In order to further probe the genetic requirements contributing to the D2 form of glaucoma, we created congenic strains of mice on the B6 genetic background. Our approach for transferring the D2-derived Tyrp1^band Gpnmb^R150X mutations to the B6 background was to first create separate congenic strains for each of the respective chromosomal regions (Figure 1). Successive rounds of backcrossing utilizing markers flanking Tyrp1 (Figure 1A) and Gpnmb (Figure 1B) yielded the single-congenic strains for Tyrp1^b (B6.D2-Tyrp1^b/Sj; here after abbreviated B6.Tyrp1^b) and Gpnmb^R150X (B6.D2-Gpnmb^R150X /Sj; here after abbreviated B6.Gpnmb^R150X). The final sizes of the N10 congenic intervals were determined to be 14–36 cM (23.6–63.2 Mb) for B6. Tyrp1^b and 9–15 cM (10.0–36.8 Mb) for B6. Gpnmb^R150X. After each single N10 congenic strain was established, mice were intercrossed to generate a double-congenic stock homozygous for both congenic intervals (B6.D2-Tyrp1^bGpnmb^R150X /Sj; here after abbreviated B6.Tyrp1^b Gpnmb^R150X or referred to as B6 double-congenic).

Cohorts of N10 mice of all three genotypes (B6.Tyrp1^b, B6. Gpnmb^R150X, and B6.Tyrp1^bGpnmb^R150X) were aged and assayed for relevant phenotypes. As expected, congenic mice homozygous for the Tyrp1^bmutation had brown coat colors 3435, whereas the Gpnmb^R150X mutation had no effect on coat color in any of the stocks examined. Clinical analysis of iris phenotypes in aged cohorts of the single-congenic B6.Tyrp1^b (Figs. 1C to 1E), single-congenic B6.Gpnmb^R150X (Figs. 1F to 1H), and double-congenic B6.Tyrp1^b Gpnmb^R150X (Figure 2) strains of mice indicated that the B6 background was permissive to the iris disease. Iris phenotypes of the single-congenic strains closely resembled our previous observations of mice mutant for only Gpnmb or Tyrp1 in varying genetic contexts 222336. Furthermore, the timing and severity of pigment dispersing iris disease in the double-congenic B6.Tyrp1^bGpnmb^R150X mice is essentially the same as in D2 mice mutant for the same Tyrp1 and Gpnmb alleles 19.

We next took advantage of the B6 background to perform strain controlled genetic epistasis experiments that tested mechanisms contributing to the Tyrp1 and Gpnmb mediated iris disease. A significant result in understanding the Tyrp1 and Gpnmb mediated iris disease is that modulating pigment production with tyrosinase (Tyr, a gene essential for pigment production) and hypopigmentation mutations completely rescues iris atrophy and pigment dispersion 2223. These experiments strongly implicated melanosomal functions of Tyrp1 and Gpnmb in pathogenesis. However, they were not absolutely complete, since only the effect of hypopigmentation mutations and not Tyr were assessed for Gpnmb. This is important because hypopigmentation mutations often affect multiple types of organelles 3738 whereas Tyr has no known function in any organelles other than melanosomes.

To assess the importance of pigmentary pathways on the B6 background, we used the Tyr^c-2J mutation that arose spontaneously in the B6 strain 3940. We generated multiple cohorts of homozygous, Tyr^c-2J albino mice that were also homozygous for both Tyrp1^b and Gpnmb^R150X (Figure 3), or also homozygous for only Tyrp1^b or Gpnmb^R150X (data not shown). Aged cohorts of these albino mice were compared with normally pigmented mice of matched genotypes. All mice homozygous for the Tyr^c-2J mutation had albino coats (Figure 3A) and albino irides (Figure 3B–G). Importantly, the Tyr^c-2J mutation rescued all observable Tyrp1^b – and Gpnmb^R150X-associated iris phenotypes in double-congenic B6.Tyrp1^b Gpnmb^R150X mice (Figure 3C, E, and 3G) and single-congenic B6.Tyrp1^b (n = 28 eyes, 12+ months) and B6.Gpnmb^R150X mice (n = 8 eyes, 12+ months). These results convincingly indicate that melanin production is a necessary component of the pathological events initiating iris disease on a B6 background, and they confirm a melanosomal component for both the Tyrp1 and Gpnmb mutant phenotypes.

Having established that B6.Tyrp1^bGpnmb^R150X mice exhibit a pigment-dispersing iris disease that is clinically, and mechanistically, similar to that in D2 mice, we next tested whether the B6 genetic background would influence the impact of this insult on IOP. Although the timing and severity of iris disease was similar, IOP elevation was significantly muted in B6.Tyrp1^bGpnmb^R150X mice compared with D2 mice (Figure 4). We compared the IOP distributions of these strains in relation to IOP values of 19 mmHg and 21 mmHg. As a strain, young D2 mice have lower IOP than young B6 mice and 19 mmHg is > 2 SD above the mean of young pre-glaucomatous D2 mice. The value of 21 mmHg is glaucoma relevant, as individuals with an IOP of 21 mmHg are commonly regarded as glaucoma suspect 41. Strikingly, a far smaller proportion of B6.Tyrp1^bGpnmb^R150X mice were detected with IOP > 21 mmHg compared with D2 mice (Figure 4A to 4D). For example, at 7–8 months only 2 % of B6.Tyrp1^bGpnmb^R150X mice had IOP > 21 mmHg, compared with 16 % of D2 mice (P = 0.007,χ² comparing number of mice with IOP < 19 mmHg, 19–21 mmHg, and > 21 mmHg for each strain at this age). Similarly, at 9–10 months, only 11% of B6.Tyrp1^bGpnmb^R150X had IOPs >21 mmHg, compared with 42 % of D2 (P < 0.0001, similar χ² comparison for 9 to 10-months mice). The same trend was observed considering all mice, irrespective of age (P < 0.0001, χ²). Further demonstrating their lower susceptibility to developing high IOP, pressure elevation was delayed in B6.Tyrp1^bGpnmb^R150X mice compared with D2 mice (Figure 4E and 4F). Whereas the greatest proportion of D2 mice had IOP > 21 mmHg (42%) at 9–10 months, the greatest proportion of B6.Tyrp1^bGpnmb^R150X mice (16%) had IOP > 21 4 months later at 13–14 months of age (P < 0.0001, χ² comparing B6.Tyrp1^b Gpnmb^R150X at 13–14 months with D2 at 9–10 months). Therefore, as a population, B6.Tyrp1^bGpnmb^R150X mice are less susceptible to IOP elevation than D2 mice.

The relatively low susceptibility of B6.Tyrp1^b Gpnmb^R150X mice to IOP elevation suggested that this strain may experience less glaucomatous nerve damage than D2 mice. In D2 mice, severe glaucomatous damage with substantial axon loss is detectable in over 50% of 11-month-old, 70% of 13-month-old and 80% of 19-month-old mice 4243. In contrast, severe damage was almost never detected in B6.Tyrp1^b Gpnmb^R150X mice (Figure 5). At ages when 50–80% of D2 mice have severe damage, only one of 109 B6.Tyrp1^bGpnmb^R150X nerves was severely affected (56 of these nerves were from mice >18 months old). The distributions of nerve-damage level were almost identical in B6.Tyrp1^b Gpnmb^R150X and wild-type B6 mice (Figure 5), indicating that the degree of damage in B6.Tyrp1^b Gpnmb^R150X mice represents standard aging changes for this strain background. To maximize the likelihood of detecting glaucoma, optic nerves from the mice with the highest IOPs were examined (Figure 5C and 5D), but none had a level of damage similar to the glaucomatous damage of D2 mice. Additionally, no significant difference in the number of healthy axons was detected between normal B6 and B6.Tyrp1^isa Gpnmb^R150X mice aged either 4–5 or 22–27 months (Figure 5G). Combined, the analysis of optic nerves from these cohorts of B6.Tyrp1^isa Gpnmb^R150X mice convincingly shows that on the B6 genetic background, the Tyrp-1 and Gpnmb-mediated iris disease does not result in glaucoma.

The current experiments demonstrate that melanosomal defects associated with melanin synthesis play a key role in the initiation of both Tyrp-1 and Gpnmb-mediated iris disease. Melanosomal synthesis of melanin involves a series of reactions creating the biopolymer melanin from the precursor tyrosine 45. Many of the indole-quinone intermediates of these reactions are cytotoxic, and the mechanisms regulating their cytotoxicity are only partially understood 4647. Our results indicate that mutations in Tyrp1 and Gpnmb act in the melanosome, where they promote cytotoxicity in a manner that can be rescued by halting production of these toxic intermediates. Based on these findings, it will be important to examine other genes encoding melanosomal proteins that may participate in these events and that may also cause similar diseases of the iris.

The experiments described above demonstrate that the severity and timing of the pigment-dispersing iris disease of B6 double-congenic mice is essentially the same as that of D2 mice. This suggests that the dispersed pigment and iris debris similarly insult the ocular drainage structures of both strains. Nevertheless, the B6 double-congenic mice were much less susceptible to IOP elevation than D2 mice. Probably because of their resistance to IOP elevation, the B6 double-congenics were not found to develop glaucoma. In fact, the degree of optic-nerve damage observed was largely similar to age-matched wild-type B6, indicating that the small degree of optic neuropathy observed was likely an independent phenomenon, such as the result of the normal aging process or a low-penetrance developmental abnormality. Experiments to understand this strain difference in IOP elevation are underway.

These experiments also suggest that the cytotoxicity mediated by Tyrp1 and Gpnmb does not kill the RGCs, and that iris disease and RGC death can be dissociated from each other. Therefore, RGC death in D2 mice appears to be a pressure-dependent process resembling the pathology of human glaucoma. In support of this, RGC death in this model is not uniformly distributed 4849, as would be expected if damaged by an iris-derived diffusible toxin. Rather, it appears that RGCs are uniquely susceptible retinal neurons, and the pattern of damage is consistent with an insult to RGC axons in the optic nerve 1749.

The broader implications of the strain-specific phenotypes may be important for studies of the heredity and mechanism of human pigmentary glaucoma. In humans, PDS/PG exhibits strong hereditary associations 505152. A few families with clear inheritance of PDS are reported. However, the low incidence of PDS in family members of the majority of affected individuals suggests a more complex etiology, possibly influenced by multiple genetic and environmental factors. In some families, linkage was initially reported between the disease phenotype and markers on 7q35-q36 50. However, causative mutations at this locus have thus far not been found. Analysis of candidate genes suggested by experiments with D2 mice (such as TYRP1 and GPNMB) have also failed to identify any clear mutations in these families 53. Thus, the disease in these families may be more complex than initially suspected. The lack of identified mutations in GPNMB and TYRP1 in these families does not preclude the involvement of these genes (or genes of similar function) in patients of other families or in the more common, more complex cases. It remains to be determined whether PDS/PG in other patient populations is caused by more tractable loci, or whether the identification of genes associated with human PDS/PG may simply necessitate approaches that can overcome genetic complexity.

From a mechanistic view, the dichotomy between the differing responses of the B6 double-congenics versus D2 mice is analogous to the differing responses of people with PDS. Many human eyes tolerate significant amounts of dispersed pigment without developing glaucoma 67. Indeed, fewer than half of human eyes with pigment dispersion progress to pigmentary glaucoma 89101112. This, along with the expected minor degree of aqueous humor outflow obstruction due to pigment itself 5455, suggests that individual differences in the ocular reaction to dispersed pigment are important in determining whether pigment dispersion progresses to pigmentary glaucoma. The D2 and B6 congenic strains provide a tractable resource for mechanistically deciphering the ocular reactions that determine whether IOP becomes harmfully elevated after the drainage structures are insulted by dispersed pigment and iris debris. Future studies of these reactions will not only improve our mechanistic understanding of the disease, but may also suggest new classes of genes worthy of analysis among human PDS/PG patients.

Our results highlight the profound importance of genetic background to complex diseases such as glaucoma. In addition to understanding human glaucoma, it is important to consider effects of genetic background when using animal models. As a genetic resource, these congenic mouse strains will help to address this issue. Until now, a strain-matched control for Tyrp1 – and Gpnmb-mediated disease has been problematic. For instance, although any standard inbred mouse strain that does not develop glaucoma can be utilized as a control for D2 mice, this experimental design does not distinguish whether any observed differences relate to the disease state of the eye, or are merely previously unappreciated strain-dependent background effects. Given the complexity of these glaucoma phenotypes, experiments comparing between different strain backgrounds or using segregating backgrounds can be misleading 1356. Importantly, the congenic mice generated here can be utilized with strain-matched normal B6 mice to perform appropriately controlled experiments.

Pressure-independent RGC death has been reported in C57BL/6 mice and may depend on environment or substrain used. In wild-type C57BL/6 mice from the supplier Harlan (C57BL/6Hsd), an age-related reduction in the number of RGCs has been described 48. By 18 months, the number of RGCs was reduced to approximately 46% that of young mice. In our current study, this age-related decline was absent in both the double-congenic B6.Tyrp1^b Gpnmb^R150X mice and normal B6 controls that all had a C57BL/6J background. Similarly, we did not detect glaucomatous RGC loss in another experiment where C57BL/6J mice were also aged to 19+ months 57. This difference in RGC death between studies may be explained by potential environmental influences, differences in analysis protocols, or genetic differences in the substrain utilized.

Mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). All animals were treated according to the guidelines of the Association for Research in Vision and Ophthalmology for use of animals in research. All experimental protocols were approved by the Animal Care and Use Committee of The Jackson Laboratory. Data for the congenic B6 strains are compared with those with a D2 background. Many of the data for the D2 mice (more than 1500 mice) have been previously reported but some new IOP data are included here. The congenic B6 mice were bred, aged and analyzed at the same time as over 300 D2 mice. The D2 mice were maintained on an NIH 31 diet with a fat content of 6%. To avoid obesity, mice with a B6 background had to be maintained on essentially the same NIH 31 diet but with a 4% fat content. Our studies have shown that D2 IOP and other glaucoma phenotypes do not differ with this small dietary difference (unpublished data). For all mice, the diet was provided ad libitum, and the water was acidified to pH 2.8–3.2. Mice were housed in cages containing white-pine bedding and covered with polyester filters. The environment was kept at 21°C with a 14-hour light:10-hour dark cycle.

Chromosomal intervals of approximately 6–7 cM were selected for the regions flanking Tyrp1 (D4Mit178 to D4Mit327) and Gpnmb (D6Mit74 to D6Mit355). Mice were reiteratively bred to B6 mice and each successive generation genotyped to select breeders heterozygous for the respective D2 regions. This process was continued for 10 generations of backcrossing. At the 10th generation, mice were intercrossed and each region bred to homozygosity. Within the respective congenic intervals, the presence of the Tyrp1^b allele was confirmed by assaying for a D2-derived polymorphism creating a TaqI restriction-enzyme site in exon 4 35, and the presence of the Gpnmb^R150X mutation was confirmed by assaying for a PvuII restriction-enzyme site created by the D2-derived mutation 23. The sizes of the congenic intervals were determined using genomic DNA prepared from homozygous mice at the 10th generation of backcrossing and assayed using polymorphic microsatellite markers with standard PCR conditions.

IOP was measured using the microneedle method as previously described in detail 5859. All cohorts included male and female mice. Each mouse used in this study had only a single IOP measurement. Thus, each age group shown consists of cohorts of different individual mice. Because the IOPs of B6 mice are very consistent, B6 mice were interspersed with experimental mice during all experiments as a methodological control to ensure proper equipment calibration and performance. Statistical comparisons of IOP profiles generated from these strains utilize nonparametric data comparisons, with data grouped according to empirically set points relevant to these studies: 19 mmHg (i.e. > 2 SD above the mean for young pre-glaucomatous D2 mice) and 21 mmHg (a value commonly regarded as glaucoma suspect in humans) 41.

Eyes were examined with a slit-lamp biomicroscope and photographed with a 40× objective lens. Phenotypic assessment of iris disease was determined by indices of iris atrophy and dispersed pigment following previously described criteria 18222360. Transillumination defects were also measured (transillumination is an assay of iris disease whereby reflected light passing through depigmented areas of iris tissue are visualized as red light). However, because transillumination is influenced by multiple strain-specific traits not controlled in these studies (such as iris thickness), it was not the primary means of following disease progression across different genetic backgrounds.

Optic nerve cross sections were examined for glaucomatous damage using a modified para-phenylenediamine (PPD) staining protocol to stain the myelin sheath of all axons and the axoplasma of damaged axons 61. Optic nerves were prepared for analysis by overnight fixation in phosphate-buffered glutaraldehyde/paraformaldehyde mixture at 4°C followed by overnight treatment in osmium tetroxide at 4°C. After osmification, nerves were rinsed twice for 10 minutes in 0.1 M phosphate buffer, once in 0.1 M sodium-acetate buffer, and dehydrated in graded ethanol concentrations. After embedding this tissue in resin, 1-μm-thick sections were stained in 1% PPD for 35–40 minutes.

Stained sections were compared using a previously described qualitative grading scale (including assessments of the approximate number of healthy axons, number of damaged axons, and scarring associated with gliosis) and used to perform quantitative axon counts 4243. The grading scale has previously been validated by comparisons to axon counts 4243. Quantitative axon counts imaged 18 non-overlapping fields at 1000× magnification, which were evenly spread throughout the cross-section of the nerve; the sum area of these fields was equal to 10% of the total nerve cross-sectional area. The majority of optic nerves were collected from mice within 48 hours of IOP measurement. Some mice with elevated IOPs were additionally age determined prior to assessment of the optic nerve. Each age group investigated contained samples from males and females, as well as left and right nerves.

The official gene abbreviations, full gene names, alleles, and stocks utilized in this study include the following. The full name for the Gpnmb gene is glycoprotein (transmembrane) nmb. The ipd allele of Gpnmb results from the R150X premature stop codon mutation, Gpnmb^R150X 23. The full name for the Tyrp1 gene is tyrosinase-related protein 1. The b allele of Tyrp1 encodes two amino-acid substitutions, compared with the C57BL/6J-derived allele 35. This D2 allele is also referred to as isa in the mouse genome database and elsewhere 2223. The DBA/2J stock utilized here was stock number 000671 from The Jackson Laboratory. The full name for the Tyr gene is tyrosinase. The c-2J mutation spontaneously arose within a C57BL/6J colony, resulting in a R77L amino acid substitution and also changing alternative splicing of the tyrosinase pre-mRNA 3940. The Tyr^c-2J stock utilized here was C57BL/6J-Tyr^c-2J/J (stock 000058 from The Jackson Laboratory).