shRNA knockdown of guanylate cyclase 2e or cyclic nucleotide gated channel alpha 1 increases photoreceptor survival in a cGMP phosphodiesterase mouse model of retinitis pigmentosa.

In vertebrate rods, dark and light conditions produce changes in guanosine 3',5'-cyclic monophosphate (cGMP) and calcium (Ca(2+) ) levels, which are regulated by the opposing function of several proteins. During the recovery of a bright flash, guanylate cyclase (GUCY) helps raise cGMP to levels that open cGMP-gated calcium sodium channels (CNG) to increase Na(+) and Ca(2+) influx in the outer segment. In contrast, light activates cGMP phosphodiesterase 6 (PDE6) causing rapid hydrolysis of cGMP, CNG closure, and reduced Na(+) and Ca(2+) levels. In Pde6b mouse models of retinitis pigmentosa (RP), photoreceptor death is preceded by abnormally high cGMP and Ca(2+) levels, likely because of continued synthesis of cGMP by guanylate cyclases and unregulated influx of Ca(2+) to toxic levels through CNG channels. To reverse the effects of Pde6b loss of function, we employed an shRNA knockdown approach to reduce the expression of Gucy2e or Cnga1 in Pde6b(H620Q) photoreceptors prior to degeneration. Gucy2e- or Cnga1-shRNA lentiviral-mediated knockdown GUCY2E and CNGA1 expression increase visual function and photoreceptor survival in Pde6b(H620Q) mice. We demonstrated that effective knockdown of GUCY2E and CNGA1 expression to counteract loss of PDE6 function may develop into a valuable approach for treating some patients with RP.

Introduction

Photoreceptor cell death is the dominant pathological feature in many retinal diseases, including retinitis pigmentosa (RP) and age-related macular degeneration (AMD) [1-3]. About 36,000 cases worldwide of simplex and familial RP are caused by loss of function mutations in cGMP phosphodiesterase (PDE6) [4-7]. Although significant advances in our understanding of RP have been made, the exact interplay between defective PDE6 and the onset of RP pathogenesis remains poorly understood.

Several Pde6b mouse models of RP, including Pde6b, Pde6b and Pde6b, have been used to study the mechanisms involved in the disease [8-10]. Loss of PDE6 enzyme activity has been shown to result in increased levels of cGMP in Pde6bmice [11-15] and high Ca2+ in Pde6b mice [8]. It has been hypothesized that Pde6b loss of function leads to high cGMP and consequently excessive Ca2+ influx, which is toxic to photoreceptors [8, 15–17].

During rod phototransduction, cGMP and Ca2+ are regulated by PDE6, guanylate cyclase (GUCY2), and cGMP-gated Na+/Ca2+ (CNG) channels. The level of cGMP is controlled by opposing activities of PDE6 and GUCY2 and, at sufficiently high levels, CNG channels open to allow increased Ca2+ influx into the rod outer segment (ROS). Thus, during photoexcitation and recovery, changes in cGMP and Ca2+ levels occur in parallel. Light stimulates PDE6 activation, rapid reduction in cGMP, closure of CNG channels and reduced Ca2+ influx. After a light flash, PDE6 is inhibited, and cGMP and Ca2+ are restored to nominal levels by GUCY2 activity and opening of CNG channels, respectively [18-36].

Knowledge gained from the Pde6b models has been used for testing different drug and gene therapies [8, 37–43]. Attempts to prevent or block photoreceptor death performed with Ca2+ channel blockers has showed limited efficacy. d-cis-Diltiazem was tested for the ability to reduce Ca2+ concentrations without affecting cGMP and was reported to prevent the loss of both rods and cones in Pde6 mice [44, 45]. However, the extent of protection originally reported has been controversial [46-49]. The specificity of d-cis-diltiazem for rod CNG channels has been demonstrated to be lower than l-cis-diltiazem, suggesting this isoform might have a greater effect. However, significant rescue of Pde6b mutant photoreceptors performed with l-cis-diltiazem is uncertain [44-50].

Gene therapy approaches to increase photoreceptor survival has also been employed. The common denominator in these studies is the use of a vector designed to express PDE6β in Pde6b mutant photoreceptors. Different vectors based on adenovirus, adeno-associated virus, lentivirus, simian immunodeficiency virus and herpes simplex virus have been tested for their ability to delay degeneration. Although significant morphological and functional restoration has been reported, incomplete long-term rescue was achieved in all cases [8, 38–41, 43].

In this report, we took an alternative approach to remedy the condition of excess cGMP and Ca2+ in Pde6b mutant mice. We employed lentivirus shRNA technology to knockdown the expression of Gucy2e and Cnga1 to rescue retinal degeneration. We injected viral particles sub-retinally in newborn (P5) Pde6b mutant and C57BL/6J mice, measured changes in gene expression, and photoreceptor function and survival. Our results demonstrate this approach is a viable alternative to previously tested pharmacological and gene therapy methods.

Materials and methods

Mouse lines and husbandry

Mice were used in accordance with the Statement for the Use of Animals in Ophthalmic and Vision Research of the Association for Research in Vision and Ophthalmology, as well as the Policy for the Use of Animals in Neuroscience Research of the Society for Neuroscience. Pde6b mice used in this experiment were bred from a colony of mice that has been previously reported [8, 51]. All Pde6b mice analysed in this study are homozygotes and will be referred to as Pde6b mutants or Pde6b mice. Pde6b and Pde6b strains are in the C3H background and age-matched C57BL/6J (B6) mice were used as controls (The Jackson Laboratory, Bar Harbor, ME, USA).

Histochemical analyses

Mice were sacrificed and haematoxylin and eosin retinal sections were obtained as described [19, 20, 52, 53]. The number and morphology of photoreceptors of lentiviral shRNA injected eyes were compared to control eyes. Briefly, quantification of photoreceptor nuclei was conducted on several sections containing the optic nerve as follows: the distance between the optic nerve and the ciliary body was divided into four quadrants and three rows of nuclei were counted within each single quadrant. Averages and standard deviations were calculated from 10-30 animals for each time-point. The corneal scar at the 6 o’clock position allowed the identification of the injected inferior retinal half of the sample analysed. Sectioning proceeded along the long axis of the segment so that each section contained upper and lower retina as well as posterior pole.

Immunoblot analysis

Retinas were homogenized in 10% sodium dodecyl sulphate (SDS) by brief sonication and denatured at 100°C for 5 min. Following centrifugation, total protein content per sample was measured by the DC Protein Assay method (Bio-Rad Laboratories, Hercules, CA, USA). Proteins were separated by SDS polyacrylamide gel electrophoresis. Samples were then transferred to nitrocellulose membranes, which were blocked in 3% bovine serum albumin (Santa Cruz Biotechnology, Santa Cruz, CA, USA), 150 mmol/l NaCl, 100 mmol/l Tris (pH 7.4) and 0.5% Tween-20 (BSA-TTBS). Membranes were incubated with either rabbit antibody to the GUCY2E (1:500, kindly provided by Alexander Dizhoor), CNGA1 (1:12, kindly provided by Robert Molday), mouse monoclonal IgG1 to rhodopsin (1:10,000, 1D4, sc-57432; Santa Cruz Biotechnology) or mouse monoclonal IgG2b to α Tubulin (1:500, 6-11B-1, sc-23950; Santa Cruz Biotechnology) antibodies in BSA-TTBS. After washing in TTBS, filters were incubated with either goat anti-rabbit conjugated horseradish peroxidase secondary antibodies (1:10,000, sc-2004; Santa Cruz Biotechnology) or goat anti-mouse IgG- horseradish peroxidase secondary antibodies (1:10,000, sc-2005; Santa Cruz Biotechnology). After washing, antibody complexes were visualized by chemilluminescence detection (Immobilon Western, Millipore Corporation, Billerica, MA, USA) and Kodak BioMax film (Kodak, Rochester, NY, USA).

Transduction of lentiviral vectors

To knockdown Gucy2e or Cnga1 expression, we injected shRNA lentivirus (1.5 μl, ∼2 × 107 ml transducing units (TU) per ml in DMEM with 10% heat-inactivated fatal bovine serum and penicillin-streptomycin) sub-retinally into the right eye of Pde6b mice at P5 (n = 75 per gene) and in age-matched control C57BL/6J (n = 75). Virus particles were injected at the 6 o’clock position and at 1.5 mm from the limbus, producing a sub-retinal bubble in mid-periphery retina. The left eye was injected with CMV::EGFP (1.5 μl, 2 × 107 TU/ml) lentivirus or saline sub-retinally and used as control. Anaesthesia and surgery were performed as described [8].

shRNA vectors deliver a short 21-nucleotide stem hairpin RNA duplexes, designed to decrease the expression of retinal Gucy2e and Cnga1. The lentiviral transduction particles were made from sequence-verified shRNA lentiviral plasmid vectors for mouse genes (SHVRS, Mission® Lentiviral Transduction Particles; Sigma-Aldrich, St. Louis, MO, USA). Clone sets injected consist of four Gucy2e and five Cnga1 individual clones targeting different regions of the mRNA for each gene. Clone sets used for this analysis are shown in Table 1. Employing different clones allow us to screen for the most effective shRNA for gene knockdown. We expect at least a knockdown efficiency of >70% with one construct from each gene targeted [54].

Table 1

Table shows the shRNA lentiviral plasmids clone sequence, clone ID and titre used to suppress the expression of guanylate cyclase 2e (Gucy2e) (A) and cyclic nucleotide gated channel alpha 1 (Cnga1) (B)

A
TRC shRNA target set clone sequence	Clone ID	Titre
CCGGGCCTAGAGTTAGAGTAGTGATCTCGAGATCACTACTCTAACTCTAGGCTTTTT	NM_008192.1-717s1c1	1.7 × 107 TU/ml
CCGGCGGTGCCCATGATGTCTATAACTCGAGTTATAGACATCATGGGCACCGTTTTT	NM_008192.1-2757s1c1	1.8 × 107 TU/ml
CCGGCCTGTTCAAGAGCATCAACAACTCGAGTTGTTGATGCTCTTGAACAGGTTTTT	NM_008192.1-2418s1c1	2.3 × 107 TU/ml
CCGGGCCTGAATCGTGACTTTGGTTCTCGAGAACCAAAGTCACGATTCAGGCTTTTT	NM_008192.1-257s1c1	2.6 × 107 TU/ml
B
TRC shRNA target set clone sequence	Clone ID	Titre
CCGGCCTGACAAACTAAGGGCAGAACTCGAGTTCTGCCCTTAGTTTGTCAGGTTTTTG	NM_007723.1-1382s1c1	2.0 × 107 TU/ml
CCGGCCGATATGTTTGTACGAACAACTCGAGTTGTTCGTACAAACATATCGGTTTTTG	NM_007723.1-648s1c1	2.6 × 107 TU/ml
CCGGGCTATCAAACAGTACATGAATCTCGAGATTCATGTACTGTTTGATAGCTTTTTG	NM_007723.1-1256s1c1	2.9 × 107 TU/ml
CCGGGCCTGTATCTTGGCTGAATATCTCGAGATATTCAGCCAAGATACAGGCTTTTTG	NM_007723.1-1934s1c1	3.4 × 107 TU/ml
CCGGGCTGTTAAGGATCTCTCGAATCTCGAGATTCGAGAGATCCTTAACAGCTTTTTG	NM_007723.1-832s1c1	3.5 × 107 TU/ml

TRC: the RNAi consortium.

Electroretinograms (ERGs)

Visual function was evaluated as described [55-60]. ERGs were performed weekly from P35 to P90 to assess global retina function in injected and control eyes. We measured ERG b-wave enhancement of rod, maximal and cone responses from shRNA-lentivirus transduced C57BL/6J and Pde6b mice. Enhancement is defined as the difference in maximum ERG responses of transduced and control fellow eyes, in μV.

Results

The efficiency of shRNA knockdown to ameliorate degeneration was assessed by biochemical, histological and physiological measurements. Initially, four Gucy2e and five Cnga1 shRNA clones were pre-screened for the ability to rescue degeneration. Each clone was injected into the sub-retinas of a litter of pups. At P56, injected and control retinas were compared histologically and the ability of the clone to rescue degeneration was determined. The most efficient shRNA-Gucy2e and shRNA-Cnga1 clone was selected for further study (Table 1).

Biochemical assessment: effective and specific knockdown of GUCY2E and CNGA1 expression

The ability of shRNA-Gucy2e and shRNA-Cnga1 viral clones to knockdown protein expression was tested in C57BL/6J mice. Protein levels in retinal lysates were tested by immunoblotting 15 and 60 days after injection. A significant difference in GUCY2E and CNGA1 expression levels was observed between control and shRNA-Gucy2e and shRNA-Cnga1 retinas, respectively (Fig. 1A and B). In contrast, expression of the rod-specific protein, GNAT1, was not significantly different between experimental and control retinas. shRNA-Cnga1 did not significantly reduce GUCY2E expression.

Fig 1

Immunoblotting of retinal lysates from P60 wild-type C57BL/6J mice transduced with shRNA-Gucy2e (A, +), shRNA-Cnga1 (B, +) or shRNA-GFP control lentivirus (A and B, −). The antibodies used were anti-GUCY2E, anti-GNAT1 and anti-CNGA1. GNAT1 is a rod-specific protein, which serves as a control for photoreceptor protein content. Bottom panel is α-tubulin (TUB), which controls for protein loading (25 μg per lane).

Histological assessment: C57BL/6J photoreceptors are not detectably affected by shRNA-Gucy2e and shRNA-Cnga1 transduction

To determine if reduction of GUCY2E and CNGA1 expression results in changes in photoreceptor numbers or morphology, sections from shRNA-Gucy2e and shRNA-Cnga1 injected and control C57BL/6J retinas were stained by haematoxylin and eosin and compared. Quantification of photoreceptor nuclei was conducted on several sections; retinae transduced by shRNAs have nine rows of photoreceptor nuclei lining areas from the optic nerve to ora serrata, similar to control injected eyes. Ten retinae were analysed for each viral vector. C57BL/6J shRNA transduced retinae showed no statistically significant differences in retinal structure compared to controls (Table 2). Moreover, neither noticeable morphological changes nor abnormal staining patterns were observed between injected or control retinas (Fig. 2A–D).

Table 2

Quantitative analysis of photoreceptor nuclear rows

A	Total AVG	Total S.D.	B	Total AVG	Total S.D.
shRNA-Gucy2e	9.31	0.255	shRNA-Cnga1	8.91	0.265
CONTROL	9.23	0.391	CONTROL	8.09	0.201
No. of C57BL/6J Eyes	10		No. of C57BL/6J Eyes	10
P-value and statistical significance			P-value and statistical significance
The two-tailed P-value = 0.3614			The two-tailed P-value = 0.2419
No statistically significant difference			No statistically significant difference
C	Total AVG	Total S.D.	D	Total AVG	Total S.D.
shRNA-Gucy2e	4	1.8	shRNA-Cnga1	4.4	0.9
CONTROL	1.1	0.3	CONTROL	1.2	0.3
No. of Pde6bH620Q Eyes	30		No. of Pde6bH620Q Eyes	30
P-value and statistical significance			P-value and statistical significance
The two-tailed P-value equals 0.0211			The two-tailed P-value is less than 0.0001
Statistically significant			Very highly statistically significant

The average (AVG) number of photoreceptor nuclei in each transduced eye was counted. The results are shown as means and standard deviation (S.D.).

Fig 2

shRNA knockdown has no detectable gross effects on photoreceptor survival or structure. Lentiviral shRNA targeting of Gucy2e (A) and Cnga1 (C) in C57BL/6J mice exhibit identical retinal structure as control retinas (B, D) injected with saline. Eight to ten rows of photoreceptor nuclei were seen in shRNA transduced and saline injected retinae.

Morphological rescue after transduction of Pde6b photoreceptors after shRNA-mediated knockdown of GUCY2E or CNGA1

Because shRNA knockdown of GUCY2E expression is not associated with gross photoreceptor degeneration in P60 control mice, we tested the ability of the shRNA-Gucy2e vector to rescue Pde6b mutant photoreceptors from degeneration. Lack of outer segments along with a single row of photoreceptor nuclei is typically observed at eight weeks in Pde6b mutants because of the natural retinal degeneration process. In contrast, Pde6b retinae transduced by shRNA-Gucy2e showed the number of photoreceptor nuclei was visibly higher than controls (Fig. 3A and B). Pde6b retinae transduced by shRNA-Gucy2e have on average four rows of photoreceptor nuclei lining areas, whereas Pde6b retinae transduced by shRNA-GFP or saline exhibited a single row of photoreceptor nuclei without outer segments. Calculated t-tests for total averages in Gucy2e retinas versus controls gave P-values of 0.0211.

Fig 3

Lentiviral shRNA targeting of Gucy2e (A) and Cnga1 (C) in Pde6b mice. At P56, the control retinae (B, D) show a single row of photoreceptors with scattered outer segments, as is typically observed in Pde6b mutants [8] (yellow arrows). In contrast, surrounding the injection site, there are rod outer segments and eight rows of photoreceptor nuclei after shRNA-Cnga1 and five rows of nuclei after shRNA-Gucy2e transduction (yellow brackets).

To evaluate whether reducing Cnga1 expression rescued degeneration, we compared histological sections from injected and control Pde6b eyes at P56. shRNA-Cnga1 injected eyes showed regions containing OS and an average of five rows of photoreceptor nuclei, whereas control retinas showed a single row of photoreceptor nuclei with scant OS at P56 (Fig. 3C and D). Differences were quantified in both individual quadrant and total averages (Table 2D). Calculated total averages of Cnga1 transduced versus control retinas gave very highly statistically significant P-values of 0.0001. Together, these results show photoreceptor survival in Pde6b mutants is improved for 2 months after shRNA-Gucy2e or shRNA-Cnga1 transduction.

Functional assessment: effect of Gucy2e and Cnga1 knockdown on photoreceptor activity in C57BL/6J controls and Pde6b mice

We measured ERG responses in C57BL/6J mice to assess the effect of shRNA-Gucy2e or shRNA-Cnga1 transduction on global retinal function between 1 and 3 months (Figs 4 and 6). In general, ERG function was depressed in eyes transduced with either shRNA-Gucy2e or shRNA-Cnga1, compared to control eyes transduced with shRNA-GFP or injected with saline. This reduction was observed in comparisons of response traces between injected and control eyes (Fig. 4) and from comparisons of averaged b-wave amplitudes for isolated rod, maximal cone-rod and isolated cone responses (Fig. 6). Although reductions in responses were noted, there was not a complete extinction of signals from the shRNA-transduced retinas up to P90.

Fig 4

Maximal dark-adapted ERG traces of C57BL/6J (wild-type) mice transduced with shRNA-Gucy2e (1) and shRNA-Cnga1 (2), and corresponding control eyes at postnatal days 35, 60 and 90. ERGs were performed on both eyes simultaneously; each colour trace represented recordings from an individual subject and corresponding control.

Fig 6

ERG b-wave enhancement of rod-specific, maximal and cone responses from shRNA-lentivirus transduced C57BL/6J and Pde6b mice. The upper panels display rod, maximal and cone ERG b-wave enhancement from C57BL/6J mice aged 1–3 months after transduction with shRNA-Gucy2e (left) or shRNA-Cnga1 (right). Lower panels show rod-specific, maximal and cone ERG b-wave enhancement from Pde6b mice aged 1–3 months after transduction with shRNA-Gucy2e (left) or shRNA-Cnga1 (right) (White = 30 days postnatal; light shaded = 60 days postnatal; Black solid = 90 days postnatal).

Next, we measured global ERG responses in Pde6b mice after shRNA-Cnga1 or shRNA-Gucy2e transduction between 1 and 3 months (Figs 5 and 6). In contrast to the effect on C57BL/6J retinas, knockdown of both genes produced enhancement of Pde6b retinal function, compared to fellow control mutant eyes. In particular, enhancement from shRNA-Cnga1 transduction was detected from P35 to P90, whereas improved photoreceptor function from shRNA-Gucy2e was observed only at P35 (Fig. 6).

Fig 5

Functional rescue of neuronal signaling in Pde6b retinas transduced with shRNA- Gucy2e (1) and shRNA-Cnga1 (2). Each mutant received a sub-retinal injection of shRNA-Gucy2e (1) and shRNA-Cnga1 (2) lentivirus in the right eye and saline in the left eye (control) at P5. Maximal dark-adapted ERGs were then performed simultaneously on both eyes of Pde6b mice at postnatal days 35, 60 and 90. Significant a-wave improvement are seen in shRNA transduced eyes at day 35. Each colour trace depicted recordings from an individual subject and corresponding control.

For shRNA-Gucy2e, the largest difference between experimental and control eyes were in maximal mixed rod-cone responses at P35: 105 μV versus 15 μV, respectively (Fig. 5). A modest improvement in b-wave peaks was also detected in isolated rod and isolated cone responses (Fig. 6). However, mice tested at P60 and P90 showed no enhancement (Figs 5 and 6).

Improved retinal function was observed from P35 to P90 in shRNA-Cnga1 transduced Pde6b mutant retinas. There was preservation of maximal ERG a-wave responses up to 60 days (Fig. 5) and isolated rod, maximal rod-cone and cone responses ERG b-wave responses up to 90 days (Fig. 6). To confirm that this effect was not because of non-specific expression of shRNAs or surgical injury response, we also tested mice injected with shRNA-GFP and saline. The ERGs from these mice did not reveal any statistical significant difference between injected eyes and non-injected eyes (data not shown).

Our interpretation of preserved rod-specific ERGs is that Gucy2e and Cnga1 knockdown compensates for sub-normal PDE activity to prevent rods from dying because of cGMP/Ca2+ toxicity. Based on our studies of the Pde6b mouse, mutant retinas express PDE6 with lower than normal specific activity and show diminished isolated rod-specific ERG responses, before significant degeneration [8]. Lowering the activity of GUCY2 in rods, through knockdown of Gucy2e, may reduce the built-up of cGMP to a level that can be managed by the endogenous Pde6b mutant enzyme to rescue cells from dying. Lowering the activity of CNG in Pde6b rods, through knockdown of Cnga1 may limit the level Ca2+ in Pde6b to a level that permit longer rod survival and function.

Discussion

We show delivery of shRNAs by lentiviral sub-retinal transduction results in reduction of GUCY2E and CNGA1 expression and significant morphological and functional rescue of Pde6b photoreceptors. These data support the model that high levels of intracellular cGMP and/or Ca2+ plays a role in the PDE6-associated RP.

In normal photoreceptors, the concentration of cGMP is controlled by the balance between its synthesis by GUCY2 and its hydrolysis by PDE6 [37-43]. GUCY2 is encoded by Gucy2e and Gucy2f, which function together to control cGMP production in photoreceptors [61]. These genes exhibit functional redundancy, as the Gucy2e/Gucy2f double mutant exhibits more severe phenotypes than either single mutant alone. For example, although full field ERGs are reduced in Gucy2e homozygotes and normal in Gucy2f mutants, no ERG responses were detected in Gucy2e/Gucy2f double homozygotes. Similarly, ONL thickness is significantly reduced by 6 months in Gucy2e/Gucy2f double homozygotes, in contrast to ONL thickness of Gucy2e mutants [61-63]. Our observation of Gucy2e knockdown in control mice associated with reduced ERG responses, in the absence of significant rod degeneration during our assay period, is consistent with the previously described Gucy2e phenotype [61].

CNGB1 and CNGA1 constitute CNG in rods. Because rods in Cngb1 mutants show reduced sensitivity to light [64], we expected to see reduced ERG responses in the shRNA-Cgna1 transduced control mice. Loss of the CNG complex could, by itself, lead to slow degeneration of rods over a period of four months [65]. However, because shRNA knockdown generally phenocopies a hypomorphic allele, we did not expect shRNA-Cnga1 to induce significant degeneration in the 2-month period of our experiments.

High intracellular cGMP levels are toxic to photoreceptors [3, 43, 66]. In Pde6b mice, PDE6 activity is undetectable and cGMP is significantly elevated above controls [43]. In this model, cGMP accumulates to high levels because GUCY2 continue to synthesize cGMP at a basal rate in the absence of PDE6 activity. In Pde6b mice, PDE6 specific activity is sub-normal suggesting that cGMP metabolism is abnormal in these mice as well [8]. Because PDE6 and GUCY2 directly catalyse reactions involving cGMP, we expected suppression of GUCY2E expression would result in increased Pde6b photoreceptor survival. Our results showing increased photoreceptor survival after Gucy2e knockdown support cGMP as an initiator of the retinal degeneration process in Pde6b mutant mice.

However necessary elevated cGMP may be in initiating cell death in these mouse models, other downstream changes are likely to be involved and may be more directly responsible for provoking degeneration. Light-sensitive conductance in photoreceptors is controlled by the effect of cGMP on Ca2+ influx through regulation of CNG channels [67-70]. Because abnormal Ca2+ levels have been implicated in causing photoreceptor cell death [15], toxicity associated with cGMP may be secondary to abnormal regulation of CNG channels [8, 16]. Indeed, our data show down-regulation of CNG channels by shRNA is associated with increased photoreceptor survival.

There are several possible reasons that may explain why the shRNA-Cnga1 approach rescued photoreceptors, compared to published reports performed with Ca2+ channel blockers that did not produce rescue [46-49]. First, the mouse line Pde6b is a partial loss of function mutant exhibiting a less severe rate of degeneration. In contrast, the Pde6b allele, which is a total loss of function mutation, was used for the calcium blocker studies. This may not be a sufficient reason, as we tested l-cis-diltiazem in Pde6b mice at three different doses (54, 108 and 216 mg/ kg/day) over the course of 2 weeks and did not detect changes in retinal degeneration relative to that in controls (unpublished results). A second possibility is that there is lower bioavailability Ca2+ channel blockers after systemic delivery compared to local sub-retinal transduction of shRNA-Cnga1. Finally, the mechanisms of action of the two approaches are different: drug binding to Ca2+ channel blocker versus RNAi-mediated reduced Ca2+ channel expression might also account for the difference in apparent effectiveness.

To date, success of viral gene therapies to rescue vision has been limited to non-photoreceptor specific diseases such as Rs1, RPE65-related early onset-retinal dystrophy, or photoreceptor diseases involving minimal cell death [71-75]. Previous attempts to rescue Pde6b mutant photoreceptors performed with viral PDE6 expression vectors has had limited success [37–41, 43]. Although future work is necessary, this approach may not provide enough long-term wild-type PDE6 activity to block excess Na+/Ca2+ entry [38-43]. Our work suggests that downstream interventions may augment this approach. Furthermore, because cGMP and Ca2+ may be a common alteration leading to cell death in other signal-dependent neurodegenerative diseases, future manipulation of cGMP or cation levels with shRNA may allow for the conversion of a progressive degeneration into a stationary disease.

Both RP and the atrophic (dry) form of AMD are characterized by an initial loss of rod photoreceptors. AMD manifests clinically by abnormal kinetics in dark adaptation and decreased rod-mediated visual function, followed by the death of cones and the retinal pigment epithelium. Rods are also affected earlier than cones in normal aging. Because reduced PDE6 function is a common denominator in many cases of photoreceptor degenerations, it may be possible to develop treatments based on decreasing cGMP production by GUCY2 and influx of Ca2+/Na+ through CNGA1. Ultimately, the results from these studies will open the possibility of genetic modification in Irish Setters with PDE6B-related degeneration and humans with photoreceptor diseases.