Severe pulmonary disease in an adult primary ciliary dyskinesia population in Brazil

Primary Ciliary Dyskinesia (PCD) is underdiagnosed in Brazil. We enrolled patients from an adult service of Bronchiectasis over a two-year period in a cross-sectional study. The inclusion criteria were laterality disorders (LD), cough with recurrent infections and the exclusion of other causes of bronchiectasis. Patients underwent at least two of the following tests: nasal nitric oxide, ciliary movement and analysis of ciliary immunofluorescence, and genetic tests (31 PCD genes + CFTR gene). The clinical characterization included the PICADAR and bronchiectasis scores, pulmonary function, chronic Pseudomonas aeruginosa (cPA) colonization, exhaled breath condensate (EBC) and mucus rheology (MR). Forty-nine of the 500 patients were diagnosed with definite (42/49), probable (5/49), and clinical (2/49) PCD. Twenty-four patients (24/47) presented bi-allelic pathogenic variants in a total of 31 screened PCD genes. A PICADAR score > 5 was found in 37/49 patients, consanguinity in 27/49, LD in 28/49, and eight PCD sibling groups. FACED diagnosed 23/49 patients with moderate or severe bronchiectasis; FEV1 ≤ 50% in 25/49 patients, eight patients had undergone lung transplantation, four had been lobectomized and cPA+ was determined in 20/49. The EBC and MR were altered in all patients. This adult PCD population was characterized by consanguinity, severe lung impairment, genetic variability, altered EBC and MR.


Results
Demographics and clinical characteristics. Of the 500 patients registered at the Bronchiectasis Outpatient Service over a two-year period, 55 fulfilled the eligibility criteria for PCD assessment, and 49 were diagnosed with PCD. The six patients excluded presented normal nNO and ultrastructure and no laterality disorders. The characteristics of these patients are shown in the Supplementary File (Table S1).
The mean ± SD age of the patients was 41.6 ± 12.9 y, ranging from 21 to 77 y (28 M:21 F). The mean ± SD BMI was 23.8 ± 3.6 kg/m 2 (ranging from 16.8 to 32.8 kg/m 2 ). Furthermore, 45/49 (92%) of the patients were Caucasians and four patients were Afro descendants. Consanguinity was present in 27/49 (55%) of the patients, and there were eight sibling groups in this population (n = 18). The place of birth of the patients was divided between the southeast and northeast regions of the country, but consanguinity was more frequent in patients from the northeast part of the country (31%). Relatives with similar respiratory commitment were described in 28/49 (57%) patients.
Thirty-five PCD patients and all controls underwent CBP evaluation. All the PCD patients presented altered exams, except for two (just one test: movement altered - Table 3).
Non-PCD bronchiectatic patients presented normal ultrastructure evaluations. Among the PCD cases, the brushing nose technique to collect cilia cells was initially successful in 46/49 cases (93.8%) with repetition of a second brush in three patients. In PCD patients, the results were as follows: dynein arm defects or dynein deficiency [DD: 20/49 (41%)], microtubular disorganisation + inner dynein arm (MTD + IDA) [ Table 3, and the complete overview is shown in the Supplementary File (Table S2). The genetic variants with proven pathogenicity screened in patients with phenotypes compatible with PCD are described in the Supplementary File (Table S3), and the variants with uncertain significance are summarized in the Supplementary File (Table S4). Related genotypes screened in patients with phenotypes compatible with primary ciliary dyskinesia are described in the Supplementary File (Table S5).
According to the first results of the iPCD cohort 3 , this patient group was categorized as follows: definite PCD diagnosis in 42/49 (86%) patients, probable PCD in 5/49 (10%) patients, and clinical PCD in 2/49 (4%) patients. Table 3 shows the characteristics of each group. The nNO was 13.3 ± 9.4 nL/min in the definite PCD group; 22.2 ± 11.5 nL/min in the probable PCD group; and 63.1 ± 21.2 nL/min in the clinical PCD group. PCD diagnostic test results are summarized in Fig. 2. Assessment of severity. Twenty-five patients (51%) had severe lung impairment (FEV 1 < 50%), with a FACED score of 3.4 ± 1.3 (range of one to six). At PCD diagnosis, eight patients had previously had lung transplants, four had been submitted to a lobectomy and one was oxygen-dependent. The more frequent ultrastructural abnormalities in these patients were as follows: DD in 11 patients, MTD + IDA in seven patients, ACP in three patients and NU in three patients. www.nature.com/scientificreports www.nature.com/scientificreports/ Inflammation and mucus rheology. EBC pH was collected from 35 patients (35/49, 71.4%), and the EBC pH of patients with PCD was 6.60 ± 0.33, which was less than that of a historical control group of healthy volunteers (7.7 ± 0.49) 14 . The mucus contact angle (53.2 ± 16.4°) was increased in relation to normal values (37 ± 2°) in 35/36 (97.2%) patients. Cough transportability (20.7 ± 5.8 mm) was below reference values (34 ± 9 mm) in 27/36 (75%) patients 15 . Viscosity values were 22.4 ± 7.7 cP, whereas plastic viscosity values were 10.9 ± 7.7 cP (no reference values available) ( Table 4).

Discussion
In this study, we report the diagnostic assessment and the clinical, functional and genetic characteristics of an adult PCD population of 49 individuals in São Paulo, Brazil. Fifty-one percent of the patients presented severe functional impairment in adult life, corroborating the idea that PCD is not a mild disease 12 . This is the first characterization of a group of adult PCD patients in this country.
The newly diagnosed PCD patients in this study represented ten percent of the patients in the Bronchiectasis Outpatient Service of the largest tertiary care hospital in Brazil, which is in keeping with the frequency of other studies 16 , showing that PCD is not as rare as previously thought in Brazil. This population had a mean age of 41.6 ± 12.9 years, and 92% were Caucasians with relatively high rates (55%) of consanguinity, which remains frequent in this country. Situs inversus (50%) was the more frequent laterality disorder, in accordance with previous reports 17 .
The diagnosis of PCD remains challenging since none of the available tests can be used as a stand-alone test 18 . Referral centres differ in the combination of five tests used to assess diagnosis: nNO, HSVM, TEM, IFM and genetic tests 18 . We were able to perform all of these tests on 53% of the patients, and ultrastructure and/or genetic screening were performed on all patients. Few studies have evaluated the five tests in the same patient group 19,20 . Furthermore, differences between the North American 21 and European diagnostic consensus 22 increase the complexity of the diagnostic approach. In our setting, we consider the North American consensus more feasible since it requires fewer replicates of examinations. IFM contributed to diagnosis in 68% of the patients, suggesting that in medical settings where genetic screening is not affordable, this technique could represent a viable alternative. Some patients did not present positive IFM results because of inadequate samples due excess mucus or few cilia 6 .   www.nature.com/scientificreports www.nature.com/scientificreports/ Therefore, we cannot exclude the possibility that the positivity of this technique could have been higher if we had repeated the exams.
Recent data indicate that PCD affects lung function in early life 23 . At the time of PCD diagnosis, 51% of our patients presented severe lung involvement. Eight (16%) patients had undergone lung transplantation at diagnosis, 4/49 (8%) had been lobectomized and 1/49 (2%) was oxygen dependent, confirming disease severity. In addition, a high prevalence of chronic cPa was present in this population, higher than previously reported 10,24 . Although dynein cilia defect was the most frequent abnormality found in the severe patients, the MTD + IDA defect was present in the younger (≤46 y) patients. These findings suggest that the MTD + IDA defect is associated with a more rapid decline in lung function 20 . There are few available data on the effect of early PCD diagnosis on later life lung function 23,25 . Nevertheless, it is highly likely that the lack of early PCD diagnosis and the lack of long-term, adequate and intensive treatment contributed to the disease severity in these patients.
Ultrastructure cilia defect indicated the diagnosis in 81.6% of the patients, and the genetic panel tests identified 49% (24/49) of the patients with the PCD bi-allelic gene. Therefore, genetic tests can only not be used to exclude the diagnosis of PCD 26 . Moreover, we found many variants of uncertain significance that could be associated with our genetically mixed population.
We observed phenotype-genotype inconsistencies in three patients (Table 3-Legend). So far, there has been no clear relationship between ultrastructure, genotypes, and respiratory phenotypes, mainly due to the clinical and genetic heterogeneity of PCD, and some inconsistencies are difficult to explain 27 . Interestingly, one male patient that had cilia ultrastructure compatible with MTD + IDA was hemizygous to a RPGR mutation linked to the X-chromosome. Such a mutation causes retinitis pigmentosa and is rarely associated with respiratory cilia defect 28,29 .
The PCD patients, like those with other chronic inflammatory airway diseases, presented lower EBC pH levels. It is possible that pH and other exhaled compounds could be a non-invasive tool to evaluate PCD treatment 30,31 . Our results indicate that reduced mucociliary transport, chronic inflammation and repeated infections in the respiratory tract produce a thicker mucus in PCD, as shown by the higher contact angle, reduced cough transportability and/or higher viscosity. This study has several limitations. We acknowledge that our study population is small, and any conclusions should be drawn with care. However, this study represents an initial effort to adequately diagnose and fully characterize these patients in Brazil. It is possible that we have included more severe patients with bronchiectasis that were treated in a tertiary care centre. However, the prevalence of bronchiectasis in adult PCD patients seems to be very high 10 . The PICADAR scores were highly variable, with some patients presenting low scores, which could be explained by memory bias. We used an nNO handheld device for screening, which is less accurate 21 . However, in countries with limited resources, such as ours, the recommended chemiluminescence nNO analyser is generally not affordable. There is no standardization of equipment, samples, processing or analysis of cilia movement, and subtle abnormalities in CBP can be difficult to differentiate from secondary dyskinesia. In this study, CBP and nNO were used as accessory tools to strengthen the positive PCD diagnosis 18 .
In conclusion, we diagnosed and described the clinical condition of 49 adult PCD patients who were monitored at a Bronchiectasis service. This population was characterized by high consanguinity levels and severe pulmonary commitment. Genetically, a wide variability of pathogenic variants in genes related to PCD and variants of uncertain significance were found, which is likely to be a reflection of the genetically mixed population of Brazil. PCD is considered an orphan disease, as it has neither the prevalence of asthma nor the lethality of cystic fibrosis. We hope to use this series of patients to contribute to PCD awareness in our country and demonstrate the need for earlier diagnosis.  (Table S5). The complete data regarding the genotypes divided by patient are presented in Supplementary File (Table S2) and Supplementary File (Table S5). Additionally, we included information about the genetic variants with proved pathogenicity in Supplementary File (Table S3) and data regarding the genetic variants with uncertain significance to pathogenicity in Supplementary File (Table S4). The patients Br-6 (X-linked gene), Br-20, Br-22, Br-39, Br-40 and Br-41 (homozygotes to uncertain significance mutations in PCD-causing genes) were considered to have a genetic PCD diagnosis, however, more studies should be performed to reach a better conclusion. Furthermore, the results may contain genetic variants of unknown significance, and a genetic diagnosis may not be clearly established. Thus, genetic counselling is recommended.

Methods
www.nature.com/scientificreports www.nature.com/scientificreports/ 2017. This study was approved by the Ethics Committee of the institution [CAAE: 22823414.8.0000.0068]. All subjects signed written informed consent statements. eligibility. Patients characterized with idiopathic bronchiectasis after a systematic aetiology protocol evaluation were screened for this study. Our institutional protocol includes genetic and/or sweat testing for cystic fibrosis, assessment of gastroesophageal reflux disorder, immunodeficiency (HIV and immunoglobulins) tests, alpha-1 anti-trypsin serum levels, rheumatological antibodies, white blood cell counts and sputum cultures (aerobic, fungi and mycobacteria). Patients were selected to enrol in the PCD diagnosis effort if they presented at least one of the following conditions: laterality disorders or productive chronic cough associated with recurring lower respiratory infections with or without upper respiratory infections and predominance of tomographic findings (bronchiectasis and tree in bud opacities) in the lower, middle and lingula lobes. Patients were submitted to at www.nature.com/scientificreports www.nature.com/scientificreports/  www.nature.com/scientificreports www.nature.com/scientificreports/ least two of the following diagnostic tests: nNO production rate measurement, cilia movement evaluation, ciliary ultrastructure evaluation, IFM and genetic tests.
We also evaluated a group of healthy volunteers for nNO measurements and cilia movement to validate our findings. Individuals were excluded from the control group if they had experienced respiratory symptoms in the previous month and/or had a history of smoking.
Clinical characterization. We collected the following information: demographic variables including age, sex, self-reported race, place of birth, parental consanguinity, symptoms in relatives, body mass index (BMI), and pulmonary function tests. We obtained data on the presence of chronic Pseudomonas aeruginosa colonization (two or more isolates of the same organism at least three months apart in one year), spermiogram, and previous surgery interventions and/or oxygen dependence. Patients were further characterized by PICADAR 32 and FACED scores 33 . In a subset of patients, exhaled breath condensate and mucus rheology were performed. Diagnostic assessment. Nasal nitric oxide production rate (nNO). The nNO production rate measurements were made using a NIOX MINO (AEROCRINE AB ® , Solna, Sweden) device, according to the manufacturer's instructions 34 , in patients who had fasted for at least eight hours and were free from acute respiratory disease [34][35][36] . See the Supplementary File for methodological details. Control groups and patients with no PCD and details regarding the protocol were registered.
Cilia beat frequency and pattern (CBF and CBP) and ciliary ultrastructure. Cilia were collected to study movement and ultrastructure procedures. The detailed method is described in the Supplementary File 13 .
Evaluation of CBF and CBP. Several strips of ciliated epithelium movement were recorded for each patient. The recorded cilia cell movement videos were studied a second time, and the CBF and CBP were classified as follows: recognisability of regular forward and recovery strokes (normal), static cilia, almost static cilia with minimal residual movement, stiff beating due to a reduced bending capacity/amplitude, and abnormal circular beating 37,38 . Cilia movement with no agreement with any previous description but without an effective stroke were considered altered. Only strips of ciliated epithelium without damaged epithelium and no isolated cilia cells were evaluated. CBF was evaluated according to previous studies 13 , and the final results of 10 measurements were recorded. In our study, we assumed CBP to be more important than CBF for evaluating and defining movement 39 . Therefore, if the CBP was altered, we assumed the final result of the movement to be altered.
Analyses of ciliary Ultrastructure by TEM. The collected material was immersed in 2% glutaraldehyde and processed according to standardized norms 40 using cross-section thicknesses of 50 nm. Quantitative and qualitative analyses were conducted. At least 100 cross-sections of cilia were evaluated. High-quality cross-sections were assessed for the presence of dynein arms. Details of the TEM evaluation are described in the Supplementary File [41][42][43] .
Genetics analysis. Detailed methodologies used for the genetic analysis are described in the Supplementary File. DNA extraction was performed using the FlexiGene DNA Kit extraction kit (Qiagen ® , Valencia, CA, 91355, USA). After DNA extraction, the sample was quantified in Qubit 2.0 (Life Technologies ® , São Paulo/SP, Brazil) and then submitted to panel sequencing.
Pulmonary function test (PFT). Spirometry (Koko Legend, Inspire Health Inc., Longmont, USA) was performed according to the recommendations of the American Thoracic Society and the European Respiratory Society. Data were interpreted based on the methods proposed by Pereira and collaborators who examined the Brazilian population 44 . The forced vital capacity (FVC) and FEV 1 were considered PFT parameters.
PCD diagnosis and severity assessment. The possibility of having PCD was considered for patients who had altered results in at least two of the following tests: nNO production rate measurement, evaluation of ciliary movement (CBP), ultrastructure analyses of the cilia (TEM), IFM and genetic tests related to PCD gene sequencing. Patients with confirmed ciliary ultrastructural defect and/or with bi-allelic causing-PCD gene variants were definitively diagnosed with PCD 3,4,39 . Probable PCD was assumed for patients with only one abnormal test, such as altered movement (CBP) and/or low nNO production rate compared to the control the group, but all typical clinical symptoms were present. Patients with negative or ambiguous tests but strong clinical characteristics were defined as a clinical PCD diagnosis 3 .
We considered patients to have severe disease when the forced expiratory volume in one second (FEV 1 ) was < 50% of that predicted.