Platelet count was measured using an automated cell counter (Sysmex KX-21 N, Norderstedt, Germany). Platelet-rich plasma (PRP) and platelet-poor plasma (PPP) were obtained by centrifugation of citrate-anticoagulated blood samples. Using the APACT 4004 (LABiTec, Ahrensburg, Germany), platelet aggregometry analyses were performed after stimulation with collagen (2 and 10 μg/ml; Takeda, Linz, Austria), adenosine diphosphate (ADP; 4 and 10 μmol/L; Sigma-Aldrich, St. Luis, MO, United States), epinephrine (8 and 16 μmol/L; Sanofi-Aventis, Frankfurt, Germany) and ristocetin (1.2 mg/ml; American Biochemical and Pharmaceutical LTD., Frankfurt, Germany).

Flow cytometry analyses were performed using FACSCalibur (Becton Dickinson, Heidelberg, Germany) (Lahav et al., 2002). Diluted PRP aliquots (5 × 10⁷ platelets/ml) were fixed and stained with FITC-labeled monoclonal surface antibody against CD41 (GPIIb/IIIa-complex), CD42a (GPIb/IX) and CD42b (GPIb) (Coulter, Immunotech, Marseille, France). FITC-labeled anti-VWF (Bio-Rad AbD Serorech, Puchheim, Germany) and Alexa Fluor 488-labeled anti-fibrinogen (Invitrogen, Waltham, MA United States) was used to stain the platelets. In the presence of 1.25 mM Gly-Pro-Arg-Pro (Bachem, Bubendorf, Switzerland) diluted PRP (5 × 10⁷ platelets/ml) was stimulated with a number of concentrations of thrombin (0, 0.05, 0.1, 0.2, 0.5, and 1 U/ml; Siemens Healthineers, Marburg, Germany) to conduct the CD62 and CD63 expression analyses. Additionally, the platelets were stained with monoclonal FITC-labeled anti-CD62 (P-selectin) and anti-CD63 antibodies (lysosomal membrane-associated glycoprotein 3, LAMP-3; Immunotech, Marseille, France). Data of patients and controls (day control and 20 independent measurements from 10 controls as mean ± standard error of the mean, SEM) were analyzed using GraphPad Prism software (version 8, San Diego, CA, United States).

Informed consent for molecular genetic analysis was obtained for each patient and the investigated family members. To extract genomic DNA from EDTA blood, we used standard procedures and the Blood and Cell Kit by Qiagen (Qiagen GmbH, Hilden, Germany). For index patients panel sequencing (95 genes including all 11 HPS genes; Supplementary Material S1 gene list) was performed using a custom-designed Nextera Rapid Enrichment Kit (Illumina) followed by sequencing on a MiSeq (Illumina). The average sequencing depth overall genes for the 3 patients investigated was 98% for 20x and 91% for 100x, respectively. SeqPilot (JSI medical systems) was used for data analyses. The variants were exported and filtered by allele frequency and serious consequences. We utilized supporting software ALAMUT^®(v.2.15), pathogenicity prediction (SIFT, MutTaster, PolyPhen2, and CADD), occurrence in population and disease databases (HGMD public version, Huizing HPS Mutation update (Huizing et al., 2020)) in order to classify the variants. These analyses were performed in accordance with the ACMG (American College of Medical Genetics) guidelines (Richards et al., 2015). For segregation analysis Sanger sequencing was conducted.

Total RNA was isolated from washed human platelets using 2 ml TRIzol^® reagent (Thermo Fisher Scientific). Single strand cDNA synthesis was generated with SuperScript^® III Reverse Transcriptase (Thermo Fisher Scientific). Amplification was performed using specific primers (fw: TGC​AGC​AGG​ATT​TCA​CCT​CC; rev: ATC​TGC​TCC​AGC​ATG​TCC​AC) covering the coding region from exon 1 to exon 9 of DTNBP1.

Platelet-rich plasma (PRP) of the patients and a healthy volunteer was obtained by centrifugation of citrate-anticoagulated whole blood (120 × g for 10 min). Platelets were then isolated from PRP by gel-filtration using a Sepharose CL-2B (GE Healthcare Life Sciences) column, eluted with Tyrode buffer (140 mM NaCl, 2.7 mM KCl, 0.42 mM NaH₂PO₄, 12 mM NaHCO₃, 5.5 mM glucose, and 5 mM HEPES; pH 7.4). Platelets were centrifuged at 1,200 × g for 10 min and lysed in lysis buffer (5 mM Tris-HCl, pH 7,4, 50 mM NaCl, 0,25 mM MgCl₂, 0.1% Nonidet P40, 10% glycerol including protease inhibitor cocktail; Roche cOmplete, Merck). Protein content was determined with Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) and adjusted to 10 µg. Denatured platelet lysates were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE, Invitrogen) and blotted onto Hybond-P polyvinylidene difluoride membrane (PVDF, Amersham, GE Healthcare Life Sciences). Membranes were blocked with 5% milk powder in TBST (20 mM Tris, 140 mM NaCl, 0.1% Tween; pH 7.6) probed with anti-dysbindin antibody (dilution 1:2.000; Abcam) and detected using horseradish peroxidase (HRP)-conjugated goat anti-rabbit (dilution 1:10.000; Cell Signaling Technology) and enhanced chemiluminescence solution (Amersham detection reagent, GE Healthcare Life Sciences). For loading control, the same blot was incubated with anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase; dilution 1:300.000; Abcam) and HRP-coupled goat anti-mouse (dilution 1:10.000; Dianova).

NK cell degranulation was analyzed as described (Bryceson et al., 2012). Briefly, fresh NK cell degranulation was assessed by stimulation of isolated peripheral blood mononuclear cells (PBMCs) with K562 target cells followed by flow cytometric analysis of CD107a surface expression. For evaluation of activated NK cell and CTL degranulation, PBMCs were cultured in the presence of Phytohemagglutinin (PHA) and Interleukin-2 (IL-2) for 48 h at 37°C prior to stimulation with K562 target cells and anti-CD3/CD28 beads.

All three index patients presented with normal platelet counts (234 × 10⁹/L, 173 × 10⁹/L, and 189 × 10⁹/L, respectively). The in-vivo bleeding time (Ivy) was severely prolonged (IP1 and IP3 > 15 min, IP2 > 8 min (norm < 6 min)). Platelet aggregation was severely impaired after stimulation with collagen (2 μg/ml) and epinephrine (8 μmol/L) for all index patients. Impaired aggregation after stimulation with low dose ADP (4 μmol/L) was seen in IP1 and IP2. Agglutination after stimulation with ristocetin (1.2 mg/ml) was normal (Table 1).

Flow cytometry analysis revealed severely reduced CD63 expression for all index patients (Figure 1). All of them showed normal values for expression of CD62, CD42a, CD42b, CD41, fibrinogen binding, and VWF-binding (data not shown). IP1’s sister presented with only slightly reduced CD63 expression, while her mother`s expression was borderline low. Flow cytometry was not performed for the father. Six of IP2`s children presented with borderline low and two with a slightly reduced CD63 expression.

In IP1 two novel heterozygous missense variants in HPS3 (NM_032383.3) were identified (c.65C > G; p.Pro22Arg and c.1193G > A; p.Cys398Tyr). Family genotyping confirmed compound heterozygosity. Father and sister of IP1 are heterozygous carriers of the c.65C > G variant, whereas the mother is a carrier of the c.1193G > A variant. Since the index patient lacked apparent cutaneous albinism we confirmed both variants to be germline by genotyping buccal swab DNA. The c.65C > G variant is absent from GnomAD v2.1 (https://gnomad.broadinstitute.org/) and EVS v.0.0.30 (https://evs.gs.washington.edu/EVS/) databases (ACMG Classification: uncertain significance [PM1, PM2]). The variation shows a moderately conserved nucleotide (phyloP: 4.79 [−19.0, 10.9]. In silico pathogenicity prediction (PP) is predominantly disease causing (SIFT, tolerated; MutationTaster, disease causing; PolyPhen2, probably damaging, CADD score 29.1). The second variant c.1193G > A is reported once in the GnomAD database in a heterozygous state and absent from EVS (ACMG Classification: uncertain significance [PM1, PM2, PP3]). In silico PP is concordant disease causing and CADD score is 29.1.

In IP2 a novel homozygous missense variant (c.760G > T; p.Val254Phe) in HPS5 (NM_181507.1) was identified. The patient’s husband showed wild type sequence at this position. NGS and Sanger sequencing revealed that all of her eight children were heterozygous carriers for this HPS5 variant as expected. The variant c.760G > T is reported six times in the GnomAD database in a heterozygous state (all counts in South Asian population) and absent from EVS (ACMG Classification: uncertain significance [PM1, PM2, PP3]). In silico PP is concordant disease causing and CADD score is 28.1. Additionally we identified a variant (NM_018668.5:c.1780A > G) in the gene VPS33B in a heterozygous state. This variant was also detected in the NGS analysis of her daughter. According to OMIM alterations in the gene VPS33P are autosomal recessive associated with ARC (Arthrogryposis, renal dysfunction, and cholestasis) and α-granule deficiency.

IP3 showed a novel homozygous deletion of exon 6 in DTNBP1 (Figure 2B). Exon 6 could not be amplified from genomic DNA using PCR analysis. We performed cDNA sequencing out of platelet derived mRNA and confirmed the absence of exon 6. Interestingly, the small exon 7 was spliced out. It seems that the spliceosome used the acceptor splice site of exon 8. This would result in an in-frame loss of 52 amino acids in the middle of the expected protein (p.H120_A171del). (Figure 2C). To investigate if the aberrant mRNA is leading to detectable dysbindin we performed Western blot analysis. Dysbindin, which is encoded by DTNBP1, was not detectable in platelets (Figure 2D). The antibody used recognizes the C-terminal sequence within human dysbindin (aa 251-301). However, it is possible that the protein will fold incorrectly and in addition protein degradation will occur. Genetic alterations of all patients are summarized in Table 2.

We analyzed the degranulation capacity of NK cells and cytotoxic T lymphocytes (CTL) from IP3 using CD107 degranulation assays. “Fresh” NK cell degranulation in response to K562 target cell stimulation was slightly impaired compared to more than 30 historical controls and additional day controls in two independent experiments (Figure 3A). Following IL-2 stimulation for 48h, the degranulation of patient NK cells increased to normal values. Moreover, degranulation of PHA/IL-2 blasts following stimulation with anti-CD3/28 was in the normal range (Figure 3B).

For IP1 (HPS-3) two compound heterozygous missense variants in HPS3 with predominantly and concordant disease-causing prediction were identified. Both variants are classified as VUS according to ACMG guidelines. The patient presented without cutaneous albinism, however, significantly hypopigmentation of the retina was detected in a specialized ophthalmological investigation. So far, the young girl exhibits mild bleeding symptoms. Taking these findings together these compound heterozygous variants most probably cause the patient’s mild phenotype. Such a mild phenotype has been described also in other patients with HPS-3 (Liu et al., 2021).

IP2 (HPS-5) carries a homozygous missense variant in HPS5 (c.760G > T). The PP is concordant disease-causing. According to the ACMG criteria the variant classifies VUS. The patient does not exhibit apparent albinism, however, a life-long history of bleeding diathesis. None of her 8 children (all heterozygous carriers) are clinically affected as severely as their mother (IP2), although some show very minor bleeding symptoms, such as gingival bleeding and easy bruising. Because the patient lives outside of Germany, a specialized opthalmological investigation has not been performed to date. IP2 suffered from a cystic lesion of her gall bladder, therefore the gall bladder was removed. Clinically, she did not show any symptoms of colitis and the ultrasound investigations did not reveal any signs of colitis. Besides the HPS5 variant, which we detected homozygous in the severely affected mother (IP2), we identified a heterozygous VUS in the gene VPS33B. According to OMIM alterations in the gene VPS33B are autosomal recessive associated with ARC (Arthrogryposis, renal dysfunction, and cholestasis). IP2 and the daughter who we investigated with NGS are carrier of the VPS33B variant; however, they do not show any symptoms of ARC. The son investigated with NGS presented wild type at the position. We did not investigate the other siblings for occurrence of the VPS33B variant. VPS33B is a protein essential for alpha granule biogenesis (Lo et al., 2005). Thrombin induced alpha-granule specific membrane protein P-selectin (CD62) exposure measured by flow cytometry was normal compared to healthy controls. However, a synergistic effect along with the homozygous HPS5 variant, which is also classified as VUS is conceivable. We excluded Chediak-Higashi syndrome an autosomal recessive disease associated with alterations in the gene LYST. One of the features is a partial or severe reduction of dense-granules. LYST is included in our NGS gene panel and the analysis did not show any pathological findings. Inclusions in polymorph nuclear leukocytes have typically been reported in the blood smear of patients with Chediak-Higashi syndrome. IP2 did not show any inclusions in polymorph nuclear leukocytes and no signs of immunodeficiency.

Both HPS3 and HPS5 gene products are part of the BLOC-2 complex and patients primarily present with mild bleeding symptoms (Huizing et al., 2020; Nurden et al., 2020) and not necessarily with noticeable hypopigmentation (Liu et al., 2021). Complications usually did not include pulmonary fibrosis and immunodeficiency. Granulomatous colitis has been reported in about 10–20% of patients (Hussain et al., 2006). Most of the reported variants in HPS3 and HPS5 are variants with serious consequences like deletions, duplications, and variants affecting splicing (Huizing et al., 2020; Liu et al., 2021), only a few missense variants have been described (Huizing et al., 2004; Michaud et al., 2017; Lasseaux et al., 2018). It has also been described that two different nonsense variants in HPS3 lead to different degrees of severities of OCA, suggesting a wide spectrum. One patient presented with a mild and two brothers exhibited a clear OCA (Sandrock-Lang et al., 2017; Lecchi et al., 2020). Both IP1 and IP2 show a mild phenotype, especially concerning the albinism. No further complications, like pulmonary fibrosis or immunodeficiency, were clinically observed in IP1 and IP2, so far.

The homozygous deletion of exon 6 in DTNBP1, found in IP3, has not been previously described. This deletion is pathogenic according to the ACMG guidelines due to the serious consequence. In addition, we showed the absence of dysbindin in platelet lysate. Although this patient has exhibited the characteristic HPS symptoms since birth, the genetic defect was detected rather late in life. Even if the symptoms of OCA are obvious, the diagnosis of HPS can be delayed if the bleeding diathesis is not recognized as part of the disease: the patient was 60 years old at the time when HPS-7 was diagnosed. To our knowledge only seven patients with HPS-7 have been reported worldwide comprising 4 nonsense or frameshift variants in DTNBP1 (Li et al., 2003; Lowe et al., 2013; Bryan et al., 2017; Bastida et al., 2019). Due to the small number of patients, a definite statement regarding additional complications cannot be made. So far, only one patient has been reported developing granulomatous colitis (Lowe et al., 2013). The DTNBP1 encoded Dysbindin-1 is involved in neurotransmission regulation and neurodevelopment (Tang et al., 2009; Ghiani et al., 2010) and seems to be involved in the etiology of Schizophrenia (Cheah et al., 2015; Wang et al., 2017). IP3 showed no symptoms of schizophrenia or other psychiatric disorders, however mild deficits in social interaction or depressive-like emotion are difficult to identify.

IP3 shows a history of severe and recurring airway infections which may possibly hint to immunodeficiency. The most obvious immune defect in lysosomal trafficking disorders is impaired NK cell and CTL cytotoxicity (Fischer et al., 2007). Strong degranulation defects predispose to hemophagocytic lymphohistiocytosis (HLH), a severe disorder of hyperinflammation, while the clinical consequences of milder defects are poorly characterized. A previous report has concluded that CTL cytotoxicity is normal in “sandy” mice (a model of HPS-7) (Bossi et al., 2005), which we could confirm in our patient. However, as previously documented in patients and mice with Chediak-Higashi-Syndrome (Jessen et al., 2011) the fresh NK cell degranulation assay is more sensitive than the CTL assay in detecting subtle impairments of the lytic machinery. Moreover, the degranulation defect could also affect other immune cells such as neutrophils or mast cells (Ramadass and Catz 2016), which could make an additional contribution to the observed infection susceptibility.

Two patients with HPS-9 (subunit of BLOC-1) have been reported presenting immunodeficiency (Badolato et al., 2012; Okamura et al., 2019). This might point to a possible complication in patients with BLOC-1 variants. However, only a few patients with BLOC-1 deficiencies have been described at all. More data are required to understand the role of BLOC-1 proteins in the immune system.