Background

Hermansky Pudlak syndrome is a rare hereditary multisystem disorder first described in 1959. More than 50% of all worldwide cases are identified in individuals from Puerto Rico where HPS has an estimated frequency of 1:1800 [1, 2]. Clinically the syndrome is characterized by oculocutaneous albinism, a bleeding diathesis due to platelet storage deficiency [1, 2], and other manifestations which may include neutropenia, a granulomatous colitis, or pulmonary fibrosis.

Genotypic analysis over the last decade allowed differentiating at least 10 separate forms of HPS, due to mutations in different genes [3]. All entities have in common defects in intracellular protein trafficking and the biogenesis of lysosome-related organelles like melanosomes or platelet dense granules [2].

Pulmonary fibrosis has not been described in HPS-3 and HPS-5 through HPS-10, which are all very rare. HPS-1 is the most common subtype and characteristically develops a severe and progressive pulmonary fibrosis in almost all cases. Usually middle-aged adults and rarely late adolescents are affected by fibrosis; however, so far, children are not described [2,3,4]. HPS-4 has been documented in less than 10 patients, few of which had pulmonary fibrosis [5, 6].

The HPS-2 subtype is also very rare, with less than 40 cases reported worldwide [4, 7,8,9,10]. HPS-2 is caused by mutations in the AP3B1 gene, inherited in an autosomal recessive way and distinguished from the other forms of HPS by the presence of neutropenia that can lead to severe respiratory infections and that is responsive to granulocyte colony-stimulating factor [1]. Among the few patients described, development of an interstitial lung disease (ILD) has been mentioned in 30 to 50% [1, 7]; details on the pulmonary phenotype have been described in four cases [4]. Potential mechanisms causing pulmonary disease in HPS-2 are poorly understood. It has been suggested that altered AP3B1 gene product within alveolar epithelial type II cells leads to defective intracellular processing of surfactant proteins B and C (SP-B, SP-C), endoplasmic reticulum-stress, apoptosis, and a fibrotic lung phenotype [1].

The aim of this study was to describe the pulmonary phenotype of HPS-2 in children and to further investigate the presence and the possible role of cellular stress and apoptosis in patient-derived material.

Methods

Patients, diagnosis, and follow up

Patients were recruited from the chILD-EU register and biobank and the kids’ lung register collecting diffuse parenchymal lung diseases [11, 12]. Among the children included between 2009 and 2017, seven children were diagnosed with HPS. A 0.4-year old infant with HPS-1 referred for the assessment of potential pulmonary involvement had no pulmonary symptoms and was excluded from this study. All the other cases were HPS-2.

The diagnosis of HPS-2 was based on typical clinical symptoms and proven by genetic analysis (Table 1). Mutational analysis was performed by Sanger sequencing. Routine clinical evaluation in different European centers was performed; data were collected retrospectively and prospectively following the inclusion into the study.

Table 1 Baseline demographics and genetics
Full size table

Lung function testing was done according to standards set previously in children old enough to perform spirometry [13]. CT images of the chest were evaluated for the presence of parenchymal abnormalities (like mosaic attenuation, ground glass opacity, consolidation, linear opacity, septal thickening, reticular opacity, nodular opacity, honeycombing, emphysema, cysts, bleb or bulla) and airway abnormalities (tree-in-bud, bronchiectasis, bronchial wall thickening) on a lobar basis, counting lingula as the separate lobe [14]. Also the presence of pneumothorax, pleural thickening, pleural effusion and enlarged hilar or mediastinal lymph node were evaluated. The image analysis was performed blinded by a pediatric radiologist with expertise in chest imaging.

Bronchoscopy and bronchoalveolar lavage (BAL)

Flexible Bronchoscopy including BAL (mostly of the middle lobe) were performed if clinically indicated using 3 times 1 ml warmed normal saline per kilogram body weight. BAL was examined cytologically and microbiologically.

Lung biopsies and histological investigations

Lung biopsies available were peer-reviewed independently and blinded by a pathologist specialized in pulmonary pathology. Lung tissue of patient 3 was analyzed by Western blotting under reducing and denaturing conditions using sodium dodecyl sulfate–polyacrylamide gel electrophoresis followed by electroblotting and immunostaining for pro-SP-C (Merk Millipore, Darmstadt, Germany), ATF6, β-actin (abcam, Cambridge, UK), and cleaved caspase-3 (Cell Signaling, Gaithersburg, USA). Blotted membranes were developed with the ECL Plus chemiluminescent detection system (Amersham Biosciences, Amersham, UK). Immunohistochemistry was performed on lung tissue fixed in 4% formaldehyde on serial sections with the AP Fast Red kit (Zytochem Systems, Berlin, Germany) after antigen retrieval by microwaving in 10 mM sodium citrate buffer, pH 6.0. Hemalaun was used as counter-stain. Slides from patient 2 were also available for immunostaining for pro-SP-C and cleaved caspase-3, as described above. As controls, lung sections from 3 different organ donor lungs were used.

Ethics, consent and permissions, consent to publish, declarations and statement

Informed consent to report individual patient data was obtained by all patients old enough to consent, and their parents or guardians. The study was approved by the ethics committee of the Ludwig-Maximilian University of Munich (EK 111-13).

All supporting data have been entered in the additional material (Additional file 1).

Result

Clinical course and mutations detected

The six patients included had a mean age at diagnosis of 8.83 years (Additional file 1: Detailed description of the individual cases. Subjects 1 to 6). On average, lung symptoms started 3.3 years before diagnosis (Table 2, Additional file 1: Table S1). Most frequent signs and symptoms reported at follow-up were dyspnea, coughing, need for oxygen, tachypnea, and clubbing. At last follow-up on average 3.17 years (range 2-6) after diagnosis, pulmonary symptoms were noted in all patients. Overall, the frequency of all pulmonary complaints increased over time. Non-pulmonary signs and symptoms are listed in Additional file 1: Table S1.

Table 2 Pulmonary signs and symptoms at diagnosis and last follow-up
Full size table

Patients 1 and 5 were siblings and had the same homozygous frameshift mutation. All other children had different mutations in AP3B1, two of the variations were described before. All other mutations are likely disease-causing mutations, as they are predicted to result in truncated AP3B1 either through the introduction of a premature stop signal, by omission of a regular stop codon or by skipping of an important structural domain for adaptor protein 3 (AP-3 complex) formation. (Table 1, Fig. 1). Although a history of consanguinity could not be obtained in all cases, only one patient was compound heterozygous for the mutations, supporting the autosomal recessive pattern of inheritance (Table 1). Of interest, 5 of 6 patients were girls. Siblings with the same mutation had different clinical courses. However, the small size of this cohort precludes definite conclusions about genotype-phenotype associations in patients with HPS-2.

Fig. 1
figure 1

AP3B1 mutations and their predicted effect on protein structure. Homozygous or compound heterozygous mutations found in individuals with HPS-2 are listed and the expected consequence for protein structure is illustrated. Colored segments represent relevant functional protein domains and regions of altered amino acid sequence after reading frame shift (grey). Genetic nomenclature refers to the respective entries in GenBank, NCBI. (*) NM_003664.4 (**) NG_007268.1 Abbreviations: aa = amino acid, cphet = compound heterozygous, fs = frame shift

Full size image

Complications related to the respiratory tract

In 2 patients scoliosis was noticed at diagnosis, together with psychomotor retardation. Scoliosis developed during follow-up in 2 additional patients. In 3 patients, spontaneous pneumothorax occurred during follow-up. Pleurodesis was necessary for 2 patients (Table 2). In patients old enough to perform lung function testing, a slightly impaired diffusion capacity for carbon monoxide was noted. Initially, spirometry was in the mild abnormal range with combined restrictive- obstructive ventilation disorder, with changing course over time (Table 3, case descriptions in Additional file 1 and Additional file 2).

Table 3 Lung function measurements, lung biopsy results, treatments, and overall outcome
Full size table

Chest imaging

CT scans were performed at the time of diagnosis and at follow-up in subjects 1, 3 and 6 (Fig. 2, Additional file 3: Figure S1, Additional file 4: Figure S2, Additional file 5: Figure S4, Additional file 6: Figure S5, Additional file 7: Figure S6). For subject 4, CT was performed only at the time of diagnosis and for subject 2 only at follow-up. Age at first CT scan was 6.4 years (range 2.3-12) and 12.9 years (range 5.3-15.4) at follow-up. Leading radiographic pattern at diagnosis was ground-glass opacity, which had a patchy distribution, occurring in almost each lobe in all patients. In patient 1 also reticular opacities and traction bronchiectasis were detected at diagnosis. Patient 3 had bronchial wall thickening in the initial CT scan. There was a rapid increase in reticular pattern and traction bronchiectasis between initial and follow-up CT in all subjects with two examinations. Honeycombing and cysts were newly detectable in 3 patients after an observation time of 5.2 years. No nodular opacity, consolidation, tree-in-bud, emphysema, or air trapping were detected (Additional file 1: Table S2, Fig. 2, Additional file 3: Figure S1, Additional file 4: Figure S2, Additional file 5: Figure S4, Additional file 6: Figure S5, Additional file 7: Figure S6). There was no predominance of one side; upper lobes showed the most frequent affection.

Fig. 2
figure 2

CT images of subject 3. a, b, c CT images at 8.7 years of age (column a) and follow-ups at age of 14.5 and 14.8 (column b and c). Leading pattern is GGO on both sides at initial scan (black arrowheads) and slight increase in reticular markings (encircled) and bronchial wall thickening (white arrowheads) at follow-up. Last follow-up showed distinctive pneumothorax and pleural effusion on the left and subpleural blebs (black arrows) in both lung apices

Full size image

BAL and histology

Four of 6 patients had a BAL at diagnosis. One child had very mild eosinophilia and neutrophilia. Patient 6 had severe neutrophilia, despite peripheral blood neutropenia, suggesting compartmentalized capacity to mobilize neutrophils into the alveolar space (Additional file 1: Table S3). No pathogenic bacteria were recovered; patients were not on antibiotic treatment at the time of diagnosis.

Three of the 6 patients received a lung biopsy for diagnosis. Due to bleeding diathesis and low thrombocyte values, thoracoscopic biopsies were performed instead of transbronchial biopsies. In patient 2, the histological pattern of cellular non-specific interstitial pneumonitis (cNSIP), usual interstitial pneumonia- like features (UIP-like), and of desquamative interstitial pneumonitis (DIP) were noted and correlated to the pattern identified on chest CT (Additional file 4: Figure S2). Patient 3 had an UIP-like pattern of patchy dense fibrosis with subpleural cystic areas, pleural fibrosis, and pleural blebs. Histology showed hyperplastic and vacuolated type II pneumocytes and ceroid containing macrophages as typical features of HPS (see Fig. 3a-d). Patient 6 underwent lung biopsy at the age of 6 years. Comparable to patient 2 the histological pattern of cNSIP with areas of DIP were noted. Moreover, lymphofollicular hyperplasia with few lymphoid follicles was detectable. Typical vacuolated type II pneumocytes cells were found ubiquitously but only few ceroid containing macrophages (Additional file 8: Figure S7).

Fig. 3
figure 3

Pulmonary histopathology in subject 3. a-d Histological pattern of patient 3 shows patchy dense interstitial and pleural fibrosis with subpleural cysts/ blebs bordered by normal lung parenchyma at the right margin (c: HE × 10, d: fibrosis and cysts × 20) and small interstitial aggregates of ceroid macrophages with pale brown pigment in the cytoplasm (arrow) as a characteristic histological finding in HPS (D × 400). e Immunohistochemistry identifying alveolar type II cells by staining of proSP-C (SP-C), showing hyperplastic and vacuolated type II pneumocytes as another typical feature of HPS (see also in panel d). Increase in the apoptosis marker cleaved caspase-3 in alveolar epithelial type II cells. Representative images of immunohistochemistry for cleaved caspase-3 and proSP-C performed on serial paraffin sections of HPS-2 and organ donor lungs. Shown here are high magnification images (40×), indicating same type II cells stained for both cleaved caspase-3 and SP-C dying type II cells in HPS-2 patient lungs

Full size image

Treatment

Because of neutropenia and resulting immune deficiency, chronic antibiotic treatment in 4 and Granulocyte-Colony Stimulating Factor, (G-CSF) treatment in 5 patients were initiated. The latter resulted in a lower normal range of neutrophil peripheral blood counts. Pirfenidone was used in 2 patients but was stopped after 3, respectively 24 months, due to gastrointestinal side effects and lack of sufficient clinical improvement (Table 3).

Alveolar epithelial cellular stress and apoptosis

In patient 3 with frozen tissue available obtained from 2 different parts of the lung, additional mechanistic investigations were performed and compared to healthy tissue obtained from unused lung of organ donors. The myofibroblast marker α- smooth muscle actin (SMA) was increased in one specimen supporting increased fibrosis (Fig. 4). Following whole lung tissue analysis, we further analyzed type II pneumocyte apoptosis, an important event in the pathogenesis of interstitial lung diseases. Serial sections revealed several type II pneumocytes to be positive for the apoptosis marker, cleaved caspase 3, indicating their apoptosis (Fig. 3e, Additional file 4: Figure S2H). Two types of cellular stress, the endoplasmic reticulum (ER) stress and autophagy can be differentiated in type II pneumocytes of HPS type 1 patient lungs [15, 16]. We thus analyzed the ER stress markers, GRP78 and p50 form of ATF6, which were markedly increased in one and moderately in the other specimen of the lungs of the HPS-2 patient (Fig. 4). Similarly, the autophagosomal marker microtubule-associated protein 1 light chain-3β, LC3B in its lipidated form (LC3BII) was elevated compared to two of the control lungs, one also gave an increased signal, along with a concomitant increase in the autophagy substrate protein, p62. This proved a defective autophagy pathway in the lung of this HPS-2 patient (Fig. 4a).

Fig. 4
figure 4

Activation of ER-stress and autophagy in subject 3. a Western blot images showing increased expression of the endoplasmic reticulum (ER) chaperone GRP78, ER stress marker AFT6, and importantly the p50 (cleaved form) of ATF6 in HPS-2 patient lungs. In addition, the autophagy marker LC3B (active lipidated form is LC3BII) and SQSTM1/p62 are concomitantly increased, indicating defective autophagy in HPS-2 patient lungs (HPS-2 = patient 3, samples from two different sites of a lung, controls = organ donors). b Representative Western blot images for the myofibroblast marker alpha-SMA and loading control, beta-actin in lung homogenates of patient 3 (HPS-2, duplicate lanes with samples from two different sites of a lung), and 3 different organ donors (controls, one lane each). c Increase in ER stress marker ATF6 in alveolar epithelial type II cells in HPS-2 patient lungs. Representative images of immunohistochemistry for ATF6 and proSP-C (SP-C, alveolar epithelial type II cells marker) performed on serial paraffin sections of HPS-2 and organ donor lungs. Shown here are low and high magnification images, using 5×, 10× and 40× objective for both HPS-2 and donor lung sections. Far right images indicate same alveolar epithelial type II cells stained for both ATF6 and SP-C indicating many alveolar epithelial type II cells positively stained for ATF6. Scale bar: as indicated in the images. d Control sections represent ATF6 and SP-C stainings in donor lung sections, where almost no ATF6 is detected in SP-C positive alveolar epithelial type II cells

Full size image

Discussion

Children suffering from HPS-2 may present with a severe and progressive chronic pulmonary phenotype. Severe lung fibrosis may develop until early adulthood; associated complications including pneumothorax, recurrent infections, and the development of scoliosis were key features identified. Together with few forms of ATP-binding cassette sub-family A member 3 (ABCA3) [17,18,19], SFTPC [20, 21], and MARS deficiency [22], this condition operates under the few clearly and rapidly fibrosing diffuse parenchymal lung diseases in childhood.

Clinically HPS-2 in children is diagnosed by the combination of albinism, bleeding diathesis, and neutropenia. The evolution of symptoms in 4 children started with epistaxis or bleeding diathesis as initial symptoms during the first 2 years of life. However, respiratory symptoms were present in all our children but obviously were rated as too non-specific or developed too insidious to contribute to the diagnosis at age 5 years. At that time, half of our patients already had clubbing, dyspnea, and oxygen demand. Seventeen of the 22 cases under 18 years of age in the largest series on HPS-2 so far had respiratory symptoms, although not further specified [7]. Tachypnea and wet coughing developed somewhat later and suggest secondary chronic bronchitis or suppurative lung disease. The latter may be due to additional immune deficiency from neutropenia in HPS-2 subjects, which may aggravate early respiratory affections in these children. Three of our 6 patients developed pneumothoraces, adding substantially to pulmonary morbidity. Subpleural lung fibrosis, in particular cysts or honeycombing, predisposes to such a complication, which is very unusual in children at this young age. Among 4 children with HPS-2, Gochuico et al. described one child with 6 recurrent pneumothoraces [4].

The natural history of HPS-2 lung disease differs from that of HPS-1 pulmonary fibrosis, which usually affects middle-aged adults and not children [23]. Based on published reports, patients generally first manifest symptoms of HPS pulmonary fibrosis in middle age, however, anecdotal experience includes rare patients with HPS-1 beginning to develop ILD in late adolescence [24]. Carmona-Rivera described a 16-year-old boy with no pulmonary symptoms in HPS-1 [25]. Characteristic pulmonary CT findings known in adults with HPS-1 are increased reticular opacities, thickened interlobular septa, and ground-glass infiltrates in addition to fibrotic changes, including traction bronchiectasis, subpleural cysts and honeycombing. These imaging findings evolve over time, starting in early adulthood, as in HPS-1 patients < 20 years usually no CT changes are noted, in those between 20 and 29 years minimal changes were identified, with increasing symptoms in patients 30 years and older [26, 27]. In HPS-1, high-resolution CT abnormalities inversely correlated with percentage of forced vital capacity and were useful in defining the progression of interstitial disease [27]. In our patients with HPS-2, patchy ground-glass opacity predominated at diagnosis during early childhood and a reticular pattern evolved rapidly over time. At follow-up most of the children developed the CT findings characteristic for HPS fibrosis in adults. Thus, compared to patients with HPS-1, in HPS-2 not only pulmonary symptoms as described above but also CT abnormalities were detected very early. Of interest and in contrast to our observations, two children with HPS-2 have been described in literature before with bilateral ground-glass opacity, thickening of interlobular septa, and interstitial reticulations (4 and 8 vs. 14 years of age) and 3 more children suffering from HPS-2 had changes in HRCT all with improvement over time [4, 9].

In all children, the diagnosis of HPS-2 was verified genetically (Table 1). A pair of siblings had the same mutation (patient 1 and 5), however their clinical course was different. This was also due to the fact that the pulmonary phenotype of the second child was much more and earlier in focus after the other died. A frameshift mutation in the C-terminal region of AP3B1 resulted in loss of the stop codon, prolonging translation into the 3’UTR region. Although an alternative in-frame stop codon is available further downstream, increased distance to the original stop codon may predispose transcripts to nonstop-mediated decay mechanisms [28]. Alternatively, the translated protein product might be subject to proteolytic breakdown due to misfolding, defective assembly, or intrinsic conformational instability [29]. Patient 2 displayed a larger genomic deletion which includes parts of introns 14, 15, and exon 15. This specific region has shown to be essential for correct assembly of the AP3-complex [30]. In all other individuals, single point mutations (patient 3) or frameshift mutations caused by base pair deletions about 10-120 base pairs upstream (patient 4, 6) lead to premature stop codons, translation termination, and possibly activation of the nonsense-mediated decay pathway (Fig. 1).

The pulmonary fibrosis in patients with HPS may be preceded by a macrophage-mediated alveolar inflammation, as BAL fluid contains increased numbers of constitutively activated macrophages [26]. Only one of the subjects with BAL had increased macrophage counts (Additional file 1: Table S3). All our patients had progressively fibrosing lung disease already during childhood. Our histological data prove that pulmonary fibrosis affects children with HPS-2. This is in contrast to HPS-1 and 4, where the development of pulmonary fibrosis starts in middle-age adults and children are only rarely affected [24]. Overall rapidly progressive fibrosing lung disease in childhood is extremely rare. The development of fibrosis might be facilitated by the patients’ neutropenia and natural killer - and T-cell dysfunction and the resulting susceptibility to severe recurrent chest infections. Such an observation is consistent with the HPS-2 animal model where environmental lung injury by silica or bleomycin aggravates fibrosis [26]. Therefore, preventive measures as vaccination and aggressive antibiotic treatment are warranted. Taking our limited observation length into account, the clinical course was not stable in most children. Despite intense symptomatic treatments after diagnosis, we saw deterioration and development of complications in several patients. Four patients remained unchanged, one patient improved, one patient (subject 5), however, died from respiratory insufficiency.

The histology of HPS-2 interstitial lung disease is not widely explored, as the diagnosis may now be done genetically. Patient 2 had a combination of NSIP and UIP-like pattern with dense fibrosis in peribronchiolar and subpleural distribution, as well as a DIP- like areas with intraalveolar aggregates of alveolar macrophages, very similar to other descriptions published [4]. Lung biopsy of patient 3 was dominated by a patchy dense fibrosis with UIP-like pattern comparable to patient 2. Lung biopsy of patient 6 did not show areas of dense fibrosis but demonstrated a cellular NSIP pattern. Additionally, there was a mild lymphoid hyperplasia with few lymphoid follicles with germinal centers that could probably be interpreted as post-infectious changes.

Apoptosis of type II pneumocytes, in addition to ER stress and defective autophagy, was observed in a HPS-2 patient lung. This finding is in line with the previously reported observations of cellular stress and apoptosis of type II pneumocytes in several interstitial lung diseases. More studies are needed to determine if defective autophagy or ER stress underlie type II pneumocyte apoptosis and are subsequently responsible for fibrotic remodeling in the HPS-2 patient lung.

Conclusion

We highlight that patients with HPS-2 in contrast to the other forms of HPS must be considered a severely and rapidly fibrosing lung disease already during early childhood which requires effective antifibrotic treatment. The latter is not yet available in childhood. At diagnosis, half of the subjects had clinical signs of chronic hypoxemia. Histology remains of importance in those extreme rare diseases, which should be made based on the clinical phenotype and confirmed by targeted genetics. During the course, which may be complicated by recurrent pneumothoraces and scoliosis, a comprehensive multidisciplinary team approach is needed. Currently early symptomatic pulmonary care including consequent antibiotic treatment of pulmonary infections and vaccinations. Nutritional management and early orthopedic treatment of scoliosis may help to improve the outcome of children with HPS-2. No specific antifibrotic treatment is available in children.