Javier Corral, Centro de Hemodonación, Ronda de Garay s/n, 30003, Murcia, Spain. E-mail: firstname.lastname@example.org
Summary. Hermansky–Pudlak syndrome (HPS) is a rare autosomal recessive disorder, which is genetically heterogeneous. In humans, mutations associated with this syndrome have been identified that affect four genes, most of them located in the HPS-1 gene. We evaluated the clinical, molecular, platelet ultrastructure and platelet function data obtained from one Spanish HPS patient and his relatives. The proband was compound heterozygous for a de novo nonsense mutation (Arg-131Stop), which has not been described previously, and for a common frameshift mutation (insC974). These two mutations were also identified by reverse transcription polymerase chain reaction (RT-PCR) in half the RNA, supporting the premise that they have minor effects on either transcription or RNA stability. The patient had an almost complete absence of platelet-dense granules. Accordingly, his platelets showed a small aggregatory response, reduced CD63 surface expression after platelet activation and minor serotonin uptake. Interestingly, despite the absence of clinical symptoms, two relatives carrying only one HPS-1 mutation (insC974) presented a decreased content of platelet-dense granules and showed significant reductions in platelet aggregation, expression of CD63 after platelet activation and serotonin uptake. Data show that the presence of a single mutation affecting one allele of the HPS-1 gene might have relevance in the organogenesis of platelet-dense granules, affecting platelet function. However, these functional defects were not of a great enough magnitude to have clinical significance and, thus, these subjects were clinically asymptomatic.
Hermansky–Pudlak syndrome (HPS) had been considered to be an autosomal recessive disorder, characterized by oculocutaneous albinism, a storage pool deficiency and lysosomal accumulation of ceroid lipofuscin (King et al, 1995). The storage pool defect arises from defects of multiple cytoplasmic organelles, including melanosomes, lysosomes and platelet-dense granules. Dense bodies are scarce or absent in the platelets of HPS patients, reducing the secondary aggregation response (Gahl et al, 1998; Huizing et al, 2001).
The molecular basis of HPS has only recently begun to be unravelled. The synthesis and/or processing of melanosomes, lysosomes and platelet-dense granules are under common and complex genetic control. In the mouse, at least 15 genes are involved in a phenotype resembling human HPS (Novak et al, 2002). The disease is likewise genetically heterogeneous in humans. Four distinct genes have been involved in HPS (HPS-1, HPS-2, HPS-3 and HPS-4) (Anikster et al, 2001; Feng et al, 2002; Suzuki et al, 2002). However, most of the mutations identified in HPS patients are located in the HPS-1 gene. To date, 14 allelic mutations in HPS-1 have been reported (Albinism Database), with two populations displaying a founder effect. A 16-bp duplication within exon 15 has been found exclusively in patients from north-west Puerto Rico (Oh et al, 1998), and the insC974 mutation, apparently the most common HPS-1 gene mutation among non-Puerto Rican patients, has been identified in several families in a region of the Swiss Alps (Oh et al, 1998).
Such molecular heterogeneity has led HPS to be considered to encompass several genetically autosomal recessive disorders, which share the main clinical manifestations of hypopigmentation and a platelet storage pool deficiency as another symptom unequally represented among patients (Huizing et al, 2000). In this context, the study of these diseases, their causative genes and the phenotype associated is of enormous interest to both cell biology and medicine. Importantly, this disease has mainly been studied from a dermatological perspective, and few studies have evaluated the haematological aspects associated with HPS.
Patients and methods
Patients and healthy subjects
We studied a 10-year-old male who presented with severe mental retardation but symptoms suggesting HPS, namely oculocutaneous albinism tyrosinase positive, bleeding diathesis and a prolonged bleeding time. There was no abnormality in coagulation factors, and appropriate coagulation and von Willebrand tests discarded other haemostatic defects. Laboratory studies revealed no abnormalities in blood cell counts, number of platelets or liver function. Electron microscopy of platelets confirmed an almost absence of platelet-dense bodies. His parents reported no consanguinity, and there was no clinical family history. However, we performed a familial study including three generations (Fig 1).
We also included 200 non-related blood donors representative of the general population of our Mediterranean area in the study. Moreover, in all functional and ultrastructural studies, a pertinent control was always tested in parallel.
The patient's relatives and blood donors were fully informed about the aim of this study, which was performed according to the Declaration of Helsinki as amended in Edinburgh 2000.
Genetic analysis of the HPS-1 gene
All coding exons of the HPS-1 gene (exons 3–20) and the flanking intron region were amplified from genomic DNA by the polymerase chain reaction (PCR) (Table I) and purified from 1·5% agarose gels using Ultraclean Gel Spin (MoBio, Solana Beach, CA, USA). The sequence reaction was performed with the ABI Prism Big Dye Terminator Cycle sequencing kit on an automated sequencer type 377 (Perkin-Elmer Applied Biosystems, Warrington, Cheshire, UK) with forward and reverse primers used for amplification. The allele-specific restriction assay or single-strand conformation polymorphism (SSCP) analysis was performed to confirm the mutations identified in this study.
Table I. Oligonucleotides and conditions used for PCRs.
C (25 s), the annealing temperature shown above (1, 50
° C; 2, 52
° C; 3, 70
° C; 4, 58
° C; 5, 63
° C; 6, 60
C) for 15 s and 72
C (30 s), followed by 72
C for 1 min.
Moreover, RNA was isolated from buffy coat using the SV Total RNA isolation system (Promega, Madison, WI, USA). Two pairs of primers were designed in order to perform reverse transcription PCR (RT-PCR) (see Table I). The amplified product was purified and sequenced as indicated above.
Platelet functional studies
Platelet aggregation. Blood samples were collected by venepuncture in trisodium citrate (1:10), and the corresponding platelet-rich plasma (PRP) was adjusted to a cell count of 250 000 platelets/µl. We evaluated the aggregation response of these samples to different doses of distinct platelet agonists: ADP (1 and 10 µmol/l), collagen (1, 5 and 10 µg/ml), epinephrine (1 µmol/l), arachidonic acid (1 mmol/l; Menarini Diagnostics, Florence, Italy) and ristocetin (1 mg/ml; Sigma-Aldrich Química, Madrid, Spain). Changes in light transmission resulting from platelet aggregation were recorded for a total time of 5 min using an Aggrecorder II aggregometer (Menarini Diagnostics).
Flow cytometric analysis of the main platelet glycoproteins (GP). The platelet surface expression of GP Ia, GP IIIa and GP Ibα was analysed after platelet immunofluorescent staining with the specific monoclonal antibodies (mAbs): anti-GP Ia, CD49b conjugated to fluorescein isothiocyanate (*FITC) (clone 10G11 from PeliCluster, Amsterdam, The Netherlands); anti-GP IIIa, CD61*FITC (Becton Dickinson, San Jose, CA, USA); and anti-GP Ibα, CD42*FITC (Becton Dickinson). Parallel immunostaining with an irrelevant isotype, murine IgG*FITC (Becton Dickinson), served as a negative control. These assays were performed in diluted PRP (20 000 platelets/µl) essentially as detailed elsewhere (Corral et al, 2000). Fluorescent flow cytometric analysis was performed using a FACScan analyser (Becton Dickinson). For each sample run, data acquisition of 5000 events was gated on forward and side-angle light scatter with gains adjusted to include the platelet population. Then, the fluorescence of stained platelets was analysed with the cellquest™ software (Becton Dickinson) to obtain both the percentage of positively stained cells, and their mean fluorescence intensity (MFI).
Liberation of platelet granule components. The release and expression on the platelet surface of two proteins contained in platelet granules, CD63 (present in dense granules and lysosomal membranes) and CD62 (stored in the platelet alpha granules), was evaluated by flow cytometric analysis after the activation of platelets with 1 U/ml thrombin, as described previously (Lozano et al, 1997). Briefly, washed platelets (5000 platelets/µl) were activated with thrombin (1 U/ml; Calbiochem, Darmstadt, Germany) or buffer during 15 min at room temperature, and immunolabelled with the mAbs, CD62 conjugated to phycoerythrin (*PE) and CD63*FITC (Becton Dickinson). Analysis of surface expression was performed with cellquest™ software.
Uptake of serotonin. The ability of platelets to take up and store radioactive serotonin ([14C]-5-hydroxytryptamine; [14C]-5HT) in dense granules was tested according to standard procedures (Holmsen & Dangelmaier, 1989).
Platelet ultrastructure. The number of dense granules per platelet was calculated using electron microscopy on whole mounts. Basically, drops of PRP were layered on Formvar-coated grids for 10 s to allow platelets to settle on the surface. The excess of PRP was removed by washing with distilled water, and grids were finally air dried and conveniently stored. These methods have been described in detail elsewhere (Witkop et al, 1978; Pujol-Moix et al, 2000).
Genomic results. Sequence analysis of the HPS-1 gene from the proband revealed two heterozygous point mutations. One of them, located in exon 11, was an insertion of a cysteine at position 974 (insC974) (nucleotide 1 begins at the first nucleotide of codon 1, according to the sequence of Bailin et al, 1997). This alteration is the most common HPS-1 gene mutation among non-Puerto Rican patients and is responsible for a frameshift (Fig 2A). The second point mutation identified was located in exon 5 and consisted of a cysteine to thymine change at position 391 (C391T), resulting in a nonsense mutation affecting arginine 131 (Arg-131Stop). This alteration has not been reported previously. The patient also presented a heterozygous polymorphism in exon 9, responsible for the missense change Gly-283Trp. This polymorphism has been found previously in 4% of the Caucasian population (Albinism Database) and was present in our control population at a frequency of 9·5%.
Familial studies revealed the presence of heterozygous insC974 in the probands of the mother and maternal grandfather (Fig 1). In contrast, the patient's father did not carry the Arg-131Stop point mutation identified in the patient. However, the presence of the Gly-283Trp polymorphism in heterozygous state in his father, and the complete compatibility between the father and the proband when analysing the SE33 and D1S80 markers, assured his paternity.
Thus, the patient appeared to be compound heterozygous, with the maternal allele inherited containing the insC974 mutation and a de novo mutation in exon 5 of the paternal allele.
RNA analysis. Reverse transcription polymerase chain reaction (RT-PCR) analysis confirmed the presence of both mutations in the patient's RNA (Fig 2A and C). The similar intensity of wild-type and mutant alleles suggested that both mutations did not affect transcription rate or RNA stability. Moreover, as expected (Bailin et al, 1997), we observed the previously described alternative splicing of exon 9 in approximately 50% of RNA from the propositus, the parents and two non-related controls (data not shown). These results suggest that this alternative splice, and the absence of exon 9, might play a minor role, if any, in the HPS phenotype.
InsC974 in the general population. As exon 11, especially the region flanking codons 321–324, has been revealed as a mutational hot-spot, we speculated that mutations affecting this region, and especially insC974, might be present in the normal asymptomatic population. In fact, this alteration has been reported as the most frequent in the HPS-1 gene among non-Puerto Rican patients. Thus, we screened the presence of this mutation in 200 controls by SSCP (Fig 2B). However, we did not find this mutation, or other mutations affecting this region, in 400 alleles.
Ultrastructural analysis. In order to evaluate the ultrastructural role of the mutations identified in the HPS-1 gene, we analysed the content and features of platelet-dense bodies by electron microscopy. The proband displayed almost complete absence of these organelles when compared with a normal individual (0·5 granules/platelet versus 5 granules/platelet respectively). His father, carrying only the exon 9 polymorphism, showed an amount equal to that of a control (5·1 granules/platelet). In contrast, his mother, bearing the insC974 in the heterozygous state, presented a significant reduction in the number of platelet-dense bodies (3·5 granules/platelet). Figure 3 shows representative slides of these studies.
Platelet aggregation. Figure 4A summarizes the aggregatory response to different agonists in members of the studied family. The father and maternal grandmother exhibited an aggregation pattern similar to that achieved by controls. Thus, their response to all assayed agonists was considered to be normal. In contrast, the proband showed an aggregation profile that was greatly reduced when his platelets were stimulated with ADP, epinephrine and collagen. Moreover, his platelets failed to respond to arachidonic acid (1 mmol/l). However, the ristocetin-induced agglutination was normal. Interestingly, subjects carrying insC974 in the heterozygous state (mother and maternal grandfather) also showed modifications in the aggregation pattern. Thus, their platelets did not aggregate in response to arachidonic acid and exhibited a significant reduction in the aggregatory response to ADP, collagen and epinephrine. As expected, the ristocetin-induced agglutination of these subjects was normal.
Platelet glycoprotein expression. In order to assess whether the observed differences in platelet aggregation were caused by significant differences in the expression of the main platelet receptors, we tested the level of expression of GP Ibα, GP Ia and GP IIIa on the platelet surface. As shown in Fig 4B, the expression of these glycoproteins was similar in the proband and in his parents.
CD63 and CD62 expression. Thrombin stimulation of platelets leads to the activation and release of platelet granule content. In order to test the content of dense and alpha granules, we evaluated by flow cytometric analysis the platelet release of CD62 and CD63, two platelet granule markers, upon stimulation with 1 U of thrombin. As indicated in Fig 4C, in all cases, secretion of CD62 was normal. However, the proband showed a significant reduction in CD63 release, whereas his mother showed a moderate reduction, and his father presented normal levels of CD63 on the platelet surface.
The proband's platelets had severely reduced [14C]-5HT uptake (36·6%). Similar to other tests affected by the number of platelet-dense bodies, the mother, carrying insC974, showed a moderate reduction in [14C]-5HT uptake (68%), but his father's uptake was comparable with controls (75% versus 73·7% respectively) (Fig 4D).
At least 200 non-Puerto Rican HPS patients are likely to exist throughout the world (Huizing et al, 2000), but this is probably an underestimation, because of a low index of suspicion for this disorder. The disease represents a heterogeneous entity from both a genetic and a clinical point of view (Gahl et al, 1998; Oh et al, 1998; Huizing et al, 2001). The molecular and clinical characterization of new cases of HPS will help in understanding a still largely unknown disease. We performed molecular, structural and functional characterization of a Spanish patient with HPS. The proband was compound heterozygous for a nonsense point mutation (Arg-131Stop) and a common frameshift (insC974), molecular features similar to that observed in two HPS patients described previously: one Italian/German/Ukrainian patient presented insC974 and Glu-133Stop (Shotelersuk et al, 1998), and one Japanese patient reported insG962 and ISV5 G + 5A (Horikawa et al, 2000). Despite having similar molecular defects, these double heterozygous patients displayed different clinical features. The Spanish and Japanese patients showed a severe clinical phenotype, whereas the Italian had mild symptoms. These data support the premise that HPS is a complex and heterogeneous disease, in which additional genetic and environmental factors could influence the severity of the disease caused by identical or similar mutations in the HPS-1 gene.
The Arg-131Stop change is a de novo mutation, which appeared in the paternal allele of the child. It was located adjacent to previously described mutations (Glu-133Stop and ISV5 G + 5A), supporting the idea that this region could be a second mutational hot-spot in the HPS-1 gene. The mutation identified in the maternal allele (insC974) was located in the HPS-1 region containing more mutations among non-Puerto Rican HPS patients, supporting the hypothesis that this region is the major mutational hot-spot of the HPS-1 gene. According to our data, these mutations do not affect the RNA level and, therefore, their pathological consequence could be associated with a smaller and/or different C-terminal end. To date, we do not know whether frameshift or nonsense mutations affecting the HPS-1 gene are associated with the disease by loss or gain of function but, with HPS being a clinical recessive disorder, the loss of function is probably the effect of most mutations affecting this gene. Certainly, all mutations identified result in the loss of the C-terminal end of the protein, a region that contains a poly l-lysine (PLL) stretch, a putative melanosomal localization signal (Jimbow et al, 2000) that has been reported to be critical for optimal protein function (Oh et al, 2000) (Fig 5).
HPS has been considered as a typical autosomal recessive disorder: the presence of at least two mutations affecting both alleles (in the homozygous or compound heterozygous state) is required to develop the clinical phenotype of HPS. Therefore, heterozygous carriers are free from clinical symptoms. Unfortunately, there are few data about the structural or functional features of subjects with a single severe mutation affecting the HPS-1 gene. Here, we identified two relatives of one HPS patient, carrying a single insC974, who presented a significant reduction in the number of platelet-dense granules and consequential impaired platelet function (Figs 3 and 4). Our results suggest that the presence of a single mutation affecting one allele of the HPS-1 gene might have relevance in the organogenesis of platelet-dense granules, affecting platelet function. However, these functional defects were not of a great enough magnitude to have clinical significance and, thus, these subjects were clinically asymptomatic. This fact and the high frequency of mutations located in the main hot-spot of the HPS-1 gene among HPS patients from different origins encouraged us to test for the presence of mutations affecting exon 11 in the normal population by SSCP; however, we found no mutations in 400 alleles. Thus, despite being a hot-spot, the prevalence of mutations affecting this region in the normal population is low.
In conclusion, our study identifies a new mutation in the HPS-1 gene involved in Hermansky–Pudlak syndrome and supports the clinical, ultrastructural and functional heterogeneity of mutations affecting the HPS-1 gene.
We acknowledge the co-operation of Dr N. Pujol-Moix in the analysis of microscopy slides. We thank all staff of the Centro Regional de Hemodonación de Murcia. The authors appreciate the co-operation of all members of the family included in the study, who contributed to our knowledge of HPS. R. González-Conejero and J. Corral are contratados de Investigación Ramón y Cajal, Universidad de Murcia. This project has been supported by grants FIS 01/1249 and Fundación Séneca PI-81/007/79/FS/01.