Hermansky-Pudlak Syndrome: Mutation Update

Abstract

Hermansky–Pudlak syndrome (HPS) is a group of ten autosomal recessive multisystem disorders, each defined by deficiency of a specific gene. HPS-associated genes encode components of four ubiquitously expressed protein complexes: Adaptor Protein (AP)-3 and Biogenesis of Lysosome-related Organelles Complex (BLOC)-1 through −3. All individuals with HPS exhibit albinism and a bleeding diathesis; additional features occur depending on the defective protein complex. Pulmonary fibrosis is associated with AP-3 and BLOC-3 deficiency, immunodeficiency with AP-3 defects, and gastrointestinal symptoms are more prevalent and severe in BLOC-3 deficiency. Therefore, identification of the HPS subtype is valuable for prognosis, clinical management and treatment options. The prevalence of HPS is estimated at 1–9 per 1,000,000. Here we summarize 264 reported and novel variants in ten HPS genes and estimate that ~333 Puerto Rican HPS subjects and ~385 with other ethnicities are reported to date. We provide pathogenicity predictions for missense and splice site variants and list variants with high minor allele frequencies (MAF). Current cellular and clinical aspects of HPS are also summarized. This review can serve as a manifest for molecular diagnostic and genetic counseling aspects of Hermansky-Pudlak syndrome.

INTRODUCTION

In 1959, two Czechoslovakian clinicians, Frantisek Hermansky and Paulus Pudlak, described what is now called Hermansky-Pudlak syndrome (HPS) in two unrelated individuals with a bleeding disorder associated with oculocutaneous albinism (Hermansky & Pudlak, 1959). Over the last six decades, the syndrome has expanded to a disorder with 10 distinct genetic causes, with over 715 cases reported worldwide and a vastly improved understanding of clinical symptoms, pathomechanism and therapeutic approaches (Bowman, Bi-Karchin, Le, & Marks, 2019; Gahl et al., 1998; Huizing, Helip-Wooley, Westbroek, Gunay-Aygun, & Gahl, 2008; Huizing, Malicdan, Gochuico, & Gahl, 2017 Oct 26 [Updated 2000 July 24]).

HPS (MIM# 203300) is a genetically heterogeneous autosomal recessive multisystem disorder characterized by oculocutaneous albinism, a bleeding diathesis, and, in some cases, granulomatous colitis, neutropenia, or a fatal pulmonary fibrosis (Gahl et al., 1998; Huizing et al., 2008; Huizing et al, 2017 Oct 26 [Updated 2000 July 24]). These features result from defects in lysosome-related organelles (LROs), such as melanosomes in melanocytes and delta granules in platelets (Bowman et al., 2019; Dell’Angelica, 2004; Marks, Heijnen, & Raposo, 2013; Raposo, Marks, & Cutler, 2007). The ten described human HPS subtypes (HPS-1 through HPS-10), are each associated with a specific gene defect (Table 1). Orthologs of these ten genes also cause HPS in mice and other animal models (Table 1) (Huizing et al., 2008; Li et al., 2004).

Table 1:

Overview of Hermansky-Pudlak Syndrome Subtypes, (Candidate) Genes and Protein Complexes

HPS Subtype	Gene Name 1 (alternative names)	Protein Complex	Human Locus	mRNA; GeneID 2 (# exons # splice variants)	Protein ID (# amino acids; molecular weight)	Reported Cases	Human Pathogenic Variants	Animal Model 3
HPS-1	HPS1 (BLOC3S1)	BLOC-3	10q24.2	NM_000195; ID: 3257 (20 exons; 18 variants)	NP_000186 (700 aa; 79.3 kD)	~148 4 ~261 (dup16-bp) 5	76 4	pale ear (m)
HPS-2	AP3B1 (ADTB3)	AP-3	5q14.1	NM_003664; ID: 8546 (27 exons; 2 variants)	NP_003655 (1094 aa; 121.3 kD)	~35	29	pearl (m), Ap3b1LN (m)
HPS-3	HPS3 (BLOC2S1)	BLOC-2	3q24	NM_032383; ID: 84343 (17 exons; 2 variants)	NP_115759 (1004 aa; 113.7 kD)	~42 6 ~72 (del 3.9-kb) 7	37 6	cocoa (m)
HPS-4	HPS4 (BLOC3S2)	BLOC-3	22q12.1	NM_022081; ID: 89781 (14 exons; 11 variants)	NP_071364 (708 aa; 76.9 kD)	~41 8	34 8	light ear (m)
HPS-5	HPS5 (BLOC2S2)	BLOC-2	11p15.1	NM_181507; ID: 11234 (23 exons; 3 variants)	NP_852608 (1129 aa; 127.4 kD)	~29	31	ruby-eye-2 (m), snow white (z), casper (sb)
HPS-6	HPS6 (BLOC2S3)	BLOC-2	10q24.32	NM_024747; ID: 79803 (1 exon; 1 variant)	NP_079023 (775 aa; 83.0 kD)	~46 9 ~20 (1065insG) 10	45 9	ruby-eye (m) no privacy (x)
HPS-7	DTNBP1 (BLOC1S8, HPS7)	BLOC-1	6p22.3	NM_032122; ID: 84062 (10 exons; 5 variants)	NP_115498 (351 aa; 39.5 kD)	8	4	sandy (m)
HPS-8	BLOC1S3 (HPS8, RP, BLOS3)	BLOC-1	19q13.32	NM_212550; ID: 388552 (2 exons; 1 variant)	NP_997715 (202 aa; 21.3 kD)	3 6 (448delC) 11	4	reduced pigmentation (m)
HPS-9	BLOC1S6 (HPS9, PLDN, BLOS6)	BLOC-1	15q21.1	NM_001311255; ID: 26258 (5 exons; 3 variants)	NP_001298184 (177 aa; 20.3 kD)	3	2	pallid (m)
HPS-10	AP3D1 (HPS10, ADTD)	AP-3	19p13.3	NM_001261826; ID: 8943 (32 exons; 2 variants)	NP_001248755 (1215 aa; 136.7 kD)	4	2	mocha (m)
-	BLOC1S4 (CNO, BLOS4)	BLOC-1	4p16.1	NM_018366; ID: 55330 (1 exon; 1 variant)	NP_060836 (217 aa; 23.3 kD)	-	-	cappuccino (m)
-	BLOC1S5, MUTED, BLOS5	BLOC-1	6p24.3	NM_201280; ID: 63915 (5 exons; 3 variants)	NP_958437 (187 aa; 21.6 kD)	-	-	muted (m)
-	BLOC1S1, BLOS1	BLOC-1 (BORC) 12	12q13.2	NM_001487; ID: 2647 (4 exons; 1 variant)	NP_001478 (153 aa; 17.2 kD)	-	-	Blos1Ell-Cre/loxp (m) Blos1nestin-Cre/loxp (m) bloc1s1ihb815 (z)
-	BLOC1S2, BLOS2	BLOC-1 (BORC)	10q24.31	NM_173809; ID: 282991 (5 exons; 6 variants)	NP_776170 (142 aa; 16.0 kD)	-	-	Bloc1s2−/− (m) bloc1s2ihb818 (z)
-	BLOC1S7, BLOS7 SNAPIN, SNAPAP	BLOC-1 (BORC) 12	1q21.3	NM_012437; ID: 23557 (4 exons; 1 variant)	NP_36569 (136 aa; 14.9 kD)	-	-	snapin−/− (m)
-	AP3M1, Mu3A	AP-3	10q22.2	NM_207012; ID: 26985 (10 exons; 5 variants)	NP_996895 (418 aa; 46.9 kD)	-	-	-
-	AP3S1, Sigma3A	AP-3	5q22.3-q23.1	NM_001284; ID: 1176 (6 exons; 6 variants)	NP_001275 (193 aa; 21.7 kD)	-	-	-
Total:						385 non-PR 13 333 PR	264 variants

¹The commonly used HPS-subtype-related gene name is grey highlighted, alternative names are listed in brackets.

²Genbank accession numbers of the mRNA encoding the longest isoform (often transcript variant 1), its number of exons, GeneID, and the number of predicted protein-encoding splice variants of each HPS gene. As of November 2019.

³Reported vertebrate HPS animal models: m, mouse; sb, stickleback (gasterosteus aculeatus); x, xenopus tropicalis (frog); z, zebrafish (danio rerio), see (Bowman et al., 2019; Huizing et al., 2008) for details and references. See (Bowman et al., 2019) for invertebrate HPS models.

⁴The ~148 reported non-Puerto Rican HPS-1 cases include 11 novel NIH cohort cases. The 76 reported HPS1 variants include 5 novel variants from the NIH HPS cohort.

⁵The ~261 reported HPS-1 cases with the northwest Puerto Rican founder variant c.1472_1487dup16-bp is an estimate and includes 166 cases from the NIH cohort and ~ 95 cases from the literature (Oh et al., 1996; Oh et al., 1998; Santiago Borrero et al., 2006). The number of worldwide cases homozygous for the HPS1 c.1472_1487dup16-bp variant is estimated to be ~ 400 (Santiago Borrero et al., 2006).

⁶The ~42 reported HPS-3 cases include those reported with the Ashkenazi-Jewish founder variants (7 cases) as well as those novel from the NIH cohort (9 cases). The 37 reported HPS3 variants include 11 novel variants from the NIH cohort.

⁷The ~72 HPS-3 cases with the central Puerto Rican founder 3.9-kb del variant (NM_032383.5:c.-2993_217+692del) include 63 cases from the literature (Anikster et al., 2001; Santiago Borrero et al., 2006) and 9 additional cases from the NIH cohort.

⁸ The ~41 reported HPS-4 cases include 4 novel cases from the NIH cohort. The 34 reported HPS4 variants include 4 novel NIH cohort variants.

⁹The ~46 reported HPS-6 cases include one novel case from the NIH cohort. The 45 reported HPS6 variants include 2 novel variants from the NIH cohort.

¹⁰The ~20 reported HPS-6 cases with the c.1065insG variant are part of an extended Israeli Muslim Bedouin family (Schreyer-Shafir et al., 2006).

¹¹One Pakistani family with 6 affected HPS-8 cases homozygous for c.448delC was reported (Morgan et al., 2006).

¹²BORC = BLOC-one-related complex (Pu et al., 2015).

¹³PR = Puerto Rican.

HPS is a rare disorder with an estimated worldwide prevalence of 1–9 per 1,000,000 individuals (Christensen, Wagner, Coleman, & Appell, 2017; Huizing et al, 2017 Oct 26 [Updated 2000 July 24]). However, the prevalence per subtype can differ due to founder mutations. HPS-1 is more common in Puerto Rico, particularly in the northwestern part of the island where about 1 in 1,800 people are affected and carry the same homozygous mutation (Witkop, Almadovar, Pineiro, & Nunez Babcock, 1990). HPS-1 has also been reported in a small isolate in a Swiss village (Oh et al., 1998; Schallreuter, Frenk, Wolfe, Witkop, & Wood, 1993) and one in Japan (S. Ito et al., 2005). HPS-3 is common in central Puerto Rico, where about 1 in 4,000 individuals are affected (Anikster et al., 2001; Santiago Borrero et al., 2006). Individuals with HPS have been described in many other regions, including China, India, South America and Western Europe (Arcot Sadagopan et al., 2017; Carmona-Rivera et al., 2011; Hermos, Huizing, Kaiser-Kupfer, & Gahl, 2002; Wei et al., 2016).

The protein products of the HPS genes assemble in four multi-subunit complexes, each involved in distinct steps of membrane trafficking and/or component sorting required for LRO biogenesis (Table 1) (Bowman et al., 2019; Dell’Angelica, 2004; Huizing et al., 2008). The adaptor protein-3 (AP-3) complex consist of 4 subunits and includes the protein products of AP3B1, which is mutated in HPS-2 (Dell’Angelica, Shotelersuk, Aguilar, Gahl, & Bonifacino, 1999), and AP3D1, mutated in HPS-10 (Ammann et al., 2016). Biogenesis of Lysosome-related Organelles Complex (BLOC)-3 consists of the HPS1 and HPS4 proteins (Martina, Moriyama, & Bonifacino, 2003), defective in disease subtypes HPS-1 and HPS-4, respectively (Oh et al., 1998; Suzuki et al., 2002). BLOC-2 consists of HPS3, HPS5 and HPS6 (Di Pietro, Falcon-Perez, & Dell’Angelica, 2004), whose defects cause subtypes HPS-3, HPS-5 and HPS-6 (Anikster et al., 2001; Huizing et al., 2001; Q. Zhang et al., 2003). BLOC-1 consists of 8 subunits (Falcon-Perez, Starcevic, Gautam, & Dell’Angelica, 2002; Starcevic & Dell’Angelica, 2004), including DTNBP1, BLOC1S3, and PLDN, defective in HPS-7, HPS-8 and HPS-9, respectively (Badolato et al., 2012; Li et al., 2003; Morgan et al., 2006). No defects in humans are reported in the other five BLOC-1 or two AP-3 subunits, but some are defective in HPS-like animal models (Table 1) (Bowman et al., 2019); suggesting that these subunits are candidates for additional human HPS subtypes. Of note, BLOC-1 shares 3 subunits (BLOC1S1, BLOC1S2 and SNAPIN) with a BLOC-1 related complex (BORC) (Table 1), which has a distinct function and is likely necessary for life (Pu et al., 2015). Defects in either one of these 3 subunits are lethal or very deleterious in mice and are unlikely to be identified in individuals with HPS.

The HPS clinical spectrum is similar in subjects with defects in genes encoding different subunits of the same AP-3 or BLOC complex. Therefore, HPS clinical features and cell biology are best understood in the context of BLOC‐1, BLOC‐2, BLOC‐3 or AP‐3 disease rather than in the context of each individual gene product (Bowman et al., 2019; Huizing et al., 2008).

Identification of the HPS subtype in each subject is important for several reasons. First, it is clinically valuable for prognosis, clinical management, and consideration of eventual treatment options (Table 2). For example, the fatal pulmonary fibrosis occurs in BLOC-3 and AP-3 deficiency, immunodeficiency is associated with AP-3 defects, while BLOC-2 deficiency results in a relatively milder phenotype without life-threatening features. Second, cell biologists can utilize the association of specific HPS defects in cells and tissues from HPS patients, mice and other animal models to study and understand LRO biology. Third, understanding LRO biology provides insights into the pathomechanism of each HPS subtype that may lead to prospects for development of novel therapies not only for HPS, but also for other LRO disorders.

Table 2:

Hermansky-Pudlak Syndrome Main Clinical Features

Clinical Feature	Manifestations	LRO Defect (Cell Type)	Deficiency of HPS Complex	Prevention/Therapy
Cutaneous Albinism	white/light hair, hypopigmented and sun-sensitive skin 1	melanosome (skin melanocytes)	AP-3, BLOC-1, -2, -3	Preventive care: sun avoidance, sun protection (sunscreen, hat, clothing), periodic skin cancer screening
Ocular Albinism	horizontal nystagmus, decreased visual acuity, pale fundus, foveal hypoplasia, iris transillumination 2	melanosome (retinal pigment epithelial cells)	AP-3, BLOC-1, -2, -3	Preventive care: sun avoidance, eye protection (sunglasses, hat) Symptomatic care: vision corrective glasses, ophthalmologic care
Bleeding Diathesis	easy bruising, epistaxis, menorrhagia, gingival bleeding, colonic bleeding, prolonged bleeding after trauma or surgery or postpartum 3	delta granule 3 (platelets)	AP-3, BLOC-1, -2, -3	Symptomatic care: local pressure on wounds, topical thrombin, 1-desamino-8D-arginine vasopressin (DDAVP) and other pro-coagulant drugs Therapeutic: platelet transfusion
Pulmonary Fibrosis (PF)	nonproductive cough, exertional dyspnea, diffuse rales, hypoxia	lamellar body 4 (type II alveolar epithelial cells)	BLOC-3, AP-3 4	Preventive care: avoidance of tobacco products Symptomatic care: supplemental oxygen for hypoxemia, pulmonary rehabilitation Therapeutic: lung transplantation
Enterocolitis	abdominal pain, cramps, fever, weight loss, malabsorption, frequent watery and bloody diarrhea.	unknown LRO-membrane formation 5	BLOC-3, BLOC-2, (BLOC-1) 5	Therapeutic: corticosteroids, non-steroidal immunomodulator drugs, anti-tumor necrosis factor-alpha drugs (effective for only some subjects)
Neutropenia	immunodeficiency	lytic and azurophil granules (neutrophils)	AP-3 (BLOC-1) 6	Therapeutic: granulocyte colony-stimulating factor (G-CSF) 7
Recurrent Infections	frequent viral and bacterial infections	LRO-related granules (dendritic cells, natural killer cells)	AP-3	Therapeutic: Not prevented by G-CSF therapy 7

¹Sunburn, photo-aging of the skin, solar keratosis and melanocyte nevi are common in HPS and patients are at risk of developing squamous cell carcinoma, basal cell carcinoma, and melanoma (Toro et al., 1999).

²HPS visual acuity is generally stable at 20/200 (legally blind in the United States) or worse. Most HPS patients exhibit nystagmus resulting from abnormal crossing of the optic nerve fibers. Iris transillumination is when a light is shone into the pupil is transmitted back through the iris because of a lack of iris pigmentation (Schneier & Fulton, 2013; Summers et al., 1988).

³Absent platelet delta granules (determined by whole mount electron microscopy) is a diagnostic hallmark of HPS. Bleeding tendency varies widely between HPS patients. Due to absent delta granules, a secondary platelet aggregation response cannot occur (Huizing et al, 2017 Oct 26 [Updated 2000 July 24]).

⁴Apart from type II epithelial cell defect, aberrant alveolar macrophage or mast cell function has been suggested to underlie HPS-PF (Kirshenbaum et al., 2016; Mahavadi et al., 2010; Nakatani et al., 2000; Rouhani et al., 2009). Onset of PF is in childhood in AP-3 deficiency (Gochuico et al., 2012) and middle age (30–50 years) in BLOC-3 deficiency (Huizing et al, 2017 Oct 26 [Updated 2000 July 24]). AP-3 related PF has not been described in HPS-10 patients (Ammann et al., 2016; Mohammed et al., 2018). There is no approved medical therapy for HPS PF. Lung transplantation may be considered (El-Chemaly et al., 2018; Gahl et al., 2002; Huizing et al, 2017 Oct 26 [Updated 2000 July 24]; Lederer et al., 2005).

⁵HPS colitis involves intestinal granulomas, erosions and inflammatory cells, and resembles Crohn’s disease. The underlying cause remains unknown. Abnormal endosomal (LRO-related) membrane formation was suggested, leading to ceroid lipofuscin formation, abnormal autophagy and phagocytosis, inflammation (Felipez et al., 2010; Sofia et al., 2017). Some BLOC-2 or BLOC-3 deficient cases develop colitis (Huizing et al, 2017 Oct 26 [Updated 2000 July 24]; Hussain et al., 2006). One BLOC-1 deficient case (HPS-7) developed Crohn’s colitis in adulthood (Lowe et al., 2013). It is unknown if colitis occurs in AP-3 deficiency.

⁶Immunodeficiency was reported in two unrelated individuals with HPS-9 (BLOC-1 deficiency) (Badolato et al., 2012; Okamura et al., 2018) and needs consideration in future BLOC-1 deficient individuals.

⁷G-CSF therapy was only used in HPS-2 patients (AP3B1 deficiency) (Ammann et al., 2016; Fontana et al., 2006). While G-CSF restores neutrophil numbers, it does not prevent recurrent infections caused by defects in innate immunity in HPS-2 (Fontana et al., 2006).

In this report, we provide an overview of the HPS-related genes, their functions and clinical consequences when deficient. We summarize reported human variants in each HPS-associated gene and list unreported variants identified in an HPS patient cohort evaluated at the National Institutes of Health (NIH). This review can serve as a reference for molecular diagnostic aspects of Hermansky-Pudlak syndrome.

HPS SUBTYPE-SPECIFIC MUTATION UPDATE

We searched online literature databases for reported pathogenic variants in 10 HPS-related human genes (as of November 2019). We also list unreported pathogenic gene variants identified in our NIH HPS patient cohort, enrolled in a protocol entitled, “Clinical and Basic Investigations into Hermansky-Pudlak Syndrome” (ClinicalTrials.gov Identifier NTC00001456). Table 1 provides an overview of the HPS subtypes, features of each gene and protein, numbers of reported pathogenic variants, an estimate of reported subjects and reported vertebrate models per subtype. Table 2 lists subtype-specific features, LRO defects, and therapeutic options. Tables 3–8 and Figures 1–7 provide HPS gene-specific pathogenic mRNA and protein variants. Footnotes under each Table describe additional variant-specific information. Pathogenicity predictions of missense and splice site variants are listed in Supp. Tables S1 and S2. Supp. Table S3 lists frequently occurring (mostly missense) variants with a high minor allele frequency (MAF > 0.001) that should be considered as possible polymorphisms when encountered in HPS genetic analyses. Supp. Table S4 lists reported variants in human HPS genes associated with traits other than HPS, which should be considered when these variants are found in future HPS cases.

Table 3:

HPS1 Pathogenic Gene Variants Associated with Hermansky-Pudlak Syndrome Type 1 (HPS-1)

#	mRNA NM_000195.5	Amino Acid NP_000186.2	Exon/Intron	Variant Type 1	Ethnic Background 2	References and Footnotes
1	del exon 2 3	-	Exon 2	Indel	-	(Lasseaux et al., 2018) 4
2	c.2T>A	p.Met1Lys 5	Exon 3	Start-loss	-	(Lasseaux et al., 2018) 4,5
3	c.9delC	p.Cys3Trpfs*26	Exon 3	Frameshift	Chinese	(Power et al., 2019)
4	c.34dupG	p.Glu12Glyfs*12	Exon 3	Frameshift	-	(Lasseaux et al., 2018) 4
5	c.81delG	p.Leu28*	Exon 3	Nonsense	Korean	(Sim et al., 2019)
6	del 121-bp 3,6	p.Pro41Aspfs*12	Intron 3/Exon 4	Indel	Pakistani	(Yousaf et al., 2016) 4,6
7	c.97_100delTCAG	p.Ser33Argfs*18	Exon 3	Indel	English, Irish, German, Scottish	(Sandrock et al., 2010) 7
8	c.166_168delATC	p.Ile56del	Exon 4	Indel	Afghan	(Oh et al., 1998)
9	c.212_215delGCTT	p.Cys71Serfs*52	Exon 4	Indel	-	(Lasseaux et al., 2018) 4
10	c.217delT	p.Ser73Profs*51	Exon 4	Frameshift	-	(Lasseaux et al., 2018) 4
11	c.255+5G>A	IVS4+5G>A (p.Tyr81Leufs*38)	Intron4	Splice site	Iranian	(Ghafouri-Fard et al., 2016)
12	c.288delT	p.Asp97Thrfs*27	Exon 5	Frameshift	Japanese	(Ito et al., 2005; Spritz & Oh, 1999)
13	c.316C>G	p.Arg106Gly 8	Exon 5	Missense	Chinese	(Wei et al., 2016) 4
14	c.344T>C	p.Leu115Pro	Exon 5	Missense	Arabic	(Khan et al., 2016) 4
15	c.355delC	p.His119Thr*5	Exon 5	Frameshift	German, Polish, Russian	(Hermos et al., 2002; Sandrock et al., 2010)
16	c.391C>T	p.Arg131*	Exon 5	Nonsense	Caucasian, Chinese, Spanish	(Arcot Sadagopan et al., 2017; Gonzalez-Conejero et al., 2003; Hermos et al., 2002; Wei et al., 2011)
17	c.397G>T	p.Glu133*	Exon 5	Nonsense	German, Italian, Ukrainian	(Hermos et al., 2002; Shotelersuk et al., 1998)
18	c.398+2T>C	IVS5+2T>C	Intron 5	Splice site	Mexican	novel 4,9,10
19	c.398+5G>A	IVS5+5G>A	Intron 5	Splice site	Chinese, Indian, Japanese	(Furuhashi et al., 2014; Horikawa et al., 2000; Ito et al., 2005; Li et al., 2016; Mai et al., 2019; Natsuga et al., 2005; Oh et al., 1998; Suzuki et al., 2004; Tanaka et al., 2015; Vincent et al., 2009; Wei et al., 2016) 4,11
20	c.418delG	p.Ala140Argfs*35	Exon 6	Frameshift	European	(Hermos et al., 2002)
21	c.461G>A	p.Trp154*	Exon 6	Nonsense	Dutch	(Thielen et al., 2010)
22	c.467_476del10	p.Tyr156Cysfs*16	Exon 6	Indel	Honduran, Salvadoran	(Carmona-Rivera, Golas, et al., 2011)
23	c.505G>A	p.Glu169Lys	Exon 6	Missense- Splice site	Arabic	(Khan et al., 2016) 4,12
24	c.507G>A	p.Glu169Glu 13	Exon 6	Splice site	African-American	(Merideth et al., 2009) 13
25	c.507+1G>A	IVS6+1G>A	Intron 6	Splice site	Japanese	(Lasseaux et al., 2018; Natsuga et al., 2005) 14
26	c.517C>T	p.Arg173*	Exon 7	Nonsense	Chinese	(Wei et al., 2016) 4
27	c.532dupC	p.Gln178Profs*4	Exon 7	Frameshift	Japanese	(Ito et al., 2005; Iwakawa et al., 2005)
28	c.610G>T	p.Glu204*	Exon 7	Nonsense	Spanish	(Sanchez-Guiu et al., 2014)
29	c.640delC	p.His214Thrfs*117	Exon 7	Frameshift	Chinese	(Wei et al., 2019) 4
30	c.695C>T	p.Ala232Val	Exon 8	Missense	Arabic	(Khan et al., 2016) 4
31	c.716T>C	p.Leu239Pro	Exon 8	Missense	Dutch, German, Irish	(Hermos et al., 2002; Lasseaux et al., 2018; Thielen et al., 2010) 4,15
32	c.868–2A>G	IVS9–2A>G	Intron 9	Splice site	Chinese	(Wei et al., 2019) 4
33	c.937G>A	p.Gly313Ser	Exon 10	Missense- Splice site	Puerto Rican	(Carmona-Rivera, Hess, et al., 2011; Lasseaux et al., 2018) 4,16
34	c.956delA	p.Glu319Glyfs*12	Exon 11	Frameshift	Chinese	(Wei et al., 2019) 4
35	c.962delG	p.Gly321Alafs*10	Exon 11	Frameshift	Ukrainian	(Oh et al., 1998)
36	c.962dupG	p.Thr322Hisfs*131	Exon 11	Frameshift	Japanese	(Horikawa et al., 2000)
37	c.972delC	p.Met325Trpfs*6	Exon 11	Frameshift	African-American, Chinese, Japanese, Mexican, Northern European, Puerto Rican	(Carmona-Rivera, Golas, et al., 2011; Carmona-Rivera, Hess, et al., 2011; Hermos et al., 2002; Lasseaux et al., 2018; Merideth et al., 2009; Oh et al., 1996; Oh et al., 1998; Shotelersuk et al., 1998; Wei et al., 2016) 4,17
38	c.972dupC	p.Met325Hisfs*128	Exon 11	Frameshift	Chinese, Japanese, Northern European, Swiss	(Hermos et al., 2002; Lasseaux et al., 2018; Oh et al., 1996; Oh et al., 1998; Okamura et al., 2019; Wei et al., 2010; Wei et al., 2011; Wei et al., 2019) 4,,17,18
39	c.988–1 G>T	IVS11–1G>T	Intron 11	Splice site	Indian	(Vincent et al., 2009) 19
40	del13,966-bp/ins49-bp 3	p.Gln329fs	Intron 11- Exon 20	Indel	Northern European	(Griffin et al., 2005)
41	c.1080C>G	p.Ser360Arg	Exon 12	Missense	Canadian, German, Irish, Scottish, Swedish, Ukrainian	novel 4,9,20
42	c.1132_1138delATCAACC	p.Ile378Trpfs*4	Exon 12	Indel	Chinese	(Wei et al., 2019) 4
43	c.1189delC	p.Gln397Serfs*2	Exon 13	Frameshift	American, Hispanic, Northern European, Russian, Ukrainian	(Doubkova et al., 2019; Griffin et al., 2005; Hermos et al., 2002; Lasseaux et al., 2018; Oh et al., 1998; Sandrock et al., 2010; Shotelersuk et al., 1998) 4,17,20–24
44	c.1228A>T	p.Lys410*	Exon 13	Nonsense	Ukrainian	novel 9,21
45	c.1276_1279dupGGAG	p.Asp427Glyfs*27	Exon 13	Indel	Chinese	(Wei et al., 2019) 4
46	c.1294_1298delATGGAinsT	p.Met432Serfs*42	Exon 13	Indel	Mexican	novel 4,9,10
47	c.1323dupA	p.Gln442Thrfs* 11	Exon 13	Frameshift	Japanese	(Oh et al., 1996)
48	c.1342T>C	p.Trp448Arg	Exon 14	Missense	Pakistani	(Yousaf et al., 2016) 4
49	c.[1375delA; c.1388C>A]	p.Ser459Valfs*16	Exon 14	Frameshift	Northern European	(Hermos et al., 2002)
50	c.1423_1428delAAGCGG	p.Lys475_Arg476del	Exon 15	Indel	-	(Lasseaux et al., 2018) 4
51	c.1457_1460dupTTCT	p.Thr488Serfs*95	Exon 15	Indel	Chinese	(Wei et al., 2016) 4
52	c.1472_1487dup16 24	p.His497Glnfs*90	Exon 15	Indel	NW-Puerto Rican	(Hermos et al., 2002; Oh et al., 1996; Santiago Borrero et al., 2006) 25
53	c.1477delA	p.Arg493Glyfs*22	Exon 15	Frameshift	Chinese	(Power et al., 2019)
54	c.1507C>T	p.Gln503*	Exon 15	Nonsense	Caucasian	(Doubkova et al., 2019) 4
55	del exon 15	deletion	Exon 15	Indel	Chinese	(Wei et al., 2019) 4
56	del exon 15–18 3	deletion	Ex15–18	Indel	Chinese	(Wei et al., 2016) 4
57	c.1639G>T/c.1645C>T	p.Val547Leu/ p.Arg549Cys	Exon 17	Missense	Assyrian, English, German, Irish	(Nazarian et al., 2008) 15,26
58	c.1691delA	p.Lys564Argfs*22	Exon 17	Frameshift	Japanese	(Ito et al., 2005)
59	c.1744–2A>C	IVS17–2A>C	Intron17	Splice site	Caucasian, English, German, Irish	(Hermos et al., 2002; Lasseaux et al., 2018; McElvaney et al., 2018; Oetting & King, 1999) 4,22,27
60	c.1749G>A	p.Trp583*	Exon 18	Nonsense	Japanese, Arabic	(Ito et al., 2005) 4,28
61	c.1763T>C	p.Leu588Pro	Exon 18	Missense	Japanese	(Okamura et al., 2019) 4
62	c.1787G>T	p.Gly596Val	Exon 18	Missense	Japanese	(Okamura et al., 2019; Takeuchi et al., 2014) 4
63	c.1857+2T>C	IVS18+2T>C	Intron 18	Splice site	Irish	(McElvaney et al., 2018) 27
64	c.1858–1G>A	IVS18–1G>A	Intron18	Splice site	Dutch, French, German, Irish, Native American	novel 9,29
65	c.1887delC	p.Val630Serfs*95	Exon 19	Frameshift	Chinese	(Wei et al., 2011)
66	c.1932delC	p.Tyr645Thrfs*80	Exon 19	Frameshift	Chinese	(Wei, Lian, Wang, & Li, 2009; Wei, Zang, Zhang, Yang, & Li, 2015; Wei et al., 2019) 4
67	c.1937A>G	p.Tyr646Cys	Exon 19	Missense	English, Irish, Scottish	novel 9,15,24
68	c.1941–2A>G	IVS19–2A>G	Intron 19	Splice site	Japanese	(Okamura et al., 2019) 4
69	c.1996G>A	p.Glu666Lys	Exon 20	Missense	Korean	(Sim et al., 2019)
70	c.1996G>C	p.Glu666Gln	Exon 20	Missense	-	(Lasseaux et al., 2018) 4
71	c.1996G>T	p.Glu666*	Exon 20	Nonsense	Scottish	(Oh et al., 1998)
72	c.2003T>C	p.Leu668Pro	Exon 20	Missense	Chinese, Japanese	(Ito et al., 2005; Iwata et al., 2017; Kanazu et al., 2014; Mai et al., 2019; Okamura et al., 2019; Wei et al., 2016) 4
73	c.2010_2037del28	p.His671Trpfs*45 29	Exon 20	Indel	-	(Lasseaux et al., 2018) 4,30
74	c.2037_2064del28	p.Leu680Glyfs*36 29	Exon 20	Indel	-	(Girot et al., 2019) 30
75	c.2037_2068delinsCTGG	p.Leu680Trpfs*36 29	Exon 20	Indel	-	(Lasseaux et al., 2018) 4,30
76	c.2056C>T	p.Gln686*	Exon 20	Nonsense	Pakistani	(Yousaf et al., 2016) 4

¹When deletion/insertion is 1 nucleotide it is named Frame shift, when larger it is named Indel.

²Extracted from literature reference. ‘-‘ = unreported.

³The nomenclature of these HPS1 variants are included in this Table as reported, see reference for each specific variant for more details.

⁴At least one of the reported cases with this variant was identified by next generation sequencing.

⁵This variant likely leads to a loss of protein translation at the start codon of the longest splice variant of HPS1 (NM_000195.5). It is also predicted to affect splicing, as it is located at the exon 2–3 splice junction (Supplemental Table S2).

⁶NC_000010.11:g.10:98435762–98435882 (GRCh38): Genomic 121-bp deletion, including a part of intron 3 and exon 4 (Yousaf et al., 2016).

⁷Two unreported siblings from the NIH HPS cohort with this c.97_100delTCAG variant were of English-Irish-Scottish descent.

⁸Gray highlight: missense variant. See Supplemental Table S1 for pathogenicity predictions.

⁹novel = previously unreported variant detected in the NIH HPS cohort.

¹⁰This novel HPS1 variant was found heterozygous by next generation sequencing in 2 unreported siblings of Mexican descent from the HPS cohort. They were compound heterozygous for c.398+2T>C and c.1294_1298delATGGAinsT. This splice site variant is predicted to delete the splice junction of exon 18/intron 18 (Supplemental Table S2).

¹¹This variant was reported to result in skipping of exon 5 (Suzuki et al., 2004), and is a frequent variant in Japanese HPS patients (Ito et al., 2005).

¹²This variant occurs 3-bp from a splice junction and is predicted to affect the splice site (Supplemental Table S2). No experimental evidence is available (Khan et al., 2016). An alternative intronic splice site, inserting 43-bp of intron 6 sequence may be used as reported for variant c.507G>A occurring in the same codon (Merideth et al., 2009).

¹³This (silent) HPS1 variant p.Glu169Glu results in a splice defect (Merideth et al., 2009).

¹⁴This variant is reported to result in use of an alternative intronic splice donor site, 44-bp into intron 6, resulting in a frameshift of the coding region (Natsuga et al., 2005).

¹⁵In vitro studies showed that the HPS1 protein with this missense variant was unstable (Carmona-Rivera et al., 2013).

¹⁶This (missense) HPS1 variant occurs at the 3’ splice junction of exon 10, resulting in a cryptic intronic splice site and an aberrantly spliced mRNA that includes 144-bp intronic sequence, producing 11 novel amino acids followed by a stop codon (Carmona-Rivera, Hess, et al., 2011).

¹⁷This HPS1 frameshift variant occurs with a high prevalence in HPS-1 subjects of various ethnic backgrounds.

¹⁸This variant c.972dupC was reported as an ethnic founder variant in a small isolate in a Swiss village (Oh et al., 1998; Schallreuter et al., 1993)

¹⁹This variant was reported to result in in-frame skipping of exon 12 and removing 56 amino acids from the protein (Vincent et al., 2009).

²⁰This HPS1 variant was identified in one unreported subject of Canadian-German-Irish-Scottish-Swedish-Ukrainian descent from the NIH HPS cohort. This subject is compound heterozygous for c.1080C>G and c.1189delC.

²¹This HPS1 variant was identified in one unreported subject of Ukrainian descent from the NIH HPS cohort. This subject is compound heterozygous for c.1189delC and c.1228A>T.

²²This HPS1 variant was identified homozygous in one unreported subject of German descent from the NIH HPS cohort.

²³This HPS1 variant was identified in one unreported subject of German-English-Irish descent in the NIH HPS cohort. This subject is compound heterozygous for c.1189delC and c.1744–2A>C.

²⁴This HPS1 variant was found heterozygous in one unreported subject of English-Irish-Scottish in the NIH HPS cohort. This subject is compound heterozygous for c.1189delC and c.1937A>G.

²⁵This HPS1 16-bp duplication (c.1472_1487dup16-bp) is originates from a genetic isolate in northwest Puerto Rico (Oh et al., 1996; Santiago Borrero et al., 2006).

²⁶These 2 missense variants occur heterozygous on the same allele in two HPS siblings of our NIH cohort, their cells showed aberrant BLOC-3 assembly (Nazarian et al., 2008). Both missense variants are predicted to be deleterious to protein function (Supplemental Table S3). In vitro studies showed that the HPS1 protein with the p.Val547Leu variant was unstable and prevents proper BLOC-3 formation (Carmona-Rivera et al., 2013). No HPS1 coding/splice site variant was detected on the other allele, but this allele appeared to be subject to non-sense mediated mRNA decay (on cDNA analysis), indicating a likely (intronic) gene-truncation variant on this allele.

²⁷This HPS1 variant was identified in a subject with of Irish descent with HPS clinical features and accelerated pulmonary fibrosis. He was compound heterozygous for c.1744–2A>C (predicted to cause exon skipping (Oetting & King, 1999)) and c.1857+2T>C (predicted to result in use of alternative intronic splice site 4 base-pairs into intron 18, resulting in a frameshift of the coding region) (Supplemental Table S2) (McElvaney et al., 2018).

²⁸This HPS1 variant was identified homozygous by next generation sequencing in one unreported subject of Arabic descent in the NIH HPS cohort.

²⁹This HPS1 variant was found homozygous in one unreported subject of Dutch-French-German-Irish-Native American descent in the NIH HPS cohort. This novel splice site variant c.1858–1G>A, is predicted to create an alternative splice site 1-bp into exon 18, resulting in a frameshift of the coding region (Supplemental Table S2).

³⁰These Indels occur in the same region and result in a loss of the HPS1 termination codon (codon #701) and extension of the translated HPS1 protein.

Table 8:

HPS6 Pathogenic Gene Variants Associated with Hermansky-Pudlak Syndrome Type 6 (HPS-6)

No	mRNA NM_024747.5	Amino Acid NP_079023.2	Exon/Intron 1	Variant Type 2	Ethnic Background 3	References and Footnotes
1	c.62_63insCGGCG	p.Leu22Glyfs*33	Exon 1	Indel	-	(Lasseaux et al., 2018) 4
2	c.60_64dupGCGGC	p.Leu22Argfs*33	Exon 1	Indel	Chinese, Japanese, Portuguese	(Bastida et al., 2019; Okamura et al., 2018; Wei et al., 2016; Wei et al., 2019) 4
3	c.87_108dup22	p.Ser37Leufs*146	Exon 1	Indel	Czech, Eastern/Northern European, German, Polish	(Radke et al., 2013; Summers & Schimmenti, 2014) 5
4	c.141_143delinsG	p.Pro49Trpfs*126	Exon 1	Indel	-	(Lasseaux et al., 2018) 4
5	c.155delT	p.Val52Glufs*6	Exon 1	Frameshift	Chinese	(Wei et al., 2016) 4
6	c.206_210dupGGGCC	p.Trp71Glyfs*158	Exon 1	Indel	Chinese	(Wei et al., 2019) 4
7	c.223C>T	p.Gln75*	Exon 1	Nonsense	Italian	(Huizing et al., 2009)
8	c.233C>G	p.Pro78Arg 6	Exon 1	Missense	Japanese	(Okamura et al., 2019) 4
9	c.238dupG	p.Asp80Glyfs*96	Exon 1	Frameshift	Dutch, German	(Huizing et al., 2009)
10	c.275T>A	p.Leu92Gln 7	Exon 1	Missense	-	(Lasseaux et al., 2018) 4,7
11	c.288G>A	p.Trp96*	Exon 1	Nonsense	Arabic	(Khan et al., 2016) 4
12	c.337C>T	p.Arg113Trp	Exon 1	Missense	-	(Lasseaux et al., 2018) 4
13	c.383T>C	p.Val128Ala	Exon 1	Missense	Caucasian	(Han et al., 2018) 4
14	c.448_505dup58	p.Glu169Glyfs*26	Exon 1	Indel	Caucasian	novel 4,8,9
15	c.455C>G	p.Ser152*	Exon 1	Nonsense	-	(Lasseaux et al., 2018) 4
16	c.503_504delTG	p.Leu168Argfs*7	Exon 1	Indel	Chinese	(Wei et al., 2019) 4
17	c.779G>A	p.Gly260Glu	Exon 1	Missense	Punjabi Afghan	(Hull et al., 2016) 4
18	c.815C>T	p.Thr272Ile	Exon 1	Missense	Dutch, German	(Huizing et al., 2009)
19	c.823C>T	p.Pro275Ser	Exon 1	Missense	Pakistani	(Yousaf et al., 2016) 4
20	c.877C>T	p.Glu293*	Exon 1	Nonsense	-	(Shamseldin et al., 2017) 4
21	c.895C>T	p.Arg299Trp	Exon 1	Missense	Chinese	(Wei et al., 2016) 4
22	c.896G>C	p.Arg299Pro	Exon 1	Missense	-	(Lasseaux et al., 2018) 4
23	c.905G>A	p.Gly302Asp	Exon 1	Missense	-	(Lasseaux et al., 2018) 4
24	c.902dupT	p.Thr303Hisfs*64	Exon 1	Frameshift	Russian-Palestinian	(Hull et al., 2016) 4
25	c.913C>T	p.Gln305*	Exon 1	Nonsense	English, German, Scottish	(Huizing et al., 2009)
26	c.1065dupG 10	p.Leu356Alafs*11	Exon 1	Frameshift	Israeli Bedouin	(Schreyer-Shafir et al., 2006) 10
27	c.1083dupC	p.Gly362Argfs*5	Exon 1	Frameshift	Russian-Palestinian	(Hull et al., 2016) 4
28	c.1114 C>T	p.Arg372*	Exon 1	Nonsense	Irish, Native American (Cherokee), Scottish	(O'Brien et al., 2016)
29	c.1234C>T	p.Gln412*	Exon 1	Nonsense	Italian	(Huizing et al., 2009)
30	c.1235_1239dupAGCGG	p.Arg414Serfs*15	Exon 1	Indel	Chinese	(Wei et al., 2019) 4
31	c.1372delG	p.Glu458Serfs*8	Exon 1	Frameshift	Chinese	(Wei et al., 2016) 4
32	c.1387C>T	p.Arg463*	Exon 1	Nonsense	-	(Lasseaux et al., 2018) 4
33	c.1513C>T	p.Gln505*	Exon 1	Nonsense	Chinese	(Wei et al., 2016; Wei et al., 2019) 4
34	c.1644delA	p.Gly550Glufs*2	Exon 1	Frameshift	Arabic	(Khan et al., 2016) 4
35	c.1711_1712insAG	p.Cys571*	Exon 1	Indel	Czech, Eastern/Northern European, German, Polish	(Radke et al., 2013; Summers & Schimmenti, 2014) 5
36	c.1714_1717delCTGT	p.Leu572Alafs*40	Exon 1	Indel	Belgian	(Lasseaux et al., 2018; Zhang et al., 2003) 4
37	c.1819C>T	p.Arg607*	Exon 1	Nonsense	Chinese	(Lasseaux et al., 2018; Wei et al., 2019) 4
38	c.1865_1866delTG	p.Leu622Argfs*12	Exon 1	Indel	German, Irish	(Huizing et al., 2009)
39	c.1898delC	p.Pro633Leufs*76	Exon 1	Frameshift	Japanese	(Miyamichi et al., 2016) 4
40	c.1919_1920delTC	p.Val640Glyfs*29	Exon 1	Indel	German-Caucasian	(Andres et al., 2017) 4
41	c.2038C>T	p.Gln680*	Exon 1	Nonsense	Japanese	(Miyamichi et al., 2016; Okamura et al., 2018; Okamura et al., 2019) 4, 11
42	c.2189dupC	p.Leu731Serfs*28	Exon 1	Frameshift	Caucasian	novel 4,8,9
43	c.2207T>C	p.Leu736Pro	Exon 1	Missense	-	(Lasseaux et al., 2018) 4
44	del19,972-bp 12	-	Exon 1	Indel	English, German, Scottish	(Huizing et al., 2009)
45	del exon 1 12	-	Exon 1	Indel	-	(Lasseaux et al., 2018) 4

¹The HPS6 gene consists of 1 exon.

²When deletion/insertion is 1 nucleotide it is named Frameshift, when larger it is named Indel.

³Extracted from literature reference. ‘-‘ = unreported.

⁴At least one of the reported individuals with this variant was identified by next generation sequencing.

⁵The subject described in these references was also seen at NIH and is of Eastern/Northern European (Czech, German, Polish) descent. This subject is compound heterozygous for c.87_108dup22-bp and c.1711_1712insAG.

⁶Gray highlight: missense variant. See text and Supplemental Tables for pathogenicity description.

⁷This variant is listed as a variant of uncertain significance with a high MAF in the ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/).

⁸novel = previously unreported variant and/or ethnicity, detected in the NIH HPS cohort.

⁹This HPS6 variant was identified by exome sequencing in one unreported subject of Caucasian descent in the NIH HPS cohort. This subject is compound heterozygous for c.448_505dup58-bp and c.2189dupC.

¹⁰This frameshift variant was previously described as c.1066_1067insG (p.Leu356Argfs*11) (Schreyer-Shafir et al., 2006).

¹¹This HPS6 variant c.2038C>T (p.Gln680*) appears to be a Japanese variant, as it occurs in 5 Japanese subjects and is not reported in dbSNP/ExAc databases.

¹²The nomenclature of these HPS6 variants are included in this Table as reported, see reference for each specific variant for more details.

Figure 1:

Distribution of HPS1 Gene Variants

Figure 7:

Distribution of BLOC1S8 (HPS-7), BLOC1S3 (HPS-8), BLOC1S6 (HPS-9), and AP3D1 (HPS-10) Gene Variants

The variant nomenclature in all Tables conforms to human genome variation society (HGVS) recommendations (den Dunnen et al., 2016). The longest mRNA splice variant of each gene is used for variant nomenclature and the GenBank accession number is indicated in each Table. Some previously reported variants are re-named to conform to the current nomenclature convention. Pathogenicity predictions of missense variants (Supp. Table S1) follow the American College of Medical Genetics (ACMG) Standards and Guidelines for interpretation of sequence variants (Richards et al., 2015).

We deposited all unreported variants in the Leiden Open Variation Database 3.0 (http://www.lovd.nl/) (Fokkema et al., 2011) and in ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/) (Landrum et al., 2018). Other databases with variable HPS-related information exist, including the Albinism Database (http://www.ifpcs.org/albinism/), AP3B1base (http://structure.bmc.lu.se/idbase/AP3B1base/), Retina International Mutation Database (http://www.retina-international.org/sci-news/databases/mutation-database) and Oculocutaneous albinism Database (https://ghr.nlm.nih.gov/condition/oculocutaneous-albinism).

HPS-1 (Table 3)

HPS1, also called BLOC3S1, is the first identified HPS-associated gene (Oh et al., 1996), causing HPS Type 1 (HPS-1; MIM# 203300) when defective. HPS1 is located on chromosome 10q24.2; its longest mRNA transcript contains 20 exons and codes for a 700-amino acids protein (~79.3-kD). At least 18 HPS1 protein-coding mRNA transcript variants are predicted; expression patterns and function of these variants remain unknown. The HPS1 protein interacts with HPS4 in the BLOC-3 complex (Carmona-Rivera, Simeonov, Cardillo, Gahl, & Cadilla, 2013; Martina et al., 2003). Pale ear is the murine orthologue of human HPS-1 (Li et al., 2004).

Identification of the HPS1 gene was aided by linkage analysis in northwest Puerto Rican individuals with HPS, who have a homozygous 16-bp duplication (c.1472_1487dup16-bp, p. p.His497Glnfs*90) in HPS1 (Fukai, Oh, Frenk, Almodovar, & Spritz, 1995; Oh et al., 1996). There are currently ~261 Puerto Rican subjects with the HPS1 16-bp duplication founder variant reported in the literature (including the NIH cohort); it is estimated that ~400 such cases exist (Santiago Borrero et al., 2006; Witkop et al., 1990). HPS-1 has also been reported in a small isolate in a Swiss village (c.972dupC, p.Met325Hisfs*128) (Oh et al., 1998; Schallreuter et al., 1993) and one in Japan (c.398+5G>A) (Ito et al., 2005). Apart from the Puerto Rican founder population, there are an additional ~137 HPS-1 cases reported, plus 11 unreported cases from our NIH cohort (Table 1).

We report a total of 76 HPS1 variants associated with the HPS-1 phenotype (Table 3), including 5 novel variants identified in our NIH HPS cohort. The variants are located throughout the entire gene and include 1 (1%) start-loss, 20 (26%) frameshift, 15 (20%) missense, 11 (14.5%) nonsense, 18 (24%) insertions and/or deletions, and 11 (14.5%) splice site variants (Figure 1). All reported HPS1 splice site variants are predicted and/or reported to cause aberrant splicing (Supp. Table S2). Of the 15 reported HPS1 missense variants, 5 occur at an intron/exon boundary and 3 are predicted and/or reported to affect splicing (Supp. Table S2). Of note, one nonsynonymous (silent) variant, p.Glu169Glu (c.507G>A), was reported to result in a splicing defect in two African-American brothers (Merideth et al., 2009). The start-loss variant p.Met1Lys (c.2T>A) likely leads to a loss of protein translation at the start codon of the longest splice variant of HPS1 (NM_000195.5), it is also predicted to affect splicing, as it is located at the exon 2–3 splice site (Supp. Table S2). The MAFs of all HPS1 pathogenic missense variants are very low or not reported, supporting possible pathogenicity (Supp. Table S1). Pathogenicity of some HPS1 missense variants was experimentally assessed (Supp. Table S1). Of the 15 reported HPS1 missense variants, 7 were predicted to be pathogenic (P) or likely pathogenic (LP) by ACMG standards and guidelines (Richards et al., 2015) (Supp. Table S1), while 8 others were classified as variants of uncertain significance (VUS), which should be taken into account when these variants are found in future HPS cases.

Four HPS1 variants (c.1286G>A, p.Arg429His; c.1395G>A, p.Trp465*; c.1888G>A^, p.Val630Ile; c.1915G>A, p.Gly639Ser) listed in Supp. Table S4 were found heterozygous in nextgen sequencing studies of probands with other traits than HPS (Stearman et al., 2019; Abouelhoda et al., 2016). Although none of these variants were reported in HPS subjects, they should be taken into account when these variants are found in future HPS cases.

The dbSNP/gnomAd/ClinVar databases list 10 frequently occurring (MAF > 0.001) HPS1 missense variants, with predicted benign or unknown pathogenicity, which should be considered as likely non-pathogenic polymorphisms when encountered in HPS1 genetic analyses (Supp. Table S3).

There are 3 HPS1 frameshift variants that occur with high prevalence in HPS-1 subjects of various ethnic backgrounds and have relatively high MAFs; they are c.972dupC, p.Met325Hisfs*128 (ClinVar MAF 0.000317; frequent in Europeans) and c.972delC, p.Met325Trpfs*6 (ClinVar MAF 0.00002; frequent in Europeans and South Asians) in exon 11 and c.1189delC, p.Gln397Serfs*2 (ClinVar MAF 0.000067; frequent in Europeans) in exon 13. In fact, in our NIH cohort of non-Puerto Rican HPS-1 cases, 19 (39%) of 49 cases carry at least one of these two variants. Hence, analyses of exon 11 and exon 13 of HPS1 could be considered before proceeding to more laborious and costly sequencing techniques in non-Puerto Rican individuals suspected of having HPS-1 disease.

HPS-1 is identified worldwide in individuals with a large spectrum of ethnic backgrounds and is the HPS subtype with the most described cases, even excluding Puerto Rican cases (Table 1). HPS-1 (together with HPS-4) displays the most severe phenotype. In individuals with HPS-1, cutaneous albinism is more profound (higher degree of hypopigmentation of skin and hair) and the ocular findings are more severe than in other subtypes (Huizing et al, 2017 Oct 26 [Updated 2000 July 24]). Of note, Hps1 (and Hps4) variants in mice appear to only have a mild effect on pigmentation and bleeding (Novak, Hui, & Swank, 1984), for unknown reasons. In mice, Hps1 variants appear to impact pigmentation tissue-specific; melanosomes in hair follicles are less affected (i.e. pigmented hair) than in interfollicular melanocytes (i.e. less pigmented skin), perhaps partially explaining the more severe skin pigmentation phenotype in human patients (Nguyen & Wei, 2007). Virtually all HPS-1 subjects develop pulmonary fibrosis by middle age, some develop granulomatous colitis and a majority of female subjects have menorrhagia (Table 2).

HPS-2 (Table 4)

Table 4:

AP3B1 Pathogenic Gene Variants Associated with Hermansky-Pudlak Syndrome Type 2 (HPS-2)

#	mRNA NM_003664.4	Amino Acid NP_003655.3	Exon/Intron	Variant Type 1	Ethnic Background 2	References and Footnotes
1	c.2T>G	p.Met1Arg 3	Exon 1	Start-loss	Australian	(Cetica et al., 2015)
2	c.62delG	p.Gly21Valfs*20	Exon 1	Frameshift	-	(Jessen et al., 2013)
3	c.155_158delAGAG	p.Glu52Alafs*11	Exon 2	Indel	Caucasian, English	(Wenham et al., 2010)
4	c.177delA	p.Lys59Asnfs*5	Exon 2	Frameshift	-	(de Boer et al., 2017)
5	c.305T>C	p.Leu102Pro 4	Exon 4	Missense	-	(Jessen et al., 2013)
6	c.716G>A	p.Trp239*	Exon 7	Nonsense	Moroccan	(de Boer et al., 2017)
7	c.904A>T	p.Arg302*	Exon 8	Nonsense	-	(Enders et al., 2006)
8	c.1063_1064delCAinsTATCAATATC	p.Gln355Tyrfs*6	Exon 10	Indel	Italian	(Fontana et al., 2006)
9	c.1095+5G>A	IVS10+5G>A	Intron10	Splice site	Mexican	(Chiang et al., 2010)
10	c.1168–1G>C 6	IVS11–1G>C	Intron11	Splice site	Dutch	(Dell'Angelica et al., 1999; Gochuico et al., 2012) 5,6
11	c.1473+6T>C	IVS14+6T>C	Intron14	Splice site	-	(Clark et al., 2003)
12	c.1525C>T	p.Arg509*	Exon 15	Nonsense	Cajun, Houma Indian	(Huizing et al., 2002)
13	c.1619dupG	p.Ala541Serfs*25	Exon 15	Frameshift	-	(Clark et al., 2003)
14	g.del8168-bp 7	del exon 15	Introns14+15, Exon 15	Indel	Turkish	(Jung et al., 2006) 7
15	g.del1872-bp 7	del exon 15	Introns14+15, Exon 15	Indel	-	(Hengst et al., 2018) 7
16	del exon 16	-	Exon 16	Indel	-	(Jessen et al., 2013)
17	c.1739T>G	p.Leu580Arg	Exon 16	Missense	Dutch	(Dell'Angelica et al., 1999)
18	c.1754delT	p.Val585Glufs*6	Exon 16	Frameshift	Caucasian	(de Boer et al., 2017)
19	c.1789dupA	p.Ile597Asnfs*12	Exon 16	Frameshift	Italian	(Fontana et al., 2006)
20	c.1839_1842delTAGA	p.Asp613Glufs*38	Exon 17	Indel	-	(de Boer et al., 2017; Hengst et al., 2018; Jung et al., 2006)
21	c.1975G>T	p.Glu659*	Exon 18	Nonsense	Cajun, Houma Indian	(Huizing et al., 2002)
22	c.2041G>T	p.Glu681*	Exon 18	Nonsense	-	(Ammann et al., 2017; Jessen et al., 2013)
23	g.del624-bp 8 c.del2077_2164	p.Glu693Valfs*13	Intron18, Exon 19	Indel	Maltese	(Wenham et al., 2010)
24	c.2546T>G	p.Leu849*	Exon 22	Nonsense	-	(Hengst et al., 2018)
25	c.2702C>G 9	p.Ser901Cys	Exon 23	Missense Splice site	Caucasian	(de Boer et al., 2017) 9
26	c.2770delC	p.Leu924Phefs*3	Exon 23	Frameshift	-	(Jessen et al., 2013)
27	c.2944delC	p.Leu982Cysfs*19	Exon 25	Frameshift	-	(Hengst et al., 2018)
28	c.3222_3223delTG 10	p.Lys1076Asnfs*60	Exon 27	Indel-Stop-loss	United Arabic Emirates	(Hengst et al., 2018; Jessen et al., 2013; Kurnik et al., 2013) 10
29	inv(5)p15.1-q14.1 11	-	-	Indel-chrom. inversion	Lebanese	(Jones et al., 2013) 11

¹When deletion/insertion is 1 nucleotide it is named Frame shift, when larger it is named Indel.

²Extracted from literature reference. ‘-‘ = unreported.

³This variant likely leads to a loss of protein translation in the start codon of the longest splice variant of AP3B1 (NM_003664.4).

⁴Gray highlight: missense variant. See Supplemental Table S1 for pathogenicity predictions.

⁵At least one of the reported cases with this variant was identified by next generation sequencing.

⁶This variant was originally described as del63-bp in the patients’ cDNA (Dell'Angelica et al., 1999), but later found to be due to a gDNA splice site variant, skipping the 63-bp exon 12 (Gochuico et al., 2012).

⁷Described as g.151312_159483del8172-bp (NG_007268) (Hengst et al., 2018). It is possible that del8168-bp reported by (Jung et al., 2006) is the same deletion.

⁸This deletion was reported as: NC_00005.8:g.180242–180866del.

⁹This nucleotide change activates a cryptic donor splice and causes a deletion of 112bp within exon 23 on the mRNA level, resulting in a frame shift and a premature termination codon p.Val900Thrfs*63 (de Boer et al., 2017).

¹⁰This frameshift in the AP3B1 C-terminal coding region results in a prolonged altered protein, beyond the termination codon, with 42 additional C-terminal amino acids compared to the wild type protein (Kurnik et al., 2013).

¹¹Chromosomal inversion breakpoints occur within the AP3B1 gene (Jones et al., 2013).

HPS-2 (MIM# 608233) is caused by biallelic pathogenic variants in AP3B1, located on chromosome 5q14.1, and encoding the β3A subunit of AP3 (Dell’Angelica et al., 1999). The longest AP3B1 mRNA transcript contains 27 exons and is translated into the 1094-amino acid protein AP3B1 (AP-3 β3A; ~121-kD). Two AP3B1 protein-coding mRNA transcripts are predicted, varying at the N-terminus, with unexplored expression and functional significance. Vertebrate models of HPS-2 include the pearl and Ap3b1^LN mice (Yang et al., 2000; Li et al., 2004). With the recognition that HPS-2 is caused by deficiency of AP-3, HPS became an informative disorder for intracellular vesicle/membrane formation and trafficking (Dell’Angelica et al., 1999). HPS cells and models have since been used extensively for LRO-related cell biology (Huizing et al., 2008).

There are currently ~35 reported HPS-2 cases of various ethnic backgrounds, including Caucasian, Chinese, Lebanese and Mexican. A total of 29 AP3B1 pathogenic variants associated with HPS-2 have been described (Table 4); they are located throughout the gene, including 1 (3%) start-loss, 7 (24%) frameshift, 3 (10.5%) missense, 6 (21%) nonsense, 9 (31%) insertions and/or deletions (including a chromosomal inversion), and 3 (10.5%) splice site variants (Figure 2). There are no known frequently occurring pathogenic variants in AP3B1, nor any apparent founder mutations. All 3 reported AP3B1 splice site variants are predicted and/or reported to cause aberrant splicing (Supp. Table S2). Of the 3 reported pathogenic missense variants, two are likely pathogenic (p.Leu580Arg, p.Ser901Cys), and one is a VUS (p.Leu102Pro).

Figure 2:

Distribution of AP3B1 Gene Variants

Next generation sequencing of different cohorts of subjects with primary immunodeficiency or hemophagocytic lymphohistiocytosis (HLH) reported least 8 AP3B1 variants (Supp. Table S4) in a heterozygous state with or without a variant in another (synergistic) gene. These findings suggest that heterozygous AP3B1 variants may contribute to an immunologic phenotype (Chi et al., 2018; Gallo et al., 2016; Gao, Zhu, Huang, & Zhou, 2015; Miao et al., 2019; Mukda et al., 2017; Tesi et al., 2015; Xu et al., 2017). These variants have not been reported in HPS subjects but were included in this report because they may cause HPS when occurring in a homozygous or compound heterozygous state.

The dbSNP/gnomAd/ClinVar databases list 7 frequently occurring (MAF > 0.001) AP3B1 missense variants and two in-frame 3-bp deletions, with predicted benign or unknown pathogenicity, which should be considered as likely non-pathogenic polymorphisms when encountered in AP3B1 genetic analyses (Supp. Table S3).

Apart from the hypopigmentation and bleeding diathesis of HPS-2, affected individuals are also at risk for developing interstitial lung disease and pulmonary fibrosis in childhood (Gochuico et al., 2012; Hengst et al., 2018); in addition, immunodeficiency associated with neutropenia is the most prevalent clinical feature (Fontana et al., 2006). The immunodeficiency, an impairment of cytotoxic activity, results from T-lymphocyte and/or natural killer cell dysfunction and can present with variable features, from mild recurrent bacterial and viral infections to severe hemophagocytic lymphohistiocytosis (HLH) (Gil-Krzewska et al., 2017; Jessen et al., 2013). These features led to discovery of involvement of AP-3 in different trafficking processes. AP-3 is involved in neutrophil formation (Badolato & Parolini, 2007; Massullo et al., 2005); AP-3 deficient cells mislocalize the neutrophil granule proteins myeloperoxidase and elastase and the lysosomal membrane protein CD63 (de Boer et al., 2017; Di Pietro et al., 2006; Jung et al., 2006; Meng et al., 2010). The AP-3 immunodeficiency also involves defective AP-3 mediated lytic granule exocytosis in natural killer (NK)-cells and cytotoxic T-cells (Clark et al., 2003; Fontana et al., 2006; Gil-Krzewska et al., 2017; Jung et al., 2006). AP-3 deficient dendritic cells showed impaired toll-like receptor recruitment (Mantegazza et al., 2012; Sasai, Linehan, & Iwasaki, 2010), leading to defects in interferon production and antigen presentation in these cells from HPS-2 subjects (Prandini et al., 2016). AP-3-dependent inflammasome positioning and activation was shown in dendritic cells from HPS-2 mice (Mantegazza et al., 2017).

Remarkably, genetic testing in all reported HPS-2 subjects was performed by AP3B1 gene-specific Sanger sequencing. Individuals with HPS-2 that present to an immunologist with (severe) immunodeficiency may escape diagnosis due to emphasis on their immunodeficiency, other mild manifestations (e.g., hypopigmentation, ocular findings, bleeding diathesis) may be overlooked, there may be unfamiliarity with HPS-2, and costs and lack of availability of AP3B1 genetic testing may provide obstacles to diagnosis. However, the recent significant number of heterozygous AP3B1 variants identified by next-generation sequencing in cohorts with immunodeficiency disorders (Chi et al., 2018; Gallo et al., 2016; Gao et al., 2015; Miao et al., 2019; Mukda et al., 2017; Tesi et al., 2015; Xu et al., 2017) emphasizes the importance of including AP3B1 in immunodeficiency-related gene panels and may result in the diagnosis of additional HPS-2 cases.

HPS-3 (Table 5)

Table 5:

HPS3 Gene Variants Associated with Hermansky-Pudlak Syndrome Type 3 (HPS-3)

No	mRNA NM_032383.5	Amino Acid NP_115759.2	Exon/Intron	Variant Type 1	Ethnic Background 2	References and Footnotes
1	c.-2993_217+692del 3	-	5’UTR Intron 1	Indel	Central Puerto Rican	(Anikster et al., 2001; Torres-Serrant et al., 2010) 3,4
2	c.15C>G	p.Tyr5*	Exon 1	Nonsense	Dutch, German	novel 5,6
3	c.87dupG	p.Arg30Alafs*2	Exon 1	Frameshift	Japanese	(Saito et al., 2019)
4	c.319C>T	p.Arg107*	Exon 1	Nonsense	Japanese	(Okamura et al., 2019) 7
5	c.437_439delGAG	p.Gly146del	Exon 1	Indel	Japanese	(Okamura et al., 2019) 7,8
6	c.712+2T>C	IVS2+2	Intron 2	Splice site	Chinese	(A. Wei et al., 2016) 7
7	c.726_727insTGCCTTACATC	p.Ile243Cysfs*41	Exon 3	Indel	Puerto Rican	novel 4,5
8	c.728_729insA	p.Ser244Phefs*4	Exon 3	Frameshift	Italian, Sicilian	(Boissy et al., 2005) 9
9	c.851_852delGA	p.Arg284Lysfs*11	Exon 3	Indel	Portuguese	novel 5,11
10	c.868C>T	p.Gln290*	Exon 3	Nonsense	Arabic	(Khan et al., 2016) 7
11	c.885–1G>A	IVS3–1G>A	Intron 3	Splice site	Libyan	(Thielen et al., 2010)
12	c.1012G>T	p.Glu338*	Exon 5	Nonsense	Caucasian	novel 5,12
13	c.1107_1119del13insC	p.Pro370_Ser373del	Exon 5	Indel	French-Canadian	novel 5,13,14
14	c.1153_1160del8	p.Val385Lysfs*2	Exon 5	Indel	Middle-Eastern	(Trujillano et al., 2017) 7
15	c.1163+1G>A	IVS5+1G>A	Intron 5	Splice site	Ashkenazi Jewish	(Huizing et al., 2001)
16	c.1189C>T	p.Arg397Trp	Exon 6	Missense	Canadian, Caucasian, Chinese, German, Japanese, Polish, Russian, Swiss	(Huizing et al., 2001; Nazarian et al., 2008; Okamura et al., 2019; Wei et al., 2016) 7,12,15
17	c.1195A>G/ c.1199_1200insATTGC	p.Ser399Gly/ p.Ala401Leufs*16	Exon 6	Indel	English, German, Irish, Scottish, Cherokee	novel 5,16
18	c.1291delC	p.Leu431Phefs*3	Exon 7	Frameshift	Japanese	(Okamura et al., 2019) 7,8
19	c.1426dupA	p.Ile476Asnfs*8	Exon 8	Frameshift	Japanese	(Saito et al., 2019)
20	c.1509G>A	p.Met503Ile	Exon 8	Missense-Splice site	Pakistani	(Yousaf et al., 2016) 7,17
21	c.1555_1595dup41	p.Leu533Phefs*10	Exon 8	Indel	Chinese	(Power et al., 2019)
22	c.1673T>C	p.Leu558Pro	Exon 8	Missense	-	(Lasseaux et al., 2018) 7
23	c.1691+1G>A	IVS9+1G>A	Intron 9	Splice site	French-Canadian	novel 5,13
24	c.1691+2T>G	IVS9+2T>G	Intron 9	Splice site	Ashkenazi Jewish	(Huizing et al., 2001)
25	c.1838C>G	p.Ser613*	Exon 10	Nonsense	Chinese	(Wei et al., 2019) 7
26	c.1870G>T	p.Glu624*	Exon 10	Nonsense	German, Irish	novel 5,18
27	c.2208_2209delTC	p.Gln737Alafs*20	Exon 12	Indel	Chinese	(Wei et al., 2016) 7
28	c.2464C>T	p.Arg822*	Exon 13	Nonsense	Dutch, German, Portuguese, Spanish	(Bastida et al., 2019) 6,7,11
29	c.2482–2A>G	IVS13–2A>G	Intron 13	Splice site	Irish/German	(Huizing et al., 2001)
30	c.2589+1G>C	IVS14+1G>C	Intron14	Splice site	German/Swiss	(Huizing et al., 2001)
31	c.2589+1G>T	IVS14+1G>T	Intron14	Splice site	German, Irish	novel 5,18
32	c.2628delT	p.Ile877Phefs*25	Exon 15	Frameshift	-	(Lasseaux et al., 2018) 7
33	c.2733delG	p.Leu912*	Exon 15	Frameshift	English, German, Irish, Scottish, Cherokee	novel 5,16,19
34	c.2739_2742delGAGA	p.Glu913Aspfs*14	Exon 15	Indel	-	novel 5,14
35	c.2771delA	p.Asn924Ilefs*4	Exon 15	Frameshift	Turkish	(Sandrock-Lang et al., 2017)
36	c.2805G>A	p.Trp935*	Exon 16	Nonsense	Chinese	(Wei et al., 2016) 7
37	c.2888–1612G>A21	IVS16–1612G>A p.Glu963Alafs*24	Intron16	Splice site	English, Irish	(Huizing et al., 2001) 20

¹When deletion/insertion is 1 nucleotide it is named Frame shift, when larger it is named Indel.

²Extracted from literature reference. ‘-‘ = unreported.

³This HPS3 3.9-kb deletion occurs within two Alu repeats, encompassing exon 1 and originates from a genetic isolate of central Puerto Rico (Anikster et al., 2001). Current nomenclature (as annotated in ClinVar) for this deletion is NM_032383.5(HPS3):c.-2993_217+692del or NC_000003.12:g.149126714_149130632del (GRCh38) or NG_009847.1:g.2131_6049del.

⁴This HPS3 variant was identified in one unreported female subject of Puerto Rican descent in the NIH cohort. This subject was compound heterozygous for c.726delinsTGCCTTACATC and the Central Puerto Rican founder variant g.del3.9-kb.

⁵novel = previously unreported variant detected in the NIH HPS cohort.

⁶This HPS3 variant was identified in one unreported male subject of Dutch-German descent in the NIH HPS cohort. This subject was compound heterozygous for c.15C>G and c.2464C>T.

⁷At least one of the reported cases with this variant was identified by next generation sequencing.

⁸The HPS3 variant c.437_439delGAG occurred compound heterozygous with c.1291delC in a subject of Japanese descent who also had non-segmental vitiligo (Okamura et al., 2019).

⁹This HPS3 variant c.728insA was identified homozygous in one unreported female subject of Sicilian decent in the NIH HPS cohort.

¹⁰Gray highlight: missense variant. See Supplemental Table S1 for pathogenicity description.

¹¹This HPS3 variant was identified in one unreported male subject of Portuguese decent in the NIH HPS cohort. This subject was compound heterozygous for c.851_852delGA and c.2464C>T.

¹²This HPS3 variant was identified in one unreported female subject of Caucasian descent in the NIH HPS cohort. This subject was compound heterozygous for c.1012G>T and c.1189C>T.

¹³This HPS3 variant was identified in one unreported female subject of French-Canadian descent in the NIH HPS cohort. This subject was compound heterozygous for c.1107_1119del13insC and c.1691+1G>A.

¹⁴This HPS3 variant was identified in one unreported female subject of in the NIH cohort (referred by Dr. Doherty, Carilion Clinic, Roanoke, VA). This subject was compound heterozygous for c.1107_1119del13insC and c.2739_2742delGAGA.

¹⁵This HPS3 variant c.1189C>T was identified homozygous in one unreported female subject of Canadian-Polish-Russian decent in the NIH cohort. Cells of this patient showed destabilized BLOC-2 assembly, likely due to pathogenicity of this variant (Nazarian et al., 2008).

¹⁶This HPS3 variant was identified in one unreported female subject of English-German-Irish-Scottish-Cherokee descent in the NIH HPS cohort. This subject was compound heterozygous for c.1195A>G/1199insATTGC and c.2733delG.

¹⁷This (missense) HPS3 variant c.1509G>A; p.Met503Ile occurs at the exon8/intron 8 splice site junction and may affect splicing. This variant occurs homozygous in 4 subjects of a consanguineous Pakistani family (Yousaf et al., 2016).

¹⁸This HPS3 variant was identified in one unreported male subject of German-Irish descent. This subject was compound heterozygous for c.1870G>T and c.2589+1G>T.

¹⁹Subject HPS34 of English-Irish descent in (Huizing et al., 2001) was reported heterozygous for c.2887+2500G>A. We found c.2733delG to be the second HPS3 variant in this subject.

²⁰This HPS3 variant was reported in an alternative nomenclature as c.2887+2500G>A. This intronic variant introduces a new consensus splice site that results in insertion of 89-bp (a ‘pseudoexon’) in the patient’s cDNA (Huizing et al., 2001; Vorechovsky, 2010).

HPS-3 (MIM# 614072) is caused by biallelic pathogenic variants in HPS3, also called BLOC2S1, located on chromosome 3q24. The longest HPS3 mRNA transcript contains 17 exons, encoding a 1004-amino acid protein (~113.7-kD). Two HPS3 protein-coding mRNA variants are predicted, but their expression and functional significance remain unexplored. The HPS3 protein interacts with the HPS5 and HPS6 proteins in BLOC-2 (Di Pietro et al., 2004). The mouse model cocoa is the murine orthologue of human HPS-3 (Suzuki et al., 2001).

The HPS3 gene was identified through homozygosity mapping in a genetic isolate of HPS originating in central Puerto Rico. These individuals are homozygous for a 3.9-kb deletion in HPS3 (NM_032383.5:c.−2993_217+692del or NG_009847.1:g.2131_6049del), detectable with a multiplex PCR amplification assay (Anikster et al., 2001). There are ~63 reported cases homozygous for this deletion (Anikster et al., 2001; Santiago Borrero et al., 2006), and there are an additional 9 such unreported subjects in our NIH cohort. One in 14,000 individuals of central Puerto Rican descent are estimated to be homozygous for this deletion. The carrier frequency in central Puerto-Rico is ~1:32 (Santiago Borrero et al., 2006) and 1:85 in all of Puerto Rico (Torres-Serrant, Ramirez, Cadilla, Ramos-Valencia, & Santiago-Borrero, 2010).

We report 42 non-Puerto Rican HPS-3 subjects, including those described with an Ashkenazi-Jewish founder variant (c.1691+2T>G; 7 cases) (Huizing et al., 2001) and including 9 novel cases from our NIH cohort (Table 5). Apart from central Puerto Rican and Ashkenazi-Jewish cases, HPS-3 subjects with a variety of other ethnic backgrounds are reported, including Arabic, Chinese, northern and southern European, and Pakistani.

We report 37 HPS3 pathogenic variants, including 11 novel variants from our NIH cohort (Table 5). These pathogenic variants are found throughout the gene and include 7 (19%) frameshift, 3 (8 %) missense, 8 (22%) nonsense, 10 (27 %) insertions and/or deletions, and 9 (24%) splice site variants (Figure 3). All HPS3 splice site variants are predicted and/or experimentally demonstrated to cause aberrant splicing (Supp. Table S2). One intronic variant, c.2888–1612G>A (originally reported as c.2887+2500G>A) introduces a new consensus splice site, resulting in cryptic exon activation and insertion of a 89-bp pseudo-exon in the cDNA, leading to a frameshift and premature protein termination (p.Glu963Alafs*24) (Huizing et al., 2001; Vorechovsky, 2010). All 3 reported HPS3 missense variants have a low MAF, one (c.1189C>T, p.Arg397Trp) is predicted likely pathogenic and the other two are predicted VUS (Supp. Table S1).

Figure 3:

Distribution of HPS3 Gene Variants

Next generation sequencing of a cohort of individuals with schizophrenia identified a de novo HPS3 missense variant c.796G>A (p.Glu266Lys) in one proband, suggesting an association of this variant with the schizophrenia phenotype (Fromer et al., 2014) (Supp. Table S4).

This rare and benign variant has not been reported in HPS subjects. In fact, no other variants in any BLOC-2 subunit (HPS3, HPS5, HPS6) have been reported to be associated with a neuronal phenotype in individuals with or without HPS or in animal models. The dbSNP/gnomAd/ClinVar databases list 5 frequently occurring (MAF > 0.001) HPS3 missense variants (Supp. Table S3), with predicted benign or unknown pathogenicity, that should be considered as likely non-pathogenic polymorphisms when encountered in HPS3 genetic analysis.

Subjects with HPS-3 have relatively mild clinical features, including minor hypopigmentation of the hair, skin, and retina. Visual acuity is often only slightly affected, and bleeding tendency is also mild, although some female subjects have significant menorrhagia. Some subjects only appear hypopigmented when compared to their siblings or other family members. Of note, pigmentation and bleeding are also only mildly affected in BLOC-2 (HPS3, HPS5, HPS6) mouse models (Novak et al., 1984). In addition, the bleeding diathesis in BLOC-2-deficient individuals might be exacerbated by disruption not only of platelet dense granules but also of altered maturation of Weibel-Palade bodies in endothelial cells, which are involved in release of von Willebrand factor (Ma et al., 2016).

Pulmonary fibrosis or immunodeficiency has not been reported in HPS-3 (nor in any other BLOC-2 deficient) subjects, and some BLOC-2 deficient individuals develop granulomatous colitis. The diagnosis of HPS-3 may be elusive in some patients due to the mildness of their symptoms.

HPS-4 (Table 6)

Table 6:

HPS4 Pathogenic Gene Variants Associated with Hermansky-Pudlak Syndrome Type 4 (HPS-4)

No	mRNA NM_022081.5	Amino Acid NP_071364.4	Exon/Intron	Variant Type 1	Ethnic Background 2	References and Footnotes
1	c.45G>A	p.Trp15*	Exon 3	Nonsense	Uruguayan, Japanese	(Carmona-Rivera, Golas, et al., 2011; Okamura et al., 2018) 3
2	c.47delA	p.Asn16Ilefs*11	Exon 3	Frameshift	Uruguayan	(Carmona-Rivera, Golas, et al., 2011)
3	c.57delT	p.Leu20Phefs*7	Exon 3	Frameshift	French, German, Irish, Northern European	(Suzuki et al., 2002) 4
4	c.123T>A	p.Tyr41*	Exon 4	Nonsense	Japanese	(Okamura et al., 2018) 3
5	c.148C>T	p.Gln50*	Exon 4	Nonsense	Chinese	(Wei et al., 2019) 3
6	c.272T>C	p.Leu91Pro 5	Exon 4	Missense	Turkish	(Bastida et al., 2019) 3
7	c.276+5G>A	IVS4+5G>A	Intron 4	Splice site	Pakistani	(Yousaf et al., 2016) 3
8	c.357C>G	p.Tyr119*	Exon 5	Nonsense	Dutch, English, Irish, Polish, Slovak	novel 6,7
9	c.412G>T	p.Glu138*	Exon 6	Nonsense	Indian	(Anderson et al., 2003)
10	c.416G>A	p.Trp139*	Exon 6	Nonsense	Chinese	(Power et al., 2019)
11	c.430G>T	p.Glu144*	Exon 6	Nonsense	Indian	(Arcot Sadagopan et al., 2017) 3
12	c.461A>G	p.His154Arg	Exon 6	Missense	Caucasian, Japanese	(Anderson et al., 2003; Saito et al., 2013) 8,9
13	c.541C>T	p.Gln181*	Exon 7	Nonsense	Southern Italian	(Suzuki et al., 2002)
14	c.554G>A	p.Arg185His	Exon 7	Missense	Indian	(Arcot Sadagopan et al., 2017) 3
15	c.596+1G>A	IVS7+1G>A	Intron 7	Splice site	Japanese	(Okamura et al., 2019) 3
16	c.597–2A>T	IVS7–2A>T	Intron 7	Splice site	-	(Jones et al., 2012) 3
17	c.630dupC	p.Ala211Argfs*47	Exon 8	Frameshift	Chinese	(Wu et al., 2019)
18	c.649C>T	p.Arg217*	Exon 8	Nonsense	Ashkenazi Jewish, English, Polish	(Anderson et al., 2003; Lozynska et al., 2018)
19	c.664G>T	p.Glu222*	Exon 8	Nonsense	Indian	(Anderson et al., 2003)
20	c.706+1G>A	IVS9+1G>A	Intron 9	Splice site	Dutch, English, Irish, Polish, Slovak	novel 6,7
21	c.730C>T	p.Gln244*	Exon 10	Nonsense	Japanese	(Araki et al., 2014)
22	c.803G>A 10	p.Arg268Lys	Exon 10	Missense-Splice Site	-	(Lasseaux et al., 2018) 3,10
23	c.949_972dup24	p.Ala317_Glu324dup	Exon 11	Indel	Dutch	(Suzuki et al., 2002)
24	c.1132C>T	p.Gln378*	Exon 11	Nonsense	-	novel 11
25	c.1318C>T	p.Gln440*	Exon 11	Nonsense	Turkish	(Sandrock-Lang et al., 2018)
26	c.1546C>T	p.Gln516*	Exon 11	Nonsense	Caucasian	novel 6,8
27	c.1547_1548delAG	p.Gln516Argfs*42	Exon 11	Indel	Indian	(Arcot Sadagopan et al., 2017) 3
28	c.1713+5G>C	IVS11+5G>C	Intron 11	Splicing	Chinese	(Wei et al., 2019) 3
29	c.1856C>T	p.Pro619Leu	Exon 13	Missense	-	(Lasseaux et al., 2018) 3
30	c.1858C>T	p.Gln620*	Exon 13	Nonsense	-	(Sakata et al., 2013)
31	c.1891C>T	p.Gln631*	Exon 13	Nonsense	German South Tirol	(Suzuki et al., 2002)
32	c.1897_1898dupCG	p.Ser634Alafs*3	Exon 13	Indel	Japanese	(Okamura et al., 2019) 3
33	c.2054delC	p.Pro685Leufs*17	Exon 14	Frameshift	Sri Lankan, Spanish	(Bachli et al., 2004; Bastida et al., 2019) 3
34	c.2089_2093dupAAGCA	p.Lys699Serfs*5	Exon 14	Indel	Austrian, Czech, English, German, Hungaran, Irish, Scandinavian, Swiss	(Anderson et al., 2003; Suzuki et al., 2002)

¹When deletion/insertion is 1 nucleotide it is named Frame shift, when larger it is named Indel.

²Extracted from literature reference. ‘-‘ = unreported.

³At least one of the reported cases with this variant was identified by next generation sequencing.

⁴This HPS4 variant was identified in heterozygous state in one unreported subject of French-German-Irish descent in the NIH HPS cohort.

⁵Gray highlight: missense variant. See Supplemental Table S1 for pathogenicity description.

⁶novel = previously unreported variant detected in the NIH HPS cohort.

⁷This HPS4 variant was identified in one unreported subject of Dutch-English-Irish-Polish-Slovak descent in the NIH HPS cohort. This subject is compound heterozygous for c.357C>G and c.706+1G>A.

⁸This HPS4 variant was identified in one unreported subject of Caucasian descent in the NIH HPS cohort. This individual is compound heterozygous for c.461A>G and c.1546C>T.

⁹This variant was found homozygous in two Japanese siblings with HPS and mental disorder (schizophrenia and major depression). It was suggested that HPS4 gene variants are associated with susceptibility to schizophrenia (Saito et al., 2013) and/or cognitive function (Kuratomi et al., 2013).

¹⁰This (missense) HPS4 variant c.803G>A; p.Arg268Lys (Lasseaux et al., 2018) occurs at the exon10/intron 11 splice site junction and is predicted to affect splicing (Supplemental Table S2).

¹¹This nonsense variant c.1132C>T (p.Gln378*) was found homozygous in one unreported female subject in the NIH cohort (referred by Dr. Everman, Greenwood Genetics Center, Greenville, SC).

HPS-4 (MIM# 614073) is caused by biallelic pathogenic variants in HPS4, also called BLOC3S2, located on chromosome 22q12.1. The longest HPS4 mRNA transcript contains 14 exons, encoding a 708-amino acid protein (~76.9-kD). This major transcript variant is expressed in all tissues tested. A second major mRNA variant contains 12 exons, is alternatively spliced in the 5’ region and is expressed in limited tissues; its function remains unexplored (Anderson, Huizing, Claassen, White, & Gahl, 2003). There are at least 9 additional predicted HPS4 protein-coding mRNA transcripts. The HPS4 protein interacts with the HPS1 protein in BLOC-3 (Carmona-Rivera et al., 2013; Martina et al., 2003).

Identification of the human HPS4 gene was initiated based upon Hps4 mutations in the light ear HPS mouse model (Suzuki et al., 2002). Subsequent genetic screening of the human orthologue gene, HPS4, in unclassified HPS subjects identified 7 cases with pathogenic variants (Suzuki et al., 2002). There are currently 37 HPS-4 subjects reported and there are 4 unreported subjects in our NIH cohort. HPS-4 is identified in various populations, including Ashkenazi-Jewish, Chinese, European, Indian, Japanese, Pakistani, Sri Lankan, and Uruguayan (Table 6). We report 34 HPS4 pathogenic variants, including 4 novel variants from our NIH cohort (Table 6). These variants are located throughout the HPS4 gene, with c.2089_2093dupAAGCA (p.Lys699Serfs*5) occurring frequently in individuals of European descent (Anderson et al., 2003; Suzuki et al., 2002). The HPS4 variants include 5 (12 %) frameshift, 5 (14.5%) missense, 16 (47%) nonsense, 4 (12 %) insertions and/or deletions, and 5 (14.5%) splice site variants (Figure 4). The 4 reported HPS4 splice site variants are all predicted to cause aberrant splicing (Supp. Table S2). Of the 5 reported missense variants, one is likely pathogenic (p.His154Arg) and the other 3 are classified as VUS; one of these occurs at a splice site junction and is predicted (but not demonstrated) to cause aberrant splicing (c.803G>A; p.Arg268Lys) (Supp. Table S1). No additional experimental evidence for pathogenicity is available for these missense variants.

Figure 4:

Distribution of HPS4 Gene Variants

Next generation sequencing of cohorts with non-HPS pulmonary fibrosis identified a heterozygous HPS4 frameshift variant (c.1102dupG, p.Asp368Glyfs*4) in a subject with sporadic pulmonary fibrosis (Deng et al., 2018) and a heterozygous HPS4 indel variant (c.1966_1967dupAC, p.Ala657Argfs*46) and a missense variant (c.1396C>T, p.Arg466Cys) in subjects with familial pulmonary fibrosis (Stearman et al., 2019). None of these variants has been reported in HPS subjects but are included in Supp. Table S4 since they may cause HPS when occurring in the homozygous or compound heterozygous state.

The dbSNP/gnomAd/ClinVar databases list 12 frequently occurring (MAF > 0.001) HPS4 missense variants, with predicted benign or unknown pathogenicity (Supp. Table S3); these should be considered as likely non-pathogenic polymorphisms when encountered in HPS4 genetic analysis.

Individuals with HPS-4 have a phenotype similar to that of HPS-1 subjects, including more profound cutaneous and ocular hypopigmentation than in other HPS subtypes, development of pulmonary fibrosis (at middle age) in virtually all subjects, menorrhagia in most female subjects, and occurrence of granulomatous colitis in some subjects (Anderson et al., 2003; Huizing et al., 2008). HPS-4 should be considered in individuals where HPS-1 was suspected based on clinical symptoms (without genetic confirmation). Of note, the HPS4 missense variant p.His154Arg (c.461A>G) was found homozygous in two Japanese siblings with HPS and mental disorder (schizophrenia and major depression). It was suggested that HPS4 gene single nucleotide polymorphisms variants may be associated with susceptibility to schizophrenia (A. Saito et al., 2013) and/or cognitive function (Kuratomi et al., 2013). However, there are no reports of other HPS-related BLOC-3 (HPS1 and HPS4) variants in individuals with HPS to be associated with neurological phenotypes.

HPS-5 (Table 7)

Table 7:

HPS5 Pathogenic Gene Variants Associated with Hermansky-Pudlak Syndrome Type 5 (HPS-5)

No	mRNA NM_181507.1	Amino Acid NP_852608.1	Exon/Intron	Variant Type 1	Ethnic Background 2	References and Footnotes
1	c.219G>A	p.Arg73Arg 3	Exon 3	Splice site	Turkish	(Lasseaux et al., 2018; Michaud et al., 2017) 3,4
2	c.285–10A>G	IVS4–10A>G	Intron 4	Splice site	Turkish	(Stephen et al., 2017) 4
3	c.302_305delTTTG	p.Val101Glyfs*3	Exon 5	Indel	Cuban, Venezuelan	(Carmona-Rivera, Golas, et al., 2011)
4	c.434G>A	p.Gly145Glu 5	Exon 5	Missense	German, Irish, Welsh	(Nazarian et al., 2008)
5	c.719G>C	p.Arg240Pro	Exon 7	Missense	African-French	(Lasseaux et al., 2018; Michaud et al., 2017) 4
6	c.803delC 6	p.Pro268Leufs*4	Exon 7	Frameshift	-	(Ringeisen et al., 2013) 6
7	del 1.4-kb 7,8	-	Exon 7	Indel	African-French	(Michaud et al., 2017) 4,7
8	c.814_818delATTAC	p.Ile272Serfs*8	Exon 7	Indel	-	(Lasseaux et al., 2018) 4
9	c.818_822delCTCTC	p.Thr273Lysfs*7	Exon 7	Indel	French	(Michaud et al., 2017) 4
10	c.879dupC	p.Lys294Glnfs*6	Exon 8	Frameshift	English, Irish	(Huizing et al., 2004)
11	c.888dupA	p.His297Thrfs*3	Exon 8	Frameshift	Turkish	(Korswagen et al., 2008)
12	c.1417C>T	p.Gln473*	Exon12	Nonsense	Turkish	(Lasseaux et al., 2018; Michaud et al., 2017) 4
13	c.1423delC	p.Leu475Serfs*37	Exon 12	Frameshift	Mexican	(Carmona-Rivera, Golas, et al., 2011)
14	c.1618C>T	p.Gln540*	Exon 13	Nonsense	Arabic	(Khan et al., 2016) 4
15	c.1634+1G>A	IVS13+1G>A	Intron13	Splice site	Cuban, Venezuelan	(Carmona-Rivera, Golas, et al., 2011)
16	c.1871T>G	p.Leu624Arg 9	Exon 16	Missense	Swiss	(Huizing et al., 2004; Lasseaux et al., 2018; Michaud et al., 2017) 4,9
17	c.1900delG	p.Glu634Serfs*3	Exon 16	Frameshift	-	(Lasseaux et al., 2018; Michaud et al., 2017) 4
18	c.1960A>T	p.Lys654*	Exon 16	Nonsense	English, German, Scottish	(Botero et al., 2018) 4
19	c.2026_2029delGTTA	p.Val676Metfs*8	Exon 16	Indel	Turkish	(Zhang et al., 2003)
20	c.2219T>C	p.Leu740Ser	Exon 16	Missense	-	(Lasseaux et al., 2018; Michaud et al., 2017) 4
21	c.2234T>C	p.Leu745Ser	Exon 16	Missense	Chinese	(Wei et al., 2016) 4,10
22	c.2593C>T	p.Arg865*	Exon 18	Nonsense	Dutch, English, Irish, Swedish	(Huizing et al., 2004)
23	c.2624delT	p.Leu875Cysfs*20	Exon 18	Frameshift	Dutch, English, Irish, Swedish	(Huizing et al., 2004)
24	c.2750_2751delAG	p.Glu917Valfs*14	Exon19	Indel	French	(Lasseaux et al., 2018; Michaud et al., 2017) 4
25	c.2928_2929dupGA	p.Thr977Argfs*15	Exon 20	Indel	English, Irish	(Huizing et al., 2004)
26	c.2974_2977delCTCT	p.Leu992Valfs*17	Exon 21	Indel	-	(Lasseaux et al., 2018) 4
27	c.2979_2982delTTTG	p.Cys993Trpfs*16	Exon 21	Indel	French	(Michaud et al., 2017) 4
28	c.3058+3A>G	IVS21+3A>G	Intron 21	Splice site	Turkish	(Lasseaux et al., 2018; Michaud et al., 2017) 4
29	c.3096_3098delCCT	p.Leu1033del	Exon 22	Indel	French	(Michaud et al., 2017) 4
- 8	c.3293C>T	p.Thr1098Ile 9	Exon 22	Missense	Swiss	(Huizing et al., 2004) 9
30	c.3346A>G	p.Met1116Val	Exon 23	Missense	Chinese	(Wei et al., 2016) 4,10
31	del 1.5-kb 8	-	-	Indel	-	(Lasseaux et al., 2018) 4

¹When deletion/insertion is 1 nucleotide it is named Frame shift, when larger it is named Indel.

²Extracted from literature reference. ‘-‘ = unreported.

³This (silent) HPS5 variant c.219G>A; p.Arg73Arg occurs at the exon3/intron 3 splice site junction and is predicted to affect splicing (Supplemental Table S2).

⁴At least one of the reported cases with this variant was identified by next generation sequencing.

⁵Gray highlight: missense variant. Gray highlight: missense variant. See Supplemental Table S1 for pathogenicity description.

⁶This variant was previously described as c.1081delC (Ringeisen et al., 2013).

⁷Described as Chr11:18327845 to Chr11:18329253 (Michaud et al., 2017).

⁸The nomenclature of these HPS6 variants are included in this Table as reported, see reference for each specific variant for more details.

⁹Two siblings were homozygous for 2 missense HPS5 variants (hemizygosity was excluded): p.Leu624Arg and p.Thr1098Ile (Huizing et al., 2004). The high MAF and low pathogenicity prediction (Supplemental Table S1) classifies p.Thr1098Ile as a benign SNP. Variant p.Leu624Arg classifies as a likely pathogenic missense and is also identified in other HPS-5 individuals in trans with a pathogenic variant (Lasseaux et al., 2018; Michaud et al., 2017).

¹⁰These variants were originally reported as c.1892T>C and c.3004A>G, using NM_007216 (HPS5 mRNA Variant 2) nomenclature. For this report, the nomenclature of these variants was converted to NM_181507.1 (HPS5 mRNA Variant 1).

HPS-5 (MIM# 614074) is caused by biallelic pathogenic variants in HPS5, also called BLOC2S2, located on chromosome 11p15.1. The longest HPS5 mRNA transcript contains 23 exons, encoding a 1129-amino acid protein (~127.4-kD). There are 3 HPS5 protein-coding mRNA variants described, each with alternatively spliced 5’ exons (Huizing et al., 2004). The HPS5 protein interacts with HPS3 and HPS6 proteins in BLOC-2 (Di Pietro et al., 2004). Vertebrate models of HPS-5 include the ruby-eye-2 mouse (Zhang et al., 2003), the snow white zebrafish (Daly, Willer, Gregg, & Gross, 2013), and the casper stickleback (Hart & Miller, 2017) (Table 1).

The HPS5 gene was discovered after Hps5 deficiency was identified in the ruby-eye-2 HPS murine model; subsequent sequencing of the human orthologue in unclassified HPS individuals identified one subject with a homozygous 4 base-pair deletion in HPS5 (Zhang et al., 2003). There are now ~29 HPS-5 subjects described worldwide of variable ethnic origins, including Arabic, Chinese, European, Mexican, South-American, and Turkish.

We report 31 HPS5 pathogenic variants (Table 7), including 6 (19%) frameshift, 6 (19%) missense, 4 (13%) nonsense, 11 (36%) insertions and/or deletions, and 4 (13%) splice site variants (Figure 5). The variants are located throughout the HPS5 gene, without an apparent frequently occurring variant. One HPS5 splice variant (c.1634+1G>A) results in skipping of exon 13 (Carmona-Rivera, et al., 2011), another (c.285–10A>G) activates a cryptic splice site leading to an in-frame insertion of 9-bp and reduced HPS5 protein expression (Stephen et al., 2017), and a third variant (c.3058+3A>G) is predicted to cause a splicing defect (Supp. Table S2). A silent HPS5 variant (c.219G>A, p.Arg73Arg; not listed in dbSNP/gnomAd/ClinVar) occurs at the exon 3/intron 3 splice junction and is predicted to weaken the consensus splice site significantly, causing a splice defect (Supp. Tables S1, S2) (Michaud et al., 2017). An additional 6 HPS5 missense variants are reported, three of which are predicted as likely pathogenic (p.Gly145Glu, p.Arg240Pro, p.Leu624Arg) and three as VUS (p.Leu740Ser, p.Leu745Ser, p.Met1116Val) (Supp. Table S1). Of note, two siblings of Swiss origin were reported homozygous for 2 missense HPS5 variants, p.Leu624Arg (c. c.1871T>G) and p.Thr1098Ile (c.3293C>T), and hemizygosity was excluded (Huizing et al., 2004); no pathogenicity predictions were reported at that time. The p.Thr1098Ile variant is a SNP (rs61884288) with a high MAF (0.02362) and low pathogenicity prediction (Supp.l Tables S1, S2, S3). Therefore, this variant should be considered a benign SNP. In contrast, p.Leu624Arg (rs281865102) has no reported allele frequency and is predicted likely pathogenic (Supp. Tables S1); this variant was recently identified in trans with a pathogenic variant in other HPS-5 individuals (Michaud et al., 2017). Therefore, p.Leu624Arg should be considered a pathogenic variant that likely caused the phenotype in the two reported Swiss siblings (Huizing et al., 2004).

Figure 5:

Distribution of HPS5 Gene Variants

There are 10 nonsynonymous HPS5 variants listed in dbSNP with a high ClinVar MAF (>0.001) and low pathogenicity scores (Supp. Table S3). These variants are likely benign SNPs with a small effect on protein function.

Like other BLOC-2 deficient individuals, HPS-5 subjects exhibit a relatively mild phenotype of hypopigmentation and bleeding diathesis. Pulmonary fibrosis or immunodeficiency has not been reported in HPS-5 subjects, and granulomatous colitis occurs some BLOC-2 deficient subjects. The fact that individuals with BLOC-2 deficiency can escape diagnosis or go undiagnosed for decades was illustrated by new diagnoses of HPS-5 in a 92-year-old man, the oldest reported individual with HPS, who had light skin and hair, nystagmus, decreasing visual acuity with age, and a life-long bleeding history (Ringeisen, Schimmenti, White, Schoonveld, & Summers, 2013), and in a 65 year-old man with oculocutaneous albinism and a mild bleeding diathesis (Botero et al., 2018).

HPS-6 (Table 8)

HPS-6 (MIM# 614075) is caused by biallelic pathogenic variants in HPS6, also called BLOC2S3, on chromosome 10q24.32. HPS6 is a one-exon gene, with one mRNA transcript, encoding a 775-amino acid protein (~83.0-kD). HPS6 protein interacts with the HPS3 and HPS5 proteins in BLOC-2 (Di Pietro et al., 2004). Vertebrate models of HPS-6 include the ruby-eye mouse (Zhang et al., 2003) and the no privacy frog (Nakayama et al., 2016).

The HPS6 gene was discovered by finding Hps6 mutated in the ruby-eye HPS mouse model, and subsequent sequencing of unclassified HPS cases identified one individual with a homozygous 4-bp deletion in HPS6 (Zhang et al., 2003). An extended Israeli Muslim Bedouin family had ~ 20 affected individuals homozygous for c.1065insG in HPS6 was reported (Schreyer-Shafir et al., 2006). Another ~45 subjects with HPS6 variants are reported, and there is one unreported case in our NIH cohort; they are from various ethnic backgrounds, including Arabic, Afghan, Caucasian, Chinese, Japanese and Pakistani (Table 8). The HPS6 variant c.2038C>T (p.Gln680*), is reported in ClinVar/gnomAD/dbSNP (rs1131692333) with a very low MAF (0.00001), occurs in 5 Japanese subjects (including 2 sisters) and appears to be a frequent Japanese variant. We report 45 pathogenic HPS6 variants associated with the HPS-6 phenotype (Table 8), including 2 novel variants identified in our NIH HPS cohort. The variants are located throughout the entire coding exon and include 9 (20%) frameshift, 11 (24.5%) missense, 11 (24.5%) nonsense, 14 (31%) insertions and/or deletions, and no splice site variants (Figure 6). Three of 11 reported HPS6 missense variants are predicted to be likely pathogenic; all others are classified as VUS and require experimental pathogenicity evidence and/or familial genetic testing to increase pathogenicity predictions (Supp. Table S1).

Figure 6:

Distribution of HPS6 Gene Variants

Next generation sequencing of a cohort of BRCA1 and BRCA2-negative subjects with breast cancer identified a heterozygous HPS6 stop-loss variant (c.2326T>C, p.*776Arg) in one subject, who also carried heterozygous protein damaging variants in two other genes (Shahi et al., 2019). This variant has not been reported in HPS subjects but is included in Supp. Table S4 because it may cause HPS when occurring in a homozygous or compound heterozygous state.

The dbSNP/gnomAd/ClinVar databases list 4 frequently occurring (MAF > 0.001) HPS6 missense variants, with predicted benign or unknown pathogenicity (Supp. Table S3); these should be considered when encountered in HPS6 genetic analysis.

Like the other two BLOC-2 HPS disorders (HPS-3 and HPS-5 subtypes), HPS-6 subjects exhibit a mild phenotype of hypopigmentation and bleeding diathesis. Pulmonary fibrosis and immunodeficiency have not been reported in HPS-6 and granulomatous colitis occurs in some BLOC-2 deficient subjects. As in other BLOC-2 subtypes, mildness of symptoms may prevent or delay diagnosis.

HPS-7 (Table 9)

Table 9:

Pathogenic Gene Variants Associated with Hermansky-Pudlak Syndrome Types 7 through 10 (HPS-7 - HPS-10)

#	mRNA	Amino Acid	Exon/Intron	Variant Type 1	Reference SNP (dbSNP) 2	Ethnic Background (age, gender) 3	References and Footnotes
HPS-7: DTNBP1 (BLOC1S8, Dysbindin)
	NM_032122.4	NP_115498.2
1	c.177G>A	p.Trp59*	Exon 4	Nonsense	rs727502866	Caucasian (77y, F)	(Lowe et al., 2013) 4
2	c.307C>T	p.Gln103*	Exon 5	Nonsense	rs104893945	Portuguese (48y, M), Paraguayan (6y, M) Portuguese (26y, M;56y, F) Portuguese (18y, F) Argentinian (M)	(Li et al., 2003) (Bryan et al., 2017) 5 (Bastida et al., 2019) 5 (Bastida et al., 2019) 5 unreported 5,6
3	c.771_774delCTCT	p.Asn257Lysfs*13	Exon 9	Indel	-	- (1 case)	(Lasseaux et al., 2018) 5
4	c.1017_1020delAGAG	p.Glu340Profs*44	Exon 10	Indel	-	Argentinian (M)	unreported 5,6
HPS-8: BLOC1S3 (Reduced Pigmentation)
	NM_212550.4	NP_997715.1
1	c.131C>A	p.Ser44*	Exon 2	Nonsense	rs281865115	Iranian (6y, M)	(Cullinane et al., 2012)
2	c.385_403del19	p.Ser129Glnfs*90	Exon 2	Indel	-	- (1 case)	(Lasseaux et al., 2018) 5
3	c.444_467del24	p.Gln150_Ala157del	Exon 2	Indel	rs754841982	- (1 case)	(Lasseaux et al., 2018) 5
4	c.448delC	p.Gly150Argfs*75	Exon 2	Frameshift	rs281865116	Pakistani (6 familial cases)	(Morgan et al., 2006) 7
HPS-9: BLOC1S6 (PLDN)
	NM_012388.3 8	NP_036520.1
1	c.232C>T	p.Gln78*	Exon 2	Nonsense	rs201348482	Italian (17y, F), Pakistani (4y, F) Indian (9mo, M)	(Badolato et al., 2012) 5,9 (Yousaf et al., 2016) 5 (Cullinane et al, NIH unpublished)
2	c.285_286dupTC	p.His96Leufs*22	Exon 3	Indel	-	Japanese (52y, F)	(Okamura et al., 2018) 5,9
HPS-10: AP3D1
	NM_001261826.3	NP_001248755.1
1	c.1978delG	p.Ala660Argfs*54	Exon 17	Frameshift	-	3 siblings	(Mohammed et al., 2018) 5,10
2	c.3565_3566delGT	p.Val1189Leufs*8	Exon 32	Indel	rs879255646	Turkish (3.5y, M)	(Ammann et al., 2017; Ammann et al., 2016) 5,11

¹When deletion/insertion is 1 nucleotide it is named Frameshift, when larger it is named Indel.

²Reference SNP numbers for each variant as listed in dbSNP (https://www.ncbi.nlm.nih.gov/snp; searched November 2019), none of the listed SNPs in this Table has a reported allele frequency, suggesting they are rare variants.

³Extracted from literature reference. Age (y, years at reporting) and gender (M, male; F, female) of each subject is included for future comparison of additional cases. ‘-‘ = unreported.

⁴This subject also developed granulomatous colitis (Lowe et al., 2013).

⁵The variant in this report was identified by next generation sequencing.

⁶This DTNBP1 variant was identified by exome sequencing in one unreported male subject of Argentinian descent in the NIH HPS cohort referred by Dr. Rosenzweig, NIH Clinical Center, NIH, Bethesda, MD). This subject is compound heterozygous for c.307C>T and c.1017_1020delAGAG.

⁷One extended Pakistani family with 6 affected HPS-8 cases homozygous for c.448delC was reported (Morgan et al., 2006).

⁸Both BLOC1S6 pathogenic variants were reported according to transcript variant 2 (NM_012388.3) nomenclature and are listed as such in this Table to avoid confusion. A longer transcript variant (Variant 1, NM_001311255.1), appeared recently in databases, and future reports may adjust variant nomenclature.

⁹Both the Italian and Japanese subjects had a history of recurrent leucopenia and mild thrombocytopenia, causing immunodeficiency (Badolato et al., 2012; Okamura et al., 2018). The Japanese subject developed schizophrenia in her late forties, a phenotype also associated with DTNBP1 haplotypes.

¹⁰Next generation sequencing identified this rare AP3D1 frameshift variant c.1978delG homozygous in 3 siblings with seizures, developmental delay, albinism and immunodeficiency. Twin girls died before 6 days of age and their brother died at age 2 years of pneumonia and sepsis (Mohammed et al., 2018).

¹¹Next generation sequencing identified this rare AP3D1 variant c.3565_3566delGT in a proband with albinism, neutropenia, immunodeficiency, neurodevelopmental delay, generalized seizures, and impaired hearing. Immunologic investigations excluded HLH in this subject. The proband died at age 3.5 years as result of septic pneumonia (Ammann et al., 2016).

HPS-7 (MIM# 614076) is caused by biallelic pathogenic variants in DTNBP1, also called BLOC1S8 or HPS7, located on chromosome 6p22.3. The longest DTNBP1 mRNA transcript contains 10 exons, encoding a 351-amino acid protein (~39.5-kD) called Dysbindin or HPS7. There are 5 DTNBP1 protein-coding mRNA variants predicted; their expression and functional significance are unknown. Dysbindin is a subunit of BLOC-1 (Li et al., 2003). The sandy mouse is the murine orthologue of human HPS-7 (Li et al., 2003).The DTNBP1 gene was discovered by finding Dtnbp1 mutated in the sandy mouse, and subsequent sequencing of unclassified HPS subjects identified a 48-year old Portuguese female with a homozygous nonsense variant c.307C>T (p.Gln103*) in DTNBP1 (Li et al., 2003). We list 4 DTNBP1 pathogenic variants that cause the HPS-7 phenotype (Figure 7). There are currently seven HPS-7 cases reported, all homozygous for a nonsense or a frameshift DTNBP1 variant, including a Caucasian female diagnosed at age 77 (Lowe et al., 2013) and 4 cases (including 2 siblings) of Portuguese ethnicity homozygous for c.307C>T (p.Gln103*) (Bastida et al., 2019; Bryan et al., 2017; Li et al., 2003) (Table 9). One unreported Argentinian boy in our NIH cohort (referred by Dr. Rosenzweig, NIH Clinical Center, Bethesda, MD), was compound heterozygous for c.307C>T (p.Gln103*) and the novel indel variant c.1017_1020delAGAG (p.Glu340Profs*44).

Next generation has identified heterozygous DTNBP1 variants suggested to contribute to different conditions (Supp. Table S4). Specifically, c.286G>T (p.Glu96*) was reported in a subject with idiopathic pulmonary fibrosis that had no other HPS clinical findings (Deng et al., 2018). The missense variant p.Pro272Ser (c.814C>T) was reported as a low penetrance risk for colorectal cancer (Webb et al., 2006). Several association studies have identified DTNBP1 as a risk allele for schizophrenia, including in European-Americans (Donohoe et al., 2008; Straub et al., 2002; Wang, Xu, Lazarovici, & Zheng, 2017; Zuo et al., 2009). Schizophrenic subjects have reduced hippocampus DTNBP1 mRNA expression (Weickert, Straub, Kleinman, Hyde, & Rothmond, 2006) and DTNBP1 variants are associated with a cognitive response to antipsychotic drug treatment (Scheggia et al., 2018), however, the association of DTNBP1 with schizophrenia has also been challenged (Ghiana & Dell’Angelica, 2011). Notably, no psychiatric illness in HPS-7 subjects has been reported but it is reasonable to consider such features in HPS-7 and BLOC-1 cases.

The dbSNP/gnomAd/ClinVar databases list 4 frequently occurring (MAF > 0.001) DTNBP1 missense variants of predicted benign or unknown pathogenicity (Supp. Table S3), which should be considered as likely non-pathogenic polymorphisms when encountered in DTNBP1 genetic analyses.

Due to the limited number of identified HPS-7 cases, it is difficult to determine whether these individuals are prone to complications other than albinism and a bleeding diathesis. All affected individuals, including two of advanced age, had normal pulmonary function and no signs of immunodeficiency/neutropenia. The 77-year old Caucasian woman had signs of colitis, which was not reported in the other affected subjects.

HPS-8 (Table 9)

HPS-8 (MIM# 614077) is caused by biallelic pathogenic variants in BLOC1S3, also called HPS8, RP or BLOS3, located on chromosome 19q13.32. The longest BLOC1S3 mRNA transcript contains 2 exons, encoding a 202-amino acid protein (~21.3-kD). There exists only one predicted BLOC1S3 mRNA transcript. BLOC1S3/HPS8 is a subunit of BLOC-1 (Starcevic & Dell’Angelica, 2004). Reduced pigmentation is the murine orthologue of human HPS-8 (Gwynn et al., 2004).

HPS-8 was first reported in an extended consanguineous Pakistani family with 6 affected individuals; autozygosity mapping assisted in identifying a homozygous frameshift variant, c.448delC (p.Gly150Argfs*75), in BLOC1S3 in all affected individuals (Morgan et al., 2006). Three additional HPS-8 cases with homozygous pathogenic BLOC1S3 variants have since been reported (Cullinane et al., 2012; Lasseaux et al., 2018) (Table 9, Figure 7).

The dbSNP/gnomAd/ClinVar databases list 3 frequently occurring (MAF > 0.001) BLOC1S3 missense variants, with predicted benign or unknown pathogenicity (Supp. Table S3); these should be considered as likely non-pathogenic polymorphisms when encountered in genetic analyses.

The few reported HPS-8 cases show typical HPS features including hypopigmentation and a bleeding diathesis. None has been reported to exhibit pulmonary fibrosis, granulomatous colitis, immunodeficiency or other complications. Identification of additional HPS-8 cases may confirm or broaden this phenotype.

HPS-9 (Table 9)

HPS-9 (MIM# 614171) is caused by biallelic pathogenic variants in BLOC1S6, also called HPS9, PLDN or BLOS6, located on chromosome 15q21.1. The longest BLOC1S6 mRNA transcript contains 5 exons encoding a 177-amino acid protein (~20.3-kD), called BLOC1S6, HPS9 or Pallidin. There are 3 predicted BLOC1S6 mRNA transcripts with unknown expression patterns and function. The longest variant 1 (NM_001311255.1) only recently appeared in databases, causing all previously described pathogenic variants to be attributed to mRNA splice variant 2 (NM_012388.3). Variant 1 (open reading frame 534-bp) and Variant 2 (open reading frame 519-bp) both have 5 exons and vary in their 5’ UTR and 5’ coding region in exon 1, and each initiate translation at a different start codon; exons 2–5 are identical in both variants. BLOC1S6/HPS9 is a subunit of BLOC-1 (Falcon-Perez et al., 2002; Moriyama & Bonifacino, 2002). Pallid is the murine orthologue of human HPS-9 (Moriyama & Bonifacino, 2002).

All 3 reported HPS-9 cases were identified through exome sequencing. A 17-year-old Italian female (Badolato et al., 2012) and a 4-year-old Pakistani female (Yousaf et al., 2016) were both homozygous for the same HPS9 nonsense variant: c.232C>T, p.Gln78* (NM_012388.3, transcript variant 2). A 52-year-old Japanese female was reported homozygous for c.285_286dupTC, p.H96Lfs*22 (NM_012388.3) (Okamura et al., 2018) (Table 9, Figure 7).

The dbSNP/gnomAd/ClinVar databases list one frequently occurring (MAF > 0.001) BLOC1S6 missense variant (p.Ala12Thr; MAF 0.0030), with predicted benign or unknown pathogenicity (Supp. Table S3); it should be considered as likely non-pathogenic polymorphism when encountered in genetic analyses.

All three HPS-9 subjects exhibited hypopigmentation, visual impairment and a bleeding diathesis. No pulmonary fibrosis or granulomatous colitis was reported in HPS-9 subjects. However, the Italian (17-year-old female) and Japanese (52-year-old female) subjects both had mild thrombocytopenia and recurrent leukopenia, causing immunodeficiency (Badolato et al., 2012; Okamura et al., 2018). In addition, the Japanese subject developed schizophrenia in her late forties, a phenotype previously associated with DTNBP1 haplotypes (Donohoe et al., 2008; Straub et al., 2002; Wang et al., 2017), however, this association appears controversial and needs further consideration(Ghiani & Dell’Angelica, 2011). Both immunodeficiency and schizophrenia should be considered in future evaluations of HPS-9 subjects.

HPS-10 (Table 9)

HPS-10 (MIM# 617050) is caused by biallelic pathogenic variants in AP3D1, also called HPS10 or ADTD, located on chromosome 19p13.3. The longest AP3D1 mRNA transcript contains 32 exons, encoding the 1215-amino acid protein AP3D1 (δ subunit of AP-3; ~136.7-kD). There are 2 predicted protein-coding AP3D1 mRNA transcripts without reported expression or functional data. The AP3D1 protein is a subunit of AP-3 (Ammann et al., 2016; Dell’Angelica et al., 1999). Ap3d1-deficient animal models include the mocha mouse (Kantheti et al., 1998).

There are only 2 AP3D1 pathogenic variants reported to cause the HPS-10 phenotype (Table 9, Figure 7). A homozygous AP3D1 pathogenic indel variant c.3565_3566delGT (p.Val1189Leufs*8) was reported in a Turkish boy, who died at age 3.5 years as result of septic pneumonia (Ammann et al., 2016). The boy had albinism, neutropenia, immunodeficiency, neurodevelopmental delay, generalized seizures, interstitial lung disease and impaired hearing. T cells from the AP3D1-deficient boy showed significantly decreased AP3D1 protein expression compared to healthy control T cells; protein expression levels of other AP-3 subunits (β3A, σ, and μ) were also reduced, consistent with an unstable AP-3 heterotetramer (Ammann et al., 2016). Cytotoxic lymphocytes from the subject exhibited an impaired degranulation response, similar to individuals with pathogenic variants in AP3B1. Immunologic investigations excluded HLH in the subject (Ammann et al., 2016; Enders et al., 2006; Jessen et al., 2013). A homozygous AP3D1 pathogenic frameshift variant, c.1978delG (p.Ala660Argfs*54), was identified in 3 siblings with seizures, developmental delay, albinism and immunodeficiency; twin girls died before 6 days of age and their brother died at age 2 years of pneumonia and sepsis (Mohammed et al., 2018).

Next generation sequencing of cohorts with autism spectrum disorder identified a de novo heterozygous AP3D1 splicing variant (c.273+1G>T) (Takata et al., 2018) and a de novo AP3D1 heterozygous missense variant (p.Gln406Arg) (Iossifov et al., 2014). Next generation sequencing of a schizophrenia cohort identified a de novo heterozygous AP3D1 missense variant (p.Asn605Lys) (Fromer et al., 2014). None of these variants has been reported in HPS subjects but they are included in Supp. Table S4 because they may cause HPS when occurring in a homozygous or compound heterozygous state.

The dbSNP/gnomAd/ClinVar databases list 2 frequently occurring (MAF > 0.001) AP3D1 missense variants, with predicted benign or unknown pathogenicity (Supp. Table S3); these should be considered as likely non-pathogenic polymorphisms when encountered in genetic analyses.

The AP3D1-deficient boys manifested features of albinism and immunodeficiency from birth. These are features characteristic of AP3B1 deficiency (HPS-2) and can likely be attributed to AP-3 deficiency. They did not have an overt tendency for bleeding. They also exhibited neurological findings not previously reported in other HPS subtypes, including microcephaly, severe neurodevelopmental delay, generalized seizures. Hearing impairment occurred in the Turkish boy. The mocha mouse shows a similar HPS-like phenotype with seizures and hearing loss, indicating that these features are likely due to AP3D1 deficiency. A diagnosis of HPS-10 should be considered in individuals with hypopigmentation, immunodeficiency, seizures, hearing loss and possibly other neurologic involvement.

DIAGNOSIS OF HPS

Individuals with HPS can present to different clinical specialties, including dermatology, ophthalmology, pulmonology, hematology, gastroenterology, immunology and neurology. The presence of a combination of hypopigmentation (light hair and skin color), ocular symptoms (nystagmus and decreased visual acuity) (Summers, Knobloch, Witkop, & King, 1988), and a bleeding diathesis (bruising, epistaxis, gingival bleeding, colonic bleeding, prolonged bleeding after minor surgeries) (Gunay-Aygun, Huizing, & Gahl, 2004) leads most physicians to suspect a diagnosis of HPS. The diagnosis of HPS is primarily established by clinical features and platelet phenotyping that show an absence or severe reduction of platelet dense granules, which can be demonstrated by whole mount electron microscopy (EM) (Witkop, Krumwiede, Sedano, & White, 1987). This semi-specialized method is not routinely offered by hematology services. Alternatively, demonstrating an absence of a secondary aggregation response of platelets to exogenous stimuli through platelet aggregation testing also supports the HPS diagnosis (White & Witkop, 1972). Quantification of mepacrine uptake (Billio et al., 2001), and super-resolution immunofluorescence microscopy analyses (Westmoreland et al., 2016) have also been presented as potential alternatives to whole mount EM analysis for diagnosis of dense granule deficiency.

Identification of biallelic variants in one of the 10 HPS-related genes ultimately confirms the HPS diagnosis and the HPS subtype. However, since HPS is a heterogenous genetic disorder, a molecular diagnosis of a particular subtype can be difficult to reach. There is no direct genotype-phenotype association among HPS genes or variants within a gene. When using a single gene testing strategy, we recommend sequence analysis of HPS1 and HPS4 in subjects with more severe clinical manifestations (hypopigmentation, ocular symptoms, bleeding diathesis, pulmonary symptoms) and analysis of HPS3, HPS5 and HPS6 (BLOC-2 subunits), followed by DTNBP1, BLOC1S3 and BLOC1S6 (BLOC-1 subunits) in mildly affected subjects. Subjects with immunological symptoms should be tested for AP3B1 and AP3D1 defects. In this era of next generation sequencing, the genetic diagnosis of HPS is increasingly established by testing all HPS genes simultaneously. This approach can also identify new HPS associated genes, since some subjects with HPS-related symptoms have no apparent pathologic variants in any of the 10 HPS genes. Candidate genes for new HPS subtypes may include genes affected in mouse models of HPS that do not yet have a human counterpart (Table 1), as well as proteins that interact with the BLOC and AP-3 complexes.

The inclusion of HPS-related genes in genetic screening panels of cohorts with clinical features of HPS has resulted in the recent identification of groups of undiagnosed HPS subjects, especially those with milder clinical phenotypes. Targeted sequencing of 990 cases with albinism identified 46 HPS subjects (Lasseaux et al., 2018). A similar study of 21 Arabian individuals with ocular hypopigmentation identified 10 HPS subjects (Khan, Tamimi, Lenzner, & Bolz, 2016), a study of 46 Japanese cases with (OCA-1 and HPS-1 negative) albinism identified 9 HPS subjects (Okamura et al., 2019), and a study of Chinese hypopigmentation cases identified 10 HPS subjects (Wei et al., 2019). Similarly, targeted sequencing of a cohort of 159 cases with bleeding, thrombotic, and platelet disorders identified 6 HPS individuals (Simeoni et al., 2016). These studies suggest that HPS is underdiagnosed, especially when clinical features are mild.

Other recent next generation sequencing approaches identified heterozygous variants in HPS genes that are considered risk alleles for certain conditions (Tables 3–9) these are not classified as HPS-causing pathogenic variants. Heterozygous HPS1, HPS4 and DTNBP1 variants were reported in familial (Stearman et al., 2019) or sporadic (Deng et al., 2018) pulmonary fibrosis cases (Tables 3, 6 and 9). Heterozygous AP3B1 variants have been reported in cases with primary immunodeficiency (Chi et al., 2018; Gallo et al., 2016) or HLH (Gao et al., 2015; Miao et al., 2019; Mukda et al., 2017; Tesi et al., 2015; Xu et al., 2017) (Table 4), and heterozygous variants in AP3D1, HPS3 and HPS4 were reported in cases with autism spectrum disorder or schizophrenia (Fromer et al., 2014; Iossifov et al., 2014; Takata et al., 2018).

Apart from direct sequencing of the exonic regions of HPS genes, alternative molecular methods to establish the HPS type have been used. When a subject’s mRNA has been isolated from whole blood or cultured cells, mRNA expression (i.e., by northern blot or quantitative PCR) and/or cDNA sequencing of each gene can be performed. Another advantage of mRNA availability is that the effects of splicing variants and variants suspected of causing nonsense mediated mRNA decay can be investigated (Anderson et al., 2003; Huizing et al., 2001; Huizing et al., 2004). Immunoblotting of cultured skin fibroblast or platelet rich plasma extracts has also proven helpful to determine or validate the HPS subtype. HPS mouse and human studies have shown that a defect in one HPS protein leads to destabilization of the entire protein complex, i.e., AP-3, BLOC-1, −2, −3 (Ammann et al., 2016; Dell’Angelica et al., 1999; Huizing et al., 2002; Li et al., 2003; Wei et al., 2019). Therefore, the use of immunoblotting with an antibody against one subunit of the complex (AP-3, BLOC-1,−2,−3) allows determination of which complex is defective in unclassified HPS subjects, reducing subsequent sequencing of genes encoding the corresponding subunits (Carmona-Rivera, et al., 2011; Nazarian et al., 2008; Wei et al., 2019).

EPIDEMIOLOGY OF HPS

The worldwide prevalence of HPS is estimated to be 1–9/1,000,000 (Christensen et al., 2017; Huizing et al, 2017 Oct 26 [Updated 2000 July 24]). A few ethnic founder variants occur in HPS genes, in particular in northwest region of Puerto Rico, an estimated 400 individuals are affected (~1/1,800 affected; carrier frequency 1:21) and carry a homozygous 16-bp duplication in HPS1 (c.1472_1487dup16-bp; Table 3) (Santiago Borrero et al., 2006; Witkop et al., 1990), and in Central Puerto Rico a 3.9-kb deletion in HPS3 (NM_032383.5(HPS3):c.−2993_217+692del; Table 5) has an estimated population prevalence of ~1/4000 (carrier frequency 1:32) in (Anikster et al., 2001; Santiago Borrero et al., 2006; Torres-Serrant et al., 2010). Other founder variants without frequency estimates have been reported and are discussed elsewhere in this report for each HPS subtype (Oh et al., 1998; Schallreuter et al., 1993; Ito et al., 2005; Schreyer-Shafir et al., 2006).

Compared to the ~189 reported BLOC-3 deficient individuals (not including ~261 cases with the Puerto Rican HPS1 16-bp duplication variant), there are remarkably fewer reported AP-3 deficient individuals (~35 cases), BLOC-1 deficient individuals (~24 cases) and BLOC-2 deficient individuals (~117 cases; not including ~72 cases with the Puerto Rican HPS3 3.9-kb deletion and ~20 cases with the Israeli Bedouin HPS6 c.1065insG variants) (Table 1).

BLOC-2 deficient cases may escape diagnosis because of their much milder hypopigmentation compared to BLOC-3 deficiency. This is in line with BLOC-2 deficient mice, which inhibit brownish-black eumelanin, but not reddish-yellow pheomelanin production (Hirobe, Ito, & Wakamatsu, 2013), and individuals with HPS have reduced levels of total melanin, but increased pheomelanin production compared to unaffected family members (Okamura et al., 2018), indicating that hypopigmentation might not be detected in HPS cases from fair-skinned families, particularly those with red-blonde hair.

BLOC-2 deficient individuals may get medical attention only for a bleeding diathesis and may be classified as storage pool deficiency, as illustrated by new diagnoses of HPS-5 in 92- and 65-year-old individuals with histories of excessive bleeding (Botero et al., 2018; Ringeisen et al., 2013).

Similarly, AP-3 deficient individuals may avoid an HPS diagnosis or have it delayed because a severe immune disorder dominates medical attention. It is puzzling why so few BLOC-1 deficient individuals are reported. Perhaps BLOC-1 defects are extremely rare or embryonically lethal (although mouse models of BLOC-1 deficiency are viable), or the human phenotypes may have features that are not considered compatible with the HPS phenotype. For example, the absence of platelet delta granules is currently essential for the HPS diagnosis, but a subset of subjects with BLOC-1 may have normal/decreased platelet delta granules and may therefore not be considered for a HPS diagnosis. In addition, reported brain-associated functions of BLOC-1 (A. Ito et al., 2018; Newell-Litwa et al., 2010; Spiegel, Chiu, James, Jentsch, & Karlsgodt, 2015) may underlay a neurologic phenotype of BLOC-1 deficiency. Recognition of these features may facilitate the diagnosis of affected individuals with HPS subtypes associated with BLOC-1 defects. Several association studies have identified DTNBP1 variants as risk alleles for schizophrenia (Donohoe et al., 2008; Straub et al., 2002; Wang et al., 2017; Zuo et al., 2009) and one HPS-9 subject was diagnosed with oculocutaneous albinism and developed schizophrenia in her forties; she was only diagnosed with HPS-9 at age 52 through a whole exome screen (Okamura et al., 2018). However, some reports have challenged brain-associated functions of BLOC-1 and/or a BLOC-1 association with schizophrenia (Ghiani & Dell’Angelica, 2011), suggesting that this issue needs further consideration. With the increased availability of exome/genome sequencing, more HPS subjects with mild or atypical HPS phenotypes will likely be identified and will expand the clinical spectrum of HPS.

CLINICAL MANAGEMENT AND THERAPEUTIC ASPECTS

The myriad of symptoms associated with HPS, some subtype or age specific and some life threatening, require multidisciplinary clinical care (Table 2) (Christensen et al., 2017; Seward & Gahl, 2013).

All individuals with HPS exhibit some degree of albinism, involving hypopigmentation of skin, hair or eyes. The skin is often light- and sun-sensitive and may develop solar keratoses and melanocytic nevi. Subjects are at increased risk for squamous cell carcinoma, basal cell carcinoma, and possibly melanoma. Protection or avoidance from ultraviolet radiation is critical (Toro, Turner, & Gahl, 1999). The hair varies from silvery-white to light brown and hypopigmentation may be evident when affected and unaffected family members are compared. The eyes appear light blue, light green or hazel, but iris color may be darker in mild HPS cases. Iris transillumination is found in most HPS subjects. Due to their albinism, HPS subjects have abnormal crossing of optic nerve fibers (Hoffmann, Lorenz, Morland, & Schmidtborn, 2005) and horizontal nystagmus, the retinal fundus appears pale and visual acuity ranges from 20/60 to 20/400 and can be mildly improved with refractive lenses (Summers et al., 1988). Many subjects are legally blind, i.e., have visual acuity worse than 20/200. BLOC-3 deficient subjects show more severe hypopigmentation than subjects with AP-3 or BLOC-2 defects. The albinism of BLOC-1 individuals has not been well characterized.

The bleeding diathesis of HPS varies in severity in all subtypes, and may include spontaneous bruising, prolonged epistaxis, menorrhagia, pronounced oozing after dental extractions, and excessive surgical blood loss. Topical thrombin, administration of pro-coagulant drugs, or intravenous 1-desamino-8D-arginine vasopressin may ameliorate or prevent bleeding. Platelet transfusion may also be used as prophylaxis or as treatment for bleeding in individuals with HPS (Han et al., 2018; Minkin, Bertetti, Lindsey, & Bovino, 2015; Ozgur & Yilmaz, 2015; Van Avermaete, Muys, & Jacquemyn, 2016). Avoidance of aspirin products and non-steroidal anti-inflammatory drugs is recommended.

A granulomatous colitis involving intestinal granulomas, erosions and inflammatory cells, which resembles Crohn’s disease, occurs in ~10–20% of all BLOC-2 and BLOC-3 subjects (Hussain et al., 2006). One BLOC-1 deficient individual (HPS-7) was diagnosed with Crohn’s colitis in adulthood (Lowe et al., 2013). It is unknown if colitis occurs in AP-3 deficiency. The colitis may respond to corticosteroids or anti-TNF-α drugs; surgical bowel resection is performed in refractory cases (Demirtas, Alahdab, Kani, Atug, & Imeryuz, 2019; Kouklakis et al., 2007; Mora & Wolfsohn, 2011). Abnormal endosomal membrane formation was suggested as an underlying cause for HPS colitis leading to ceroid lipofuscin formation, abnormal autophagy and phagocytosis, and inflammation (Felipez, Gokhale, & Guandalini, 2010; Sofia, Sakuraba, & Rubin, 2017). It was suggested that the presence of risk alleles in Crohn’s disease-associated genes, like NOD2 or ATG16L1, in HPS subjects may contribute to developing colitis (Lozynska et al., 2018). Low vitamin D levels, which may be a factor in HPS subjects avoiding sun exposure, could also contribute to developing colitis (Lozynska et al., 2018).

Ceroid lipofuscin, an amorphous, granular, electron-dense, autofluorescent lipid-protein material, was identified in LROs in HPS cell types, including alveolar macrophages, cells of the gastrointestinal tract, renal tubular cells, bone marrow, lymph nodes, liver, spleen, and heart (Gahl et al., 1998; Harada et al., 2014; Hermansky & Pudlak, 1959; Sparrow et al., 2010; Takahashi & Yokoyama, 1984). Ceroid lipofuscin may accumulate because cells cannot rapidly degrade mistargeted vesicular membranes. Accumulation of ceroid lipofuscin was suggested to underlie the colitis and pulmonary fibrosis in HPS, but this has not been confirmed and other pathogenic mechanisms have been proposed. End stage renal disease attributed to deposition of ceroid lipofuscin has occurred in a few HPS subjects, for which renal transplant is a treatment option (Abdullah, Davis, Quinn, & Mohan, 2018; Gordillo, Del Rio, Thomas, Flynn, & Woroniecki, 2011; Tagboto et al., 2001).

Immunodeficiency and/or neutropenia occurs in AP-3-deficient HPS, resulting in susceptibility to infections (subtypes HPS-2 and HPS-10) (Ammann et al., 2016; de Boer et al., 2017; Fontana et al., 2006; Huizing et al., 2002; Mohammed et al., 2018). No BLOC-2 or BLOC-3 deficient subjects are described with immunodeficiency, but it was reported in two unrelated individuals with HPS-9 (Badolato et al., 2012; Okamura et al., 2018) and therefore needs consideration in future BLOC-1 deficient individuals.

Manifestations of the AP-3 immunodeficiency can vary from mild recurrent viral and bacterial infections to severe hemophagocytic lymphohistiocytosis (HLH) (Dell’Acqua et al., 2019; Enders et al., 2006; Jessen et al., 2013). The neutropenia associated with AP3B1 (HPS-2) deficiency is granulocyte colony-stimulating factor (G-CSF) responsive, but G-CSF therapy has not been used in AP3D1-deficienct subjects, all of whom died before age 3.5 of pneumonia and/or sepsis without signs of HLH. Of note, G-CSF therapy restores the neutrophil numbers, but not the recurrent infections in HPS-2 subjects (Fontana et al., 2006), suggesting defects in innate immunity and bacterial antigen presentation, supported by several studies in AP3-deficient dendritic, natural killer and lymphoblastoid cells from mice and humans (Briken, Jackman, Dasgupta, Hoenig, & Porcelli, 2002; Fontana et al., 2006; Meantegazza et al., 2012; Mantegazza et al., 2017; Sugita et al., 2002; Sasai et al., 2010).

A few AP3B1-deficient subjects developed HLH, which was lethal in 2 subjects (Dell’Acqua et al., 2019; Enders et al., 2006; Fontana et al., 2006). Although the risk of HLH should be considered in AP-3 deficient subjects, preemptive hematopoietic stem cell transplantation (HSCT, therapeutic for HLH but challenging for the subject) has been deemed to not be justified. HSCT could certainly be considered after a severe HLH episode (Dell’Acqua et al., 2019).

Pulmonary fibrosis, a progressive interstitial lung disease with a variable time course, occurs in BLOC-3 (HPS-1 and HPS-4) and AP-3 (HPS-2 and HPS-10) subjects. There are no reports of pulmonary fibrosis in BLOC-1 and BLOC-2 cases. Most BLOC-3 deficient subjects develop pulmonary fibrosis in middle age (30–50 years) and progress to death within a decade (Brantly et al., 2000; Gahl et al., 1998). AP-3 deficiency-related pulmonary fibrosis is reported in a few HPS-2 cases and one HPS-10 case, and symptoms can start as early as childhood (Ammann et al., 2016; Gochuico et al., 2012; Hengst et al., 2018). While the natural history of BLOC-3-related pulmonary fibrosis has been well-documented, that of AP-3 related lung disease needs further elucidation with longitudinal data from more subjects (Gochuico et al., 2012). The exact cause of lung disease in HPS remains unknown; it was suggested that altered LRO formation within alveolar epithelial type II cells may lead to defective formation of lamellar bodies and/or intracellular processing of surfactant proteins, leading to endoplasmic reticulum-stress, apoptosis, and a fibrotic lung phenotype (Guttentag et al., 2005; Mahavadi et al., 2010; Kook et al., 2018). The fibrotic lung phenotype in BLOC-3- or AP-3-deficient mice was shown to be due to a non-hematopoietic cell type and could be averted in Ap3b1-deficient mice by re-expression of Ab3b1 specifically in lung epithelial type 2 cells (Young et al., 2012), strongly supporting the lamellar body defect as causative for lung disease, at least in mice. Abnormal alveolar macrophage or mast cell function was also suggested to underlie HPS-related pulmonary fibrosis (Kirshenbaum et al., 2016; Mahavadi et al., 2010; Nakatani et al., 2000; Rouhani et al., 2009).

No approved medical therapy for or prophylaxis against HPS-related pulmonary fibrosis exists. Maximizing pulmonary function before onset of pulmonary fibrosis by avoidance of cigarette smoke and other lung toxins, treatment of pulmonary infections, influenza and pneumococcal immunization, and regular moderate exercise are recommended. Steroids have no apparent beneficial effect. The anti-fibrotic drug pirfenidone may slow the progression of HPS-related pulmonary fibrosis in some cases (Gahl et al., 2002; O’Brien et al., 2011; O’Brien, Introne, et al., 2018), but it is not an approved therapy for HPS-related pulmonary fibrosis (O’Brien, Gahl, & Gochuico, 2018). Lung transplant is the only known treatment for pulmonary fibrosis, and several individuals with HPS1-related pulmonary fibrosis successfully underwent bilateral or single-lung transplantation (El-Chemaly et al., 2018; Gahl et al., 2002; Lederer et al., 2005).

HPS pathogenesis and therapeutic options continue to be investigated, including through the use of organoids (Korogi et al., 2019; Strikoudis et al., 2019) and explorations of gene therapy (Ikawa et al., 2015; Iyer et al., 2019; Shen et al., 2018).

CONCLUSION

HPS is a rare autosomal recessive disorder characterized by genetic and phenotypic heterogeneity. With the recent rapid evolution of affordable next generation sequencing methods, there is an increased recognition of HPS subjects and new HPS genetic subtypes (Arcot Sadagopan et al., 2017; A. Wei et al., 2016; Yousaf et al., 2016). This is also evidenced by the increased diagnosis of subjects with non-classic HPS phenotypes (i.e. HPS-2, HPS-7, HPS-8, HPS-9, HPS-10) (Ammann et al., 2016; Bryan et al., 2017; Cetica et al., 2015; Iwata et al., 2017; Okamura et al., 2018), as well as diagnosis of ‘unexpected’ HPS in cohorts with albinism (Ito et al., 2005; Khan et al., 2016), immunodeficiency (Badolato et al., 2012), ocular disease (Hull et al., 2016; Miyamichi et al., 2016), or platelet disorders (Jones et al., 2012).

These recent developments have created the need for a current overview of molecular diagnostic and genetic counseling aspects of HPS. We intend this report to serve as a reference for interpretation of molecular data for HPS. The extensive HPS mutational spectrum provides pathogenicity interpretation for future HPS-related variants (such as missense variants currently classified as VUS), phenotype-genotype relationships (in particular atypical symptoms such as autism, schizophrenia and immune deficiency), assistance in genetic counseling for affected individuals (surveillance, management of anticipated symptoms), and tools for cell biologists to elucidate pathways, investigate interactions of HPS-related proteins, and initiate therapeutic efforts.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available in the the Leiden Open Variation Database (http://www.lovd.nl/) and in ClinVar (https://www.ncbi.nlm.nih.gov/clinvar/). The data that support the findings of this study are also available from the corresponding author upon reasonable request.