Primary ciliary dyskinesia (PCD) is a rare disease [1], caused by congenital abnormalities in both structure and function of the motile cilia. It is characterized by upper and lower respiratory tract infections, an increased incidence of situs inversus (SI) and male infertility. It is a genetically heterogeneous and predominantly autosomal recessive disorder.

To date, mutations causing PCD have been described in more than 20 genes. Specific genes are linked to distinct ciliary ultrastructural abnormalities on transmission electron microscopy (TEM): mutations in DNAI1, DNAH5, TXNDC3, DNAI2, DNAL1, CCDC114 and ARMC4 cause PCD with outer dynein arm (ODA) deficiency [2–4], whereas mutations in DNAAF1, DNAAF2, DNAAF3, HEATR2, LRRC6, CCDC103, DYX1C1, ZMYND10, SPAG1 and C21orf59 cause combined inner dynein arm (IDA) and ODA deficiency [2, 5–8]. Absence or displacement of the central microtubular pair (RSPH9, RSPH4A, RSPH1, CCDC39, CCDC40) or absence of the nexin links (CCDC164) have also been linked to specific genetic defects [2, 9]. DNAH11 and recently also HYDIN and CCDC65 mutations are reported in some patients with PCD and normal ultrastructure (NU) [10–12]. The reported prevalence of PCD with NU ranges from 10% to just over 20% of all PCD cases [13, 14]. However, it has previously been shown to be up to 28% in our lab [15]. Nevertheless, DNAH11 mutations are rare in population studies [16] and the frequency of PCD with NU remains under discussion.

The diagnosis of PCD is challenging and cannot rely on signs and symptoms alone [17]. Absence of the ODA and/or IDA or abnormalities of the central and peripheral microtubules can be detected by TEM. This technique was once proposed as the gold standard for diagnosis, but will obviously miss PCD with NU. In PCD with NU, only in vitro functional evaluation of the ciliary motility can identify absent, diminished or uncoordinated movement. However, secondary ciliary dyskinesia (SCD), defined as a non-inherited dysfunction of the cilia due to respiratory infection or exposure to cigarette smoke, can complicate functional evaluation. SCD should be excluded by repeated biopsy or by cell culture [18]. Because de novo regrowth of cilia is induced, SCD can be excluded reliably, increasing specificity to diagnose PCD. Nasal nitric oxide measurements are useful as a screening test for PCD. They are low in most patients with PCD, including those with normal ultrastructure. However, it is not widely available and standardization is lacking.

The phenotypic features of children and adults with PCD have been described previously [19, 20], and attempts have been made to detect a correlation between the ultrastructural abnormality and clinical signs and symptoms [21–26]. However, only few data are available on the characteristics of PCD with NU and little is known about the clinical characteristics of these patients in comparison with those of patients with ‘classic’ PCD [11, 12, 16].

The main aim of our study was to describe clinical characteristics of a large population of patients with PCD, in relation to their ultrastructural defect. We also aimed to demonstrate the need for biopsy and in vitro ciliogenesis to confidently diagnose PCD, especially in the NU subtype. Therefore, we retrospectively compared the results from functional and structural evaluations of the cilia before and after cell culture. Genetic analysis was performed to support the diagnosis of PCD in the subgroup of patients with NU.

We first described the clinical characteristics of patients diagnosed with PCD, confirmed in the KU Leuven Laboratory of Ciliary Function from January 1990 to August 2012, with special focus on patients with the NU subtype. Additionally, we described the laboratory findings before and after cell culture. The study was approved by the Ethical Committee of the University Hospital of Leuven.

The KU Leuven Laboratory of Ciliary Function is the only facility for in-depth evaluation of ciliary structure and function by cell culture in Belgium. Patients were referred from other centers. Biopsy samples were sent for diagnostic evaluation of PCD. Screening tests such as nasal NO measurements were performed in local patients, but not in all patients referred from other centers because of lack of equipment in some referring centers.

Diagnosis of PCD by cell culture, TEM and genetics

The flow chart for the diagnostic algorithm of PCD in our center is shown in Figure 1. All patients included underwent at least one nasal mucosal biopsy for functional and structural evaluation of the respiratory cilia before and after ciliogenesis in culture, even if the evaluation before ciliogenesis was normal. In case insufficient material was available for both evaluations, all material was brought into culture.

Nasal epithelial cells were enzymatically dissociated from nasal biopsy specimens. If cilia were seen, ciliary beat frequency (CBF) and coordination of ciliary activity were evaluated. Ciliary movement was recorded with digital high-speed video imaging using a MotionScope high-speed camera mounted on an inverted Olympus microscope. CBF was calculated with Matlab software (version 6.5) (reference values obtained in our laboratory were CBF 7.9 Hz, SD 1.8 at 23°C). Coordination of ciliary motion was evaluated by directly observing movement of fluid and debris in the medium and rotation or migration of aggregates.

If sufficient material was present, TEM was performed on the initial biopsy sample in order to identify any structural abnormalities. Cross-sections were made in the ciliary shaft, neither towards the base nor towards the tip of the cilium. Per sample at least 50 transversal cross-sections were evaluated. ODA deficiency was diagnosed when < 3 ODAs were visible, partial ODA deficiency when <7 ODAs were observed. IDA deficiency was defined as <1.5 IDA per cilium. Absence or displacement of the central pair was indicative of PCD if present in >1/3 of the cilia. SCD score was calculated as a composite score of membrane abnormalities (blebs, naked cilia, compound cilia), disorganization of microtubules, abnormalities of the central microtubular pair (absent, transposed, eccentric, double) and peripheral microtubular abnormalities (absent, supplementary, dislocated) [29]. SCD score >5% was defined as abnormal.

Every biopsy sample was cultured irrespective of biopsy results [30]. Cells were first grown in a monolayer to expand the basal cell population without cilia (dedifferentiation). Cells were then transferred to a suspension medium to induce redifferentiation into ciliated epithelial cells after 2–3 weeks (ciliogenesis). Ciliary activity (coordination and CBF) and TEM were reevaluated after cell culture. Abnormalities secondary to respiratory infection and toxic substances thus disappeared and SCD and PCD could be clearly distinguished [29]. Coordination of the ciliary activity was subjectively evaluated. Normal coordination implied the observation of rotation of spheroids and movement of debris in the medium due to the beating motion of the cilia. PCD was defined as the absence of coordinated ciliary activity after ciliogenesis in culture, indicating that efficient ciliary transport was absent (see Additional files 1, 2 and 3 for examples of normal and abnormal coordination of ciliary activity). The validity of this evaluation of coordination to distinguish PCD from SCD was replicated by an independent group [31].

If patients consented and if DNA was available, genetic analysis was performed in a subgroup of patients with NU. Exome sequencing was performed to detect mutations in DNAH11 and 29 other candidate genes (Additional file 4) (DNA library preparation kit TruSeq DNA Sample Prep kit v2, Illumina; Exome capturing kit SeqCap EZ Human Exome Library v2.0, Nimblegen; Sequencing kits TruSeq PE Cluster Kit v3-cBot-HS and TruSeq SBS Kit v3-HS, Illumina; Sequencing apparatus Hiseq 2000, Illumina; Bioinformatic analysis CLC Genomics Workbench, CLC Bio). It should be noted that only a ‘partial’ exome was obtained and that any mutation located in a gene not covered by the partial exome, could therefore not be identified. Sanger sequencing, performed by amplifying and sequencing the coding exons and flanking intronic sequences of DNAH11, was performed in most patients to confirm the results of exome sequencing.

Between January 1990 and August 2012, PCD was diagnosed in 206 patients. Ninety patients were diagnosed with PCD with DD (44%)(45 ODA deficiency, 37 partial ODA deficiency and 8 IDA and ODA deficiency). PCD with NU was diagnosed in 68 patients (33%), and PCD with CP in 41 patients (20%) (22 with eccentric central pair and IDA deficiency, 19 with absence of central pair). PCD due to ciliary aplasia (n = 6) and peripheral microtubular abnormalities (n = 1) was rare. Therefore, these patients were excluded for further analysis. Of the remaining 199 patients, no clinical data could be obtained in 31. Data from 168 remaining patients were included for analysis.

We described the clinical characteristics of our total population of patients with PCD, and focused on the subgroup of patients with PCD and NU. In addition, we described the diagnostic pathway needed in order not to miss PCD with NU. Of the 206 Belgian patients with PCD, patients with NU emerged as an important subgroup.

In patients with PCD and NU, clinical characteristics were similar compared to those of patients with PCD and DD or CP. Our analysis confirmed that PCD with NU have severe clinical symptoms: low height for age, chronic cough and sputum production, abnormal lung function from an early age on and high prevalence of bronchiectasis. The prevalence of bronchiectasis and nasal polyps is higher in adults than in children, meaning that the disease is progressive over time.

In our population of PCD with NU, the incidence of SI was significantly lower than 50%, as was the case in PCD with CP. The group of PCD with CP was a combination of patients with absence of the central pair (RSPH4A or RSPH9 mutations) in whom SI did not occur and patients with displacement of the central pair (CCDC39 or CCDC40 mutations) in whom SI was found. Similarly, the group of PCD with NU could be a combination of patients with NU and patients with subtle CP. Indeed, subtle abnormalities in the central pair that cannot be visualized by TEM were recently reported by Olbrich et al. [11], who described HYDIN mutations as a cause of PCD with NU without SI. One patient in our population with NU had 2 HYDIN mutations.

Recent data of the European Taskforce on PCD demonstrated a median age at diagnosis of 5.3 years for patients with a current age below 20 years [33]. When only patients younger than 18 years were included, the age at diagnosis in our center was similar (5.1 years). However, one third of our population (59/168) was only diagnosed in adulthood. This could be an indication that the diagnosis can easily be missed or that patients with a milder form only present towards adulthood.

Our cross-sectional data did not show any difference in FEV₁ or FVC between ages. Remarkably, however, lung function was already found to be abnormal at the age of 5 years. This observation suggests that destruction of the lung parenchyma starts early, as previously suggested [34, 35]. In addition, since complications like bronchiectasis are more frequent in older patients, disease progression is most likely. Further prospective research in a larger patient population is necessary to corroborate these findings.

To demonstrate the need for cell culture to robustly diagnose PCD with NU, we retrospectively analyzed our diagnostic algorithm. We have clearly shown that PCD with NU (and also with CP) can be missed if cell culture is not an integral part of the diagnostic path. To date, a combination of functional evaluation of the ciliary motility and TEM in a fresh biopsy sample has been the gold standard to detect PCD with or without ultrastructural abnormalities. Cell culture was to be reserved to diagnose PCD in selected, difficult cases [36, 37]. We have now shown that ciliogenesis in culture is essential not to miss the subgroup of PCD with NU: in the current population, 16% of the patients with PCD and NU would have been missed (5% of the total PCD population). In a subset of patients, the features of ciliary dysmotility only became apparent after ciliogenesis in culture. Thus, cell culture increases both diagnostic specificity and sensitivity. Moreover, we have identified several new mutations in DNAH11 in patients with NU. This confirms that the culture technique is capable of correctly detecting PCD with NU. Although no functional testing was performed, most mutations were likely to be pathogenic, since most of them were either stop codon or frameshift mutations.

We agree that it is difficult to understand that coordination is normal in a fresh biopsy sample and reverts to abnormal after culture. We hypothesize that some growth factors might be absent in the suspension medium, which alters a hyperfrequent beating phenotype turn into a ciliostatic phenotype, leading to uncoordinated ciliary motion. We are confident that cell death did not occur, because cilia only reappeared on the last few days of suspension culture. Misinterpretation is unlikely, as the result could be confirmed repeatedly. Pifferi et al. have recently shown the same phenomenon of normal evaluation of ciliary coordination in the fresh biopsy and abnormal ciliary coordination after culture in one patient with 2 DNAH11 mutations [31]. Also for patients with CP, we found a high incidence of normal functional evaluation in the fresh biopsy. Because central pair abnormalities are also a sign of SCD, cell culture has an added value in PCD with CP. Therefore, a repeat biopsy with a cell culture procedure must be considered when clinical suspicion of PCD is high and an alternative diagnosis is lacking.

One of the strengths of our study is population size. A group of patients with PCD and NU has also been described by Knowles et al. [16], but this study focused on genetic diagnosis.

This study has some weaknesses. Although we did measure CBF, we did not perform a computerized three-dimensional ciliary beat pattern analysis, proposed as gold standard by some centers [22]. It is impossible to perform this technique on moving spheroids. Therefore, we were not able to link the ultrastructural abnormalities to a specific beat pattern. However, ciliary beat pattern analysis is difficult to interpret and abnormalities may be very subtle [38]. Variability of beat pattern analysis before versus after culture is another concern [38]. We did not systematically include nNO measurement as a diagnostic technique, because it was not available in most of the referring centers. To date however, nNO measurements have not been standardized and need further validation.

We did not detect PCD with isolated IDA deficiency. IDA deficiency is difficult to detect [39] and its existence is debated [40]. Moreover, no gene mutations have been described causing isolated IDA deficiency. It can be hypothesized that some of the IDA deficiencies might be misinterpreted as NU, but there was no significant difference in IDA counts between PCD with NU and control patients without PCD. The fact that we did not find isolated IDA deficiency after cell culture, might confirm that it is part of SCD, as suggested previously [40].

One could question the usefulness of TEM in the diagnosis of PCD, as it does not add much information on disease phenotype. TEM is expensive, time-consuming and difficult to interpret. Additionally, it is insufficient as stand-alone test since some abnormalities can be very subtle. However, if genetic testing becomes routine in the future, TEM morphology may guide genetic investigations. Genetic analysis will aid diagnosis and will allow to examine genotype-phenotype correlations.

We have analyzed the clinical characteristics of a large population with PCD. PCD with NU emerges as a large subgroup and does not differ considerably from other subgroups. We have shown the need for ciliogenesis in culture so as not to miss PCD with NU. The robustness of these diagnoses of PCD with NU is further proven by the consistency of the findings in repeat biopsies as well as by demonstration of mutations by genetic analysis.