Representative histologic micrographs of one rabbit from each dose group are shown in
Figure 5. No histologic damage was evident in the retina of the rabbit injected with the low dose of cefuroxime (
Fig. 5a). The retina from the experimental eye retained its normal layered structure in regions close to the injection site and in remote loci. Similar observations were seen in a total of six rabbits belonging to the low-dose group that were tested for structural retinal damage.
The experimental retina from the rabbit treated with the high dose of cefuroxime (
Fig. 5b) showed marked damage when compared with the retina of the control eye. Retinal damage was most apparent in the inferior retina probably reflecting our attempt to inject the drug in the region of the optic disc and the spread of the drug down toward the inferior retina (
Fig. 5b, top left panel). Evidence of the damage was loss of the retinal layered structure, disorganization, and retinal thinning. Remote retinal loci in the inferior retina and most of the superior retinal regions of the experimental eye treated with the high-dose cefuroxime appeared to be normal (
Fig. 5b, bottom row).
Histologic findings, similar to those presented in
Figure 5b, were observed in eight rabbits, belonging to the high-dose group, but the magnitude of retinal structural damage differed between rabbits. Variability in the degree of cefuroxime-induced damage was expected because the intravitreal injection of the drug could not be identical between different animals treated at different times.
To assess the degree of cefuroxime-induced retinal structural damage, we measured the total length of damaged retina in one section of the experimental eye in each of the eight rabbits and compared it in
Figure 6a to its dark-adapted b-wave
V max ratio. Despite technical difficulties in length measurements and the small sample size (
n = 8), the trend is clear: the larger the region of retinal damage the larger the degree of ERG deficit, as represented by a lower dark-adapted b-wave
V max ratio. The relationship between retinal damage defined functionally from the ERG and the length of damaged retina fits quite well to a linear equation (
R = 0.94), supporting the notion that the ERG responses reflect algebraic summation of contributions by small retinal regions.
We also measured retinal thickness, from the vitreoretinal junction to the tips of the photoreceptor outer segments in the control eyes and experimental eyes of both groups of rabbits. Thickness measurements were obtained in the inferior retina at a distance of 2 mm from the optic disc, a region that was consistently damaged in the eyes exhibiting cefuroxime-induced damage. In the low-dose group (
n = 6), we measured an average (±SD) retinal thickness of 168.28 ± 4.15 μm and 170.56 ± 8.83 μm in the experimental and control eyes, respectively. These values did not differ significantly (
P = 0.360), as determined by Student's paired
t-test. In the high-dose group (
n = 7), a trend was seen toward lower retinal thickness of the experimental eyes compared with the control eyes, but the average (±SD) retinal thickness of the experimental eyes, 138.93 ± 50.42 μm was not significantly different (
P = 0.059) from that of the control eyes, 164.06 ± 24.19 μm. We attribute this observation to the large variability in the degree of retinal damage, as expressed by the large standard deviation of the measurements. To test this possibility, we compared the relationship between the degree of functional damage, as assessed from the dark-adapted b-wave
V max ratio, and the retinal thickness ratio (experimental eye/control eye) in seven rabbits from the high-dose group (
Fig. 6b). The data show some variability, but clearly indicate that the larger the degree of structural damage (lower retinal thickness ratio), the larger the degree of ERG damage (lower dark-adapted b-wave
V max ratio). It should be noted that since we compared the ERG deficit to total retinal thickness, a nonrecordable ERG was expected for a retinal thickness of more than 0.
Glial fibrillary acidic protein (GFAP) is an intermediate filament that is normally expressed in astrocytes but not in Müller cells in the retina.
20 However, in a variety of retinal injuries including retinal detachment,
21 ischemia,
22 and increased intraocular pressure,
23 GFAP expression in retinal Müller cells became apparent. Therefore, GFAP expression in Müller cells is widely used as a molecular indicator for retinal stress.
Representative micrographs of experimental and control retinas of one rabbit from each dose group are shown in
Figures 7. GFAP expression in Müller cells is evident in the experimental retinas from the rabbits in both dose groups (
Fig. 7, top panels), but not in the control retinas (
Fig. 7 bottom panels). The expression of GFAP in Müller cells was more extensive in the retina exposed to the high dose of cefuroxime (
Fig. 7 top right panel) compared to that of the retina exposed to the low dose of the drug (
Fig. 7, top left panel). In the retinas of the experimental and control eyes of both groups, GFAP immunoreactivity was also found in astrocytes, when retinal sections from the central region—the region of the medullary rays—were stained for the protein (not shown here). Similar findings were observed in retinas from two rabbits belonging to the low-dose group and in those of four rabbits belonging to the high-dose group.
The finding of GFAP expression in Müller cells of the retinas exposed to high-dose cefuroxime was expected, considering the functional (ERG responses) and histologic damage that was found. However, GFAP expression in the Müller cells of the retinas exposed to the low dose of cefuroxime was not expected, considering the normal ERG responses and histologic findings at the end of follow-up.