Ethical approval for this project was obtained from the Ministry of Health Research Ethics Committee (Ethics ID: 62/2013). This study was carried out with the permission of all participating family members; informed written consent was obtained from the adult participants and from the parents/guardians/next-of-kin of children for collection of blood samples from the patients and non-affected parents as well as nasal biopsies.

Patients with a clinical picture of PCD were referred to clinics run by the Kuwait Ministry of Health for genetic assessment in collaboration with the Kuwait University Faculty of Medicine. The ethnic background of the patients included in this study include Asian Arabs descended from different tribes originating from the Arabian Peninsula (Al-Mutairi et al., 2022; Al-Mutairi et al., 2023). The inclusion criteria were clinical features typical of the hallmark disease of PCD (MIM: PS244400), including neonatal respiratory distress, chronic respiratory disease with symptoms of chronic airway infections, rhinosinusitis, otitis media, and bronchiectasis. Most patients were suspected to carry recessive disease-causing hereditary variants as more than one individual in their families exhibited similar clinical phenotypes.

Pedigrees for the families were constructed using Cyrillic version 2.1 (http://www.cyrillicsoftware.com/) as previously published (Al-Mutairi et al., 2022; Al-Mutairi et al., 2023). All radiological examinations of the patients in this study were performed at Sabah Hospital in Kuwait. The genomic DNA extractions and exome sequencing protocols were performed as described previously (Al-Mutairi et al., 2022; Al-Mutairi et al., 2023).

Genetic screening using autozygosity mapping was performed as described previously (Carr et al., 2013; Al-Mutairi et al., 2022; Al-Mutairi et al., 2023). AgileMultiIdeogram software was used to visualise the homozygous intervals using exome data for linkage, in which all the homozygous intervals were displayed against a circular ideogram for the 22 autosomal chromosomes for the related patients (Carr et al., 2013; Al-Mutairi et al., 2022; Al-Mutairi et al., 2023). Linkage analysis using autozygosity mapping was performed for all the affected individuals belonging to the same family, such as family KU-5 and family KU-9 together in one setting. This linkage was also performed for the patient belonging to family KU-2 separately. The reference genome annotation used was the Human Genome Build hg19 (UCSC genome browser) (Al-Mutairi et al., 2022; Al-Mutairi et al., 2023).

Primers for variant confirmation and segregation analyses were designed using the Primer3 software (http://frodo.wi.mit.edu/primer3/). Exon sequences were obtained from Genome Browser of the University of California, Santa Cruz (http://genome.ucsc.edu/) and RefSNP Report—dbSNP—NCBI (nih.gov) for all exons of DNAH5 as follows. For the first heterozygous variant rs761622153 in family KU-2 exon 34, the forward primer sequence used was 5′-AAT​GAG​AAA​CGT​GAC​TTC​AAC​A-3´ and reverse primer sequence used was 5′-ACT​GGG​TAA​ATG​CAG​ATA​GTG​TC-3´. For the second heterozygous variant in family KU-2 exon 72, the forward primer sequence used was 5′-TTT​GAC​CTG​GAT​TAA​TTT​AAA​ATG​G-3´ and reverse primer sequence used was 5′-CAC​AAA​AGA​GGA​CCA​TCA​CC-3´. For the first homozygous variant in family KU-5 exon 73, the forward primer sequence used was 5′-TCA​GGG​ATT​TAC​AAA​AAT​GAT​TAC​T-3´ and reverse primer sequence used was 5′-TAC​GCC​AAA​GCT​AGG​AGG​TC-3´. For the second homozygous variant in family KU-5 exon 75, the forward primer sequence used was 5′-TGA​GTT​TGT​AGC​TTG​ACA​CCA​A-3´ and reverse primer sequence used was 5′-TCT​CCC​AAC​CGA​AGA​TAT​GC-3´. For the first heterozygous variant in family KU-9 exon 10, the forward primer sequence used was 5′-ATC​AAA​GGG​TTA​TTG​GTG​GAT-3´ and reverse primer sequence used was 5′-TCC​CAA​GGA​GGG​TAC​AAA​AA-3´. For the second heterozygous variant in family KU-9 exon 77, the forward primer sequence used was 5′-AAA​CAC​TTA​CGC​CAG​AAT​CCA-3´ and reverse primer sequence used was 5′-AGT​CGA​GCT​AGC​AGC​CTG​AG-3´. The Sanger sequencing protocol was performed as described previously (Al-Mutairi et al., 2022; Al-Mutairi et al., 2023), and the chromatograms were analysed using the GeneScreen software (Carr et al., 2011a; Al-Mutairi et al., 2022; Al-Mutairi et al., 2024). The allele frequencies of the two homozygous missense variants of DNAH5 in the Arab population were estimated using the allelic discrimination test described previously (Al-Mutairi et al., 2022; Al-Mutairi et al., 2023).

The algorithms used to estimate the pathogenicity of the two missense variants, DNAH5:c.12614G>A and DNAH5:c.12947T>C, in Tables 1, 2 are derived from the extended information in the MutPred database, which estimates the pathogenicities of the identified variants based on the protein sequences and model changes with regard to the structural features and functional sites between the wild-type and mutant sequences (Li et al., 2009). All pathogenicity scores mentioned in Table 3 are from the dbNSFP v4.1a database, which is a human database of aggregated scores from a large number of precalculated scores and algorithms that have also been augmented with custom rank scores to improve comparability. The dbNSFP v4.1a database focuses on non-synonymous single-nucleotide variants (nsSNVs) (Liu et al., 2020).

Respiratory epithelial cells were obtained from the subjects by nasal brush biopsy and suspended in a cell culture medium. Samples were then spread onto glass slides, air dried, and stored at −80°C until use. The respiratory epithelial cells from the healthy control and five PCD-affected individuals highlighted in the pedigrees (Figures 1A, 3A, 6A) were dual labelled for IF analysis with antibodies against axonemal components (visualised with red-fluorescent secondary antibody) and antibodies to acetylated anti-tubulin (visualised with green-fluorescent secondary antibody) as the ciliary markers. The nuclei were stained with Hoechst 33342 (blue fluorescent) (Wallmeier et al., 2014; Al-Mutairi et al., 2023). The IF images were acquired with a Zeiss LSM 800 confocal microscope and processed using ZEN and ImageJ software (Al-Mutairi et al., 2022; Al-Mutairi et al., 2023).

The panel of primary antibodies used for the initial screening included mouse monoclonal anti-tubulin (acetylated) antibody (Sigma Aldrich, T6793) for staining the ciliary microtubules and rabbit polyclonal antibody anti-DNAH5 (1:300) (HPA037470, Atlas Antibodies). The secondary antibodies used were highly cross absorbed, including Alexa Fluor 488-conjugated goat antibodies to mouse (1:1,000) (A11029, Molecular Probes, Invitrogen) and Alexa Fluor 546-conjugated goat antibodies to rabbit (1:1,000) (A11035, Molecular Probes, Invitrogen) (Al-Mutairi et al., 2022; Al-Mutairi et al., 2023).

For the TEM analyses, nasal biopsies were obtained from the middle turbinate for the same five affected individuals. The ciliated cells were first fixed with glutaraldehyde (2.5%) in Sorensen’s phosphate buffer (pH 7.4), following which the samples were treated with osmium tetroxide (1.3%), dehydrated using graded ethanol series, and immersed in hexamethyldisilazane as the chemical drying reagent. The samples were next embedded overnight in 1,2-epoxypropan-epon- mixture (1:1) at 4°C. After polymerisation, several sections were retrieved onto copper grids and were stained with Reynold’s lead citrate. TEM was then performed using the Philips CM10 device (Wallmeier et al., 2014; Al-Mutairi et al., 2022; Al-Mutairi et al., 2023).

The KU-2 family had one affected male child KU-2.II.1 who met the inclusion criteria for PCD symptoms with situs solitus (Figure 1A); the child underwent spirometry testing at various ages and showed declining pulmonary functions with both obstructive and restrictive patterns at the age of 18 years. His diagnosis of PCD was confirmed by genetic and in vitro functional analyses of the airway cilia. The selected chest X-ray for KU-2.II.1 shows a triangular retrocardiac opacity with ‘flat-waist sign’ suggesting left lower lobe collapse (Figure 1C). Selected axial sections of the computed tomography (CT) chest study for this patient from the soft-tissue and lung windows show subsegmental right lower lobe collapse in addition to the left lower lobe collapse (arrows in Figures 1D, E).

Autozygosity mapping for KU-2.II.1 revealed a single region of homozygosity (ROH) summation at chromosome 15, as shown in Figure 1B. Filtering the variants in the exome data across this interval did not show any pathogenic variants; furthermore, since the parents are not related, this indicates that the PCD-causing variants are more likely inherited in compound heterozygous form. The analysis of exome data for deleterious variants in the known 53 PCD genes revealed that the patient carried two deleterious variants of DNAH5 inherited in the compound heterozygous fashion. The first heterozygous variant was a heterozygous non-sense mutation NM_001369.2(DNAH5):c.5503C>T; (p.Gln 1835*); rs761622153 in exon 34 from paternal inheritance; RS = 5.8 and RVIS = 0.18 indicate that this is significantly damaging. This was recently reported as a PCD-causing variant inherited in a consanguineous family from United Arab Emirates (Alsamri et al., 2021). The second heterozygous variant was a novel frameshift deletion NM_001369.2(DNAH5):c.12357_12358delTG; (p.Glu4120Alafs*4) resulting in premature termination of translation from maternal inheritance. Segregation analysis confirmed that the two heterozygous variants of DNAH5 were segregated with the disease phenotype, consistent with both being the PCD-causing variants in the KU-2.II.1 patient (Figures 2A–C).

The KU-5 family was a multiplex family with two affected females, KU-5.IV.1 and KU-5.IV.3, both of whom presented with the following symptoms: chronic productive cough, recurrent pneumonias, wheezing dyspnea, and constant nasal mucopurulent secretions with situs inverses totalis (Figure 3A). Confirmed diagnoses of PCD were made at the ages of 7 and 5 years, with both subjects having situs inversus totalis. According to the KU-5.IV.1 patient’s questionnaire, the symptoms became gradually progressive with age, and the results of spirometry tests showed that the patient had declining pulmonary function, with both obstructive and restrictive patterns at the age of 12 years. The radiological findings for the two affected siblings (KU-5.IV.1 and KU-5.IV.3) can be summarised as follows: chest X-rays demonstrate right-sided heart apex (arrow) and gastric air bubbles (arrowhead), consistent with situs inversus totalis (Figures 4A, B). In addition, KU-5.IV.1 had mild diffuse bronchiectatic changes that were visible in the medial aspects of both lower zones, as highlighted by the thin yellow arrows (Figure 4A). The main radiological findings from the high-resolution CT (HRCT) chest examination of KU-5.IV.1 were situs inversus and subsegmental atelectasis in both lungs with mild underlying bronchiectatic changes (arrows in Figures 4C, D). Furthermore, selected axial and coronal reconstruction images of the CT of the paranasal sinus (CT-PNS) revealed complete opacification of the bilateral paranasal sinuses, consistent with chronic inflammatory pansinusitis (highlighted by arrows in Figures 4E, F).

Autozygosity mappings for the two PCD individuals of family KU-5 showed a single shared interval that was identical by descent (IBD) at chromosome 5, as seen in Figure 3B. The estimated shared homozygous intervals were precisely determined using AgileVCFMapper software (chr5:11,427,782-21,860,252) and found to be almost 10.5 kB (Figure 3C). The DNAH5 location at the shared IBD interval (chr5:13,690,328-13,944,688) is highlighted with a red asterisk in Figure 3C. The exome data for the two affected siblings were filtered at the shared IBD interval at chromosome 5 for screening of the deleterious variants of DNAH5.

Two novel homozygous missense variants of DNAH5 were identified in both patients by whole exome sequencing. The first homozygous missense variant NM_001369.2(DNAH5):c.12614G>A; (p.Gly4205Glu) was identified in exon 73, as shown in Figure 5A. The second homozygous missense mutation NM_001369.2(DNAH5):c.12947T>C; (p.Leu4316Pro) was identified in exon 75, as shown in Figure 5B. Both variants have no RefSNPs (rs) numbers to date in either Ensembl or the UCSC genome browsers. Following linkage and sequencing analyses, segregation analysis was performed to confirm that both variants identified in DNAH5 were segregated with the disease phenotype. The results of segregation analyses confirmed that the parents, KU-5.III.1 and KU-5.III-2, were heterozygous carriers for both missense variants (Figures 5A, B). This implies that either of these variants could be responsible for PCD in family KU-5.

The two variants were assessed using different algorithms and statistical measurements to predict their pathogenicity. The first of these was rejected substitutions (RSs), defined as the number of substitutions expected under a neutral model minus the number ‘observed’ (estimated) for a particular alignment (Pazour et al., 2006; Al-Mutairi et al., 2022; Al-Mutairi et al., 2024); the second was polymorphism phenotyping (PolyPhen), which is a tool to predict the possible impacts of an amino acid substitution on the structure and function of a human protein (Adzhubei et al., 2013; Al-Mutairi et al., 2022; Al-Mutairi et al., 2024). The missense variant DNAH5:c.12614G>A resulting in an amino acid exchange (p.Gly4205Glu) had RS = 5.2, PPH = 3, MedPred = 81, and Varsome: likely pathogenic, all of which indicate that the variant is highly damaging. The missense variant DNAH5:c.12947T>C resulting in an amino acid exchange (p.Leu4316Pro) had RS = 4.6, PPH = 3, and MedPred = 83, all of which indicate that the second variant is also highly damaging. Together, these data for family KU-5 are unique and interesting since this is the only family in the study cohort that has two novel homozygous missense variants in DNAH5 that are not reported previously in any database.

We extended our in silico analyses to predict the pathogenicities of the DNAH5:c.12614G>A and DNAH5:c.12947T>C variants. Our findings reveal noteworthy scores using mutation prediction tools designed to assess the pathogenicity of the identified mutations, as outlined in Tables 1–3. First, the loss of the catalytic residue is critical for ciliary targeting, and the loss of integrity of the catalytic domain in multi-ciliated respiratory cell mutants is directly related to ciliary mistargeting. This involves the targeting of cargo-containing membrane carriers or vesicles to the periciliary membrane (Lu and Madugula, 2018; Al-Mutairi et al., 2022; Cilleros-Rodriguez et al., 2022; Al-Mutairi et al., 2024). Loss of catalytic residues for the two missense variants were significantly low and were considered pathogenic (Tables 1, 2). In addition, the two variants DNAH5:c.12614G>A and DNAH5:c.12947T>C had significant loss of helix scores, indicating that both variants were highly damaging (Romero Romero et al., 2018). The missense variant DNAH5:c.12614G>A had significant scores for gain in ubiquitination at Lys 4200, loss of sheet, and loss of molecular recognition feature (MoRF) binding, all of which indicate that the variant is highly pathogenic (Table 1).

Furthermore, the missense variant DNAH5:c.12947T>C causes severe loss of the catalytic residue at Leu 4316, which is also directly associated with cilia mistargeting. The DNAH5:c.12947T>C variant leads to significant loss of stability, which is a common and widely used predictive tool for estimating the pathogenicity of missense mutations and disease development. The predicted pathogenicity from single amino acid changes affecting the thermodynamic stabilities of proteins highlights the role of protein stability in genetic diseases (Blaabjerg et al., 2023; Al-Mutairi et al., 2024). Furthermore, the DNAH5:c.12947T>C variant has a significant loss of loop score, which estimates the pathogenicity of the missense variant on a phosphate-binding loop (P-loop) motif that is necessary for ATP binding (Romero Romero et al., 2018; Al-Mutairi et al., 2024). It is well known that axonemal dyneins are directly involved in driving ciliary movements by providing the sliding forces for the microtubules within the cilia, which require energy from ATP hydrolysis within the heavy chain domains of axonemal dyneins (Mitchison and Mitchison, 2010). The DNAH5:c.12947T>C variant has a significant gain of disorder score supporting the previous pathogenic value of this amino acid substitution (Table 2).

In addition, we utilised nine meta-tools available in the dbNSFP database for annotation and functional predictions of all potential nsSNVs in the human genome (Liu et al., 2011; Kopanos et al., 2019; Liu et al., 2020; Al-Mutairi et al., 2022; Al-Mutairi et al., 2024). All of these determine the pathogenicity based on combined evidence from several other in silico predictors, and all of them predict both DNAH5:c.12614G>A and DNAH5:c.12947T>C to be deleterious/pathogenic and disease-causing missense variants, as seen in Table 3. Thus, the in silico analyses indicate that each variant is potentially highly pathogenic and PCD-causing, i.e. the presence of each variant alone may be sufficient for the development of PCD.

According to the ACMG criteria, the missense variant DNAH5:c.12614G>A has been annotated as likely pathogenic and missense variant DNAH5:c.12947T>C has been annotated to have uncertain significance despite their high pathogenicity prediction scores. Subsequently, allelic discrimination assays were performed on the two homozygous missense DNAH5 variants using RT-PCR to estimate their carrier rate frequencies (DNAH5:c.12614G>A in Supplementary Figure S1 and DNAH5:c.12947T>C in Supplementary Figure S2) using a control panel composed of 100 DNA samples from normal individuals belonging to different Arabian tribes. As shown in Supplementary Figure S1, the genotype for the two PCD individuals is homozygous for the missense variant (A), and their parents are all heterozygous carriers for the wild-type allele (G) and defective allele (A), consistent with the Sanger sequencing data. Remarkably, the genotypes for the Arab control DNA panel show that all healthy individuals have the two wild-type alleles (G). Regarding the second variant, the genotype for the two PCD individuals is also homozygous for the second missense variant (C), and their parents are heterozygous carriers for the wild-type allele (T) and defective allele (C), which is again consistent with the Sanger sequencing data (Supplementary Figure S2).

The KU-9 family is a multiplex family with two affected siblings KU-9.II.1 and KU-9.II.2 (Figure 6A) who met the inclusion criteria for PCD with situs solitus but showed only obstructive spirometry patterns. The male patient KU-9.II.1 was a smoker, so it is difficult to predict whether the pulmonary dysfunction was due to PCD or smoking; the patient was also substerile and had conceived two children by IVF. The female patient KU-9.II.2 had typical PCD symptoms, and her physician referred her to a genetic clinic for a confirmed diagnosis; she also experienced fertility problems and conceived a child by IVF after 5 years of marriage. The confirmed diagnoses of PCD were made through genetic and ciliary functional analyses of the two affected siblings KU-9.II.1 and KU-9.II.2 at the ages of 32 and 27 years, respectively. The main radiological findings from the HRCT chest investigation for KU-9.II.2 demonstrate situs solitus morphology with moderate cystic bronchiectatic changes involving the lateral segment of the right middle lobe. No other abnormalities or signs of active infection were observed (Figures 6C–E).

Although family KU-9 is a multiplex family with two affected siblings, the parents were not close relatives; it is very rarely the case that a family with two affected individuals has unrelated parents. However, exome sequencing of the two siblings indicated that the family was highly inbred, with the ROH summation being 220 and 117 for KU-9.II.1 and KU-9.II.2, respectively. The ROH summation cutoff value for inbreeding is 100, so the individuals theoretically belong to a highly inbred family. It is worth noting that family KU-9 belongs to a tribe that has practised consanguineous marriages among members of the tribe for generations and is considered to be particularly isolated, i.e. they have never married outside the tribe for historical and/or cultural reasons. Linkage analysis for the two affected siblings showed multiple ROH segments consistent with the ROH scores. However, none of these were shared between the two affected siblings, except for a minute interval at chromosome 2 (Figure 6B). Screening of this interval does not show any pathogenic variants. Subsequently, two interesting ROH intervals at chromosome 1 for patient KU-9.II.1 and chromosome 5 for patient KU-9.II.2 were prioritised for screening of the pathogenic variants. The ROH highlighted with a red asterisk (Figure 6B) is across the DNAH5 locus at chromosome 5 for patient KU-9.II.2. The other ROH interval at chromosome 1 did not harbour any pathogenic variants. Interestingly, the exome data for the two affected siblings showed that both had two pathogenic heterozygous variants of DNAH5. Theoretically, autozygosity mapping can only map an IBD interval harbouring homozygous variants, and since we mapped the compound heterozygous variants in DNAH5 in the two patients of family KU-9, this might explain the reason for not detecting the shared IBD using exome data across the DNAH5 locus in the two affected siblings. The set of genetic data for family KU-9 is also unique and interesting since this multiplex family has two affected individuals who have inherited two novel compound heterozygous variants of DNAH5 causing PCD with situs solitus, which has not been reported previously.

Exome and Sanger sequencings show that the two PCD individuals shared compound heterozygous variants; they were both carriers for a missense variant leading to amino acid exchange in exon 10 of DNAH5, i.e. NM_001369.2(DNAH5):c.1206 T>A; (p.Asn402Lys); rs140782270 (Figure 7A). The allele frequency of T was 100% for all ethnic backgrounds in the Ensembl genome browser with uncertain significance. They were also carriers for a second non-sense heterozygous variant in exon 77 of DNAH5, namely, NM_001369.2(DNAH5):c.13465C>T; (p. Gln4489*), that led to premature termination of translation and the variant having no RS number till date (Figure 7B). Segregation analyses for the two heterozygous variants showed that the heterozygous missense variant c.1206 T>A had a paternal inheritance. The heterozygous non-sense variant c.13465C>T had a maternal inheritance. This confirms that the presence of two heterozygous variants in the two PCD siblings of family KU-9 is most likely responsible for developing the PCD phenotype in the compound heterozygous mode of inheritance (Figures 7A, B). In conclusion, family KU-9 is the first multiplex family in which we failed to detect a shared IBD interval between the two affected individuals; nevertheless, both PCD individuals harboured the same heterozygous pathogenic variants. In conclusion, the variants mapped for DNAH5 were pathogenic and associated with PCD. We also summarise the variants in the three Arabic families from Kuwait in Table 4.

IF and TEM analyses were performed in this study on the nasal biopsies derived from five PCD individuals with different variants of DNAH5 to identify the ultrastructural defects of the axoneme of the multi-ciliated respiratory epithelial cells. IF analyses of the primary airway epithelial cells using monoclonal antibodies showed staining for acetylated α-tubulin, which acts as an axonemal marker protein, in the five PCD individuals and in the healthy control, as seen in panel (a) of Figure 8. However, IF analysis using polyclonal antibodies directed against the ODA heavy chain DNAH5 showed complete lack of staining due to the pathogenic variants of DNAH5, resulting in loss of DNAH5 from the ciliary axoneme, as demonstrated by the absence of the DNAH5 signal in panel (b) of Figure 8 as well as absence of yellow co-staining in panel (c) of Figure 8 compared to the healthy subject who showed DNAH5 staining across the ciliary axoneme. In addition, TEM analyses confirmed the lack of ODAs (Figure 9), consistent with previously published reports (Olbrich et al., 2002; Hornef et al., 2006; Al-Mutairi et al., 2022). These analyses demonstrated that DNAH5 was absent from the ciliary axonemes of the affected individuals, thus supporting our genetic findings.