Sildenafil-Mediated Reduction in Retinal Function in Heterozygous Mice Lacking the γ-Subunit of Phosphodiesterase

Abstract

purpose. Retinitis pigmentosa (RP) is a common inherited degenerative retinal disease that has many causes including mutations in the genes coding for cyclic guanosine monophosphate (cGMP) phosphodiesterase 6 (PDE6). Sildenafil (Viagra; Pfizer Pharmaceuticals, New York, NY), a widely used medication for erectile dysfunction, is a specific inhibitor of PDE, with the potential to affect PDE6 in the retina. The purpose of this study was to investigate the retinal effects of sildenafil on knockout mice heterozygous for a mutation causing absence of the γ subunit of rod PDE6 (Pdeg ^tm1/+).

methods. Wild-type mice and Pdeg ^tm1/+ mice were subjected to electroretinography (ERG) 1 hour after exposure to one of three treatments: 1) no drug, 2) an intraperitoneal injection of sildenafil at 2 times the equivalent maximal daily recommended dosage for humans, or 3) 10 times this dosage. Control ERGs were also obtained to evaluate the reversibility of changes in retinal function after sildenafil treatment. A minimum of 48 hours elapsed between electroretinogram (ERG) recordings for drug washout and animal recovery.

results. ERGs of the Pdeg ^tm1/+ mice treated with sildenafil showed a reversible, dose-dependent decrease in a- and b-wave amplitudes. Wild-type mice treated with sildenafil did not show significant differences in either a- or b-wave amplitudes compared with untreated control animals.

conclusions. These findings suggest that sildenafil has a significant impact on retinal function in Pdeg ^tm1/+ mice and may have implications in human carriers of RP. In addition, extension of these results in other model systems could be useful in understanding the mechanisms of RP and other forms of retinal degeneration.

Retinitis pigmentosa (RP) is a group of inherited retinal diseases that causes degeneration of the photoreceptors and often progresses to blindness. Affected individuals experience night blindness, loss of peripheral vision, and decreased central vision progressing over many years. ¹ Despite an explosion of information on the molecular biology of RP, an effective treatment for most patients remains elusive. Many RP mutations have been identified in humans, including some in the cyclic guanosine monophosphate (cGMP) phosphodiesterase 6 (PDE6) genes. ^{2 3 4 5 6} PDE6 is a central enzyme in the retinal phototransduction cascade, and its catalytic site has a 45% to 48% sequence identity with PDE type 5 (PDE5), ⁷ an isoenzyme found in vascular smooth muscle in the corpus cavernosum. ^{8 9 10} As would be expected, their sequence homology leads to similarities in responses to inhibitors. ¹⁰  

Sildenafil (Viagra; Pfizer Pharmaceuticals, New York, NY) is a PDE5 inhibitor in wide use for the treatment of erectile dysfunction. ^{8 9} It is known to increase the level of cGMP, and ultimately increases penile response to sexual stimulation. ¹¹ Sildenafil preferentially inhibits PDE5 more than PDE6, with a relative mean inhibitory concentration of 1:10. ¹⁰ The potential therefore exists for this drug to alter retinal function, not only in normal subjects, but also in patients with PDE6 gene mutations such as in some forms of RP. 

Visual side effects of sildenafil in humans have been reported to include the temporary occurrence of blue-tinged or hazy vision and an increased sensitivity to light. ^{12 13 14 15 16} These phenomena increase with dosage, with the incidence approaching 50% if recommended dosages are exceeded. ¹⁷ Studies evaluating retinal function in humans by electroretinography have noted small transient changes to the electroretinogram (ERG). ^{18 19 20 21 22} Such investigations have focused on normal subjects rather than those with underlying retinal diseases. Little information is therefore available on the potential side effects of sildenafil in patients with frank RP or in carriers. 

The PDE enzyme consists of an α, a β, and two γ subunits. The first two subunits are catalytic, whereas the latter perform an inhibitory function. Pdeg ^tm1/Pdeg ^tm1 gene knockout mice have been well-characterized recently. ^{23 24 25 26} These homozygotes do not possess theγ -subunit of PDE6, resulting in low enzyme activity, leading to elevated levels of cGMP in the retina. High cGMP concentrations are thought to be the initiating factor that triggers the degeneration of the photoreceptor cells. The Pdeg ^tm1/Pdeg ^tm1 homozygotes have been shown to have retinal degeneration analogous to RP in humans with known PDE mutations. Conversely, heterozygous (Pdeg ^tm1/+) mice have been shown to have normal retinal electrophysiology and histology, similar to most autosomal recessive human RP carriers. This animal model makes it possible to test the effects of various pharmacologic agents on the retina in vivo, with potential application to the understanding and treatment of human RP. The goal of this study was to quantify the effects of sildenafil on retinal function in heterozygous Pdeg ^tm1/+ mice using electroretinography. 

Methods

Mice

All procedures concerning animals in this study were approved by the Animal Care Committee at the University of British Columbia and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Founder homozygous Pdeg ^tm1/Pdeg ^tm1 mice were obtained from Columbia University (New York, NY) and crossed with wild-type C57BL/6 (+/+) mice (Jackson Laboratories, Bar Harbor, ME) to produce heterozygous Pdeg ^tm1/+ mice. All matings and subsequent offspring were carefully monitored in our animal care facility (Jack Bell Research Building, Vancouver, Canada). Mice were reared in a 12-hour dark–light cycle, which produced an ambient light level between 60 and 100 lux in the cages. All experiments were performed between postnatal day (P)28 and P49. 

Sildenafil

Using sterile technique, 100-mg tablets of sildenafil were pulverized and then solubilized with 0.9% NaCl. The subsequent solution was then filtered with a Sterivex-GS 0.22-μm filter unit (Millipore, Bedford, MA) and stored in a sterile container at 4°C. The sildenafil (3.5 mg/ml) solution was warmed to room temperature before treatment. 

Electroretinography

Three groups of mice were tested. Each mouse in group I (+/+ wild-type) underwent ERG recordings in the following order at three time points: 1) before sildenafil treatment (n = 11), 2) 1 hour after an intraperitoneal injection of sildenafil at 2 times (2×; 2.9 mg/kg) the equivalent maximum dose by weight recommended for a 70-kg human (n = 11), and 3) 1 hour after an intraperitoneal injection of sildenafil at 10 times (10×; 14.3 mg/kg) the equivalent maximum dose by weight recommended for a 70-kg human (n = 7). 

Each mouse in group II (Pdeg ^tm1/+ heterozygotes) underwent ERG recordings at three time points: 1) before sildenafil treatment (n = 12), 2) 1 hour after an intraperitoneal injection of sildenafil at the 2× dose (n = 11), and 3) 1 hour after an intraperitoneal injection of sildenafil at the 10× dose (n = 12). Group II, Pdeg ^tm1/+ heterozygous mice (n = 7) were also retested without sildenafil treatment, after the 10× treatment to evaluate the potential reversibility of the effect of sildenafil treatment on the ERGs. 

The initial recordings obtained from groups I and II were between P28 and P36. A minimum of 48-hours was maintained between all successive ERG recordings to ensure drug washout and to allow animal recovery. 

Each mouse in group III (Pdeg ^tm1/Pdeg ^tm1 homozygotes) underwent ERG recordings at a single time point (P49) without sildenafil treatment (n = 5) as a control to verify the retinal degeneration previously reported at P56. ²³  

All ERG recordings were performed by the same investigator (DB), according to a similar method previously reported. ²⁷ Before each recording session, all mice were dark adapted for a minimum of 12 hours. Under dim illumination, mice receiving treatment were given an intraperitoneal injection of 2× and of 10× sildenafil 1 hour before ERG recordings. Just before recording, all mice were anesthetized with an intramuscular injection of a ketamine (150 mg/kg) and xylazine (7.5 mg/kg) mixture. The pupils were dilated (1% cyclopentolate hydrochloride), and a DTL electrode (X-Static silver-coated conductive nylon yarn; Sauquoit, Scranton, PA) was placed on the corneal surface, which was kept moist with 0.5% methylcellulose. A reference electrode was placed in the mouth (E5 disc electrode; Grass, Quincy, MA), and a ground electrode (E2 subdermal electrode; Grass) was positioned in the tail, after topical anesthetic (1% lidocaine jelly) application to minimize any discomfort. Mice were positioned in front of a Ganzfeld dome on a warmed stage to minimize body heat loss. Scotopic intensity response functions were generated with flashes of white light emitted by a visual electrodiagnostic system (model UTAS E-2000; LKC, Gaithersburg, MD) 20 μsec in duration, starting with a −4.3 log[ candelas]cd/m⁻² per second stimulus and progressing to the brightest flash at 0.5 log cd/m⁻² per second in 0.8-log-unit steps. Each response represented an average of two to five flashes, depending on the intensity of the stimulus. To avoid the conditioning effect previously reported to alter dark-adapted ERGs, a minimum interstimulus interval of 15 seconds was maintained. ^{28 29} All responses were amplified, filtered (low [0.5 Hz] and high [1500 Hz] cutoff), and averaged (LKC software). For ERG waveform analysis, the a-wave was measured from baseline to the first negative trough of the ERG response, and the b-wave amplitude was measured from the a-wave trough to the most positive peak of the response. At flash intensities that failed to evoke an a-wave, the b-wave was measured from baseline to the most positive peak. All implicit times were measure from flash onset to the corresponding a- and b-wave amplitudes. Statistical significance was determined using analysis of variance (ANOVA). 

Results

Decreased Retinal Function

We recorded dark-adapted ERGs from Pdeg ^tm1/+ mice and wild-type (+/+) mice 1 hour after they were treated with 2 and 10 times the maximum equivalent therapeutic dose of sildenafil per unit mass recommended for humans. ERG recordings from Pdeg ^tm1/+ heterozygotes and+ /+ wild-type mice obtained before sildenafil treatment were used as control animals. The average a- and b-wave amplitudes (Fig. 1B 1D) obtained from 2×-treated Pdeg ^tm1/+ mice were significantly (P < 0.001) reduced to 121 ± 19 μV (47% ± 6.8% of the control value) and 473 ± 67 μV (55% ± 8.1% of control value), respectively, at the brightest stimulus intensity (0.5 log cd/m⁻² per second). When 2×-treated Pdeg ^tm1/+ mice were exposed to stimuli eliciting a pure rod response (stimulus of −1.9 log cd/m⁻² per second), the average b-wave amplitude reached 206 ± 37 μV representing 49%± 7.8% of the b-wave amplitude recorded from untreated control animals at the same intensity. Thus, ERGs from the 2× Pdeg ^tm1/+ group (Fig. 2B) exhibited a marked decrease in amplitudes compared with untreated control animals (Fig. 2A) . Implicit times of the a- and b-waves of 2×-treated mice were not significantly different from control animals at either the pure rod response (b-wave P > 0.05) or at the brightest stimulus intensity (a-wave and b-wave P > 0.05). 

ERG responses in 10×-treated Pdeg ^tm1/+ mice were severely impaired compared with untreated Pdeg ^tm1/+ control animals (Figs. 1B 1D) . The average a- and b-wave amplitudes obtained from Pdeg ^tm1/+ mice were significantly (P < 0.001) reduced to 40 ± 6.5 μV (21% ± 3.4% of the control value) and 331 ± 33 μV (34% ± 4.6% of the control value), respectively, at the brightest stimulus intensity. Dim stimuli eliciting a pure rod response in 10×-treated heterozygous mice (Fig. 2C) produced a significant (P < 0.001) decrease in the b-wave amplitude to 27 ± 6.9 μV, representing 5.6% ± 1.7% of the value in untreated control animals. Implicit times of the a- and b-waves were delayed at the brightest flash intensity (a-wave: 33 ± 1.8 msec at 10× sildenafil versus 20 ± 0.8 msec for the control, P < 0.001; b-wave: 108 ± 5.9 msec at 10× sildenafil versus 81 ± 4.3 msec for the control, P < 0.01). However, no significant change was observed in b-wave implicit times at stimuli evoking a pure rod response (b-wave: 111 ± 5.3 msec at 10×, versus 98 ± 5.3 msec control P > 0.05). 

Reversibility

The reversibility of the acute effects of sildenafil was determined by retesting Pdeg ^tm1/+ mice that had previously received 2× and 10× sildenafil treatment but had been given at least a 48-hour period for recovery and drug washout. No significant differences in scotopic amplitudes or implicit times were observed at either the dimmest (−1.9 log cd/m⁻² per second) or brightest (0.5 log cd/m⁻² per second) stimulus intensities between the retested eyes and the responses previously obtained from the untreated Pdeg ^tm1/+ control animals (P > 0.05; Fig. 2D ). 

Sildenafil Effects on Retinal Function in Wild-Type (+/+) Mice

We measured ERGs in wild-type (+/+) mice 1 hour after intraperitoneal injections of 2× and 10× sildenafil (Figs. 1A 1C) . Neither treatment (Figs. 3B 3C) produced differences in a- and b-wave amplitudes (P > 0.05) compared with untreated control animals. At the two brightest stimuli, however, the b-wave amplitudes displayed a trend toward higher amplitudes with increasing sildenafil dosage. The a- and b-wave implicit times of 2×- and 10×-treated wild-type (+/+) mice showed no differences (P > 0.05) when compared with those in untreated control animals. 

Evaluation of Pdeg^tm1/Pdeg^tm1 Function

Dark-adapted ERG recordings were obtained from Pdeg ^tm1/Pdeg ^tm1 (−/−) mice 49 days after birth. As expected from previous studies performed in homozygous Pdeg ^tm1/ Pdeg ^tm1 mice, the scotopic ERGs of untreated Pdeg ^tm1/Pdeg ^tm1 mice were severely affected by the absence of the γ subunit of PDE6. ³⁰ We were unable to detect any appreciable ERG responses, even at the brightest stimulus intensity (Fig. 3D) . 

Discussion

Our study shows that sildenafil can markedly decrease retinal function in knockout mice heterozygous for the PDE6γ -subunit gene mutation (Pdeg ^tm1/+) with as little as 2× the maximum equivalent dose recommended for humans. Further decreases to both a- and b-wave amplitudes, with concurrent delays in implicit times, were observed in Pdeg ^tm1/+ mice treated with 10× sildenafil. This dose-dependent decrease was not observed in wild-type (+/+) control animals. 

We were surprised to note that wild-type (+/+) mice had a subtle dose-dependent augmentation of ERG amplitudes at brighter stimuli with sildenafil treatment. These results are in contrast to the known action of sildenafil as an inhibitor of PDE6. A similar effect was reported in wild-type dogs receiving increasing doses of sildenafil. ^{31 32} However, a moderate decline in the ERG was noted in these animals when the dosage reached 10× the maximum equivalent dose recommended for humans, a result not observed in the present study. Studies of nonspecific PDE inhibitors in cats demonstrated that lower doses slightly increase the amplitude of the rod b-wave, whereas diminished amplitudes were observed at higher doses. ³³ Because a given concentration of sildenafil inhibits PDE6 10 times less than PDE5, ¹⁰ the 2×- and 10×-treated wild-type (+/+) mice may have responded in a relatively low-dose fashion, exhibiting the ERG enhancement seen at lower doses of nonspecific PDE inhibitors. This low-dose response may also be responsible for the temporary side effects, such as increased sensitivity to light, that have been noted in patients taking sildenafil. ¹²  

The heterozygous PDE6 γ-subunit knockout mutation probably leads to a decrease in functional PDE6, creating enhanced susceptibility to the inhibitory effects of sildenafil. Higher levels of cGMP resulting from this effect may clamp the membrane potential at a depolarized voltage resulting in a partial blockage of the phototransduction cascade. Increasing the dosages of sildenafil would be expected to produce further reductions in PDE6 activity and retinal function, as we observed. Accordingly, heterozygous Pdeg ^tm1/+ mice receiving 2× and 10× sildenafil had diminished retinal function; whereas similarly treated wild-type (+/+) mice did not. Although other investigators have studied the effects of phototransduction cascade inhibitors, so far as we are aware, we are the first to report that a specific inhibitor of PDE decreases retinal function in a model of retinal degeneration. 

Although our study focused on the inhibitory effects of sildenafil, we also evaluated short-term reversibility of these effects on retinal function in heterozygous Pdeg ^tm1/+ mice. A minimum 48-hour (36–120 half-lives) washout and recovery period was used before testing for reversibility, ³⁰ and it is therefore unlikely that any drug remained at the time of retesting. Sildenafil’s impact on retinal function in Pdeg ^tm1/+ mice appeared to be temporary, although there are no long-term data. One-year ocular toxicity studies involving wild-type dogs receiving high doses of sildenafil have reported similar results, ³¹ but toxicity studies have not been performed on animals that are carriers of retinal degenerations. Repeated exposure to sildenafil in carriers may result in chronically elevated levels of cGMP, placing them at an increased risk for retinal toxicity. Long-term studies would be helpful to investigate this concern further. 

We found no appreciable rod function in untreated Pdeg ^tm1/Pdeg ^tm1 knockout mice at 7 weeks after birth. It has been reported previously that these homozygous Pdeg ^tm1/Pdeg ^tm1 mice have reduced PDE6 activity, reduced electrophysiologic responses, and histologic evidence of retinal degeneration analogous to that found in RP. ²⁴ Diminished levels of PDE6 are known to increase cGMP in Pdeg ^tm1/Pdeg ^tm1 mice, and high cGMP levels are in turn known to be toxic to the retina, ^{34 35 36} explaining both the electrophysiologic and histologic findings. ²³ Our observations of extinguished ERGs in the homozygous Pdeg ^tm1/Pdeg ^tm1 mice are in keeping with this mechanism. Given that the PDE γ-subunit deletion of the Pdeg ^tm1/Pdeg ^tm1 mouse effectively reproduces the same retinal degeneration observed in the PDE β-subunit mutant rd/rd, presumably the implications for heterozygous Pdeg ^tm1/+β -subunit carriers would be similar to the γ carriers previously described in the rd1 mouse. ²³  

At least 1 in 50 people worldwide is a carrier of RP. ³⁷ Most carriers have normal visual function and ERGs, ³⁸ but their risk for potential retinal toxicity from PDE inhibitors is presently unknown. Recent human studies found no evidence of ocular toxicity, aside from transient visual disturbances in patients taking sildenafil. ^{15 16 22 39 40} However, these studies have either excluded subjects with known visual problems, or have been unlikely to include a sufficient number of RP carriers to study this important subpopulation. Thus, the transient and potential long-term effect of sildenafil on visual function may not be fully appreciated at this time. 

Although our understanding of RP has progressed rapidly, only one half of known genes responsible for this disease have been characterized ⁴¹ (summarized on Retnet at http://www.sph.uth.tmc.edu/Retnet/home.htm; provided by the University of Houston–Texas Health Science Center). PDE α- and β-subunit mutations have been characterized in individuals with RP ^{2 3 4 5 6} ; however, γ-subunit mutations have not yet been identified in humans. ⁴² Whether people heterozygous or homozygous for α- and β-subunit mutations are susceptible to changes in retinal function from sildenafil in a fashion analogous to the Pdeg ^tm1/+ heterozygous mice remains to be seen. Moreover, because PDE6 mutations comprise a modest 3% to 4.5% of cases of autosomal recessive RP in humans, ⁵ an important issue is to determine whether mutations in other genes leading to retinal degenerations lead to the same potential susceptibility. We look forward to extension of our results in other model systems to further elucidate the physiologic risks to people heterozygous for RP. 

 Submitted for publication July 19, 2000; revised October 2, 2000; accepted October 6, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Summary of scotopic intensity response functions obtained from untreated and sildenafil-treated wild type (A, C) and heterozygous Pdeg ^tm1/+ (B, D) mice. (A, B) b-Wave and (C, D) a-wave amplitudes obtained in untreated conditions and after 2× and 10× sildenafil treatment. Sildenafil had little effect on ERG amplitudes in wild-type (+/+) mice but appeared to have a dose-dependent reduction in a-wave and b-wave amplitudes in heterozygous (Pdeg ^tm1/+) mice. All symbols represent the mean values recorded at that stimulus. Bars, SEM.

Figure 2.

 

Representative scotopic ERG intensity response functions obtained from the left eye of the same heterozygous (Pdeg ^tm1/+) mouse showing (A) untreated control conditions, (B) response to 2× and (C) 10× sildenafil treatments, and (D) retesting after washout of 2× and 10× sildenafil treatments. Mice treated with sildenafil (B, C) showed a dose-dependent decrease in a-wave (arrowhead) and b-wave amplitudes in contrast to untreated (A) values. ERG responses illustrated in the posttreatment retested mouse (D) showed little difference when compared with its control (A). A minimum 48-hour period for recovery and drug washout was maintained between each recording session. Arrow: Flash onset.

Figure 3.

 

Representative scotopic ERG intensity response functions obtained from the left eye of one wild-type (+/+) mouse showing (A) untreated control conditions (B) after 2× and (C) 10× sildenafil treatments. When different treatments were compared at the same stimulus, the a-wave (arrowhead) and b-wave amplitudes showed minor variations. A minimum 48-hour period for recovery and drug washout was maintained between each recording condition. (D) Representative scotopic ERG intensity response functions obtained from an untreated (−/−) control mouse on postnatal day 49 were not measurable. Arrow: Flash onset.

The authors thank Peter Gouras and Stephen Tsang for providing Pdeg ^tm1/Pdeg ^tm1 mice, Pierre Lachapelle for critically reading the manuscript, and Ann Lee for technical assistance.