Photoreceptor Degeneration and Loss of Retinal Function in the C57BL/6-C^2J Mouse

Abstract

purpose. The C57BL/6-c^2J (c2J) mouse strain has been shown in earlier studies to be highly resistant to light damage. Subsequent studies related this resistance to an amino acid substitution (leu450met) in a pigment epithelial enzyme (RPE65), which slowed the rate of rhodopsin regeneration. The present study was conducted to examine patterns of photoreceptor death, electrophysiological function (the ERG) and trophic factor expression over the life of the C57BL/6-c^2J retina.

methods. Observations were made on two C57BL/6J-c^2J substrains, one albino (Tyr/Tyr) and one pigmented (Tyr/ ⁺), and two nondegenerative strains, one albino (BALB/cJ) and one pigmented (C57BL/6J). Mice were raised in dim cyclic light (12 hours at 5 lux, 12 hours in the dark), and a developmental series of retinas of each strain was taken between postnatal day (P)4 and (P365⁺). Retinas were examined for cell death by using the TUNEL technique, stress-induced protein expression (FGF-2 and GFAP), and measures of retinal thickness. The dark-adapted ERG was recorded in dark-adapted conditions in early adulthood (13–15 weeks) and late adulthood (>1 year).

results. In both C57BL/6-c^2J substrains, the retina showed marked degenerative features when compared with two control strains, BALB/cJ (leucine at codon 450 in RPE65) and C57BL/6J (methionine). During development and into young adulthood, photoreceptor death rates were abnormally high, levels of two stress-inducible proteins (FGF-2 and GFAP) were abnormally high, and the ERG (electroretinogram) was significantly reduced in amplitude (<50% of values in BALB/cJ or C57BL/6J). The rate of photoreceptor death remained abnormally high into young adulthood (2–3 months) but decreased to control levels by 1 year. Accordingly, the thickness of the outer nuclear layer and the ERG were stable over the same period.

conclusions. Results suggest that a still-unidentified stress increases photoreceptor death in the C57BL/6-c^2J retina during the critical period of photoreceptor development and into young adulthood, upregulates stress-inducible factors, and markedly limits the amplitude of the ERG. These degenerative changes do not continue after early adulthood, the retina remaining stable in structure and function into late adulthood. The degenerative changes were apparent in both albino and pigmented C57BL/6-c^2J substrains. Their genetic cause remains unknown.

In a prior study, LaVail et al. ¹ screened inbred albino strains of mice from the Jackson Laboratories (Maine) for variations in the resistance of the retina to light damage. The BALB/cJ strain was found to be particularly sensitive to light damage and the C57BL/6-c^2J strain to be particularly resistant. Using these two strains, Danciger et al. ² subsequently mapped a quantitative trait locus (QTL) that could account for approximately 50% of the light-damage-resistance phenotype, to chromosome 3. The RPE65 gene mapped to this region and, given its role in the regeneration of rhodopsin in the visual cycle, ³ is considered a strong candidate for a role in resistance to light. Whereas mRNA levels of RPE65 are comparable between the two strains, sequencing has revealed a single-base-pair substitution (C to A at codon 450) in the C57BL/6-c^2J. This results in the substitution of methionine for leucine at the corresponding position in RPE65. The role of this substitution in determining resistance to light damage has been confirmed in RPE65 knockout mice, ⁴ and resistance to light damage can be shown to vary with the leu450met variation by modification of rhodopsin renewal kinetics ⁵ and with recovery from photobleaching. ⁶ Both Wenzel at al. ⁵ and Danciger et al. ² noted evidence of additional factors contributing to resistance to light damage. Danciger et al. reported several QTLs related to the resistance phenotype that were distinct from the RPE65 site. 

The present experiments were designed to assess photoreceptor death and trophic factor expression in the C57BL/6-c^2J strain during normal development. Results suggest that the C57BL/6-c^2J retina is degenerative, with abnormally high rates of photoreceptor death persisting into young adulthood and a significant reduction in the ERG. The rate of photoreceptor death declines to control levels by 1 year, and the ERG persists into late adulthood. The C57BL/6-c^2J retina may provide an example of a “stabilizing degeneration” or “stationary blindness,” in which mechanisms of retinal self-protection activated by the early degenerative process act to slow the degeneration of photoreceptors to control levels. 

Methods

Strains Studied

Retinas were examined in three strains of mice from the Jackson Laboratories (Bar Harbor, ME): an albino BALB/cJ and two strains of C57BL/6-c^2J. One c2J strain was homozygous and the other heterozygous for a G-to-T (G291T) substitution in the gene that encodes tyrosinase. ⁷ This variant results in an arginine-to-leucine substitution at codon 77 and, in the homozygote, gives an albino phenotype. ⁷ C57BL/6 mice were obtained from the Animal Resource Centre (ARC; Murdoch University, Perth, Australia). All mice used were raised in the Sydney University animal facility and reared in dim cyclic illumination (12 hours at 5 lux, 12 hours in the dark) from birth. In addition, data were available for the C57BL/6 strain, from Mervin and Stone. ⁸ All procedures were in accord with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Tissue Preparation and Labeling

Retinas were studied from a developmental series of each strain, from postnatal day (P)4 to the mature (P365⁺). Eyes were immersion-fixed in 4% paraformaldehyde for 2 hours and then cryoprotected by immersion in 15% sucrose overnight. Cryosections (20 μm) were cut and labeled with the TUNEL technique, as described previously. ⁹ The TUNEL technique detects the fragmentation of nuclear DNA characteristic of apoptosis. ¹⁰ To allow fragmenting DNA of apoptotic cells to be seen against a background of nondegenerative cells (Fig. 1) , the sections were counterstained with a DNA-specific dye, either bisbenzamide or propidium iodide. For bisbenzamide staining, sections were washed briefly in PBS, immersed in a highly dilute (<1:50,000) solution of bisbenzamide in PBS for 1 to 2 minute, and then washed briefly. For propidium iodide staining, sections were washed briefly in PBS then immersed in a 1:500 solution (0.1 M PBS) for 2 minutes at room temperature. Some sections were labeled with antibodies to FGF-2 and glial fibrillary acidic protein (GFAP), according to previously published protocols. ⁹  

Assessment of TUNEL Labeling

The frequency of TUNEL⁺ profiles in a tissue section provides an estimate of the rate of cell death in the population of cells at the time of tissue fixation. TUNEL⁺ photoreceptors were clearly identifiable at P9, at which age the separation of the neuroblast layer into the inner [INL] and outer (ONL) nuclear layers was complete, except at the most peripheral edge of the retina. ^{11 12 13} At P7, the site of the OPL could be identified by early cellular changes, and counts were recorded at this age for the developing INL and ONL. Counts were not attempted at earlier ages. Photoreceptor death rates were estimated as the frequency of TUNEL⁺ profiles in the ONL (expressed as number/per millimeter along a section through the retina), at successive developmental ages, from P9 to adulthood. 

To quantify cell death rates, TUNEL-labeled sections of retina were examined by fluorescence microscopy. Sections were analyzed which included or were close to the optic disc. Each section was scanned from one edge to the other in consecutive segments of 400 μm. The number of TUNEL⁺ profiles per segment, was recorded separately for the ONL and INL. For each animal, we examined two sections, and for each age of each strain, 3 to 10 animals were examined. Cell death trends did not differ detectably between the albino and pigmented c2J strains, and quantitative data from the two substrains were pooled. 

Assessment of FGF-2 and GFAP Immunolabeling

For quantitative comparison of FGF-2 and GFAP labeling, sections were immunolabeled with both antibodies, in the same labeling runs and with identical incubation and processing times. When imaged by confocal microscopy, the photomultiplier was set to match the ranges of labeling to be studied, and the settings were held constant. To minimize cross-talk between the two labels, sequential imaging protocols were used to record FGF-2 labeling (using the green-emitting Alexa488 fluorophore conjugated to goat anti-mouse IgG (H+L) and GFAP labeling (using the red-emitting Alexa 598 conjugated to goat anti-rabbit IgG [H+L]; Molecular Probes, Portland, OR). To assess FGF-2 labeling, the analysis tool of NIH Image (available by ftp at zippy.nimh.nih.gov/ or at http://rsb.info.nih.gov/nih-image; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD) was used to measure mean signal brightness of labeling over the full thickness of the ONL and INL, in two areas (one in the superior retina, one in the inferior) from one section of each eye studied. Measures from three animals of each strain were pooled. To assess GFAP labeling, a transect was defined along the length of the retina in the INL. The transect was 250 μm (512 pixels) long and crossed a series of Müller cell processes. Expression of GFAP in Müller cell processes then appeared as a series of peaks along the length of the transect. Measures were made in one section of each eye studied, in both the superior and inferior midperipheral retina, in three animals of each strain. 

Retinal Thickness Measurements

Measurements of ONL, INL, and inner to outer limiting membrane (ILM-OLM) thicknesses were taken on digital images of cryosections labeled with DNA-specific dyes (propidium iodide, bisbenzamide). Measurements were made on one section of each eye examined. In each section a total of four measurements were made, at 400 and 1600 μm from each edge. The measurements were then averaged for each section. 

Electroretinography

At both P100 and P365, electrophysiological recordings were taken in 3 to 7 animals from each of the mouse strains. Methods used were according to previously published protocols. ¹⁴ Briefly, mice were dark adapted overnight and placed under dim red illumination. Anesthesia was achieved with intramuscular injections of ketamine (60 mg/kg) and xylazine (10 mg/kg) and pupils were dilated with 0.5% tropicamide. Corneal hydration and electrical contact were maintained with synthetic tears (Viscotears; Carbomer 940 2 mg/g; CIBA Vision, Baulkham Hills, NSW, Australia). Body temperature was maintained close to 37°C with an electric blanket controlled by feedback from a rectal temperature probe (Harvard Apparatus Inc., Holliston, MA). The ERG was recorded between a platinum wire touching the cornea and an Ag/AgCl pellet electrode (Clarke E206; SDR Clinical Technology, Middle Cove, NSW, Australia) in the mouth. Animals were placed in the center of a custom-made Ganzfeld dome (approximately 60 cm in diameter) with the flash source positioned at 45° above the animal’s head. 

The flash stimulus was provided by a flash unit (model 70; Metz GmBH, Zirndorf, Germany) and flash intensity was attenuated over a 7-log-unit range with neutral density filters. To minimize the cone contribution to the ERG, all stimulus flashes were delivered with a filter (Wratten 47; Eastman Kodak, Rochester, NY) in place. Stimuli were controlled, recorded, and displayed on a computer workstation (MacLab/200 system and Scope software; ADInstruments, Castle Hill, NSW, Australia). Responses were band-pass filtered at 0.3 to 200 Hz. A 50-Hz notch filter was used to minimize mains noise. With attenuations between 1.4 and 7.0 log units, two to three responses were averaged with an interstimulus interval of between 20 (ND 7.0) and 120 (ND 1.4) seconds. At the lowest attenuation (i.e., the brightest flash), a single response was recorded after an interval of at least 2 minutes from any prior flash. 

The final estimate of flash output (with the Wratten 47 filter; Eastman Kodak) in photoisomerizations per rod per flash was 1.09 × 10⁷ (see Yu et al. ¹⁴ for methods used to convert irradiance to illuminance units). In practice, the brightest flash used was with attenuation by the 0.7-ND filter (φ = 2.18 × 10⁶). This was of sufficient intensity to elicit saturated a-wave responses. 

The a-wave amplitude was measured from baseline to the a-wave trough and implicit time (latency) was measured to the trough peak. The b-wave amplitude was determined from a-wave trough to b-wave peak. 

Results

Photoreceptor Death and Survival in C57BL/6-c^2J Retina

Representative TUNEL-labeled regions of BALB/cJ and C57BL/6-c^2J retinas are shown in Figures 1A 1B 1C 1D 1E 1F 1G 1H at ages P17 to 21, P30, P80, and P365. At these ages, TUNEL⁺ cells were almost exclusive to the ONL, indicating that most dying cells were photoreceptors. TUNEL⁺ ONL cells were most frequent in the P17 to P21 age range and in the C57BL/6-c^2J strain. Also demonstrated is the good preservation of the laminar structure in the C57BL/6-c^2J retina at 12 months (Fig. 1H) . 

Quantitative Comparisons of Cell Death

The pattern of cell death in the C57BL/6-c^2J retina, from early postnatal life to young adulthood, is shown in Figures 2A and 2B , with data from the nondegenerative albino BALB/cJ mouse shown for comparison. At the earliest ages measured (P9, P11), TUNEL⁺ profiles were more frequent in the INL than the ONL (Figs. 2A 2B) , in both C57BL/6-^C2J and BALB/cJ retinas. Thereafter, cell death in the INL declined rapidly to near zero at P23. Cell death in the ONL was initially low in both strains, rose to a peak at P15 to P19, and then declined toward adult levels. The rapid diminution of cell death in the INL and the subsequent wave of death in the ONL have been reported previously in the C57BL/6J mouse ⁸ and in other mammals. ^{9 15} Statistically, the frequency of TUNEL⁺ profiles was significantly higher in the C57BL/6-c^2J retina than in the BALB/cJ retina in the INL at P8 and P12 and in the ONL at P9 and from P15 through to P65 (P < 0.02; two-tailed t-test). In relative terms, the frequency of TUNEL⁺ ONL profiles was approximately twice as high in the C57BL/6-c^2J retina as in the BALB/cJ retina during the period of peak cell death (P15–P19); this ratio rose to 11-fold higher at P65. 

The C57BL/6-c^2J is a variant of C57BL/6J stock. Figure 2C compares ONL death during postnatal life in the BALB/cJ (data from Fig. 2B ) and the C57BL/6J (data from Ref. ¹⁶ ) strains. There is considerable congruence between the two strains in the time course of photoreceptor death, and in the absolute rates observed (both peak at 10 to 12 TUNEL⁺ profiles per millimeter). Differences between TUNEL⁺ profile frequencies were not significant, except that the P26 value for the C57BL/6J mouse was significantly (P < 0.05) higher than the values for the BALB/cJ at P23 and P30. 

The C57BL/6-c^2J data in Figure 2B combine measurements made on two substrains: one homozygous for the albino mutation (Tyr/Tyr) and one heterozygous (Tyr/ ⁺) and therefore pigmented. ONL data for the two substrains are separated in Figure 2D . The wave of photoreceptor death appeared to occur 2 to 3 days earlier in the homozygote than in the heterozygote, but the frequency of TUNEL⁺ profiles in the ONL peaked at 20 to 25/mm in both substrains. 

Duration of Photoreceptor Degeneration

In the young adult (Fig. 3A) , the frequency of TUNEL⁺ profiles in the INL was low in all strains (<1 TUNEL⁺ profile per millimeter). TUNEL⁺ frequencies in the ONL were several-fold higher in C57BL/6-c^2J strains than in either control, and the differences were highly significant (P < 0.0025; two-tailed t-tests). Data for 1-year-old animals were available for the BALB/cJ and C57BL/6-c^2J strains (Fig. 3B) . The frequency of TUNEL⁺ profiles in the INL remained low. The frequency of TUNEL⁺ cells in the ONL changed little in the BALB/cJ strain, remaining at low values (<1/mm). In both C57BL/6-c^2J substrains, however, TUNEL⁺ frequencies fell markedly from more than 4/mm to less than 1/mm. This eliminated any difference in photoreceptor death rates between the C57BL/6-c^2J strains and the BALB/cJ strain or the adult C57BL/6J strain. ¹⁶ The degenerative process in the C57BL/6-c^2J strain thus seems to stop between 2 and 12 months, leaving many surviving photoreceptors (Fig. 1H) . This confirms the conclusions of Danciger et al. ² and LaVail et al. ¹⁷ that the late adult retina in this strain is not actively degenerating. 

FGF-2 and GFAP Protein Levels

In the control C57BL/6J strain, FGF-2 levels in midperipheral retina were low, except in the cell bodies of Müller cells in the INL (Fig. 4A) and in the nuclei of astrocytes at the ILM (confirming previous reports in unstressed retina ^{16 18} ). GFAP levels were high only in astrocyte processes at the ILM (right panel in Fig. 4A , confirming Ref. ¹⁶ ). In the C57BL/6-c^2J mouse (Fig. 4B) , FGF-2 levels were raised in Müller cell somas in the INL and in the cytoplasm of photoreceptors in the ONL. GFAP was prominent in the radially oriented processes of Müller cells (Fig 4B , right). In the BALB/cJ mouse (Fig. 4C) , both FGF-2 and GFAP levels were raised, compared with the C57BL/6J. 

When FGF-2 signal levels in the ONL and INL were quantified (Figs. 5A 5B) FGF-2 levels in the C57BL/6-c^2J strain proved significantly higher than in the stock C57BL/6J strain in both the INL and ONL (P = 0.001, P = 0.00004, respectively; two-tailed t-tests). FGF-2 levels were also significantly greater in the BALB/cJ strain than in the C57BL/6J strain (P = 0.000036) in the ONL, but not in the INL (P = 0.07). 

Because the upregulation of GFAP in Müller cells occurred within the narrow processes, we devised the semiquantitative analysis shown in Figures 5C and 5D . For three animals in each group, GFAP immunofluorescence was quantified along a transect that passed along the INL (e.g., between the black and white arrows in Fig. 4B ). A GFAP⁺ Müller cell process shows in these graphs as a peak. GFAP levels were markedly higher in the C57BL/6-c^2J strain than in the C57BL/6J stock (Fig. 5C) . The technique also detected (Fig. 5D) the raised level of GFAP in the BALB/cJ strain apparent to inspection of Figure 4C . 

The Thickness of the C57BL/6-c^2J Retina

The thickness of the retina and two of its layers was assessed in three groups of animals aged P56 to P83, P125 to P153, and older than P365. Three measures of thickness were recorded: the distance between the ILM and the OLM, the thickness of the INL and the thickness of the ONL. Present data (Table 1) show no trend for the C57BL/6J-c^2J retina to thin with age in any of these parameters (confirming previous reports ^{1 2} ). There was also no trend for the ONL to be thinner in the C57BL/6-c^2J retina than in either control. To control for the possibility that the thickness of the C57BL/6J-c^2J ONL was artifactually increased by the obliquity of the section, the ONL-to-INL thickness ratio was calculated. Where degeneration occurs preferentially in photoreceptors, this ratio should decrease, independent of the angle of section. As Table 1 shows, this ratio was not lower in the C57BL/6-c^2J retinas than in the age-matched controls. 

The ERG in the C57BL/6-c^2J Retina

Representative responses to a series of flashes of increasing intensity are shown in Figure 6A 6B 6C 6D , for young adults of each of the four strains studied. Responses from the C57BL/6J-c^2J strains (Figs. 6B 6C) show an a-wave at high intensities, oscillatory potentials and a b-wave, but were smaller in amplitude than their comparably pigmented control strains (C57BL/6J, BALB/cJ; Figs. 6A 6D ). Responses for old (>P471) adults are shown Figures 6E 6F 6G for the C57BL/6-^C2J substrains and the BALB/cJ strain. Amplitude data over the full flash intensity series are summarized in Figure 7 in groups of animals of each of the strains studied. 

In the young adults, a-wave amplitudes in the albino (BALB/cJ) and pigmented (C57BL/6J) control strains were significantly larger than in either C57BL/6J-c^2J strain. The amplitude of the a-wave was greater in the albino control strain (BALB/cJ) than in the C57BL/6J-c^2J strain over the full intensity range used, and the difference was significant (P < 0.05; 2-tailed t-test) at flash intensities above ND 4.3 (φ = 546). The a-wave amplitude in the pigmented control (C57BL/6J) was also greater than in the C57BL/6J-c^2J strains at intensities higher than ND 3.3 (P < 0.004). We also noted a difference between the two control strains, the a-wave being smaller in the C57BL/6J control than in the BALB/cJ at lower intensities (P < 0.05 at intensities less than ND 3.7), and larger at higher intensities (P < 0.02 at intensities greater than ND 3.0). The amplitude of the b-wave was also greater in the control strains than in the C57BL/6J-c^2J strains (Fig. 7B) and the differences between each control and each C57BL/6J-c^2J strain were significant (P < 0.009 for all intensities). The differences between the two C57BL/6J-c^2J strains were not significant. The b-wave was larger in the C57BL/6J pigmented control than in the BALB/cJ albino, the difference being significant at the two highest intensities used (ND 1.4, 0.7). The peak latency of the a-wave varied with coat color, being shorter in the two albino strains (BALB/cJ and C57BL/6J-c^2J Tyr/ ⁺; Fig. 7C ) at higher intensities. The albino/pigmented differences were significant (P < 0.001) at stimulus intensities of ND 3.3 and greater. Differences between the two albino strains were not significant. Latencies were shorter in the pigmented C57BL/6J control than in the C57BL/6J-c^2J Tyr/ ⁺, and the differences were significant (P < 0.05) at intensities of ND 5 and greater. 

In brief summary of the young adult data, the amplitude of a- and b-waves was lower in C57BL/6J-c^2J strains than in the control animals, independent of coat color, whereas the latency of the a-wave varied with coat color, being shorter in the albino, independent of the C57BL/6J-c^2J mutation. 

In the older adults, the amplitude of the a-wave in the albino BALB/c control was significantly reduced from its young adult values (P < 0.05 on a two-tailed t-test for attenuation of 4 log units or less) and was no longer greater than in the C57BL/6-c^2J strains. The b-wave was correspondingly reduced in (Fig. 7E) . By contrast, the amplitudes of the a- and b-wave in the C57BL/6-c^2J were not significantly lower in the older group, when compared with young adults of the same strain (compare Figs. 7D with 7A and 7E with 7B ; P > 0.05 at all attenuations). The relationship between coat color and a-wave latency noted for the young adults (shorter in albinos) also held for the older animals (Figs. 7C 7F) . 

Discussion

Duration of Degeneration of the C57BL/6-c^2J Retina

The present data show several degenerative features in the C57BL/6-c^2J retina. In the juvenile and young adult, photoreceptor death occurred at significantly higher rates than in BALB/cJ or C57BL/6J controls (Figs. 2 3) , the expression of stress-inducible proteins (FGF-2, GFAP) increased (Figs. 4 5) , and the ERG was markedly reduced (Figs. 6 7) . Present data also provided evidence that by late adulthood (12 months) the process of degeneration had stabilized. Photoreceptor death rates were as low as in BALB/cJ or C57BL/6J strains (Fig. 3) , the ONL was not significantly thinned compared with the BALB/cJ, the ERG had not declined from young adulthood, and the retina appeared structurally intact (Fig. 1) . 

Although most photoreceptor degenerations are progressive, nonprogressive (stationary) forms have long been recognized. Many are heritable and involve loss of rod vision and are therefore known as congenital stationary night blindness (CSNB). The genetic causes of CSNB vary and occur through autosomal dominant, autosomal recessive, and X-linked modes of inheritance. ¹⁹  

Two causative genes (CACNA1F and NYX) for X-linked CSNB have now been identified through positional cloning strategies and are known as the complete (CSNB1) and incomplete (CSNB2) forms, respectively (based on the extent of b-wave involvement). The calcium channel (CANCA1F) gene encodes the retina-specific calcium channel α1-subunit ^{19 20 21} and may have a role in mediating glutamate release from photoreceptors. ²² The nyctalopin (NYX) gene is thought to have a role in transmission from photoreceptors to ON-bipolar cells (depolarizing bipolar cells) or a role in organization of protein complexes involved in this pathway. ^{23 24 25 26} Other forms of CSNB described in human pedigrees cosegregate in an autosomal dominant manner and include mutations in the β-subunit of rod cGMP, ²⁷ the α-subunit of rod transducin, ²⁸ and the rhodopsin molecule. ^{29 30 31} CSNB is thus a heterogeneous category of diseases, with variability in the mode of inheritance, the gene involved, and in the mutations that occur in the same gene. For example, more than 20 different mutations in CACNA1F have been described as causing CSNB. ¹⁹  

Three murine models of CSNB have been reported. The nob mouse defect, recently attributed to a mutation in NYX, ²⁶ is thought to provide a model for the complete form of X-linked CSNB (CSNB1). In the nob mouse, ³² retinal structure and the a-wave of the electroretinogram appear normal, but the b-wave is greatly reduced, indicating a failure of photoreceptor transmission to rod bipolars. The second model is an arrestin knockout, which Chen et al. ³³ suggest is a model of human Oguchi disease, also considered a form of CSNB. The third model expresses the G90D rhodopsin mutation reported in humans by Sieving et al. ³¹ It is characterized by raised scotopic thresholds and minimal rod cell loss. ³³ Given that G90D rhodopsin was shown to activate transducin independent of light, ³⁴ “constitutive activation” is proposed as the mechanism by which the rod photoresponse is desensitized. ³³  

Present data suggest that the C57BL/6-c^2J mouse is a naturally occurring model of CSNB. It appears to be distinct from the nob mouse, in which the a-wave is unaffected but the b-wave is severely reduced and inheritance is X-linked, and distinct from the arrestin knockout, in which the degeneration of the ONL is progressive. It shares similarities with the G90D rhodopsin mutation in that visual thresholds are elevated and cell loss is minimal. The mechanisms leading to the c2J phenotype are yet to be investigated. 

Patterns of Age-Related Photoreceptor Death

Several recent reports have compared photoreceptor death in C57BL/6J and BALB/cJ strains, ^{2 36 37} noting that rod loss occurs throughout adult life in both strains but is faster in the BALB/cJ. This difference in the rate of photoreceptor loss has formed a phenotype for the identification of genes controlling the difference. ³⁷ Present data for the C57BL/6-c^2J strain suggest a distinctive pattern of age-related photoreceptor death, characterized by high early rates and low adult rates of photoreceptor death, early decline, and late stability of the dark-adapted ERG and early upregulation of protective factors. This distinct phenotype may be of use in the identification of genes controlling the stability of photoreceptors. 

Cause of C57BL/6-c^2J Degeneration

The C57BL/6-c^2J variant was first recognized as an albino variant in the Jackson Laboratories, ³⁸ and became of interest when LaVail et al. ¹ reported that its photoreceptors are more resistant to light damage that those of other albino forms, such as the BALB/cJ mouse. That resistance is independent of coat color (the tyrosinase mutation), however, and has been shown to arise in large part from a leu450met substitution in RPE65, which originates from the C57BL/6J background, ^{2 5} not from a mutation unique to the C57BL/6-c^2J strain. The present finding that the C57BL/6-c^2J retina is degenerative from the critical period to young adulthood and then becomes highly stable is novel. Both pigmented (Tyr/ ⁺) and albino (Tyr/Tyr) substrains of the C57BL/6-c^2J mouse were tested. Both showed excess photoreceptor degeneration, raised levels of stress-induced proteins, early loss of ERG amplitude, and late stability. These patterns of photoreceptor instability and stability in the C57BL/6-c^2J retina thus appear to arise from a mutation that is independent of the tyrosinase mutation and remains unknown. 

³  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported by Retina Australia, the National Health and Medical Research Council of Australia, and the Medical Foundation of the University of Sydney.

 Submitted for publication August 6, 2003; revised December 26, 2003; accepted January 9, 2004.

 Disclosure: A. Bravo-Nuevo, None; N. Walsh, None; J. Stone, 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.

 

TUNEL-labeled cells in the mouse retina (green), with normal nuclear DNA labeled with bisbenzamide (red). (A, B) Juvenile (P17–P30) BALB/cJ, (C) young adult (P80), and (D) a mature (12 month) BALB/cJ retinas. (E–H) Corresponding age series from the pigmented C57BL/6-c^2J (Tyr/ ⁺) strain.

Figure 2.

 

Time course of naturally occurring cell death, estimated as TUNEL⁺ profiles per millimeter. *Time points at which the frequency of TUNEL⁺ profiles differed significantly (P < 0.01 on a one-tailed t-test) between the strains. (A) Comparison of the nondegenerative BALB/cJ strain and the C57BL/6-c^2Jsubstrains in the INL. (B) As in (A), but in the ONL. *Time points at which the frequency of TUNEL⁺ profiles in the INL or ONL was significantly (P < 0.05; two-tailed test) higher in the C57BL/6-c^2J retina than in the BALB/cJ retina. (C) Comparison of TUNEL-labeling of the ONL in two nondegenerative strains (BALB/cJ and C57BL/6J; data from Mervin and Stone ⁸ ). The data obtained differed significantly at one time point. (D) Comparison of ONL labeling in the two C57BL/6-c^2J substrains.

Figure 3.

 

Photoreceptor death rates in the four strains examined, estimated as TUNEL⁺ profiles per millimeter of the ONL. Each histogram bar represents the mean ± SD (error bar) of measurements in a group of animals from each strain. (A) In the young adult (2 months old): groups sizes were BALB/cJ, 6; C57BL/6J, 7; C57BL/6-c^2J Tyr/Tyr, 7; C57BL/6J Tyr/ ⁺, 6. (B) In late adulthood (12 months): group sizes were BALB/cJ, 6; C57BL/6-c^2J Tyr/Tyr, 10; C57BL/6J Tyr/ ⁺, 4.

Figure 4.

 

Immunohistochemical labeling for FGF-2 (green) and GFAP (red) in the mouse retina. Each row of images represents one region of retina, with the FGF-2 and GFAP signals combined (left) and separated (middle and right). (A) Midperipheral regions of (A) C57BL/6J, (B) C57BL/6-c^2J, or (C) BALB/cJ retina.

Figure 5.

 

Quantification of immunolabeling. (A, B) The intensity of FGF-2 labeling in the ONL was measured over the full thickness of the INL (A) and ONL (B) in three animals from each of the C57BL/6J, C57BL/6-c^2J, and BALB/cJ strains. In the INL, labeling intensity was significantly (P < 0.05 in a t-test) greater in the C57BL/6-c^2J strain than in the C57BL/6J strain (A, *). In the ONL, labeling intensity in both the C57BL/6-c^2J and BALB/cJ strains was significantly greater than in the C57BL/6J strain. (C, D) Labeling intensity for GFAP measured along a 250 μm transect chosen to cross Müller cell processes at right angles, where they cross the INL. A GFAP⁺ process shows as a peak in the trace. Traces from three different animals are averaged for each strain. GFAP-labeling was consistently higher in the C57BL/6-c^2J strain than in the stock C57BL/6J strain (C). Labeling was more weakly upregulated in the BALB/cJ strain than in the C57BL/6J strain (D).

Table 1.

 

Thickness Measurements in the Retinas of the Mouse Strains Studied

Strain	BALB/cJ	C57BL/6-c2J	C57BL/6
Aged P56-83	n = 12	n = 14	n = 4
ILM-OLM	146 ± 26.03	123 ± 28.85	168 ± 6.91
ONL	56 ± 7.96	52 ± 10.62	53 ± 2.08
INL	37 ± 10.43	29 ± 7.66	32 ± 2.76
ONL/INL	1.51	1.79	1.66
Aged P125-53	n = 3	n = 6
ILM-OLM	131 ± 2.87	136 ± 22.02
ONL	55 ± 4.27	53 ± 4.27
INL	31 ± 3.74	30 ± 6.09
ONL/INL	1.77	1.77
Aged P > 365	n = 6	n = 12
ILM-OLM	145 ± 13.3	142 ± 13.68
ONL	57 ± 6.86	64 ± 10.18
INL	32 ± 1.15	34 ± 6.03
ONL/INL	1.78	1.88

 Measures were made of the thickness of the whole retina, from the inner to the outer limiting membranes (ILM-OLM), and of the INL and ONL. Data are mean thicknesses expressed in micrometers ± SD.

Figure 6.

 

Intensity response series recorded in dark-adapted conditions with attenuations of 7.0 to 0.7 log units. In young adults, ERG amplitude was larger in the BALB/c and C57BL/6J control mice than in the C57BL/6-c^2J substrains. By late adulthood, the BALB/c ERG had declined to the amplitude range found in the C57BL/6-c^2J substrains. (A–D) Responses from young adults of four strains, aged approximately 15 weeks. The responses shown for the C57BL/6-c^2J substrains are small within the range observed. (E–G) Responses of old adults (P470). The responses shown for the C57BL/6-c^2J substrains are large within the range observed.

Figure 7.

 

Amplitude and latency data, for groups of animals (n = 5–7 for all groups except the late adult C57BL/6-c^2J Tyr/ ⁺, for n = 3). Error bars show ± 1 SD. (A–C) Young adults, aged approximately 15 weeks (P100). (D, E) Old adults, aged more than P470 (P471–P542). Data were available for the BALB/cJ and C57BL/6-c^2J strains.