Intravitreal injection of ciliary neurotrophic factor (CNTF) causes peripheral remodeling and does not prevent photoreceptor loss in canine RPGR mutant retina

Abstract

Ciliary neurotrophic factor (CNTF) rescues photoreceptors in several animal models of retinal degeneration and is currently being evaluated as a potential treatment for retinitis pigmentosa in humans. This study was conducted to test whether CNTF prevents photoreceptor cell loss in XLPRA2, an early onset canine model of X-linked retinitis pigmentosa caused by a frameshift mutation in RPGR exon ORF15. Four different treatment regimens of CNTF were tested in XLPRA2 dogs. Under anesthesia, the animals received at different ages an intravitreal injection of 12 microg of CNTF in the left eye. The right eye served as a control and was injected with a similar volume of phosphate buffered saline (PBS). Ocular examinations were performed regularly during the treatment periods. At termination, the dogs were euthanatized, eyes collected and the retinas were processed for embedding in optimal cutting temperature (OCT) medium. The outer nuclear layer (ONL) thickness was evaluated on H&E sections and values in both CNTF- and PBS-treated eyes were compared. Morphologic alterations in the peripheral retina were characterized by immunohistochemistry using cell-specific markers. Cell proliferation in the retinas was examined on semi-thin plastic sections, and by BrdU pulse-labeling and Ki67 immunohistochemistry on cryosections. All CNTF-treated eyes showed early clinical signs of corneal epitheliopathy, subcapsular cataracts and uveitis. No statistically significant difference in ONL thickness was seen between the CNTF- and PBS-injected eyes. Prominent retinal remodeling that consisted in an abnormal increase in the number of rods, and in misplacement of some rods, cones, bipolar and Müller cells, was observed in the peripheral retina of CNTF-treated eyes. This was only seen when CNTF was in injected before the age at which the canine retina reaches full maturation. In XLPRA2 dogs, intravitreal injections of CNTF failed to prevent photoreceptors from undergoing cell death in the central and mid-peripheral retina. CNTF also caused ocular side-effects and morphologic alterations in the periphery that were consistent with cell dedifferentiation and proliferation. Our findings suggest that some inherited forms of retinal degeneration may not respond to CNTF's neuroprotective effects.

Figure 1

Schematic of a retinal section showing the location of the sites S1, S2, and S3. S1: 2000 ± 500 μm from the optic disc; S3: 2000 ± 500 μm from the ora serrata; S2: equidistant (± 500 μm) from S1 and S3.

Figure 2

Values (mean, and range) of intraocular pressure in CNTF- and PBS-treated eyes. Data is provided for dogs from treatment Group #3 (XLPRA2; n=3), and rcd1 control group (n=3) that were injected at 7 and 10 weeks of age. Pre-injection values of intraocular pressure are indicated at time 0. In the days following both CNTF injections a decrease in IOP was observed. The intraocular hypotension was transient, since IOP values returned to pre-injection values within approximately 20 days.

Figure 3

Ocular lesions caused by intravitreal injection of CNTF. (A) Corneal epithelial haze (arrowheads), and posterior subcapsular cataracts (arrows) 7 weeks post-injection. A corneal haze and sutural cataract had been detected in this dog 11 days following the first CNTF injection. (B) The earliest lens changes involved the extremities of the posterior Y sutures (arrows). (C) Immature cataract with anterior and posterior cortical opacification.

Figure 4

Histological lesions of the cornea and lens present after intravitreal injection of CNTF. (A) Normal aspect of the corneal epithelium in a PBS-injected eye. Note the presence of normal basal cells (arrows). (B) Cornea of the contralateral eye that was injected with CNTF. Note the thinning of the corneal epithelium due to the loss of the basal and suprabasilar (wing) cells. (C) Cataract caused by CNTF. Swelling and disorganization of lens fibers was observed in the posterior cortex, and was associated with the retention of lens fiber nuclei (arrows). (D) High magnification of the posterior subcapsular region of the same lens as in (C). Note the presence of large vacuoles that contain eosinophilic material (asterisks), and swollen lens fibers. H&E stain; scale bars: (A, B) 20 μm; (C) 80 μm; (D) 40 μm.

Figure 5

Scatter plots comparing the ONL thickness (mean number of rows of nuclei at S1 and S2 sites combined) along the four meridians of eyes injected with CNTF or PBS. No statistically significant difference in ONL thickness was seen in any of the XLPRA2 treatment Groups #1 – #3. CNTF caused a statistically significant increase in ONL thickness in rcd1. *: < 0.05; **: < 0.001; ND: not determined; OS: left eye; OD: right eye.

Figure 6

Illustration of the effect of CNTF on outer nuclear layer thickness in the mid-peripheral retina (Site S2) of a 14 week-old XLPRA2 dog (treatment Group #3) (A₂), and a 14 week-old rcd1 dog (rcd1 control group)(B₂). Contralateral eyes served as controls and were injected with PBS (A₁, B₁). CNTF caused an increase in the number of photoreceptor nuclei in rcd1 (compare B₂ to B₁), but no differences were seen in XLPRA2 (compare A₂ with A₁). CNTF often caused the vitreal border of the ONL to appear wavy (A₂, B₂). On these cryosections, RPE separation from the neuroretina was an artifact. H&E stain; scale bars: 20 μm.

Figure 7

Rate of photoreceptor cell loss in the XLPRA2 retina as a function of age. ONL thickness was measured as the number of rows of photoreceptor nuclei. The figure illustrates for the superior meridian the expected decline in ONL thickness in the absence of a photoreceptor rescue effect for each treatment Group #1, #2, and #3 between the time of 1^st injection and termination. Double headed arrows and numbers show the expected change in number of nuclei between these 2 time points for each treatment group. Note that treatment Groups #2 and #3 received a second injection, respectively, at 8 and 10 weeks of age. Modified (with permission) from Figure 2A in Beltran et al. IOVS, 2006; 47: 1669–1681.

Figure 8

Immunohistochemical localization of CNTFRα in the retinas of normal, rcd1, and XLPRA2 dogs of different ages. Strong immunoreactivity for CNTFRα was seen in the GCL, INL and inner segments (IS, arrows) in the normal and mutant retinas at 4 weeks of age (A₁–A₃). At 8 weeks of age, intense labeling is seen throughout the IS of rods and cones in the normal retina (B₁). In the mutant retinas, labeling persists in the IS of both rods and cones in XLPRA2 (B₃), but is essentially limited to cone IS in rcd1 which is the predominant cell class remaining in the PRL at this stage of disease (B₂). (C₁–C₃) At 24 weeks of age, CNTFRα-positive cone IS and somas are still seen in rcd1 and XLPRA2 despite substantial loss of photoreceptors. Negative controls (A₄, B₄, C₄) were XLPRA2 retinas that were treated with the omission of the primary antibody. Scale bars: 20 μm.

Figure 9

Illustration of the morphologic changes observed in the peripheral retina of XLPRA2 dogs following intravitreal injection with CNTF. An example from each of the 4 treatment groups is shown at the site where the measured ONL thickness was maximal. This same location in the contralateral (PBS-injected eye) is shown for comparison. The boxed areas in the low power photographs are shown in higher magnification. (A₁, A₂) Inferior periphery (1,000 μm from the ora serrata) in an 8 week-old dog (Z263; treatment Group #1). (B₁, B₂) Superior periphery (3,900 μm from the ora serrata) in a 12 week-old dog (Z259; treatment Group #2). (C₁, C₂) Temporal periphery (550 μm from the ora serrata) in a 14 week-old dog (Z272; treatment Group #3). (D₁, D₂) Superior periphery (680 μm from the ora serrata) in a 15.6 week-old dog (Z288; treatment Group #4). Note the prominent increase in ONL thickness in dogs from Groups #1–3 (A₂, B₂, C₂), and the misplacement of rod-like nuclei in the INL (A₂, B₂; arrows). No significant difference in ONL is seen in Group #4 (D₁, D₂). A loss of the photoreceptor layer (PRL) is seen in Groups #1–3, which causes the external limiting membrane to be in direct contact with the retinal pigment epithelium (A₂, B₂, C₂; asterisks). Cryosections; H&E stain; scale bars: 100 μm.

Figure 10

Immunohistochemical characterization of CNTF-mediated remodeling of the peripheral retina of XLPRA2. (A₁, A₂, B₁, B₂) Double immunofluorescence labeling of rods and cones with, respectively, anti-rod opsin (green) and anti-cone arrestin (red), in a 12 week-old dog, 8 weeks after the first CNTF injection (Group # 2; A₁, A₂), and in a 5 week-old animal one week post-CNTF injection (B₁, B₂). No labeling is seen at the level of the PRL due to the loss of inner and outer segments (A₂; asterisks). The majority of the cells found in the ONL are rod opsin-positive due to the increased mislocalization caused by CNTF (A₂). Rod opsin immunoreactive cells are seen misplaced in the INL (arrows), and cone arrestin-labeled somas are present within the thickened ONL (arrowheads). One week following CNTF injection, there is a significant decrease in rod opsin and cone arrestin expression (compare B₂ with B₁ which were taken at the same exposure settings). Digitally increasing the red fluorescent signal in an area of (B₂) shows the displacement of cone somas (arrowheads) down their respective axons (arrows; B₃). (C₁, C₂) Double immunofluorescence labeling of ON bipolar cells with anti-Goα (red), and rod bipolar cells with anti-PKCα (green) in a 12 week-old dog 8 weeks after the first CNTF injection (Group # 2). DAPI (blue) was used as a nuclear counterstain. Rod bipolar cells are co-labeled with both antibodies and appeared yellow-orange, whereas ON cone bipolar cells were only labeled with anti-Goα and appeared red. Note the neuritic sprouting extending into the ONL (arrows), and the presence of Goα- and PKCα-positive somas in the ONL (arrowheads) of the CNTF-treated eye. (D₁, D₂) Immunofluorecence labeling of horizontal and amacrine cells with anti-calretinin in a 12 week-old dog (Group #2) 8 weeks after the first CNTF injection. There are radial extensions into the ONL of neurite sprouts originating from the horizontal cells (arrows) in the CNTF-treated eye. (E₁, E₂) Immunofluorescence labeling of Müller cells with anti-CRALBP in a 12 week-old dog (Group #2) 8 weeks after the first CNTF injection. Müller cell somas are seen misplaced into the ONL (arrowheads) in the CNTF-treated eye. Note the intense normal CRALBP labeling in the RPE. Scale bars: 20 μm.

Figure 11

Changes in the ONL of the inferior periphery in an 8 week-old XLPRA2 dog following intravitreal injection of CNTF (PBS in the contralateral eye) at 4 weeks of age. Note the increase in ONL thickness (Compare A and B). The 1 μm thin epoxy resin sections enabled counts of individual cells in the ONL (C₁, C₂). Rod-like nuclei were seen in the INL (arrows), and larger euchromatic nuclei with a cone-type morphology were found in deeper layers of the ONL (arrowheads). Azure II-methylene blue stain with PPDA counterstain; scale bars: (A, B) 100 μm; (C₁, C₂) 20 μm.

Figure 12

Evidence for cell proliferative events in the peripheral ONL of XLPRA2 following intravitreal injection of CNTF. (A–D) One week after CNTF injection there was increased ONL thickness in a 5 week-old dog. Note also the extended loss of the PRL (B, and C₂, asterisks). (D₁–D₃) Fluorescence immunohistochemical labeling for BrdU (green) shows incorporation in the ciliary epithelium of the pars plana (D₁, D₂, arrowheads), and in cells located at the retinal margin (D₁, D₂, arrows) in both the CNTF- and PBS-injected eye. (D₃) BrdU incorporation was detected in cells located in the INL (arrowhead), GCL and NFL in both eyes, but few BrdU-positive cells were seen only in the ONL of the CNTF-treated eye (arrow). (E₁, E₂) Fluorescence immunohistochemical labeling for the nuclear cell marker of proliferation Ki67 (red). (E₂) Ki67-positive cells were seen exclusively in the ONL of the CNTF-treated eye (arrows), but Ki67 positive cells were found in other retinal layers in both PBS-and CNTF-injected eyes. Scale bars: (A, B) 100 μm; (C₁-E₂) 20 μm.