Graphical Abstract

Overview of the study. PCD: primary ciliary dyskinesia; FEV1: forced expiratory volume in 1 s; ODA: outer dynein arm; MCC: multiciliated cell.

Abstract

Background Primary ciliary dyskinesia is a genetic disorder caused by aberrant motile cilia function that results in defective ciliary airway clearance and subsequently leads to recurrent airway infections and bronchiectasis. We aimed to determine: how many functional multiciliated airway cells are sufficient to maintain ciliary airway clearance?

Methods To answer this question we exploited the molecular defects of the X-linked recessive primary ciliary dyskinesia variant caused by pathogenic variants in DNAAF6 (PIH1D3), characterised by immotile cilia in affected males. We carefully analysed the clinical phenotype and molecular defect (using immunofluorescence and transmission electron microscopy) and performed in vitro studies (particle tracking in air–liquid interface cultures) and in vivo studies (radiolabelled tracer studies) to assess ciliary clearance of respiratory cells from female individuals with heterozygous and male individuals with hemizygous pathogenic DNAAF6 variants.

Results Primary ciliary dyskinesia male individuals with hemizygous pathogenic DNAAF6 variants displayed exclusively immotile cilia, absence of ciliary clearance and severe primary ciliary dyskinesia symptoms. Owing to random or skewed X-chromosome inactivation in six female carriers with heterozygous pathogenic DNAAF6 variants, 54.3±10% (range 38–70%) of multiciliated cells were defective. Nevertheless, in vitro and in vivo assessment of the ciliary airway clearance was normal or slightly abnormal. Consistently, heterozygous female individuals showed no or only mild respiratory symptoms.

Conclusions Our findings indicate that having 30–62% of multiciliated respiratory cells functioning can generate either normal or slightly reduced ciliary clearance. Because heterozygous female carriers displayed either no or subtle respiratory symptoms, complete correction of 30% of cells by precision medicine could improve ciliary airway clearance in individuals with primary ciliary dyskinesia, as well as clinical symptoms.

Shareable abstract

Restoring cilia function in 30–62% of defective respiratory cells is sufficient to improve ciliary clearance in PCD individuals https://bit.ly/45PQQg9

Introduction

The respiratory epithelium consists of different types of cells: non-ciliated cells (e.g. ionocytes, goblet and club cells) and ciliated cells. Coordinated beating of multiple motile cilia is essential for ciliary clearance of the upper and lower airways and represents an effective defence mechanism to prevent airway damage [1, 2]. Respiratory cilia characteristically beat with a forward effective stroke and a backward recovery stroke in the same plane along the cell surface. Primary ciliary dyskinesia (PCD), listed in the Online Mendelian Inheritance in Man catalogue of human genes and genetic disorders under the number 244400, is a rare genetic heterogeneous disorder caused by aberrant motile cilia function. Dysfunction of motile cilia in the airway results in defective ciliary clearance of the upper and lower airways and subsequently leads to recurrent airway infections, chronic inflammation, bronchiectasis and progressive lung failure [2, 3]. Outer dynein arm (ODA) complexes attached at 24 nm intervals to the peripheral microtubule doublets contain molecular motors that drive and regulate ciliary motility, while the inner dynein arms (IDAs) regulate the ciliary and flagellar beating pattern. Dynein axonemal assembly factors (DNAAFs) are responsible for the cytoplasmic preassembly of ODAs and IDAs before transport to the ciliary axoneme [4, 5]. We and others have shown that pathogenic variants in DNAAF6 cause the absence of all ODAs from the ciliary axonemes (shown by the absence of dynein axonemal heavy chain 5 (DNAH5)) as well as IDA defects (shown by the absence of dynein axonemal light intermediate chain 1 (DNALI1)) and therefore result in cilia immotility [6–8]. No causative treatment is thus far available and treatment efforts are aimed at hydration of mucus and early eradication of bacterial airway infections. Azithromycin maintenance therapy currently stands as the only evidence-based medical treatment within this context [9]. Precision medicine approaches in PCD, including gene therapy, mRNA transcript therapy and read-through therapy, are currently being explored [10].

However, it is not known how many multiciliated airway cells have to be completely corrected in order to restore ciliary airway clearance. To answer this very important medical and biological question, we exploited the molecular defects caused by pathogenic variants in the DNAAF gene DNAAF6/PIH1D3, responsible for a severe type of PCD characterised by immotile cilia in affected male individuals [6, 7]. Here, aberrant cytoplasmic assembly of axonemal ODAs and IDAs responsible for cilia beat generation and regulation causes complete ciliary immotility. We chose to explore the molecular defects underlying this X-linked model of PCD because X-chromosome inactivation (XCI) has evolved to equalise X-linked gene expression between female XX and male XY individuals to enable dosage compensation in marsupial and placental mammalians [11]. Failure to induce XCI in XX human embryos shows that a double dose of certain X-located genes results in early lethality during development [12]. In somatic tissues, the choice of which X chromosome to inactivate is generally random (figure 1). Because XCI is stably maintained, the same chromosome is inactivated in all progeny cells. Therefore, female individuals are mosaics of two populations of cells that differ in the X-chromosome that is active. Thus, in female individuals with heterozygous pathogenic DNAAF6 variants, half of all ciliated respiratory cells are lined with immotile cilia, whereas male individuals with hemizygous pathogenic DNAAF6 variants display exclusively immotile cilia in the airways (figure 1). However, skewed XCI can occur [13, 14] and even result in disease. For instance, in some female carriers of a dystrophin gene mutation, skewed inactivation (>90%) results in muscular dystrophy [15–17].

FIGURE 1

Random X-chromosome inactivation (XCI) in female carriers of heterozygous pathogenic DNAAF6 variants results in two populations of respiratory ciliated cells. The differentiated respiratory epithelium consists of different type of cells: non-ciliated cells (e.g. ionocytes, goblet and club cells) and ciliated cells. DNAAF6 is located on the X chromosome. In contrast to female individuals, in male individuals there is no XCI. Male individuals with a hemizygous pathogenic DNAAF6 variant display exclusively immotile cilia in the airways, resulting in primary ciliary dyskinesia. In contrast, due to XCI, female individuals have a mosaic of two populations of cells that differ in the X-chromosome that is active. Thus, in female individuals with a heterozygous pathogenic DNAAF6 variant, it is expected that half of all ciliated respiratory cells are lined with immotile cilia. In this study, we performed detailed in vitro and in vivo studies to assess the impact on ciliary clearance.

Here, we report five male PCD individuals with hemizygous and six female individuals with heterozygous pathogenic DNAAF6 variants from four distinct families (figure 2). We identified three distinct large genomic deletions with variable sizes, a 10 kB deletion (exons 2–4 of DNAAF6), a 28 kB deletion (exon 1 of DNAAF6) and a 400 kB deletion (all DNAAF6 exons), as well as the pathogenic DNA variant c.266c>A resulting in premature stop of translation (p.Trp89*). All male DNAAF6-affected individuals presented with typical PCD symptoms and variant respiratory cilia were immotile due to absence of ODAs and IDAs. Interestingly, we observed in the female heterozygous carriers that, due to random XCI, 38–70% of multiciliated respiratory cells displayed immotile cilia due to aberrant ODAs (absence of DNAH5) and IDAs (absence of DNALI1). We carefully examined the clinical phenotype of the female carriers of the heterozygous pathogenic DNAAF6 variants and performed in vivo functional ciliary clearance analyses using radiolabelled tracer and in vitro ciliary clearance assays in air–liquid interface (ALI) cultures of ciliated respiratory cells. Hemizygous male PCD individuals did not have any in vitro or in vivo ciliary clearance. However, despite 38–70% of the respiratory ciliated cells carrying immotile cilia, female carriers with heterozygous pathogenic DNAAF6 variants showed no or only mild respiratory symptoms. In vitro and in vivo ciliary clearance assays were either normal or showed a subtle reduction in ciliary clearance. We conclude that it is not necessary to have 100% of respiratory ciliated cells functioning to ensure effective ciliary clearance. A range of 30–62% of functioning cells is sufficient to provide either normal or only a subtle reduction in in vitro and in vivo ciliary clearance.

FIGURE 2

Pathogenic variants in DNAAF6 located on the X chromosome were identified in four families. a) In family OP-1826/OP-1796 we found a 28 kb deletion (del) including the first exon of DNAAF6 and the first two exons of NUP62CL. b) In family OP-3577, a 10.6 kb deletion comprising DNAAF6 exons 2–4 was identified in OP-3577 I2. Segregation analyses revealed that neither her brother (OP-3577 I3) nor grandmother (OP-3577 02) carry the deletion, indicating that this is a de novo variant. OP-3577 I2 reported difficulties in getting pregnant and had one spontaneous abortion in the sixth week of pregnancy. c) In OP-3141 we identified a 400 kb deletion comprising DNAAF6, NUP62CL and RBM41. OP-3141 II1 is the fourth of six children. He has three older and one younger sister. The only brother died 10 days after birth owing to cardiorespiratory failure of unknown cause. His oldest sister gave birth to two male individuals: the first born died 3 months after birth, the second born died 4 months after birth. Both died of unknown causes with cardiorespiratory failure. With another partner, she gave birth to a healthy boy. The second oldest sister has three healthy children, two girls and a boy. The third oldest sister gave birth to a healthy boy. The younger sister has no children. d) In family OP-2835, a transition from G to A (c.266G>A) was identified in DNAAF6 exon 4, predicting premature stop of translation (p.Trp89*). Segregation analyses revealed that the mother (OP-2835 I2) and the grandmother (OP-2835 02) carry this variant in a heterozygous status whereas the father (OP-2835 I1) is unaffected. The great-grandmother of OP-2835 II1 had a spontaneous abortion (SAB) and gave birth to OP-2835 02, who inherited the DNAAF6 variant. The great-grandmother died owing to complications of an ectopic (ECT) pregnancy. OP-2835 02 gave birth to three daughters. The oldest one has two healthy children, a girl, who already has another healthy daughter, and a boy. The second oldest one, OP-2835 I2, is a carrier of the DNAAF6 variant and has two children. She had one ectopic pregnancy with a previous partner and with her current partner two spontaneous abortions and an ectopic pregnancy. OP-2835 II2 was her fourth pregnancy and is also a carrier of the DNAAF6 variant. In addition, she gave birth to the index primary ciliary dyskinesia (PCD) male individual OP-2835 II1, who was her sixth pregnancy and who carries the hemizygous DNAAF6 variant. The younger sister had one abortion and no children. PCD-affected males are shaded black and unaffected siblings are shaded white. Heterozygous female carriers are indicated by a central dot. Genotypes are indicated in the pedigrees.

Material and methodsPCD-affected individuals and families

Signed and informed consent was obtained from individuals fulfilling the diagnostic criteria of PCD [18] and from family members according to protocols approved by the institutional ethics review board of the University Hospital Muenster (Ethik-Kommission Westfalen Lippe in Muenster, reference number 2015-104-f-S) and Copenhagen (Danish Data Protection Agency 2014).

Summary of applied methods

Please see the supplementary file for a detailed method section.

In this study, we performed mutational analyses using a customised PCD gene panel and whole exome sequencing (WES) and whole genome sequencing as well as Sanger sequencing to identify families with male PCD individuals carrying disease-causing hemizygous DNAAF6 DNA variants (figure 2, supplementary figures S1–S4). Genetic variants were evaluated according to American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) guidelines [19]. We only report disease-causing (pathogenic) variants (class 5) according to ACMG/AMP guidelines. Further genetic analyses identified female heterozygous carriers of the DNAAF6 variants (figure 2). Clinical findings such as laterality status and lung function and family history were ascertained (table 1).

TABLE 1

Clinical findings of female heterozygous pathogenic DNAAF6 variant carriers and PCD males with hemizygous pathogenic DNAAF6 variants

To understand the molecular defects in the male hemizygous and female heterozygous DNAAF6 variant individuals we performed various experiments in nasal ciliated airway cells (native cells after brushing), comprising transmission electron microscopy (TEM), high-speed video microscopy analysis (HVMA) and high-resolution immunofluorescence (IF) microscopy. We performed IF using antibodies directed against the ODA components DNAI1, DNAH5, DNAH9 and DNAI2 and the IDA component DNALI1. We counted a total of 500 cells per IF staining for male (n=1) and female (n=2) controls (OP-1796 I2, OP-1796 II1, OP-1826 I2, OP-1826 II1, OP-3577 I2 and OP-3577 II1). For OP-2835 II1, 200 cells and for OP-2835 I2 and OP-2835 II2 a total of 75 cells per IF staining and individual were counted. In addition ALI cultures from male hemizygous and female heterozygous DNAAF6 variant individuals were established and the ciliary beat frequencies, ciliary particle transport, high-resolution IF and immunoblot analyses were performed. For ciliary particle transport measurements (in vitro ciliary clearance assay), fluorescent beads (0.5 and 2.0 μm in diameter) were mixed with pre-warmed cell culture medium and added to the apical compartment of respiratory cells in ALI culture (day 30 after airlift). Transport of fluorescent nanoparticles by ciliary beating was recorded using the Nikon Eclips Ti-S microscope (×20 objective lens) equipped with the NIS-Elements Advanced Research software (20 s; 7.5 frames·s−1). In total, 30 videos per individual were analysed for statistical evaluation and 253 particles were tracked per video on average from three different cell culture inserts from each individual. Because two different bead sizes (0.5 and 2 µm) were used for the analyses, values were normalised against the mean value of the healthy control group. Tracking videos were evaluated using the NIS-Elements Advanced Research software (version 4.51.000) and NIS Advanced 2D Tracking plug-in to generate polargraphs and to determine the speed (μm·s−1). In parallel to each recorded tracking video, a corresponding differential interference contrast (DIC) video was taken by an additionally equipped Basler sc640–120fm monochrome high-speed video camera (recording 125 frames·s−1) to determine the cell condition as well as the ciliary beating pattern. DIC videos were evaluated using SAVA software [20]. Particle transport was measured in 18 healthy female individuals and one healthy male individual (age 25–60 years). The results of the experiments were quantified and statistically analysed using R (www.r-project.org).

To measure pulmonary ciliary clearance in vivo, we performed pulmonary radioaerosol mucociliary clearance (PRMC) measurements in one male PCD individual and three female carriers. For PRMC measurement, insoluble technetium-99m labelled nanocolloids were aerosolised and inhaled by 20 slow inspirations and successive forced expirations. Initial lung deposition and subsequent pulmonary mucociliary clearance immediately after inhalation of the radiolabelled tracer were assessed by repeated dynamic and static scintigraphic acquisitions for 2 h. In addition, a static measurement after 24 h and krypton-81m ventilation imagery of the ventilated lung area was performed. Whole-lung retention 1 h and 2 h after tracer inhalation were compared to predicted values calculated from penetration index, age and sex using previously published reference equations [21]. Tracheobronchial velocity (TBV) (bolus transport) was measured from dynamic acquisitions within the first hour. Involuntary cough was monitored by trained staff and significant coughing rendered the test inconclusive to avoid false positive PRMC results due to cough clearance. If a subject coughed during the test, voluntary cough clearance was assessed by 1 min of rigorous coughing after 2 h by analysing both the central part and the total lung regions [22].

ResultsHigh-resolution IF microscopy analyses

We performed high-resolution IF microscopy to analyse the ciliary defects at the cellular level in females with heterozygous and males with hemizygous pathogenic DNAAF6 variants, respectively (figures 3 and 4). We used antibodies directed against ODA components DNAI1, DNAH5, DNAH9 and DNAI2. Analyses of ciliated respiratory cells from female heterozygous carriers (OP-1796 I2, OP-1796 I3, OP-1826 I2, OP-3577 I2, OP-2835 I2 and OP-2835 II2) revealed that 54.3±10% (range 38–70%, table 1) of ciliated respiratory cells showed abnormal localisation of the ODA components DNAI1, DNAH5, DNAH9 and DNAI2 (table 1, figures 3 and 5, and supplementary figures S5, S9, S11 and S13) when compared to control respiratory cells (1.8±0.67%), consistent with random XCI. However, some female carriers (e.g. OP-1769 I3) had a very low proportion of normal multiciliated cells (30%), whereas others (e.g. OP1796 I2) had a high proportion (62%), indicating that those carriers had a skewed XCI. In accordance with reported findings [6–8], we found no ODAs in analysed ciliated respiratory cells (99.8±0.37%) from male PCD individuals carrying hemizygous variants in DNAAF6 (OP-1796 II1, OP-1826 II1, OP-3577 II1 and OP-2835 II1; figures 4 and 5, and supplementary figures S6, S10 and S12). Next, we analysed the localisation of the IDA component DNALI1. In female carriers, 45.2±13% (range 32–66%, table 1) of the respiratory cells showed abnormal localisation of DNALI1 along the ciliary axoneme when compared to control respiratory cells (2.4±1.6%) (figures 3 and 5, and supplementary figure S7). As expected, we observed an absence of DNALI1 in almost all respiratory cells (99.6±0.4%) from male PCD individuals carrying hemizygous variants in DNAAF6 (figures 4 and 5, and supplementary figure S8). Statistical analysis using the ANOVA two-tailed test showed a significantly increased count of ODA-defective and IDA-defective cells in respiratory cell samples from female carriers with a heterozygous DNAAF6 pathogenic variant when compared to healthy control respiratory cells (p<0.001). In addition, there was a significantly increased count of cells with normal ODA/IDA composition in respiratory cell samples from female carriers with a heterozygous pathogenic DNAAF6 variant when compared to male PCD individuals with a hemizygous pathogenic DNAAF6 variant (p<0.001). For male (n=1) and female (n=2) controls and for OP-1796 I2, OP-1796 I3, OP-1796 II1, OP-1826 I2, OP-1826 II1, OP-3577 I2 and OP-3577 II1, a total of 500 cells per IF staining were counted. For OP-2835II1, 200 cells were counted and for OP-2835 I2 and OP-2835 II2, 75 cells per IF staining were counted.

FIGURE 3

Heterozygous DNAAF6 pathogenic variants in female individuals result in a mosaic of cells with and without axonemal outer and inner dynein arm defects, respectively. a–c) Respiratory epithelial cells from female control individuals and female carriers with a heterozygous pathogenic DNAAF6 variant were double-labelled with antibodies directed against the outer dynein arm protein DNAH5 (green) and GAS8 (red). Both proteins colocalise (yellow) along the ciliary axonemes in all cells from the unaffected control (a). In female carriers with a heterozygous pathogenic DNAAF6 variant, DNAH5 showed a normal localisation in some cells or was completely absent or severely reduced in the ciliary axonemes of other (b, c). d–f) Respiratory epithelial cells from a female control and female carriers with a heterozygous pathogenic DNAAF6 variant were double-labelled with antibodies directed against acetylated tubulin (green) and the inner dynein arm protein DNALI1 (red). Both proteins colocalise (yellow) along the ciliary axonemes in all cells from the unaffected control (d). In female carriers with heterozygous pathogenic DNAAF6 variant, DNALI1 showed a normal localisation in some cells or was completely absent or severely reduced in the ciliary axonemes of other cells (e, f). Nuclei were stained with Hoechst 33342 (blue). g) Schematic summarising the immunofluorescence microscopy results. In female control cells, cytoplasmic preassembly of dynein arms was not altered and all cells showed normal axonemal composition of inner (DNALI1) and outer (DNAH5) dynein arms. In contrast, female carriers with a heterozygous pathogenic DNAAF6 variant show a mosaic of cells with normal axonemal composition of dynein arms (blue) and abnormal composition dynein arms (yellow), respectively. Scale bars: 10 μm. DIC: differential interference contrast.

FIGURE 4

Hemizygous pathogenic variants in DNAAF6 result in outer and inner dynein arm defects in primary ciliary dyskinesia (PCD)-affected males. a–c) Respiratory epithelial cells from control individuals and male PCD individuals OP-1826 II1 and OP-1796 II1 with a hemizygous pathogenic DNAAF6 variant were double-labelled with antibodies directed against the outer dynein arm protein DNAH5 (green) and GAS8 (red). Both proteins colocalise (yellow) along the ciliary axonemes in cells from the unaffected control (a). In cells from PCD individuals OP-1826 II1 (b) and OP-1796 II1 (c), DNAH5 was completely absent or severely reduced from the ciliary axonemes of all cells analysed. d–f) Respiratory epithelial cells from a control and male PCD individuals OP-1826 II1 and OP-1796 II1 with hemizygous pathogenic DNAAF6 variants were double-labelled with antibodies directed against acetylated tubulin (green) and the inner dynein arm protein DNALI1 (red). Both proteins colocalise (yellow) along the ciliary axonemes in all cells from the unaffected control (d). In cells from the male PCD individuals OP-1826 II1 (e) and OP-1796 II1 (f), DNALI1 was completely absent or severely reduced in the ciliary axonemes of all cells analysed. Nuclei were stained with Hoechst 33342 (blue). g) Schematic summarising the immunofluorescence microscopy results. In all control cells (blue), cytoplasmic preassembly of dynein arms was not altered and all cells showed normal axonemal composition of inner (DNALI1) and outer (DNAH5) dynein arms. In contrast, in male PCD individuals with hemizygous DNAAF6 pathogenic variants, all cells show abnormal composition of dynein arms (yellow). Scale bars: 10 μm. DIC: differential interference contrast.

FIGURE 5

Pooled statistical analysis of immunofluorescence (IF) microscopy findings using antibodies directed against outer dynein arm (ODA) component DNAH5 (a) and the inner dynein arm (IDA) protein DNALI1 (b) in control respiratory cells, cells from female carriers with a heterozygous pathogenic DNAAF6 variant (OP-1796 I2, OP-1796 I3, OP-1826 I2, OP-3577 I2 and OP-2835 I2) and male primary ciliary dyskinesia (PCD) individuals with a hemizygous pathogenic DNAAF6 variant (OP-1796 II1, OP-1826 II1, OP-3577 II1, OP-2835 II1 and OP-2835 II2). a) Analyses of ciliated respiratory cells of females with a heterozygous pathogenic DNAAF6 variant revealed that 54.3±10% (range 38–70%, table 1) show abnormal axonemal localisation of DNAH5 when compared to control respiratory cells (1.8±0.67%). In male PCD individuals with a hemizygous pathogenic DNAAF6 variant, almost all respiratory cells showed absence of DNAH5 from the ciliary axonemes (99.8±0.37%). b) Analyses of ciliated respiratory cells of female carriers with a heterozygous DNAAF6 pathogenic variant revealed that 45.2±13% (range 32–66%, table 1) show abnormal localisation of DNALI1 when compared to control respiratory cells (2.4±1.6%). In male PCD individuals with hemizygous pathogenic DNAAF6 variants, almost all respiratory cells showed axonemal absence of DNALI1 (99.6±0.4%). For male (n=1) and female controls (n=2) and for OP-1796 I2, OP-1796 I3, OP-1796 II1, OP-1826 I2, OP-1826 II1, OP-3577 I2 and OP-3577 II1, a total of 500 cells per IF staining were counted. For OP-2835 II1, 200 cells were counted. For OP-2835 I2 and OP-2835 II2, 75 cells were counted. Error bars denote standard deviation. p-values were generated by the ANOVA two-tailed tests. ***: p<0.001.

TEM

To further analyse the ciliary defect resulting from heterozygous pathogenic variants in DNAAF6, we performed TEM with samples from OP-1796 I3. In total, 581 ciliary axonemes were analysed. Our findings confirm that heterozygous pathogenic DNAAF6 variants result in a mosaic of cells with and without axonemal ODA defects (figure 6 and supplementary figure S14). Furthermore, ultrastructural analyses found ciliary cross-sections without ODAs (n=391, 67%) but also ciliary cross-sections with normal ODA localisation (n=190, 33%), consistent with our IF microscopy findings (table 1) and skewed XCI in OP-1796 I3.

FIGURE 6

Heterozygous pathogenic DNAAF6 variants in female carriers result in a mosaic of cells with and without axonemal outer dynein arm (ODA) defects. Transmission electron microscopy analysis of ciliary cross-sections from a) a control and b) female individual OP-1796 I3 carrying a heterozygous DNAAF6 variant show the absence of ODAs in more than half of the ciliated cells (n=391 of 581 sections, 67%) from the female carrier. Red arrows mark ODAs. A total of 581 ciliary cross-sections from OP-1796 I3 were analysed. White scale bars: 500 nm; black scale bars: 250 nm.

HVMA

HVMA of nasal respiratory epithelial cells to assess ciliary beating in male PCD individuals carrying hemizygous DNAAF6 variants showed that all respiratory cells were lined with immotile cilia (supplementary videos 1 and 2), consistent with the observed axonemal ODA defects by IF microscopy (figure 4). In the female carriers, patches with motile and immotile cilia were observed (supplementary videos 3–7 and supplementary table S2), consistent with variable XCI and our IF microscopy findings. In OP-1796 I3, 340 of 597 (57%) multiciliated cells carried immotile cilia, consistent with skewed XCI and also supported by IF and TEM findings. In the other female carriers who were analysed by HVMA (OP-2835 II2 and OP-1826 I2), patches of multiciliated cells with motile and immotile cilia were observed. However, due to the low number of cells counted, no quantification was performed.

Measurement of ciliary clearance in vitro and in vivo

Next, we tested whether respiratory cilia of DNAAF6-mutant cells can generate a directed fluid flow. To mimic the process of ciliary clearance in vitro, we added fluorescent particles to the apical compartment of respiratory cells in ALI culture from healthy control individuals; female carriers OP-1796 I2, OP-1796 I3, OP-1826 I2 and OP-3577 I2 with the heterozygous pathogenic DNAAF6 variant; and PCD male individuals OP-1796 II1, OP-1826 II1, OP-2835 II1 and OP-3141 II1 carrying hemizygous pathogenic DNAAF6 variants. Consistent with PCD and aberrant ciliary clearance, mutant cilia from PCD male individuals with hemizygous pathogenic DNAAF6 variants were not able to transport fluorescent particles along the surface of the differentiated epithelium (figure 7 and supplementary figure S15) and ciliary clearance significantly differed from healthy controls (p=0.0003). In contrast, respiratory cilia from healthy controls and carriers of heterozygous pathogenic DNAAF6 variants were able to generate a directed fluid flow (figure 7). However, the velocity of this flow was significantly slower than that of the healthy controls (p=0.009). Two female carriers (OP-1796 I2 and OP-3577 I2) had in vitro ciliary transport velocities in the lower normal range also consistent with IF findings. OP-1796 I3 showed the lowest in vitro ciliary transport velocity (figure 7), consistent with skewed XCI that was also supported by IF, TEM and HVMA findings. OP-1826 I2 also had low in vitro ciliary transport velocity, also consistent with IF findings.