Analysis of Ocular Hypopigmentation in Rab38 ^cht/cht Mice

Abstract

purpose. To characterize the ocular phenotype resulting from mutation of Rab38, a candidate gene for Hermansky-Pudlak syndrome.

methods. Chocolate mice (cht, Rab38 ^cht/cht ) and control heterozygous (Rab38 ^{cht/ +}) and wild-type mice were examined clinically, histologically, ultrastructurally, and electrophysiologically. Mice homozygous for both the Rab38 ^cht and the Tyrp1 ^b alleles were similarly examined.

results. Rab38 ^cht/cht mice showed variable peripheral iris transillumination defects at 2 months of age. Patches of RPE hypopigmentation were noted clinically in 57% of Rab38 ^cht/cht eyes and 6% of Rab38 ^{cht/ +} eyes. Rab38 ^cht/cht mice exhibited thinning of the iris and RPE and larger b-wave amplitudes in the scotopic range when compared with the control animals. Compared with wild-type mice, Rab38 ^cht/cht melanosomes were smaller and there were fewer in neuroectodermally derived retinal pigment epithelium; in neural crest-derived choroid melanocytes, they were smaller in size only. Mutation of both Rab38 and Tyrp1 produced mice with ocular and coat color pigment dilution greater than that seen with either mutation alone. Comprehensive clinical and pathologic analyses showed no other organ system or blood defects in Rab38 ^cht/cht mice.

conclusions. Rab38 ^cht/cht mice show ocular characteristics reminiscent of human oculocutaneous albinism, as well as iris and RPE thinning. The synergistic effects of the Rab38 ^cht and Tyrp1 ^b alleles suggest that TYRP1 is not the only target of RAB38 trafficking. This mouse line provides a useful model for studying melanosome biology and its role in human ocular diseases.

The analysis of mice that exhibit defects in coat coloration (coat color mutants) has aided in the identification of genes important in eye, skin, and hair pigmentation. ¹ Many of these genes are mutated in patients with pigmentary anomalies. Coat color mutants can exhibit a wide range of variation, including altered, dilute, or absent coloration. Skin and hair coloration result from the pigment cells in hair follicles (melanocytes) that synthesize a melanin-based pigment. Melanin is also produced in the retinal pigment epithelium (RPE) and choroid of the eye. Melanocytes in skin, hair follicles, and choroid are derived from the neural crest (NC), a transient population of stem cells that arise early in development at the dorsal neural tube. The RPE, in contrast, is derived from the neuroepithelium. 

Melanin production occurs in specialized organelles within pigment cells called melanosomes. The melanin synthetic enzymes (including tyrosinase, tyrosinase-related protein 1 (TYRP1), and dopachrome tautomerase) must be trafficked within the melanocyte to the melanosome to achieve proper melanin production. Variations in the coat color of mouse mutants can result from the absence of pigment cells (white spots), defects in melanogenic enzymes, and aberrant trafficking of melanogenic proteins to the melanosome (color alterations or dilution). 

Genes mutated in mice with coat color variations have often been associated with human pigmentary anomalies with similar phenotypes. These include pigmentary glaucoma, ^{2 3} oculocutaneous and X-linked ocular albinism, ^{4 5 6 7} Hermansky-Pudlak syndrome (HPS), ⁸ and Chediak-Higashi syndrome. HPS is an autosomal recessive disease characterized by variable oculocutaneous albinism (including foveal hypoplasia, nystagmus, and iris transillumination defects), absent platelet-dense bodies (leading to prolonged bleeding times), and sporadic lung fibrosis. ^{9 10} Genes for 16 hypopigmented mouse mutants with platelet-mediated bleeding defects have been cloned, and mutations in the human orthologous genes, most of which are involved in organelle trafficking, were subsequently found in humans with HPS. ^{11 12 13}  

Similarly, Tyrp1 ^b/b mice carry a mutation in a melanin synthetic enzyme and exhibit a brown coat on a wild-type black strain background. Patients with oculocutaneous albinism type 3 (OCA3) have homozygous TYRP1 mutations, ^{6 14} which result in moderate hypopigmentation. Of course, there can be human and mouse phenotype differences for disruption of the same gene. For example, mutation of Tyrp1 in both humans and mice results in hypopigmentation and iris transillumination. However, DBA/2J mice homozygous for the Tyrp1 ^b allele develop age-dependent iris stromal atrophy, elevated intraocular pressure, and pigmentary glaucoma, which has not been observed in humans with TYRP1 mutations. ^{2 3} This difference in known human and mouse phenotypes may be due to the nature of the mutation and whether or not the mutant TYRP1 protein can still stabilize tyrosinase. Because the mouse iris atrophy phenotype is dependent on pigment production and known human mutations induce ocular albinism, it has been suggested that the human mutations are self-rescuing with respect to iris atrophy. ^{2 3}  

The chocolate (Rab38 ^cht/cht ) mouse mutant arose spontaneously on the C57BL/6J black background as a dark brown coat color variant. A G19V point mutation in a highly conserved amino acid of Rab38 is responsible for the chocolate phenotype. ^{15 16} Rab38 ^cht/cht mice have normal blood clotting times. In addition to the chocolate mouse, RAB38 is altered in a rat coat color mutant called Ruby (red-eyed dilution, R) which has been proposed as an animal model for HPS. ¹⁷ Ruby rats have hypopigmented eyes and coat and a bleeding diathesis. The Ruby Rab38 translation-initiation codon has a missense mutation that is predicted to stop translation at the first codon and RAB38 protein is not detected. However, unlike in patients with HPS, platelet-dense granules are present with normal appearance and numbers in Fawn-hooded hypertensive rats, which also carry a first codon missense mutation in Rab38 and have a Ruby phenotype of hypopigmentation and platelet storage pool defect. ^{18 19}  

Rabs are small GTP-binding proteins involved in vesicular transport, motility, and fusion in the secretory and endocytic pathways of cells. ^{20 21} The precise function of RAB38 remains unknown, although it appears to be important in melanogenesis and necessary for proper targeting of TYRP1 protein in melanosomes. ¹⁶ Given the role of Rab proteins in trafficking and the association of alterations of Rab38 with coat color variants, Rab38 ^cht/cht mice may be a model for an HPS-like syndrome, or—like Tyrp1 ^b/b mice—may develop an age-dependent form of pigmentary glaucoma or may display ocular albinism similar to OCA3. 

In this study, the ocular phenotype in Rab38 ^cht/cht , Rab38 ^cht/+, and wild-type mice was assessed by using clinical examination, histopathology, electrophysiology, and ultrastructural techniques. Because preliminary evidence suggested that RAB38 is important in TYRP1 targeting, we examined mice homozygous for both the Rab38 ^cht and Tyrp1 ^b alleles. To investigate whether alterations in RAB38 are associated with human disease, RAB38 was sequenced in a small group of human subjects with ocular and/or systemic pigmentary abnormalities. 

Materials and Methods

Mouse Stocks, Clinical Eye, Organ, and Blood Examination

Rab38 ^cht/cht (stock no. 000976) and wild-type C57BL/6J mice (stock no. 000664) were obtained from The Jackson Laboratory (Bar Harbor, ME). Tyrp1 ^b/b mice were supplied by Lynn Lamoreaux (Texas A & M). The mice were housed according to our Institutional Animal Review Board standards with a 14-hour light–10-hour dark cycle. Both Rab38 ^cht/cht and Tyrp1 ^b/b mice are on a C57BL/6 background, minimizing the potential for phenotypic variation due to strain background. 

Clinical examination of the anterior segment was performed on gently restrained, awake mice with a slit lamp (BQ; Haag-Streit, Mason, OH) and an indirect ophthalmoscope (Keeler, Windsor, Berkshire, UK) with a 90-D condensing lens (Volk, Mentor, OH). The mice were euthanatized with carbon dioxide according to institutional guidelines. “Young” mice were defined as 2 to 3 months of age; “aged” mice were defined as older than 1 year. The fundi were also examined. 

For body organ and blood system analysis, three male and three female 5-month-old Rab38 ^cht/cht mice were compared with six, age- and sex-matched wild-type control mice. We analyzed differences in organ morphology, serum chemistries, hematocrit, and neutrophil and platelet counts between Rab38 ^cht/cht and wild-type mice. Age-related coat color changes were observed by visual comparison of subgroups of mice. 

These studies conformed to the principles for laboratory animal research outlined by the Animal Welfare Act (National Institutes of Health/Department of Health and Human Services) and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care and Use Committee. 

Histopathology, Electron Microscopy, and Melanosome Image Analysis

For light microscopy, mouse eyes were enucleated and fixed in a phosphate-buffered paraformaldehyde–glutaraldehyde mixture, according to published protocols. ²² Hematoxylin and eosin–stained methacrylate-fixed sections from the pupillary–optic nerve axis were used for histopathology. 

For electron microscopy, mouse eyes were dissected and fixed in 4% glutaraldehyde in 0.15% phosphate buffer for 1 hour at room temperature and then transferred to 4% paraformaldehyde for overnight fixation at 4°C. Ocular tissue samples were processed through ascending alcohols, propylene oxide, and a 50:50 mixture of propylene oxide and Ladd LX112 epoxy resin. They were then infiltrated with 100% LX112 and embedded in fresh resin. Samples were cut in a cryotome (Ultracut R; Leica), stained with uranyl acetate and lead citrate, and examined by electron microscope (JEM 1010; JEOL, Tokyo, Japan). 

The density and cross-sectional areas of melanosomes in the RPE and choroid of Rab38 ^cht/cht , Rab38 ^cht/+, and Rab38 ^+/+ mice were measured from randomly selected transmission electron microscopy images at 5000× magnification (AxioVision LE, ver. 4.5; Carl Zeiss Meditec, Dublin, CA). Care was taken to image comparable areas of the posterior pole between different groups of mice. At least three independent images were analyzed for each group. RPE areas analyzed per sample were between 75 and 115 μm². 

Electrophysiology

Electroretinograms (ERGs) were recorded in 2- to 3-month-old homozygous Rab38 ^cht/cht , heterozygous Rab38 ^cht/+, and wild-type (C57BL/6 Rab38 ^+/+) mice. They were dark-adapted for 12 hours before intraperitoneal anesthesia with ketamine (80 mg/kg) and xylazine (4 mg/kg). After instillation of 1% proparacaine anesthetic, the pupils were dilated with topical 0.5% tropicamide and 0.5% phenylephrine HCl. Body temperature was maintained near 38°C with a heating pad. ERGs were recorded simultaneously from both eyes, with gold wire loops placed on the cornea with a drop of methylcellulose. Gold wires were placed on the sclera at the limbus as the differential electrodes, and the ground wire was attached to the left paw. 

Scotopic ERG responses were elicited in the dark-adapted state with single xenon photostrobe flashes (PS33 Photic Stimulator; Grass-Telefactor; West Warwick, RI) delivered in a Ganzfeld light-integrating sphere, with interstimulus intervals of 3 to 60 seconds depending on stimulus intensity. The stimulus intensity range of −6.9 to +0.6 log cd-s/m² was obtained with neutral density (ND) filters (Wratten; Eastman Kodak, Rochester, NY). Responses in a frequency range of 0.1 to 1000 Hz (3-dB cutoff) were amplified 5000 times with a 60-Hz line frequency notch filter (CP511 AC amplifier; Grass-Telefactor). Photopic responses were elicited in a light-adapted state on a rod-suppressing white background of 34 cd/m², with single flashes at 2-second interstimulus intervals. Up to 20 responses were averaged at all intensities tested. The a-waves were measured from the prestimulus baseline to the initial trough. The b-waves were measured either from the baseline or from the a-wave trough when present. Implicit times were measured from flash onset to the a- and b-wave maximum. 

Intensity–response amplitude data were displayed conventionally on log–log coordinates and log-linear coordinates. Response profiles were compared across intensity range by mixed ANOVA (PROC MIXED in SAS for Windows; ver. 9.0.2; SAS Institute, Inc., Cary, NC). 

RAB38 Antibody Production and Protein Detection

Polyclonal affinity-purified rabbit anti-mouse Rab38 was prepared by Bio-Synthesis, Inc. (Lewisville, TX). Rab38 antigen sequence LESIEPDIVKPHLTS, position 188-203, was chosen to minimize any homology with other Rab proteins, using multisequence Rab alignment. ²³  

Western blot analysis was performed on immortalized melanocyte cell lines: control melan-Ink4a-1, which produces normal black pigment, and melan-Ink4a-cht5, which is homozygous for Rab38 ^cht/cht mutation. Decreasing amounts of both protein preparations were loaded on two nondenatured, precast, 4% to 20% Tris-glycine gels (Invitrogen, Carlsbad, CA) and transferred to polyvinylidene difluoride (PVDF) membranes (Invitrogen). The membranes were incubated with 1:100 dilution anti-Rab38 antibody or 1:100 dilution anti-Rab38 and 1:1000 dilution anti-tubulin antibodies (MP Biomedicals, Inc., Aurora, IL). A peroxidase-linked secondary antibody was added (GE Healthcare, Piscataway, NJ) and detected with chemiluminescence (ECL kit; GE Healthcare). 

Human Mutation Analysis

Research subjects with Hermansky-Pudlak syndrome or with a Hermansky-Pudlak–like syndrome (i.e., patients with oculocutaneous albinism and indications of a bleeding defect, but with normal-appearing platelet-dense bodies) and patients with oculocutaneous albinism or pigmentary glaucoma consented to venipuncture and research-based DNA analysis. The subjects' genomic DNA was PCR amplified exon by exon (primers available on request), with standard techniques, ²⁴ and directly sequenced. Automated sequencing was performed (CEQ 2000 sequencer, with CEQ Dye-Terminator Cycle Sequencing kit; Beckman Coulter, Fullerton, CA) according to the manufacturer's protocols. The deposited GenBank sequence (NM_022337) was used as a reference for identifying polymorphisms (http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). 

All human subject research was conducted with the approval of local Institutional Review Boards and was in compliance with the Declaration of Helsinki. 

Results

Hypopigmentation Effects

Iris: Variable Thinning and Hypopigmentation.

We clinically examined the anterior segments of the eyes of Rab38 ^cht/cht mice. Transillumination defects of the irises of each eye were graded at the slit lamp as 0 (nonexistent), +1 (peripheral; Fig. 1A ), and +2 (extending to the central iris; Fig. 1B ). Examination of the anterior segments of young Rab38 ^cht/cht mice (age 1.5–4 months, n = 34 eyes) revealed +1 and +2 iris transillumination in 79% and 21% of eyes, respectively. Transillumination was most prominent peripherally in all cases. Rab38 ^cht/+ mice did not exhibit this phenotype (Fig. 1C , age 1.5–4 months, n = 14 eyes). 

We aged a cohort of Rab38 ^cht/cht and Rab38 ^cht/+ mice until they were older than 1 year and performed clinical examinations. Of the aged Rab38 ^cht/cht mouse eyes, 74% and 26% showed +1 and +2 transillumination, respectively (age 12–24 months, n = 106 eyes). This result was not significantly different from that observed in the younger cohort of mice (P = 0.83). None of the aged Rab38 ^cht/cht mice developed iris holes or correctopia, nor was there evidence of pigment dispersion on the corneal endothelium or the anterior lens capsule. Age-matched Rab38 ^cht/+ mice continued to show no evidence of iris transillumination (n = 28). 

Although measurement of intraocular pressure was not performed, there was no clinical or histologic evidence of glaucoma in Rab38 ^cht/cht mice greater than 1 year of age, as would be observed in Tyrp1 ^b/b mice. Specifically, there was no deepening of the anterior chamber, ectasia of the cornea, or cupping of the optic nerve on clinical examination, nor was there any evidence of significant ganglion cell loss or optic nerve atrophy on histologic sections (data not shown). 

We performed histopathology on Rab38 ^cht/cht and Rab38 ^cht/+ mice in both young and old cohorts of mice. The irides of young Rab38 ^cht/cht mice showed thinning and hypopigmentation compared with age-matched Rab38 ^cht/+ control mice (Fig. 2) . Thinning was most pronounced peripherally and was variable. Different severities of thinning were often observed in the same eye when different sections of the iris were sampled. Examination of the irides of the aged cohort of Rab38 ^cht/cht mice showed no qualitative difference from those in the young cohort, consistent with our clinical observation that there was no atrophic component to this iris thinning (data not shown). 

Fundus and RPE: Focal Depigmentation and Thinning.

Patches of peripheral fundus depigmentation were noted in 17 (50%) of 34 eyes of the young cohort of Rab38 ^cht/cht mice, usually in the superior quadrant of the retina (Fig. 3A) . In Rab38 ^cht/+ mice, 1 (2%) of 42 eyes exhibited a similar phenotype, significantly fewer than in Rab38 ^cht/cht mice (P = 8.8 × 10⁻⁸). Similar areas of depigmentation were present in 30 (57%) of 52 eyes of the aged cohort of Rab38 ^cht/cht mice—a rate comparable to that in young Rab38 ^cht/cht mice (P = 0.47). Qualitatively, these depigmented areas appeared to be of similar size and character in the young and old Rab38 ^cht/cht mice. Wild-type C57BL/6J mice do not exhibit these areas of hypopigmentation (n = 50). Transmission electron micrographs of the junction between normal-appearing and depigmented RPE in Rab38 ^cht/cht mice showed focal hypoplasia of the RPE (Fig. 3B) . 

Altered Melanosomes in Choroid and RPE.

We examined RPE and choroid melanosomes in 3-month-old Rab38 ^cht/cht , Rab38 ^cht/+, and wild-type C57BL/6J eyes by electron microscopy. The RPE of Rab38 ^cht/cht mice showed fewer and smaller mature (stage IV) melanosomes than did the RPE of Rab38 ^cht/+and wild-type mice (Figs. 4A 4B 4C 5 ; Table 1 ). Those mature melanosomes that are present in Rab38 ^cht/cht mice, however, appeared capable of entering the apical villi of the cells. Immature melanosomes did not appear to accumulate in any of the samples examined. 

The melanosomes of the NC–derived melanocytes in the choroid in the Rab38 ^cht/cht mice were smaller in cross-sectional area than were those of the wild-type mice (Figs. 4D 4F 5 ; Table 1 ). The number of melanosomes, however, was unchanged by the presence of the Rab38 ^cht allele. As with RPE melanosomes, the Rab38 ^cht/+ mice showed no significant difference in the size or number of choroidal melanosomes when compared with wild-type (Figs. 4E 4F ; Table 1 ). 

Levels of RAB38 Protein.

To determine the amount of RAB38 protein in wild-type and Rab38 ^cht/cht cells, we developed an anti-RAB38 antibody. On Western blot, anti-RAB38 antibody recognizes a single, correctly sized protein band in a mouse melanocyte cell line (Fig. 6) . The mouse skin melanocyte cell lines we used are homozygous for the Ink4a null mutation, to allow for rapid immortalizing, and are from the same inbred mouse background (C57BL/6). Therefore, melan-Ink4a-1 cells and melan-Ink4a-cht5 cells should differ only at the Rab38 locus. Melan-Ink4a-cht5 cells homozygous for the Rab38 ^cht allele showed much-reduced RAB38 protein expression on Western blot analysis. 

Electrophysiology Consistent with Albinism.

Electroretinography was performed on 2- to 3-month-old wild-type (C57BL/6), Rab38 ^cht/+, and Rab38 ^cht/cht animals. All three animal groups had normal a- and b-wave morphology and timing (data not shown). Compared with wild-type animals, Rab38 ^cht/cht animals showed a statistically significant increase in b-wave amplitude under moderate-intensity, scotopic conditions (Fig. 7A) . There was no statistically significant difference in a-wave amplitude (scotopic or photopic, data not shown) or photopic b-wave (Fig. 7B) . 

Coat Color Change with Age.

Coat color pigment dilution was obvious in the light-brown coats of very young (∼1 month old) Rab38 ^cht/cht mice, unlike the dark coats of Rab38 ^cht/+and wild-type littermates. At approximately 3 months of age, the Rab38 ^cht/cht coat color had darkened, but remained distinguishable from the black coats of the Rab38 ^cht/+ and wild-type littermates (Fig. 8) . 

Tyrp1 and Rab38

Rab38 ^cht/cht mice were mated to mice homozygous for the brown (b) allele of Tyrp1, and the obligate heterozygous offspring were then intercrossed to create double-homozygous Rab38 ^cht/cht ; Tyrp1 ^b/b mice. Because both the Rab38 ^cht/cht mice and the Tyrp1 ^{b/ b} mice had a C57BL/6 background, any change in pigmentation was presumably due to the effect of the mutant alleles. Double homozygous mice (age, 2–3 months) had striking iris transillumination (n = 8 eyes) compared with age-matched Rab38 ^cht/cht and Tyrp1 ^b/b mice, as well as a lighter coat color (Fig. 9) . Histologic examination showed significantly reduced pigmentation, especially in the neuroectodermally derived pigmented tissue (e.g., the posterior pigmented layer of the iris and the RPE). 

Rab38 ^cht/cht and Human Disease

Blood and Organ Systems.

We clinically and pathologically analyzed three male and three female 5-month-old Rab38 ^cht/cht mice and compared them with six age- and sex-matched wild-type control mice (Table 2) . Organ morphology, serum chemistries, hematocrit, and platelet counts were not different between Rab38 ^cht/cht and wild-type mice. The neutrophil count in Rab38 ^cht/cht mice was significantly higher than that in wild-type mice, but within the normal range for other mouse strains (e.g., FVB). 

Sequencing of RAB38 in Patient Samples.

All three exons and the intron–exon boundaries of RAB38 were sequenced in 12 subjects with unclassified, nonsyndromic OCA phenotypes, 9 subjects with HPS, 12 with HPS-like (oculocutaneous albinism [OCA] with bleeding abnormalities, but normal platelet dense bodies), and 17 with pigmentary glaucoma. Subjects with OCA did not have coding-region mutations in the known genes for these disorders. A single-nucleotide polymorphism, c.T105C, was observed in subjects with pigmentary glaucoma, HPS and an HPS-like syndrome. This polymorphism causes no change in amino acid sequence (p.S35S) and is not predicted to alter splicing. In one individual with pigmentary glaucoma, a heterozygous c.C583A change was identified that is predicted to cause a p.P195T change. This proline is well-conserved across mammalian species (mouse, rat, chimpanzee, cow, and dog). 

Discussion

Study of mouse coat color mutants has provided many insights into human pigmentation function and disorders. The ocular hypopigmentation defect of the pigmentation dilution Rab38 ^cht/cht mouse was characterized and analyzed, to provide additional information about Rab38's possible role in human diseases such as ocular albinism, Hermansky-Pudlak syndrome, or pigmentary glaucoma. 

We examined the anterior segments, RPE, and other aspects of the eyes of Rab38 ^cht/cht mice. The results showed that Rab38 ^cht/cht mice display iris and RPE thinning and focal depigmentation as well as ocular phenotypes similar to those of human OCA. For example, electroretinography performed on young Rab38 ^cht/cht mice is reminiscent of findings in some humans with albinism. Krill and Lee ²⁵ describe supranormal ERG responses under scotopic conditions in patients with oculocutaneous or ocular albinism. Photopic and flicker fusion responses were normal, however, and carriers of X-linked albinism were not significantly different from those in control subjects. They also observe that the scotopic changes in the albino ERG tend to normalize with age, perhaps due to increased pigmentation over time. Because our measurements were conducted in the mice at 3 months of age, they are most likely to model the findings in younger patients with albinism. Krill and Lee postulate that this supranormal ERG is the result of increased internal reflection of light within the eye from reduced pigmentation. This response, however, has not been observed in all patients with albinism. ^{26 27 28}  

The ocular pigment dilution in Rab38 ^cht/cht mice is most likely due to the presence of fewer and smaller melanosomes in the RPE and/or smaller melanosomes in the choroid. Our analysis suggests that the role of Rab38 is more pronounced in neuroectodermally derived structures, such as the RPE, than in NC-derived tissues, such as the choroidal melanocytes. Although the precise reason for this difference is unclear, recent experiments by Wasmeier et al. ²⁹ have shown that Rab32 and Rab38 may have redundant roles in melanogenesis. A tissue-specific difference in the effect of the Rab38 mutation could therefore be related to a difference in Rab32 expression. Although we did not quantitate melanosome size and number in older animals, we suspect that the age-related increase in pigmentation observed in these mice is related to compensatory changes in melanocytes with time. Similar increasing pigmentation has been noted clinically in some patients with OCA. ³⁰  

In the Tyrp1 ^b/b mouse, the number of RPE and choroidal melanosomes is decreased, ³¹ and the Tyrp1 ^b/b melanosomes are smaller and rounder than in the wild-type mouse. ³² Other data suggest the Rab38 is involved in appropriate targeting of TYRP1 within melanocytes, ¹⁶ and the Rab38 ^cht/cht mouse brown coat and eye color on a black strain background resembled the Tyrp1 ^b/b phenotype. Murine Tyrp1 functions enzymatically in the synthesis of eumelanin (black melanin), and melanin formed in Tyrp1 ^b/b mutant is brown. ^{33 34 35 36} Additional functions of Tyrp1 include stabilization of the melanin-synthesizing enzyme, tyrosinase, in melanosomes, and maintenance of melanosomal structure. 

To investigate the relationship of Rab38 and Tyrp1 in vivo, we created mice homozygous for the Rab38 ^cht allele and the Tyrp1 ^b allele. The double-mutant mice showed a striking loss of pigmentation that was qualitatively greater than the sum of hypopigmentation observed in Rab38 ^cht/cht Tyrp1 ^{+/ +} and Rab38 ^{+/ +} Tyrp1 ^b/b mice. This finding suggests that there is not a simple linear relationship between RAB38 function and TYRP1 function. If the sole role of RAB38 was to target TYRP1 to melanosomes properly, we would expect that disruption of RAB38 function would have little effect on the Tyrp1 ^b/b phenotype, as both proteins would function in the same “linear” pathway to produce pigment. Our observations that the double mutants were significantly hypopigmented implies that TYRP1 is not the only protein involved in RAB38-mediated pigment, particularly in the neuroectodermally derived layers of the eye. The double mutants may have such a dramatic phenotype because of misrouting of proteins other than TYRP1 to the melanosomes. 

Tyrp1 ^b/b mice on a DBA/2J background develop iris stromal atrophy and pigmentary glaucoma with age. ^{2 3} Iris transillumination, which is observed in Rab38 ^cht/cht mice, can be seen in iris atrophy as well as in hypopigmentation. To address this, we examined the irides of an aged cohort of Rab38 ^cht/cht mice and found no qualitative difference from a young cohort. There was no evidence in young or aged Rab38 ^cht/cht of significant iris atrophy, pigment dispersion, or glaucoma. Therefore, the iris transillumination in Rab38 ^cht/cht mice was more likely due to hypopigmentation and/or hypoplasia of the iris rather than to iris atrophy. The possibility remains that a Rab38 mutation could cause a pigmentary glaucoma phenotype on other mouse genetic backgrounds and/or could act as a modifier allele of this disease. 

In one patient with pigmentary glaucoma, the Rab38 sequence showed a heterozygous c.C583A change, which is predicted to cause a p.P195T change. Whether this change is disease-related is unclear. This proline is well-conserved across mammalian species (mouse, rat, chimpanzee, cow, and dog). Although a proline-to-threonine change is likely to be significant, the fact that Rab38 ^cht/+ mice appeared identical with wild-type mice and that Rab38 ^cht/cht mice did not develop pigmentary glaucoma argues against heterozygous changes having primary pathologic consequences in humans. 

Two lines of evidence suggest that Rab38 mutation may cause human Hermansky-Pudlak syndrome. First, the rat HPS model Ruby (red-eyed dilution, R) hypopigmentation and bleeding phenotype is caused by an Rab38 mutation. No RAB38 protein is produced in the Ruby rat. ¹⁷ In addition, Rab38, like other Rabs, is probably involved in vesicular trafficking, and several known HPS genes are intracellular vesicle trafficking proteins. ^{9 16}  

We did not observe any blood or organ system defects in the Rab38 ^cht/cht mice. Neutrophil counts were significantly higher than in wild-type mice, but within the normal range for other mouse strains (e.g., FVB). Further analysis is necessary to determine whether this difference results in any impairment of the immune system. Unlike the Ruby rat, RAB38 is produced in Rab38 ^cht/cht mice, which would be predicted for a point-mutation protein, but the steady state level of mutant RAB38 is much reduced. Recent in vitro work suggests that the chocolate allele produces functionally inactive RAB38. ²⁹ These results, along with data showing that Rab38 ^cht/cht mice have normal bleeding times ^{16 37} suggest that the Rab38 ^cht/cht allele does not lead to the systemic abnormalities seen in HPS. 

Our preliminary analysis of the RAB38 gene in a small number of patients with HPS or an HPS-like condition revealed only a single nucleotide change that does not result in amino acid substitution or clearly alter a splice site. In addition, this same nucleotide change was observed in patients with pigmentary glaucoma. A detailed systemic evaluation of the Rab38 ^cht/cht mouse failed to reveal any abnormality beyond ocular and cutaneous pigmentation dilution. Specifically, the Rab38 ^cht/cht mouse did not have the platelet dysfunction that is characteristic of HPS. Future studies of the functional significance of the Rab38 ^cht allele and the human sequence change could be performed to elucidate the potential role of RAB38 in HPS. Alternatively, RAB38 could be involved in another form of syndromic albinism not covered in our patient sample. 

Sequencing of the RAB38 gene in patients with OCA who did not have mutations in the known albinism genes did not reveal any sequence changes. Suzuki et al. ³⁸ likewise did not find any RAB38 mutations in their cohort of Japanese patients with albinism. We therefore conclude that RAB38 is not a major locus for human OCA. The fact remains, however, that the Rab38 ^cht allele affects coat color; as such, it is still quite possible that RAB38—although not itself a major disease locus—modifies the phenotype in patients with pigment-related diseases. 

²  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported by the intramural program of the National Institutes of Health, the Howard Hughes Medical Institute, and National Eye Institute Grants EY01475 and EY11721. BPB is part of the Joint Physician-Scientist Development Program at the National Eye Institute and the National Human Genome Research Institute.

 Submitted for publication December 11, 2006; revised April 30, 2007; accepted July 19, 2007.

 Disclosure: B.P. Brooks, None; D.M. Larson, None; C.-C. Chan, None; S. Kjellstrom, None; R.S. Smith, None; M.A. Crawford, None; L. Lamoreux, None; M. Huizing, None; R. Hess, None; X. Jiao, None; J.F. Hejtmancik, None; A. Maminishkis, None; S.W.M. John, None; R. Bush, None; W.J. Pavan, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Rab38 ^cht/cht mice, but not Rab38 ^{cht/ +} mice, exhibited peripheral iris transillumination (TI). The degree of TI was graded as +1 (peripheral only, A), +2 (peripheral and mid-iris, B) or 0 (none, C). Slightly more than 20% of young (2–6 months of age) Rab38 ^cht/cht mice exhibited +2 TI; the remainder exhibited +1 TI. This proportion did not vary in a cohort of older mice (12–24 months), indicating that TI is probably due to variable iris hypoplasia, rather than iris atrophy. Rab38 ^cht/+ mice showed no iris TI, even at advanced ages (C, age 23 months in this instance). Similarly, wild-type mice did not exhibit iris TI (data not shown).

Figure 2.

 

Histopathology of Rab38 ^cht/cht mice (B, D) revealed iris hypopigmentation and thinning, particularly peripherally (D, arrow) when compared to that in age-matched Rab38 ^cht/+ mice (A, C). Wild-type irides were similar to those of Rab38 ^cht/+ mice (data not shown). Bars, 10 μm.

Figure 3.

 

Rab38 ^cht/cht mice exhibited peripheral patches of depigmentation (A, fundus photo). These patches were most frequently found in the superior retina. Transmission electron microscopy showed that they represented focal RPE thinning (B, arrow, border of normal–abnormal tissue dissected). Bar, 2 μm.

Figure 4.

 

Melanosomes were smaller in cross-sectional area and fewer in the RPE of Rab38 ^cht/cht mice (A) compared with wild-type (Rab38 ^{+/ +}; C); heterozygote (Rab38 ^cht/+) mice were similar to wild-type (B). Melanosomes in Rab38 ^cht/cht mice were able to enter the apical villi of RPE cells (A, arrowheads), suggesting that, unlike RAB27a, RAB38 is not necessary for transport into these structures. Melanosomes from the NC-derived choroid were also smaller in Rab38 ^cht/cht mice than in wild-type mice. Bar, 2 μm.

Figure 5.

 

The distribution of melanosome cross-sectional area in the RPE and choroid of Rab38 ^cht/cht mice demonstrates the significantly smaller areas of Rab38 ^cht/cht melanosomes relative to wild-type.

Table 1.

 

Comparison of Melanosomes in RPE and Choroid

Genotype	Melanosome Density (n/μm2)	Melanosome Area (μm2)
RPE
Rab38 cht/cht	0.48 ± 0.12 (5)*	0.06 ± 0.04 (120)*
Rab38 cht/+	1.08 ± 0.29 (4)	0.13 ± 0.11 (220)
Rab38 +/+	0.98 ± 0.08 (4)	0.12 ± 0.08 (255)
Choroid
Rab38 cht/cht	2.79 ± 0.72 (4)	0.04 ± 0.03 (748)*
Rab38 cht/+	3.83 ± 1.48 (3)	0.07 ± 0.04 (481)
Rab38 +/+	2.63 ± 0.75 (4)	0.07 ± 0.04 (646)

 Data are the mean ± SD (sample number).

*  Group mean differs from wild-type mean at P < 0.01, by one-way ANOVA with the Dunnett posttest for significance.

Figure 6.

 

An immortalized pigmented mouse melanocyte cell line melan-Melan-Ink4a-cht5 had a reduced amount of RAB38 protein compared with the normal black pigmented cell line, melan-Ink4a-1, on Western blot.

Figure 7.

 

Scotopic (A) and photopic (B) ERG b-wave amplitudes (±SEM from 2.5-month-old homozygous (Rab38 ^cht/cht ), heterozygous (Rab38 ^cht/+), and wild-type C57BL/6J (Rab38 ^{+ / +}) mice. Compared with wild-type mice, homozygous mice had significantly larger b-wave amplitudes in the scotopic range (ANOVA, P < 0.01). In the photopic range, there was no significant difference between any of the groups. Homozygous and wild-type: n = 5 animals/10 eyes; heterozygous: n = 2 animals/4 eyes.

Figure 8.

 

In one-month-old Rab38 ^cht/cht mice, the light brown coat color was easily distinguished from the black coat of wild-type and Rab38 ^{cht/ +} littermates. In 3-month-old and older Rab38 ^cht/cht mice, the darker coat was still distinguishable from the black coat of the wild-type and Rab38 ^{cht/ +} littermates.

Figure 9.

 

The Tyrp1 ^b/b allele interacted synergistically with the Rab38 ^cht/cht allele in C57BL/6 mice, to produce striking pigment dilution. Slit lamp examination of the anterior segments of (A)Rab38 ^{+/ +}, Tyrp1 ^{+/ +}; (B) Rab38 ^cht/cht , Tyrp1 ^{+/ +}; (C) Rab38 ^{+/ +}, Tyrp1 ^b/b ; and (D) Rab38 ^cht/cht , Tyrp1 ^b/b mice showed significantly more depigmentation and iris transillumination in the double-mutant homozygote than in the other three strains. This effect was observed in both the eyes and the coat color of the mice (G). (E, F) Rab38 ^cht/cht , Tyrp1 ^b/b mice showed reduced pigmentation on histologic sections, particularly in the neuroectodermally derived layers. This difference in pigmentation was evident in the posterior pigmented epithelial layer of the iris (E, arrowhead) and the RPE (F, arrowhead). Bars, 50 μm.

Table 2.

 

Results of Systemic Analysis

Test	Rab38 cht/cht	Rab38 +/+	P
RBC count	8.4	10	0.15
Hemoglobin	12.9	13.9	0.18
Hematocrit	40.2	42.7	0.17
Platelets	1016	837	0.18
Neutrophils	0.259	0.075	0.002
Cholesterol	117	85	0.058
BUN	21	33	0.03
Magnesium	1.17	1.45	0.1
Alkaline phosphatase	63.6	102	0.1
Total bilirubin	0.18	0.32	0.07
GGT	1.66	4.16	0.15

 Test data are the mean of six mice in each group. Histopathology for Rab38 ^cht/cht : two of three females had suppurative metritis, thought to be postpartum; two of six mice had suppurative otitis media; two of six had mild lymphocytic infiltrates in nasal mucosa; one of six had mild lymphocytic laryngitis; and one of six had chronic pancreatitis. Histopathology for Rab38 ^+/+: four of six had mild granulocytic cholecystitis; one of six had uterine endometrial edema; one of six had edema of tympanic membrane; and one of six had focal lacrimal gland adenitis.

The authors thank Richard King's laboratory, University of Minnesota, for clinical data; Dot Bennett for the melan-Ink4a-1 and melan-Ink4a-cht5 cells used in the study; the National Human Genome Research Institute (NHGRI) mouse core for rederivation of mouse lines; and John Hammer, Xufeng Wu, and members of the Pavan laboratory for useful discussions.