, while they represent the PCD phenotype well, are limited in their accessibility and expandability. Even though it has been demonstrated that primary cells can be expanded over 35 passages and maintained their differentiation potential [49], this requires laborious matrix-embedded cultures, which limits the cell yield. Furthermore, targeted gene editing is not possible on a clonal level. hiPSCs exhibit infinite self-renewal and can be differentiated into the cell type of interest, thereby overcoming the limitations of primary cells. Moreover, PCD-patient-specific hiPSCs can be used to study the disease

Discussion
To date, mainly primary cells or non-airway immortalized cell lines (e.g., HEK293 cell line) have been used for PCD in vitro modeling or drug screening [46][47][48]. However, both cell sources harbor severe limitations for such applications. Non-airway immortalized cell lines do not allow assessment of the therapeutic effect on a functional level, and primary airway cells, while they represent the PCD phenotype well, are limited in their accessibility and expandability. Even though it has been demonstrated that primary cells can be expanded over 35 passages and maintained their differentiation potential [49], this requires laborious matrix-embedded cultures, which limits the cell yield. Furthermore, targeted gene editing is not possible on a clonal level. hiPSCs exhibit infinite self-renewal and can be differentiated into the cell type of interest, thereby overcoming the limitations of primary cells. Moreover, PCD-patient-specific hiPSCs can be used to study the disease on different disease-implicated tissues (airway cilia, sperm cells, nodal cilia) and can be genetically modified on a clonal level, e.g., for introducing reporter systems for screening readout. Hence, hiPSCs represent a valuable tool for PCD disease modeling and screening applications.
Current differentiation protocols for the generation of airway epithelium from hiPSCs often include an intermediate 3D culture step, in which lung progenitor cells are embedded in Matrigel droplets, prior to differentiation to airway epithelium [25,26]. Hence, these protocols are lengthy (28-46 days until initiation of ciliation by airlift or supplementation with DAPT [25,26]), and scale-up of cell production is severely limited due to the small scale and laborious 3D culture. Here, were describe an efficient airway differentiation protocol that takes only 20 days until initiation of ciliation. Moreover, our protocol is not dependent on complex Matrigel-dependent organoid culture steps, hence more defined and scalable for future high throughput drug screening applications.
Sone et al. previously reported a hiPSC-based disease model of PCD-causing mutations in DNAH11, HEATR2, and PIH1D3 using a complex airway-on-a-chip technology [24]. By applying fluid shear stress via defined cell culture medium flow rates on the apical side of the forming epithelium, planar cell polarity can be effectively induced in ciliated cells. Hence, compared to simple transwell ALI cultures, the chip technology allows the application and investigation of mechanical effects on the cells. However, such culture-on-a-chip is technically complex, expensive, and also harbors several limitations. The microfluidic design of the system hampers the establishment of chip cultures in unexperienced labs, the access to the cultured epithelium, the application of standard characterization methods (e.g., TEM, staining of cross-sections), and functional high throughput readouts using plate reader technology. In particular, electrophysiological assessments such as TEER and Ussing chamber measurements, which are valuable techniques to determine epithelial quality and functionality [36,50], are not feasible in commercially available airway chips. Integration of electrodes is possible in custom-made chip systems but this approach is very laborious and requires microfabrication infrastructure and expertise [51,52]. Moreover, the submerged culture system utilized by Sone and colleagues [24] does not allow for investigation of the implication of mucus properties in PCD disease development and infection processes. Additionally, for future compound screening applications, downscaling of culture size to high or medium throughput formats is indispensable and not yet possible for chip cultures.
In contrast, transwell ALI cultures are compatible with standard readouts, easier to establish due to the commercial availability and handling simplicity, and available in various formats including 96-well plates (Corning, Corning, NY, USA #100-0419), allowing higher throughput screening of pharmacological compounds. Altogether, the airway-ona-chip system is a valuable tool to investigate specific research questions under dynamic flow conditions; however, the transwell ALI culture system remains indispensable to be used as a complementary system, in particular for screening purposes.
In the present study, we describe the establishment of a hiPSC-based PCD in vitro disease model using patient lines containing mutations in the most frequently mutated and well-described PCD-associated gene DNAH5 [41,42], and in the recently PCD-associated described gene NME5 [36], respectively. Our data show that mutated DNAH5 results in loss of DNAH5 protein, altered ODA ultrastructure, and reduced CBF, consistent with previously published iPSC-based models [23]. Additionally, we investigated bead transport on the tissue level that displayed significantly reduced MCC capacity in DNAH5-mutated cultures. Moreover, we demonstrate the application of the newly developed PCD Detect software [30] that allows the identification of ODA defects. Previous disease models mainly focused on describing a severe DNAH5 ultrastructural phenotype that exhibits a total absence of ODA [23], but mutated DNAH5 can also result in less prominent ODA defects such as dynein arm truncation [4]. Reliable analysis of the whole range of ODA phenotypes is necessary for diagnostics and experimental readouts; however, analyzing ultrastruc-ture with only subtle defects can be challenging and strongly rely on the investigators' expertise [30]. Hence, we used the PCD Detect software as an investigator-independent and time-efficient diagnostic method [30] to identify subtle ODA shortening in our hiPSC DNAH5-mutated cultures. Altogether, the detailed characterization of our DNAH5-mutated iALI cultures serves as a proof-of-concept to demonstrate that our in vitro model recapitulates the DNAH5 mut -specific phenotype and can be captured in detail by utilizing improved readout methods.
Next, we aimed at modeling the impact of mutations in the gene NME5, which was recently identified as PCD-associated. NME5 is part of the radial spoke neck; however, so far, only little is known about the function of NME5 and the pathomechanism of its mutation [36,53]. Solely Cho and colleagues studied the effect of mutant NME5 in humans and have described an abnormal axonemal ultrastructure in mutant ciliated cells (loss of the central pair; 9 + 0 pattern) [36]. Here, we investigated the effects of mutated NME5 in more detail. Our TEM results are consistent with the previously reported loss of central pair (9 + 0) in NME5-mutated cilia, and moreover, show additional ultrastructural abnormalities, such as outer doublet transposition (8 + 1, 7 + 4, 8 + 4) and presence of additional microtubules (9 + 4), which have not been described for this mutation before. Importantly, similar ultrastructural abnormalities were reported in other radial spoke protein defects [22,45,[54][55][56]. These findings suggest that NME5 is indispensable for proper radial spoke assembly and stabilization of the central pair complex.
To further investigate the effect of mutated NME5 on single-cell and tissue levels, we performed high-speed video microscopy (HSVM). Single-cell functional analysis revealed that NME5-mutant ciliated cells beat in an uncoordinated way and exhibited partially circular cilia movements that resembled 9 + 0 nodal cilia beating. These nodal-like beat kinematics have been previously reported to be a consequence of central pair and transposition defects [24,[43][44][45]. Additionally, NME5-mutated ciliated cells showed an increased variance of CBFs, compared to controls. By assessing the mucociliary transport capacity, we could additionally demonstrate that these alterations lead to a significantly diminished mucociliary clearance capacity (coverage and speed) compared to non-PCD controls (iALI and pALI). Altogether, these findings indicate that mutated NME5 not only impairs the beating synchrony and pattern of a given cell but also potentially prevents tissue-level metachronal wave formation due to divergent beat frequencies, which breaks down mucus clearance [57].
Notably, also non-PCD iALIs showed reduced MCC compared to primary cell controls (non-PCD pALI), despite comparable CBF parameters (median and variance) and synchronous cilia beating. This heterogeneity may have been neglected in other studies that used smaller fields of view for measuring bead transport; e.g., Sone et al. used a 100 µm × 100 µm fields of view (FOV) [24] compared to our 600 µm × 600 FOV. Our larger FOV allows us to obtain a more representative picture of the tissue, which might be particularly necessary to detect phenotypes on a tissue level (as in the case of NME5 mut cells). Further underlying causes for the observed differences between pALI and iALI cultures may include an insufficient maturation of the iALIs. Despite analyzing the cultures earliest at day 35 after the airlift, maturation of ciliary beating (synchrony, force generation, and cilia length) might be not fully completed. So far, the maturation of cilia beating during differentiation is not well understood, and investigation of bead transport over the course of differentiation and analysis of cilia length remains to be carried out for further understanding. In addition, the higher heterogeneity of hiPSC-derived cultures (ciliated cells appearing in clusters and/or uneven epithelium surface) compared to pALIs might reduce MCC coverage and average trajectory speed, since bead transport slows down in non-ciliated areas.

Conclusions
Taking together, here we present a hiPSC-based PCD model platform utilizing a simple air-liquid interface culture system that recapitulates mutation-specific pheno-types by detailed characterization on a molecular (mRNA, protein, ultrastructure) and functional (CBF, HSVM, bead transport) level. In addition, our system allows down-scaling of the culture format and up-scaling of cell yield, which are essential prerequisites for upcoming higher throughput screening applications.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Hannover Medical School (8457_BO_K_2019).