Abstract
purpose. Ocular melanin synthesis modulates rod photoreceptor production, because in albino eyes, rod numbers are reduced by approximately 30%. In this study, rod numbers and ocular rhodopsin concentrations were measured in intermediate pigmentation phenotypes to determine whether proportional reductions in melanin are correlated with proportional changes in rod numbers. Further, patterns of cell production and death were examined around the time of birth, when rod production peaks, to determine whether there are abnormalities in these features associated with hypopigmentation.
methods. Four mouse pigmentation phenotypes were used: fully pigmented, albino, Beige, and Himalayan. The latter two are intermediate-pigmentation phenotypes, with Beige having markedly more pigment than Himalayan. Ocular melanin concentrations were measured during development and at maturity. Rods were counted at maturity and measurements of ocular rhodopsin undertaken. Mitotic and pyknotic cells were also counted in neonates.
results. Rods and ocular rhodopsin were reduced in both Beige and Himalayan mice below levels found in fully pigmented mice, but not to levels found in albino animals. This was more marked in Himalayan than Beige mice, reflecting the lower concentration of melanin found in the former compared with the latter, both in development and at maturity. Although patterns of cell production were elevated in the hypopigmented animals, such patterns varied.
conclusions. Rod numbers are modulated within a range between that in fully pigmented and albino phenotypes by the concentration of ocular melanin. However, in these animals, there is no obvious correlation between these events and patterns of cell production and death in neonates.
The retinal pigment epithelium (RPE) is critical for normal retinal function.
1 2 When ocular melanin synthesis fails during development, profound abnormalities arise in the neural retina. These occur regardless of the reason for failure of melanin synthesis. Hence, albinism is genetically heterogeneous but manifests the same abnormalities. These include underdevelopment of the central retina, abnormal chiasmatic routing of temporal retinal axons and a reduction in rod photoreceptor numbers of approximately 25% to 30%.
3 4 Because these deficits are diverse, the aspects of development that melanin regulates must be fundamental to a wide range of retinal features.
The reduction in rod numbers in the outer retinas of individuals with albinism is cell specific, in that cones appear to be unaffected in their number or mosaic distribution. Further, the abnormality arises early and is not the result of environmental factors. The loss of rods is mirrored by a proportional decrease in rhodopsin levels.
5 6 7 It is not clear why there is such a large rod deficit. However, albino rats have abnormally elevated levels of mitosis and pyknosis throughout development that ultimately result in cellular depletion. These events peak during rod production, but are less apparent at earlier stages when cones are being produced.
8 9 10
In the current study, we sought to determine whether there is a dose relationship between the amounts of melanin and rod numbers, by quantifying ocular melanin levels in the developing and mature retina and comparing this with adult rod numbers and rhodopsin concentrations. This was undertaken in two types of hypopigmented mutant mice with intermediate pigmentation levels, along with normally pigmented and albino mice. Further, we attempted to relate these data to patterns of cell production and death during rod production. The two mutants used were Beige and Himalayan mice. Both have reduced levels of pigmentation, with that in the Himalayan being heavily reduced and that in the Beige being intermediate, between levels found in the Himalayan and the fully pigmented animal.
Methods
Four groups of mice were used in this study: pigmented C57Bl/6J, albinos (BALB/c), Himalayan (c
h/c
h), and Beige (bg/bg). Both Himalayan and Beige were on a C57Bl/6J background. Although in Beige mice the mutation is not melanocyte specific and melanin packaging and distribution are also affected, it should be stressed that the failure of melanin synthesis for any reason results in a common pattern of retinal problems.
3 Beige and Himalayan mice were chosen, because macroscopic examination of the eyes revealed the mice to be clear intermediate pigmentation phenotypes, and they were both readily available. All mice were maintained in breeding colonies. A series of independent experiments were undertaken. The number of animals used in each and their stage of development is given in the relevant sections. Animal usage conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and conformed to British Home Office regulations.
Melanin Measurement
The total amount of ocular melanin was measured by the method of sodium hydroxide solubilization.
11 Immediately after killing the mice with an overdose of pentobarbitone sodium (60 mg/kg), animals were weighed, and their eyes were removed and weighed and then placed in phosphate-buffered saline (PBS; pH 7.2). The eyes were opened and the lens and vitreous removed, but the iris and cornea were left attached. Care was taken not to damage the RPE. The remaining tissue was placed in tubes (Eppendorf, Fremont, CA) containing 100 μL of 1 M NaOH (pH >12) and 10 μL dimethyl sulfoxide (DMSO) and boiled for a maximum of 30 minutes to resuspend the insoluble mouse melanin. Samples were brought up to 500 μL with distilled water and neutralized using diluted acetic acid. Immediately after melanin solubilization, the absorption of the samples was measured at 475 nm against the blank solubilization buffer. In each experiment, eyes of age-matched albino mice were assessed as a control for evaluating the absorption of the nonpigmented tissue (background). The melanin content was calculated using a standard curve of absorption for a control pure-melanin (Sigma, Poole, UK) solution, and results were expressed in units of total melanin per eye. The melanin content of eyes from each pigmentation phenotype were measured against that found in the albino eyes. Melanin measurements were undertaken at the day (D) of birth (D1) when rod production peaks (albino,
n = 8; Himalayan,
n = 3; Beige,
n = 4; C57Bl,
n = 9) on D3 (albino,
n = 6; Himalayan,
n = 4; Beige,
n = 6; C57Bl,
n = 7), D6 (albino,
n = 18; Himalayan,
n = 3; Beige,
n = 5; C57Bl,
n = 16), D12 (albino,
n = 9; Himalayan,
n = 3; Beige,
n = 4; C57Bl,
n = 8), D24 (albino,
n = 6; Himalayan,
n = 4; Beige,
n = 4; C57Bl,
n = 10), and D210 (albino,
n = 13; Himalayan,
n = 4; Beige,
n = 4; C57Bl,
n = 10).
Rod Numbers
To estimate rod numbers in each of the four pigmentation phenotypes, mature animals (approximately 4 months old; albino animals, n = 3; Himalayan, n = 4; Beige, n = 5; C57Bl, n = 3) were killed with CO2 and perfused transcardially with PBS (pH 7.2) followed by 2% paraformaldehyde and 2% glutaraldehyde in PBS (pH 7.2). After fixation, the eyes were removed and left in the fixative overnight. The following day, the anterior chamber and lens were extracted, and the eyes placed in 1% osmium tetroxide for 1 hour. They were then washed in PBS and dehydrated through a graded series of alcohols before being embedded in resin (Historesin; Leica, Cambridge, UK). They were then cut at thicknesses of between 0.5 and 1.0 μm with glass knives, with a 1:10 series collected. These were mounted on clean slides that were left on a hot plate overnight before being stained with cresyl violet and coverslipped.
Rod outer segments were counted in central regions equidistant between the two retinal peripheries on each side in a section adjacent to the optic nerve head. In each case, outer segment numbers were counted in a sampled field that was a strip 100 μm long, along a line that ran orthogonal to the length of the outer segments and that was halfway between the outer limits of the outer nuclear layer and the RPE. This was achieved by frame grabbing images at a magnification of ×1250 and enhancing these with an image-management computer program (Matrox Inspector; Matrox Graphics, Inc., Dorval, Québec, Canada) to reduce any ambiguities. This measure was repeated for five sections in three animals.
Rhodopsin Measurement
Rhodopsin was extracted according to the method described by de Grip et al.
12 Before extraction, animals were kept in total darkness for 14 days to control for differential effects of photostasis.
13 Typically, four to five pairs of eyes of each animal phenotype were used in each experiment. All the following procedures were performed in the dark under an infrared safelight (Wratten; Kodak, Rochester, NY). Animals were killed by cervical dislocation (albino,
n = 22; Himalayan,
n = 23; Beige,
n = 16; C57Bl,
n = 15). Both eyes were immediately removed and placed in a light-protected tube (Eppendorf) containing 1 mL PBS (pH 7.4) on ice. They were then finely chopped and rotated for 1 hour at 4°C. After centrifugation for 30 minutes at 13,000 rpm, the supernatant containing the soluble proteins was discarded and the pellet resuspended in and extracted with 500 μL PBS containing 60 mM EDTA and 10%
N,
N-dimethyldodecylamine-
N-oxide detergent (30% solution, Ammonyx; Fluka Chemica/Sigma) by gentle rotation for 24 hours at 4°C.
After centrifugation again for 30 minutes at 13,000 rpm, the supernatant containing the rhodopsin extract was scanned for its absorption spectrum between 190 and 700 nm, against the extraction buffer with a spectrophotometer (model 5020; PE-Applied Biosystems, Foster City, CA). Each extract was scanned twice before and after bleaching, which was performed by exposure to 10 minutes of incandescent light and/or addition of 50 μL of a 50% hydroxylamine (Sigma) solution in PBS. No difference in the absorption spectra was observed with either of these bleaching procedures.
The concentration of rhodopsin was determined by measuring the difference of absorption (
A) of the extract at 500 nm (maximal rhodopsin absorption wavelength) and 650 nm (no rhodopsin absorption) before bleaching (intact rhodopsin) and after bleaching (background), according to the equation
\[\mathrm{Rhodopsin\ concentration}\ {=}\ (A_{500\ \mathrm{nm}}\ {-}\ A_{650\ \mathrm{nm}})\ _{\mathrm{before}}\]
\[\ {-}\ (A_{500\ \mathrm{nm}}\ {-}\ A_{650\ \mathrm{nm}})_{\mathrm{after}}/\mathrm{extinction\ coefficient\ of\ rhodopsin}\]
Results are expressed in total picomoles of rhodopsin per eye in each animal.
Mitotic and Pyknotic Patterns
To determine whether there was a simple relationship between levels of ocular pigmentation and the number of mitotic and pyknotic profiles, animals from each pigmentation phenotype were bred and collected on D1 and D3 (D1: albino, n = 3; Himalayan, n = 3; Beige, n = 6; C57Bl, n = 3; D3: albino, n = 8; Himalayan, n = 3; Beige, n = 7; C57Bl, n = 8). These were killed as described, and their eyelids opened. The heads were then placed in 2% paraformaldehyde and 2% glutaraldehyde in PBS (pH 7.2) and left overnight. The following day, the heads were dissected, leaving the eyes attached through the pallet. The tissue was then dehydrated as described and embedded in resin (Historesin; Leica). These blocks were sectioned horizontally at 5 μm, with every fifth section being saved. Sections were mounted on clean slides, left on a hot plate overnight, and then stained with cresyl violet. In each case, three sections close to the horizontal meridian were selected for analysis from three animals from each pigmentation phenotype, and the number of mitotic and pyknotic profiles were counted in these. Mitotic figures were identified as clear meta- or anaphase profiles adjacent to the RPE, which represents the ventricular margin where cell division takes place. Pyknotic profiles were present throughout the depth of the developing retina and were clearly identified as dense, round profiles that were heavily stained. Mitotic and pyknotic profiles were counted in complete retinal sections from one edge of the periphery to the other.
Results
For the purposes of comparison,
Figure 1 shows the photomicrographs of the melanin distributions in the RPE in each of the pigmentation phenotypes. In the fully pigmented animals, melanin was present continuously along the RPE
(Fig. 1A) , but was totally absent in the albino mouse
(Fig. 1B) . In the Beige mouse, melanin levels were lower than those found in the fully pigmented animal. Further, in many cases as well as there being normal melanosomes, there were giant melanosomes that occupied a large proportion of individual RPE cells
14 (Fig. 1C) . However, in all other respects, the retinas of these animals appeared to be qualitatively similar. In the Himalayan mouse, pigment was relatively uniformly distributed in the RPE, but its density was very low in all retinal regions
(Fig. 1D) . In both hypopigmented phenotypes, similar patterns of melanin distribution were found in all retinal regions. Also, the pattern of melanin distribution in the RPE was reflected in those present in the choroid.
Relative Melanin and Rhodopsin Levels and Rod Numbers in the Differentiated Retina
Levels of Ocular Melanin.
The results were standardized per eye rather than per total milligrams protein content (or weight), because it was found that, as animals matured, the amount of protein in the external tissue of the eye (non-RPE related) increased disproportionately compared with the increase in the melanin content. The total melanin content in the eyes was measured at three stages in animals after the end of cell production in the retina.
15 These were at D12, D24, and D210
(Fig. 2) . Each measurement was made against that found in age-matched albino mice.
Figure 2 shows that there was a clear positive trend in melanin accumulation in each of the three groups of pigmented mice during this period. The relative relationship between the amounts of melanin in each group remained roughly similar. The Himalayan animals had the least ocular melanin, whereas the levels in Beige mice were intermediate between those found in Himalayan and C57Bl mice. The differences between C57Bl animals and the hypopigmented groups became more marked as animals matured. At maturity, ocular melanin levels in Himalayan mice were only 17% of those found in C54Bl (significant at
P < 0.01, Mann-Whitney), whereas levels in Beige mice were only 47% (significant at
P < 0.01, Mann-Whitney). The difference between the two hypopigmented groups was also significant (
P < 0.05, Mann-Whitney;
Fig. 2 ).
Rod Numbers and Rhodopsin Levels.
Rod outer segment numbers were counted in the four different pigmentation phenotypes and compared with rhodopsin levels. One-way Kruskal-Wallis ANOVA showed significant differences in rod numbers between the four pigmentation phenotypes (
P < 0.005). There was again a clear trend of increased rod numbers with increased levels of pigmentation
(Fig. 3A) between the different mouse phenotypes. The same patterns were found when ocular rhodopsin levels were compared with melanin content, with a higher level of rhodopsin associated with increased pigmentation
(Fig. 3B) .
Rod numbers in albino mice were reduced by almost 25% compared with numbers found in C57Bl animals (significant at P < 0.01; Mann-Whitney). In Himalayan mice the reduction was 20% (significant at P < 0.001; Mann-Whitney). However, in the Beige mice a reduction of only 5% was found, which just failed to be statistically significant.
These differences in the number of rod outer segments were matched by the measurements of total ocular rhodopsin levels
(Fig. 3B) . A one-way ANOVA again revealed significant differences between the groups (
P < 0.005). The rhodopsin content of albino mice was just under half (49%) that found in C57Bl animals (significant at
P < 0.005; Mann-Whitney). There was a reduction of approximately a third (35%) in Himalayan mice and of 23% in Beige animals. These last two differences just failed to achieve statistical significance.
Melanin Levels and Patterns of Mitosis and Pyknosis in the Undifferentiated Retina
Levels of Ocular Melanin.
Levels of ocular melanin were also assessed during the early postnatal period when the retina is undifferentiated and were related to levels of mitosis and pyknosis. Postnatal D1 was selected as the starting point, because this is approximately when rod production peaks, and D6 as the last point, because rod production is largely over at this stage.
15
Levels of ocular melanin at D1, D3, and D6 are shown in
Figure 4 . The relative patterns found during this period are similar to those seen at later sages
(Fig. 2) , with there being a steady accumulation of melanin in each group. Although there was considerable relative variability in Beige animals during this period, the relative differences in melanin content between the three pigmented phenotypes followed a clear trend, and these relative differences were roughly in the range of those observed after D12.
From D1 onward, eyes of the C57Bl animals always contained more melanin than those of the other two groups, and those of Beige animals contained more than those of Himalayan mice. Only very limited amounts of melanin were present in the latter group. The relative differences were maintained at D3 and D6, although they were less marked between Beige and C57Bl eyes at D6.
The very low levels of pigmentation found in Himalayan compared with C57Bl mice at any day spanning this period (3% at D1, 7% at D3, and 13% at D6) correlates with low final rod numbers and rhodopsin contents in the adult, and are similar to data derived from the albino mice. Throughout this period, melanin levels in the Beige mice were much closer to those found in the C57Bl animals, rising to almost normal levels at D6. Again, this correlates with the almost normal levels of rhodopsin and rod numbers in these mice.
Mitotic and Pyknotic Profiles.
Mitotic and pyknotic profiles were counted at D1 and D3 to determine whether there was a simple relationship between levels of ocular melanin and cell production and death. Although there was a trend, with the hypopigmented animals tending to have more mitotic figures than the C57Bl mice, the differences were not great
(Fig. 5A) . All hypopigmented animals had more mitotic figures on D1 than the C57Bl mice, but this was not consistently the case on D3, when levels of mitosis in the Beige mice were lower than those in the fully pigmented animals. The levels of mitosis found in the Himalayan mice were markedly elevated at both stages.
Patterns of pyknosis were variable and did not follow a simple pattern during the period of examination, with more cell death occurring in the hypopigmented animals. More were found in Beige mice at both D1 and D3 than were present in the C57Bl animals. However, far less were present in the Himalayan mice. The albino mice had more on D1 but less on D3 than the pigmented animals. Hence, during this period, retinas from the hypopigmented group of animals tended to be more proliferative than in the normally pigmented animals, but there was no consistent pattern in the levels of cell death during these days.
Discussion
The results of this study clearly demonstrate a correlation between rod numbers and the amounts of ocular melanin present during the postnatal period of rod production. This finding is reinforced by the correlation between melanin at these early stages and the total amounts of ocular rhodopsin found in the adult. Although it is likely that prenatal melanin production may well also influence rod production, the clear correlation found in this study between postnatal melanin levels and rod numbers is consistent with there being a relationship between them. We have concentrated on postnatal melanin production because rod production peaks on the day of birth,
15 and also because measuring melanin levels in fetal eyes becomes increasingly difficult. Despite this, there is no evidence for differences in the time course of melanin production between the phenotypes used in this study.
14 Greater differences were found in rhodopsin levels than in the number of rods. The probable reason for this is that some rhodopsin is lost during the extraction process, no matter how carefully it is undertaken. The important point is the relative rhodopsin levels between the groups and that these reflect the relative trend found when rod counts were made.
In estimating the relative effects of melanin on the magnitude of the rod population, it is more helpful if relative differences in levels of melanin are pooled over the first 6 postnatal days and compared with those found in the C57Bl animals. When this is undertaken, it shows that adult Himalayan mice had only an average of 8% of the melanin found in C57Bl mice, whereas levels in Beige mice reached 60% of those in the fully pigmented group. Final rod numbers were 80% and 94% in Himalayan and Beige mice, respectively, when compared with that found in C57Bl animals, whereas levels of pigmentation were reduced by roughly 90% in Himalayan and 40% in Beige mice. Hence, half of the maximum level of pigmentation during development can result in almost normal rod numbers, whereas if pigmentation levels are reduced to approximately 10% of that in the normal animal, rod numbers reflect those found in albino. Hence, the minimum amount of melanin capable of significantly influencing final rod numbers must be more than 10% of that found in the normal animal.
Ilia and Jeffery
8 revealed that levels of retinal mitosis are elevated in developing albino rats compared with matched pigmented animals throughout development, and this is associated with an elevated wave of cell death that is at its maximum when rod production peaks. In this study, only two key stages were examined to determine whether mitosis and cell death were elevated. They revealed no simple relationship between levels of mitosis and melanin at the two time points examined in the hypopigmented phenotypes. However, patterns of mitosis were clearly abnormally elevated in Himalayan mice and slightly elevated in albino mice. Unfortunately, both hypopigmented mutants did not breed well, and thus it was not possible to undertake a comprehensive study of a wide range of developmental stages. Further, some caution is needed in the interpretation of these data in relation to the Beige mice, because in these animals there are differences, not only in the total amount of melanin, but also in how it is packaged.
14 Despite this, it should be emphasized that a disruption of melanin synthesis for any reason results in a common pattern of deficits.
3
Albino mammals have a range of neurologic abnormalities, including chiasmatic rerouting of many fibers from the temporal retina that should project ipsilaterally into the contralateral optic tract.
3 16 17 18 Levels of ocular pigmentation appear to have an interesting correlation with the degree of misrouting at the chiasm. Fewer temporal retinal ganglion cells adopt an abnormal chiasmatic pathway in hypopigmented mutants with relatively high levels of ocular melanin, compared with those in which levels of ocular melanin are significantly reduced,
19 20 although no serious attempt was made in prior studies to quantify ocular melanin. These results reflect those presented in this study, with levels of ocular melanin correlated with rod numbers and ocular rhodopsin content. Although it is possible that the extent of defects are in part dose dependent on levels of ocular melanin, it is not known whether this applies to the underdevelopment of the central retina found in albino animals, because this has not been systematically examined. To do this, it would be necessary to have an animal model that possesses a highly developed central retina and in which there is a range of pigmentation phenotypes. In this respect, the human would be an ideal model.
In this study, we have compared total ocular melanin with rod numbers and rhodopsin levels. Previously, no serious attempt had been made to quantify melanin levels in relation to the magnitude of retinal deficits found subjects with albinism. Our measurements have provided unambiguous results that have correlated well with rod numbers and levels of ocular rhodopsin. It would be interesting to know what contributions were made to rod numbers by the separate sources of ocular melanin, but although the RPE is a prime candidate for influencing rod production, there is no definitive study that excludes other sources, particularly the iris, which is adjacent to the germinal region of the developing retina. Hence, our results demonstrate a correlation between the global level of pigmentation of the eye and rod numbers.
It would be valuable to distinguish the relative influences of the different sources of melanin on rod production, but one critical factor limits this: The separate sources of melanin would have to be isolated during rod production, at approximately D1. Reliably harvesting the complete RPE alone, particularly in the large samples of hypopigmented mutant animals, when it can hardly be seen, is not possible.
In conclusion, although this study has clearly revealed a correlation between ocular melanin and rod numbers, the mechanism by which melanin regulates the production of a proportion of these cells is still unclear. There is evidence for elevated levels of mitosis and cell death during rod production in the rat.
8 There is also evidence that Dopa emanating from the RPE from the early stages of melanin synthesis has the capacity to regulate the cell cycle.
21 However, despite there being clear correlations between melanin and rods and the fact that agents have been identified that are known to regulate specific features of retinal development, there is still not a clear understanding of the specific route by which melanin modulates rod production.
Supported by a project grant from The Wellcome Trust.
Submitted for publication September 25, 2001; revised December 11, 2001; accepted December 18, 2001.
Commercial relationships policy: N.
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.
Corresponding author: Glen Jeffery, Institute of Ophthalmology, University College London, Bath Street, London EC1V 9EL, UK;
g.jeffery@ucl.ac.uk.
Schraermeyer U, Heimann K. Current understanding on the role of retinal pigment epithelium and pigmentation.
Pigment Cell Res. 1999;12:219–236.
[CrossRef] [PubMed]Sarna T. Properties and function of the ocular melanin: a photobiophysical view.
J Photochem Photobiol B. 1992;12:215–258.
[CrossRef] [PubMed]Jeffery G. The albino retina: an abnormality that provides insight into normal retinal development.
Trends Neurosci. 1997;20:165–169.
[CrossRef] [PubMed]Jeffery G, Darling K, Whitmore A. Melanin and the regulation of mammalian photoreceptor topography. Eur J Neurol. 1994;6:657–667.
Jeffery G, Brem G, Montoliu L. Correction of retinal abnormalities found in albinism by introduction of a functional gene in transgenic mice and rabbits.
Dev Brain Res. 1997;99:95–102.
[CrossRef] Grant S, Patel NN, Philp AR, et al. Rod photopigment deficits in albinos are specific to mammals and arise during retinal development.
Vis Neurosci. 2001;18:245–251.
[CrossRef] [PubMed]Ilia M, Jeffery G. Retinal mitosis is regulated by Dopa, a melanin precursor that may influence the time at which cells exit the cell cycle: analysis of patterns of cell production in pigmented and albino retinae.
J Comp Neurol. 1999;405:394–405.
[CrossRef] [PubMed]Harman AM, Beazley LD. Generation of retinal cells in the wallaby,
Setonix Brachyurus (Quokka).
Neuroscience. 1989;28:219–232.
[CrossRef] [PubMed]La Vail JH, Rapaport DH, Rakic P. Cytogenesis in the monkey retina.
J Comp Neurol. 1991;309:86–114.
[CrossRef] [PubMed]Donatien PD, Orlow SJ. Interaction of melanosomal proteins with melanin.
Eur J Biochem. 1995;232:159–164.
[CrossRef] [PubMed]de Grip WJ, Daemen FJ, Bonting SL. Isolation and purification of bovine rhodopsin.
Methods Enzymol. 1980;67:301–320.
[PubMed]Penn JS, Williams TP. Photostasis:. Regulation of daily photoncatch by rat retinas in response to various cyclic illuminance.
Exp Eye Res. 1986;43:915–928.
[CrossRef] [PubMed]Guillery RW, Price SD. The early development of retinal pigment in Beige mice: biological, molecular and clinical aspects of pigmentation. Proceedings of the XIIth International Pigment Cell Conference. Bagnara J Klaus SN Paul E Schartl M eds. Pigment Cell. 1985.303–312. University of Tokyo Press Tokyo.
Young RW. The life history of retinal cells.
Trans Am Ophthalmol Soc. 1983;81:193–228.
[PubMed]Guillery RW. Neural abnormalities of albinos.
Trends Neurosci. 1986;9:364–367.
[CrossRef] Drager UC, Olsen VF. Origins of crossed and uncrossed retinal projections in pigmented and albino mice.
J Comp Neurol. 1980;191:383–412.
[CrossRef] [PubMed]Thompson ID, Morgan JE. The development of retinal ganglion cell decussation patterns in postnatal pigmented and albino ferrets.
Eur J Neurosci. 1993;5:341–356.
[CrossRef] [PubMed]Balkema GW, Drager UC. Origins of uncrossed retinofugal projections in normal and hypopigmented mice.
Vis Neurosci. 1990;4:595–604.
[CrossRef] [PubMed]LaVail JH, Nixon RA, Sidman RL. Genetic control of retinal ganglion cell projections.
J Comp Neurol. 1978;182:399–421.
[CrossRef] [PubMed]Akeo K, Tanaka Y, Okisaka S. A comparison between melanotic and amelanotic retinal pigment epithelial cells in vitro concerning the effects of L-dopa and oxygen on cell cycle.
Pigment Cell Res. 1994;7:141–145.
[CrossRef] [PubMed]