Impairment of Rod cGMP-Gated Channel α-Subunit Expression Leads to Photoreceptor and Bipolar Cell Degeneration

Abstract

purpose. To determine whether alterations in rod cGMP-gated channel function mediate retinal degeneration, a transgenic approach in which the α subunit of the rod cGMP-gated channel is reduced by the expression of an antisense RNA was used.

methods. A 890-bp fragment of the 5′ mouse rod cGMP-gated channel cDNA was cloned in the antisense orientation under the control of the strong bacterial cytomegalovirus promoter. This antisense construct was used to generate transgenic mice in which the expression of the antisense transgene was measured by reverse transcription–polymerase chain reaction. Histologic, immunocytochemical, and TdT-dUTP terminal nick-end labeling (TUNEL) analyses were performed on control and transgenic retina sections to determine the effects of antisense expression on specific cell types.

results. The antisense RNA was able to inhibit the translation of the rod channel protein in an in vitro assay. Three transgenic mouse lines expressing the antisense construct were obtained. Molecular analyses showed that the antisense is expressed in the eye of each transgenic mouse line, where it reduces rod cGMP-gated channel RNA expression. Histologic and immunocytochemical data showed that transgenic retinas have a reduced number of photoreceptors and bipolar cells. TUNEL staining confirmed that photoreceptor and bipolar cells degenerate along an apoptotic pathway.

conclusions. Impairment of rod cGMP-gated channel α subunit expression leads to photoreceptor and bipolar cell degeneration. These transgenic mice are the first model of retinal degeneration caused by an alteration in the expression of the rod cGMP-gated channel. This model system can be used to test therapies designed to slow or stalled retinal degenerations.

Cyclic nucleotide-gated (CNG) ion channels were first characterized in rod photoreceptors, where they mediate the final step in the transduction of sensory stimuli into neuronal activity. ^{1 2 3 4} Subsequently, in the retina, CNG channels were identified in cone photoreceptors, ⁵ bipolar cells, ⁶ in a subpopulation of ganglion cells, ⁷ and in Müller glial cells. ⁸ In rod photoreceptors, the cGMP-gated channel plays a central role in phototransduction by controlling the flow of cations accross the outer segment plasma membrane in response to light-induced changes in cGMP. ^{9 10 11} In ON-bipolar cells, transduction of the visual signal involves a cGMP-gated channel that is linked to the metabotropic glutamate receptor (mGluR6) pathway. ^{6 12} Although the function of CNG channels in other retinal cell types is not well established, there is increasing evidence that CNG channels may be important modulators of neuronal activity in the retina. 

The native rod cGMP-gated channel is believed to be a hetero-oligomer, probably a tetramer, composed of two different subunits (α and β), each encoded by a separate gene. ^{13 14} Expression of theα subunit by itself results in functional channels that have most of the electrophysiological properties of the native rod channel. ^{3 15 16} β Subunits are incapable of forming functional channels by themselves but can modulate the channel properties of the α subunits. ^{16 17} Therefore, the α subunit constitutes the essential functional unit of the rod cGMP-gated channel. 

Dryja and coworkers ¹⁸ recently provided evidence that defects in the gene encoding the human rod cGMP-gated channel α subunit (rCNGα) may be responsible for the retinal degeneration that occurs in one form of autosomal recessive retinitis pigmentosa (RP). Five different types of mutations of the rCNGα subunit were identified in unrelated cases of this disease, in which affected individuals develop slow, progressive degeneration of photoreceptors. ^{19 20} Three of these mutations result in the expression of nonfunctional α subunits of the rod cGMP-gated channel. Although the two remaining mutant alleles encode functional channels when expressed in vitro, they failed to reach the plasma membrane. ¹⁸  

Animal models recently have been very useful in determining the role of CNG channels in other neuronal systems. Mutations in CNG channels in the nematode Caenorhabditis elegans were found to affect thermotaxis and chemotaxis as well as axon guidance. ^{21 22} Analysis of mice lacking the olfactory CNG channel α subunit demonstrated that this channel is required for the transduction of odorant stimuli. ²³ To date, however, no spontaneous or induced mouse mutations in the rod CNG channel have been reported. Therefore, a mouse model in which rCNGα function is impaired would provide a unique tool for addressing the role of the rod cGMP-gated channel in the retinal pathologies associated with one form of autosomal recessive RP. Complete loss of function of other phototransduction proteins such as rhodopsin in the null mutant ²⁴ or cGMP-phosphodiesterase β-subunit in the naturally occurring mouse rd ²⁵ caused a rapid photoreceptor degeneration, uncharacteristic of human RP. Consequently, we chose to develop an antisense strategy that may provide a mouse mutant that exhibits a slow and progressive retinal degeneration, as observed in human disease. 

To determine the role of the rCNGα gene in retinal degeneration, transgenic mice expressing an antisense RNA complementary to the endogenous rCNGα messenger RNA were generated. This antisense sequence corresponds to the N-terminal region of the rod channel protein where the highest sequence variation or the most specific part of the protein sequence is found. ²⁶ Thus, the function of the rod channel was perturbed to test the hypothesis that alteration of rCNGα gene expression is a cause of retinal degeneration. Molecular analysis showed that the antisense was expressed in the transgenic eye, where it reduced the amount of the endogenous rCNGα sense mRNA. Histologic and TdT-dUTP terminal nick-end labeling (TUNEL) analysis of the transgenic retinas revealed a slow and progressive degeneration of both photoreceptor and bipolar cells. Therefore, these transgenic mice provide a new model in which the mechanisms of retinal degeneration can be characterized. 

Materials and Methods

The use of animals in this work was in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Construction of the Transgene

Total RNA was extracted from adult mouse retina using TRIzol reagent procedure (Gibco BRL, Gaithersburg, MD). An 815-bp fragment of the 5′ mouse rCNGα cDNA was amplified using reverse transcription–polymerase chain reaction (RT-PCR). The PCR product was first subcloned into the pCR 2.1 vector (Invitrogen, Carlsbad, CA), and the clones containing the fragment in the antisense orientation were selected by restriction analysis. The antisense was then cloned as a 890-bp fragment into the HindIII/EcoRV sites of the pcDNA3 vector (Invitrogen, Carlsbad, CA) that contains the cytomegalovirus (CMV) promoter and the bovine growth hormone polyA. Transgene identity was confirmed by sequencing. Comparison of the antisense sequence with the sequences present in GenBank databases did not reveal any significant homology with any mouse proteins other than the cGMP-gated cation channel protein (GenBank accession number M84742). ²⁶  

In Vitro Expression of the Transgene Construct

Sense and antisense rCNGα RNAs were synthetised in vitro using an RNA polymerase kit (MEGAscript T7; Ambion, Austin, TX). The sense RNA was then incubated overnight at 48°C with different amounts of antisense RNA in the hybridization buffer. Proteins were translated in vitro using the rabbit reticulocyte lysate system (Promega, Madison, WI) with [³⁵S]methionine. Translation products were separated by SDS-polyacrylamide gel electrophoresis. Labeled protein bands were visualized by autoradiography. 

Generation and Identification of Transgenic Mice

The antisense construct was isolated from the pcDNA3 vector sequences by BglII/SmaI enzyme digestion (Fig. 1) and was purified by ultracentrifugation in sucrose gradients. Mice were produced by the Yale Transgenic mice unit. (SJL × C57BL6) F₁ mice were used as embryo donors, stud males, and pseudopregnant females. Transgenic mice were generated by microinjection into the male pronucleus of fertilized eggs according to described methods. ²⁷ Transgenic mice were identified by PCR or Southern blot analysis of tail genomic DNA as previously described. ²⁸  

RT-PCR and Northern Blot Analysis

RT-PCR was carried out on total RNA after treatment with RNase-free DNase to remove any genomic contamination. Transgene expression was detected using PCR primers based on the antisense sequence and on the bovine growth hormone polyA sequence (Fig. 1) . Amplification of the endogenous gene was performed using primers based on the rod CNG channel sequence. Amplification with primers forβ -actin was used as a control for the amount of cDNA. PCR products were electrophoresed in 1% agarose gel. For northern blot analysis analysis, 40 μg of total RNA was electrophoresed in 1% agarose-formaldehyde gel. PCR products or total RNA were transferred to a nylon membrane and probed with ³²P-labeled rod CNG channel or actin DNA probe. 

Histology and Immunostaining

Eyecups were fixed for several hours in 4% paraformaldehyde, rinsed with phosphate-buffered saline (PBS) and cryoprotected overnight in 30% sucrose in PBS, mounted in OCT medium, and sectioned at 20 μm in a cryostat. Retina sections were stained with toluidine blue or processed for immunocytochemistry. Sections were blocked for 1 hour in 5% normal goat serum (NGS) and then incubated overnight at 4°C in the appropriate dilutions of the primary antisera in 5% NGS. Immunolabeling was performed with several antibodies: affinity-purified rabbit anti-human Otx2 polyclonal antibody (Wikler KC, Baas D, Stull DL, et al., unpublished results), rabbit polyclonal antiserum to recoverin, ²⁹ and the monoclonal antibodies, Calbindin (Sigma, St. Louis, MO), HPC-1, ³⁰ and RET-B1. ³¹ After incubation with the appropriate secondary antibody for 30 minutes in the dark, sections were mounted in 50% glycerol in PBS. 

TUNEL Staining

TUNEL staining was carried out on tissue sections using the Apoptag in situ apoptosis detection kit (Oncor, Gaithersburg, MD), which detects apoptotic cells by direct fluorescence detection of digoxigenin-labeled genomic DNA. ³²  

Results

Construction and In Vitro Assay of the Antisense RNA Construct

A 890-bp fragment from the 5′ coding region of mouse rCNGα subunit cDNA was cloned in the reverse orientation in front of the bovine growth hormone polyA and downstream from the CMV promoter to ensure high level of transgene expression (Fig. 1) . 

To test whether this antisense construct could inhibit the expression of the rCNGα protein in vitro, we designed an assay in which different amounts of antisense RNAs were incubated with sense RNA before being translated into proteins using a reticulocyte system (Fig. 2) . Equal amounts of sense transcripts were incubated overnight at 42°C with no or increasing amounts (ratios 1:1, 1:2, and 1:3) of rCNGα antisense RNA. In the control lane (rCNG), a major band of 69 kDa, corresponding to the full length rod channel protein, is detected, together with a number of minor bands (Fig. 2) , some of which were found in control without added RNA (data not shown). When rCNGα sense RNA was incubated with increasing amounts of antisense RNA (lanes rCNG+ A′ and rCNG + A″), the major band of 69 kDa was progressively eliminated (Fig. 2 , arrow). 

Expression of Antisense RNA in Different Tissues of Adult Transgenic Mice and Reduction of Rod cGMP-Gated Channel RNA in Transgenic Mouse Eyes

Microinjection of the antisense construct into the pronuclei of mouse fertilized eggs resulted in 36 offspring. Three founder mice were identified as transgenic by Southern blot analysis and PCR detection of the transgene from tail genomic DNA (Fig. 3) . The three founder animals (F0) were successfully bred with CD1 mice to establish three transgenic mouse lines (5499, 5510, and 5518). 

To confirm the specificity of the primers that amplify the transgene, expression and detection of the transgene was first tested in HEK 293 cells transfected with either the antisense construct alone (A) or with both sense and antisense constructs (A + S) (Fig. 4) . No signal was detected in HEK cells transfected with the sense construct alone (S) or in the control tissues, indicating that these primers amplify only the transgene. Expression of the antisense construct in vivo was then verified by RT-PCR analysis of RNA extracted from different adult tissues for each of the three transgenic lines (5499, 5510, and 5518), namely eye, brain, heart, and kidney (Fig. 4) . 

Expression of the sense endogenous rCNGα RNA was then quantified in the eyes of control and transgenic mice using northern blot analysis and RT-PCR analysis. Figure 5A illustrates a northern blot analysis experiment showing the detection of a 3.2-kb signal corresponding to the single transcript of the endogenous rCNGα RNA and of a 2.1-kb signal corresponding to the actin transcript. Whereas the actin signals show the same intensity, rod channel transcript signal intensity is decreased in transgenic eyes compared to the control, particularly in line 5518 (Fig. 5A) . Measurement of the signal intensity and normalization to actin in three different experiments are presented in the graph in Figure 5B : compared to the control eye, the amount of the endogenous rCNGα RNA measured by northern blot analysis is reduced by 20% in the eye of lines 5499 and 5510 and by 50% in the eye of line 5518. Figure 6A illustrates a RT-PCR experiment where both actin and rCNGα PCR products are detected. Normalization to actin for at least 10 different mice for each line is presented in the graph in Figure 6B . Compared to the control eye, the amount of the rod channel RNA measured by RT-PCR is reduced by 40% in line 5499, 25% in line 5510, and 50% in line 5518. 

Reduction of Photoreceptor and Bipolar Cell Numbers in Transgenic Mouse Retina

Histologic and immunocytochemistry studies were performed in line 5518 where expression of rod channel is the most reduced (50%) at the RNA level. 

At the histologic level, no difference was seen between the transgenic and the control retinas younger than 3 months (data not shown). However, in mice older than 3 months, the entire transgenic retina appear thinner compared to the control retina. Comparison of transgenic retina sections (Fig. 7B ) with control retina sections (Fig. 7A) shows a reduction in the thickness of the different layers. In particular, the thickness of the outer nuclear layer (ONL) and the inner nuclear layer (INL) were reduced in the transgenic retina (Figs. 7 8) . We used RET-B1, an antibody that recognizes a membrane protein of bipolar and rod photoreceptors. ³¹ Compared to the control (Fig. 8A) , RET-B1 labeling showed a reduction in the thickness of the transgenic ONL (Fig. 8B) and a restriction of RET-B1-positive cells to the most outer portion of the INL. This suggests a smaller number of photoreceptors and bipolar cells (Fig. 8B) . 

To quantitate a decrease in thickness of the ONL in the transgenic retina, the mean ONL column height was determined by counting the number of photoreceptors spanning the ONL in toluidine blue–stained (Fig. 7) or in recoverin-labeled photoreceptors (Figs. 8E 8F) . Transgenic retinas have a significantly (P < 0.0001) thinner ONL composed of fewer (8.938 ± 0.566) photoreceptor rows than controls (11.938 ± 0.71) (Fig. 9A ). We found that in control retinas of any age, the ONL is composed of 11 to 12 rows of photoreceptors nuclei (Fig. 7A) , whereas in the transgenic 15-month-old retina, the ONL contains only 7 to 8 rows of photoreceptor nuclei (Fig. 7B) . The time course of the changing ONL thickness, between the ages of 2 weeks and 15 months, is presented in the Figure 10 , showing the slowly progressing degeneration of photoreceptors in the transgenic retina. 

The mouse INL contains bipolar (41%), amacrine (39%), Müller (16%), and horizontal cells (3%). ³³ To determine which cell types in the INL were affected by antisense expression, immunocytochemistry was performed on retina sections using antibodies that label specific cell types in the INL. Using HPC-1 and calbindin antibodies, no difference in the number of amacrine or horizontal cells was observed (data not shown). Otx2, an antibody that recognizes the paired homeodomain protein, specifically labels the ONL and bipolar cells in the INL of adult mouse retina (Wikler KC, Baas D, Stull DL, et al., unpublished results). Interestingly, the ONL of transgenic animals lacked Otx-2 staining compared with that of control. In the transgenic INL, Otx2 immunostaining showed a thinner layer of positive cells compared to that of the control retina (Figs. 8C 8D) . Positive Otx2 bipolar cells were counted in control and transgenic retina sections. Cell counts presented in the graph (Fig. 9B) showed that there is a 22% to 38% reduction in the number of bipolar cells in the transgenic mouse retina. 

The mouse retina contains at least nine types of cone bipolar cells and one type of rod bipolar cells. To test whether a specific subset of bipolar cells was affected by antisense expression, we used a recoverin antibody that only recognizes two types of cone bipolar cells in the INL in addition to all photoreceptors in the ONL. ²⁹ As observed with Otx2 immunostaining, fewer cells were stained in the transgenic retina compared to that of the control (Figs. 8E 8F) . Cell counts of INL recoverin-positive cells confirmed that the number of bipolar cells in the transgenic INL is significantly (P < 0.0001) reduced by 25% to 40% (Fig. 9C) . 

Photoreceptor and Bipolar Cells Degeneration via an Apoptotic Pathway

Because previous analyses of the retinal degeneration phenotypes in transgenic and mutant mice revealed that photoreceptor cell death occurs through an apoptotic mechanism, ^{34 35} we carried out TUNEL staining on retina sections from control and line 5518 transgenic mice. Some retina sections were incubated with DNaseI as a positive control to confirm that all cells were stained (Fig. 11 A). Few or no stained cells were found in nontransgenic retina sections (Fig. 11B) . In the transgenic retina, at the age of 2 weeks, TUNEL-positive cells were mostly found in the INL (Fig. 11C) . At later stages, stained cells were observed in the different layers of the retina (Fig. 11D , arrows). TUNEL-positive cells located in the ONL correspond to degenerating photoreceptors, whereas TUNEL-positive cells found in the INL may correspond to degenerating bipolar cells (Figs. 11C 11D) . Some stained cells also were detected in the ganglion cell layer (Fig. 11D) . Similar results were obtained in lines 5499 and 5510 (data not shown). 

Discussion

In this study, we provide the first report of a transgenic mouse model in which the expression of rCNGα is impaired using an antisense approach. The data presented here show direct evidence that a decrease in rCNGα expression induces a slow and progressive degeneration of not only photoreceptors but also of bipolar cells. These transgenic mice can be used as a new model of retinal degeneration caused by one form of RP. 

The Role of rCNGα in Retinal Degeneration: An Antisense RNA Strategy

An antisense strategy was chosen because a decrease in the expression of the rCNGα gene (knock-down) can closely mimic a slow retinal degeneration as observed in human RP disease, rather than a complete loss of function (knock-out), which may cause a rapid degeneration as seen in rhodopsin knockout. ²⁴ Furthermore, this approach allows the obtention of several lines of transgenic mice expressing different levels of rCNGα RNA (Figs. 5 6) . Such a strategy requires the use of a very strong promoter, because a sufficient amount of antisense RNA has to be expressed to inhibit the synthesis of rCNGα protein (see Fig. 2 ). The well-characterized rhodopsin promoter would restrict transgene expression to rod photoreceptors only. ²⁸ Consequently, we used the CMV promoter because its expression is not cell-specific (see Fig. 4 ). Thus, effects of the antisense construct can be studied in any cell expressing the endogenous rCNGα gene. For example, expression of the transgene was observed in the eye, brain, heart, and kidney (Fig. 4) , where the endogenous rCNGα has been detected. ^{36 37 38} Despite the widespread distribution of the transgene, the mice do not appear to suffer from any severe pathology, because they can live to be more than 1 year of age with no obvious physiological defects. The effects of the transgene in tissues other than the retina would be of great interest in determining the role of the rod cGMP-gated channel in other systems. 

Photoreceptor and Bipolar Cells Degeneration in the Mouse Transgenic Retina

The present study focused on the effects of transgene expression in the retina. Molecular data in combination with histology and cell counts show that a 50% reduction of rCNGα expression induces a loss in the photoreceptor and the bipolar cell populations in the 5518 transgenic line. No cell loss was observed in the amacrine or the horizontal cells. Therefore, photoreceptor and bipolar cell degeneration are due to a reduction in the expression of the rCNGα rather than transgene toxicity. Furthermore, the time course of the degeneration is so slow that it is not indicative of a toxic effect (Fig. 10) . Also, at 2 weeks, there is no difference between the number of photoreceptor or bipolar cells in nontransgenic and transgenic littermates. Therefore, expression of the antisense construct does not appear to affect retinal development, because all cells develop normally before the progressive degeneration of specific cell populations. 

Although Otx-2 immunolabeling is detected in the cytoplasm of control photoreceptor nuclei in the adult mouse retina (Wikler KC, Baas D, Stull DL, et al., unpublished results) (Fig. 8C) , we noticed that this staining was absent in transgenic photoreceptors (Fig. 8D) . This observation suggests that Otx-2 expression is downregulated in photoreceptors with impaired rCNGα expression. At the present time, it is not known whether there is a link between the mechanisms involved in photoreceptor degeneration and the regulation of Otx-2 transcription factor expression. 

TUNEL staining demonstrated that cells in the ONL, INL, and ganglion cell layer (GCL) degenerate via an apoptotic pathway in the transgenic mouse retina. These data are consistent with the mode of cell death in photoreceptor cells that has been observed in other mouse models of retinal degeneration. ^{34 35} However, most studies of retinal degeneration in animal models reported that cell death is restricted to the photoreceptor cells, ³⁹ whereas the data presented here also indicate a cell degeneration in the INL and in the GCL. Because immunocytochemistry and cell counts indicate a cell loss in the bipolar cell population but not in the amacrine or horizontal cells populations, the TUNEL-positive cells in the INL may correspond to bipolar cells. A previous report demonstrated the presence of the rod cGMP-gated channel in a subset of ganglion cells. ⁷ TUNEL-positive cells, which were observed in the GCL, may correspond to ganglion cells. Cell counts of ganglion cells must be done to determine the precise effect of the transgene on these cells. 

Bipolar cell degeneration observed in this study may be due to either impairment of the bipolar cGMP-gated channel protein or to transneuronal degeneration. It still remains unclear whether the cGMP-gated channel expressed in bipolar cells is identical with that of rod photoreceptors. The cGMP-gated channel expressed in bipolar cells was not recognized by an antibody directed against the cGMP-gated channel of bovine rod outer segments ⁴⁰ and has different pharmacological properties, which would suggest that it is not identical with that of rod photoreceptors. ¹² The time course of the degeneration indicates that cell death of bipolar cells parallels rather than follows that of photoreceptors. This observation suggests that at least early cell death of bipolar cells may result from the impairment of bipolar cGMP-gated channel expression. Transneuronal degeneration has been mainly observed after extensive loss of photoreceptors in severe cases of retinitis pigmentosa. This is thought to follow photoreceptor death and occurs because of the loss of synaptic input or trophic factors. ^{41 42 43} Some bipolar cells may also undergo transynaptic degeneration, since TUNEL-stained bipolar cells were detected in 1-year-old mice. 

A New Model of Retinitis Pigmentosa

This study was initiated because five different types of mutations in the rCNGα gene were found to cosegregate with autosomal recessive RP in four unrelated families. ¹⁸ Thus, defects in rod channel proteins may be responsible for the retinal degeneration that occurs in this human disease. RP patients with a mutation in the rCNGα gene have a slower retinal degeneration in comparison to other RP patients with mutation in a different gene (Berson EL, personal communication, October 1998). Data obtained from this study provide direct evidence that alteration of rCNGα expression can induce a slow and progressive retinal degeneration, such as the one observed in one case of autosomal recessive retinitis pigmentosa. 

There is still no effective treatment for retinal disease such as RP, in which the loss of retinal cells causes visual loss and eventually blindness. ^{19 20} Evaluating the retinal damage caused by the disease is a prerequisite to find new therapies. Recent morphometric analyses of eyes from RP patients demonstrated that the disease also can result in loss in inner retinal cells. ^{42 44} However, none of these studies indicate which cell type was lost in the INL. In this study, we reported loss of bipolar cells in addition to photoreceptor cells. Previous mouse models of retinal degeneration have had difficulties in providing a model whose degeneration time course mimics the one observed in human disease. Transgenic mice with impaired rCNGα expression mimic the human pathology in that the retinal degeneration is slow and progressive. They provide a model system in which novel therapies designed to slow or to stall the damage caused by one form of RP can be tested. 

This mouse model also can be useful to address the fundamental question: What are the cellular and molecular mechanisms that link the genetic defect to the death of the cell? Because the α subunit is the essential functional unit of the rod channel, alteration of its expression may lead to a decrease in channel functional activity, which in turn may induce cell degeneration. Further studies of this model may help to better understand the mechanisms of retinal degeneration. 

 Supported by the Johannsen Research Fund of Fight For Sight/Prevent Blindness America, the Foundation Fighting Blindness, National Institutes of Health Grant EY11356, and the Kemper Research Fund.

 Submitted for publication May 27, 1999; revised September 8, 1999; accepted October 5, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Diagram of the antisense transgene construct. An 890-bp fragment of the 5′ coding region from the mouse rod cGMP-gated channel cDNA has been cloned in the antisense orientation in the HindIII/EcoRV sites of the pcDNA3 vector between the cytomegalovirus (CMV) promoter and the bovine growth hormone polyA. The transgene was excised from the plasmid by BglII/SmaI digestion. The arrows numbered 1 show the localization of the PCR primers used to detect the presence of the transgene in genomic tail DNA of transgenic mice. The arrows numbered 2 show the localization of the RT-PCR primers used to detect the expression of the transgene.

Figure 2.

 

Inhibition of rod CNG channel protein expression in vitro. Autoradiograph of in vitro protein synthesis. Sense and antisense RNAs were transcribed in vitro using T7 RNA polymerase. Equal amounts of sense transcripts were incubated overnight at 42°C with increasing amounts of rod CNG channel antisense RNA (A: ratio 1:1; A′: ratio 1:2; A″: ratio 1:3). RNAs were translated in vitro using a reticulocyte lysate system and [³⁵S]methionine. Translation products were identified by SDS–polyacrylamide electrophoresis and autoradiography. In the control lane (rCNG), a major band of 69 kDa is detected, which corresponds to the full length rod channel protein. When rod CNG channel sense RNA was incubated with increasing amounts of antisense RNA (lanes rCNG + A′ and rCNG + A″), the major band of 69 kDa was progressively eliminated.

Figure 3.

 

Southern blot analysis of mouse genomic tail DNA. Genomic DNA was extracted from mouse tail biopsies and digested with HindIII/EcoRV restriction enzymes. DNA was electrophoresed on agarose gel, transferred to a nylon membrane, and probed with a ³²P DNA probe corresponding to the HindIII/EcoRV fragment of the transgene construct. Each lane (1 to 8) corresponds to tail genomic DNA of a different mouse. A band at 890 bp (white arrow) indicates the presence of the transgene in the genomic DNA of three transgenic founder mice: line 5499 (lane 3), line 5510 (lane 5), and line 5518 (lane 7).

Figure 4.

 

RT-PCR analysis of antisense expression in transgenic mouse tissues. Total RNA was extracted from eye, brain, heart, and kidney of adult transgenic mice and of nontransgenic mice as a control. RT-PCR was carried out on total RNA after treatment with RNase-free DNase to remove any genomic contamination. Transgene expression was detected using PCR primers based on the antisense sequence and on the bovine growth hormone polyA sequence (see Fig. 1 ). Amplification with primers specific for the antisense construct gave a PCR product of 733 bp. (The presence of two bands in many lanes is due to primer annealing to repeated sequences in the bovine polyA, resulting in the synthesis of two PCR products, slightly different in size.) Both bands were recognized by a probe specific for the rod cGMP-gated channel α subunit. Transfected HEK 293 cells were used as a control for the specificity of the primers. No signal was detected in the control. The antisense construct was expressed in each tissue tested for the lines 5499, 5510, and 5518.

Figure 5.

 

Northern blot analysis of sense expression in transgenic eyes. Northern blot analysis: total RNA was isolated from adult eyes for each transgenic line and for nontransgenic mice as a control. Expression of rod CNG channel endogenous mRNA was detected as a band of 3.2 kb. Expression of actin mRNA gave a band of 2.1 kb and was used as a control for the amount of total RNA. (A) Each lane corresponds to RNA extracted from the eyes of two adult mice. Normalization to actin shows a reduction in the level of rod channel expression in the three transgenic lines compared to the control (see graph shown in B). (B) This graph represents the measurement and normalization to actin of values resulting from four independent experiments. These results indicate a reduction of rod channel mRNA expression: 20% in lines 5499 and 5510, 50% in line 5518.

Figure 6.

 

RT-PCR analysis of sense expression in transgenic eyes. (A) RNA samples used in this experiment were obtained from the same two mice as those used in the northern blot analysis shown in Figure 5A . Normalization to actin indicates a reduction of the expression of the rod CNG channel mRNA in the three transgenic lines and most particularly in line 5518 (see graph in B). (B) This graph represents the measurement and normalization to actin of values resulting from several RT-PCR experiments. The results show a reduction of the rod CNG channel RNA level in the eyes of the three transgenic lines compared to the control: 40% of reduction is observed in line 5499, 25% in line 5510, and 50% in line 5518.

Figure 7.

 

Toluidine blue staining of control and 5518 transgenic retina sections. (A) Retina section stained with toluidine blue from 1-year-old nontransgenic mouse. (B) Retina section stained with toluidine blue from 1-year-old 5518 founder transgenic mouse: the whole transgenic retina looks thinner; in particular, notice the reduction in the thickness of both the outer nuclear layer (ONL) and the inner nuclear layer (INL), compared to the control section shown in (A). OS, outer segment; IS, inner segment; OPL, outer plexiform layer; IPL, inner plexiform layer; GCL, ganglion cell layer. Scale bar, 20 μm

Figure 8.

 

Immunostaining of 5518 transgenic retina sections. (A) Retina section from a 1-year-old control: RET-B1 antibody stains the entire population of bipolar cells as well as all photoreceptor nuclei and inner segments. (B) Retina section from 5518 transgenic founder. Compared to the control seen in (A), RET-B1 staining shows a dramatic reduction in the thickness of the outer nuclear layer (ONL), suggesting that many photoreceptors have degenerated. Furthermore, the bipolar cells appear to be disorganised and in a smaller number. (C) Retina section from a 1-year-old control: Otx-2 antibody stains bipolar cell nuclei in the INL and the cytoplasm of photoreceptor nuclei in the ONL. (D) Retina section from 5518 transgenic founder. Compared to the control seen in (C), Otx2 staining in the INL appears as a thinner layer, suggesting that bipolar cells have degenerated. Otx-2 staining is completely absent from the cytoplasm of photoreceptor nuclei. (E) Retina section from a 1-year-old control: Recoverin antibody recognizes the entire population of photoreceptors (ONL) and only two types of cone bipolar cell (INL). (F) Retina sections from a 1-year-old 5518 founder transgenic mouse. Compared to the control shown in (E) and as seen in (B), the outer nuclear layer (ONL) is thinner and fewer bipolar cells (INL) are stained, suggesting that both photoreceptors and bipolar cells have degenerated. Scale bar, 14 μm.

Figure 9.

 

Number of photoreceptor and bipolar cells in mouse retina. (A) Graph presenting counts of photoreceptor nuclei spanning the ONL in control and transgenic mouse retinas. The control retina has an average number of rows of photoreceptor nuclei of 11.720 ± 0.566. The transgenic retina has a significant lower number of rows of photoreceptor nuclei of 8.938 ± 0.71 (*P < 0.0001). (B) Graph presenting counts of otx-2–positive cells in control and transgenic mouse retinas. The control retina has 203.463 ± 15.165 bipolar cells in a ×40 field. The transgenic retina has only 146.417 ± 12.599 bipolar cells in a ×40 field (*P < 0.0001). (C) Graph presenting counts of INL recoverin-positive cells in control and transgenic retinas. The control retina has 41.433 ± 3.042 stained cells. The transgenic retina has 28.216 ± 3.163 stained cells (*P < 0.0001).

Figure 10.

 

Time course of photoreceptor degeneration. Graph presenting the average number of photoreceptor nuclei spanning the outer nuclear layer at different ages in control and transgenic mice. In control (⋄), this number remains constant between 11 and 12 rows of photoreceptor nuclei per ONL thickness. However, in the transgenic retina (□), the number of photoreceptors decreases slowly with age indicating a very slow degeneration.

Figure 11.

 

TUNEL staining of adult retina sections. (A) Positive control for the TUNEL staining experiment: DNase I was applied to the section. All the cell nuclei from the outer nuclear layer (ONL), from the inner nuclear layer (INL), and from the ganglion cell layer (GCL) are stained. (B) No TUNEL staining was detected on retina sections from the nontransgenic littermate. (C) Retina section from a 2-week-old 5518 transgenic mouse. A TUNEL-positive cell (arrow) is present in the inner nuclear layer, corresponding to a degenerating bipolar cell. (D) Retina section from a 14-month-old 5518 transgenic mouse. TUNEL-positive cells (white arrows) are observed in the different layers of the retina (ONL, INL, and GCL). This suggests that reduction of the channel protein synthesis may lead to apoptosis and affect different cell types throughout the retina. Scale bar, 10 μm.

The authors thank all members of the Barnstable Laboratory past and present for their help and valuable conversations; Adrienne LaRue and Stephen Viviano for outstanding technical assistance; Keely Bumsted, Charles Greer, Thom Hughes, and Ji–Ye Wei for critical reading of the manuscript; and Robert Brown for help for photography.