|Home | About | Journals | Submit | Contact Us | Français|
Retinitis pigmentosa (RP) is a genetically heterogeneous set of blinding diseases that affects more than a million people worldwide. In humans, ~5-8% of recessive and dominant RP cases are caused by nonsense mutations in the Pde6b gene coding for the ß-subunit of the rod photoreceptor cGMP phosphodiesterase 6 (PDE6-ß). The study of the disease has been greatly aided by the Pde6brd1 (rd1) mouse model of RP carrying a null PDE6ß allele. Degenerating rd1 rods were found to experience a pathological increase in intracellular calcium concentration (‘Ca overload’) when they enter the apoptotic process at postnatal day 10. A 1999 study suggested that the Ca2+ channel antagonist D-cis diltiazem delays the kinetics of rd1 rod degeneration, conferring partial rescue on scotopic vision. Subsequent reports were mixed: whereas several studies failed to replicate the original results, others appeared to confirm the neuroprotective effects of Ca2+ channel antagonists such as diltiazem, nilvadipine and verapamil. We discuss the discordant results and the controversy they caused, and suggest potential causes for the discrepancy. We also discuss potential involvement of recently identified Ca2+-dependent mechanisms that include protective calcium ATPase mechanisms, ryanodine and IP3 calcium stores, and store operated channels in Pde6brd1 neurodegeneration.
Retinitis pigmentosa (RP), a genetically heterogeneous set of blinding diseases affecting more than a million people worldwide, derives its name from an accumulation of pigment clusters in the neural retina. Early symptoms associated with the main rod-cone dystrophy variants include night blindness and loss of peripheral vision, followed by eventual loss of central vision. One of the best-characterized forms of recessive RP in humans is caused by a nonsense mutation in the gene coding for the β subunit of the rod photoreceptor cGMP phosphodiesterase 6 (PDE6b) that has been linked to 5-8% cases of RP (McLaughlin et al., 1995). The essential features of photoreceptor degeneration in human RP have been replicated in rodent models, which can be studied over much shorter time frames. The Pde6brd1 (rd1) mouse model carries a recessive nonsense mutation in PDE6b (reviewed in Farber, 1995). The mutation is characterized by rapid degeneration of rods beginning at late stages of photoreceptor differentiation, which in turn triggers apoptosis of cones. By P90, virtually all photoreceptors have disappeared (Carter-Dawson-et al., 1978; Punzo et al., 2009) except ~3% of cone perikarya in the dorsal retina (Jimenez et al., 1996). While gene therapy and pharmacological treatments have been attempted to ameliorate photoreceptor degeneration in RP models, a significant bottleneck for developing treatment has occurred due to the large number of different genes that lead to RP (>46; www.sph.uth.tmc.edu/Retnet/) and the apparent inconsistency of pharmacological studies focused on treating RP. Hopes for a possible restorative pharmacological intervention were kindled by the report that D-cis diltiazem - an antagonist of voltage-activated Ca2+ channels which is commonly used as a cardioprotectant – partially rescues rd1 photoreceptors (Frasson et al., 1999). Suppression of Ca2+ pathways that orchestrate apoptotic cascades in rods and cones promised to deliver a general therapeutic strategy that could be applied to treating many recessive retinal degenerations. Subsequent reports appeared to confirm (Read et al. 2002; Sanges et al., 2006) or repudiate (Pawlyk et al, 2002; Bush et al. 2000) the basic findings of the Frasson study. The aim of this paper is to review the literature, explore potential mechanisms involved in Pde6brd1 neurodegeneration in the light of recent findings on Ca2+ signaling in vertebrate photoreceptors and explore the possibilities offered by the pharmacological approach to prevention of photoreceptor degeneration.
The Pde6brd1 (rd1) mouse phenotype was first identified in a Harvard albino mouse strain and subsequently found in wild and laboratory mice across the United States and Europe (e.g., Pittler et al., 1993). The mutation is associated with a nonfunctional rod-specific PDE6 gene located on the mouse chromosome 5. Every mouse strain with the rd1 gene carries a murine leukemia provirus integrated into the first intron, combined with a point mutation which introduces a stop codon in exon 7 (Farber, 1995). At P10, 33% of mRNA is transcribed in Pde6brd1 rods (Viczian et al., 1991), however, the message is not expressed, presumably due to nonsense-mediated decay of PDEb6 mRNA and/or instability of the truncated PDE6b peptide (Bowes et al., 1990; W. Baehr, personal communication). Loss of the β subunit completely eliminates function of the PDE complex which consists of a catalytic dimer (non-identical α and β subunits) and two inhibitory γ subunits. PDE malfunction affects both amplitude and kinetics of the photocurrent and results in cGMP, Na+ and Ca2+ overloads within rd1 cells.
Although it is clear that rd1 phenotype is caused by mutated PDE6, the causal relationship between abnormally high [cGMP] and apoptosis is still not well understood. cGMP content in the Pde6brd1 retina increases days before signs of degeneration become apparent, reaching its maximum around the eye opening when rod degeneration is at its peak (Farber and Lolley 1974). Early signs of rd1 expression in photoreceptors such as retarded growth of OS and IS are seen 4-8 days after birth (Sanyal and Bal, 1973). Swelling of rod mitochondria and appearance of vacuoles in the IS occur at P8 (Farber and Lolley 1974). While cellular integrity and thickness of P10 rd1 ONL is comparable to the WT, OS disks show signs of disruption, chromatin is fragmented and the P10 increase in TUNEL-positive outer nuclear layer (ONL) cells is followed by rapid degeneration of rods by P14. At P18-P21, rods are gone, leaving a single row of cone perikarya (Carter-Dawson et al., 1978) and a total loss of image-forming vision by P40.
Rod photoreceptors are formed by two compartments, the OS and the IS, that communicate through a thin nonmotile cilium. PDE6 in the wild type retina is located in the OS whereas the pro-apoptotic cascades are confined to the IS. Pde6brd1 retinas express a rudimentary OS, hence it is not clear whether a messenger mechanism transmits a ‘degenerate and die’ message from the OS to the IS and/or the entire holoenzyme complex remains ‘stuck’ in the IS. Such a messenger could be the metabolic stress caused by depletion of ATP expended to combat excessive Na+ and Ca2+ influx or stress caused by accumulation of proteins in the ER. An alternative, which is not mutually exclusive, is that the signal is mediated by elevated [Ca2+] itself as Ca2+ regulates all aspects of photoreceptor function, including degeneration (Szikra and Krizaj, 2009).
There is little doubt that degenerating photoreceptors experience Ca2+ overload. Elevated cytosolic Ca2+ concentrations in rd1 rods have been measured directly with Ca2+ indicator dyes and indirectly by observing compensatory changes in Ca2+-dependent proteins. Around P10, [Ca2+]i elevations are observed in acutely dissociated rd1 rods and in cultured rd1 explants (Doonan et al., 2005; Sanges et al., 2006), accompanied by upregulated CaM kinase II and migration of phosducins & associated Gtβ1γ1 subunits (Donovan and Cotter, 2002; Hauck et al., 2005) whereas the calpain inhibitor calpastatin is strongly downregulated (Paquet-Durand et al. 2006). Other parallel signs of Ca2+-dependent changes in rd1 rods include prominent increases in m and μ calpain proteases and cleavage of caspases-3 & -12 and AIF (apoptosis-inducing factor; Sharma and Rohrer, 2004). Mitochondria are depolarized by excess Ca2+ (Doonan et al., 2005), possibly leading to activation of the permeability transition pore (He et al., 2000) and translocation of AIF and caspase 12 into the nucleus (Sanges et al., 2006). Another Ca2+ store, the endoplasmic reticulum (ER), activates pro-apoptotic Bid, Bax, PERK, IRE1α and ATF4 pathways (Lin et al., 2007). Given that rd1 rods experience an enormous degree of ER stress due to the traffic jam of proteins normally destined for the OS, it is very likely that there is a concomitant malfunction of SERCA-mediated Ca2+ sequestration and ryanodine & IP3 receptor mediated release of ER Ca2+ (see below).
In early 90-ies Mark Tso's group at the University of Illinois showed that the diphenylpiperazine flunarizine, a non-specific antagonist of Ca2+ channels (Edward et al., 1991) but not the dihydropyridine nimodipine, a specific L-type channel antagonist (Li et al., 1991), protect photoreceptors stressed by exposure to strong light. However, two L-type antagonists, D-cis diltiazem and verapamil, were soon afterwards found to inhibit photoreceptor degeneration in the Drosophila rdbB mutant (Sahly et al., 1992) and to confer partial protection on Pde6brd1 photoreceptors (Frasson et al., 1999; Takano et al., 2004; Sanges et al., 2006). These studies initiated interest in pharmacological approach to protecting vision through management of the Ca2+ overload.
The dihydropyridine diltiazem is a competitive antagonist of L-type calcium channels with an IC50 of ~5 μM (Koch and, Kaupp 1985). The D-cis stereoisomer is the active compound in several commonly used prescription drugs. Three other L-type blockers, verapamil, nilvadipine, nimodipine and nifedipine are also designated for human use.
The first important study using D-cis diltiazem was performed by Serge Picaud's group in Strasbourg and published in Nature Medicine in 1999 (Frasson et al., 1999). The protocol involved intraperitoneal (IP) injections of D-cis-diltiazem two times a day, starting at P9, reaching a daily dose of 54 mg per kg body weight, a dose ~6 fold higher than the maximal allowed human daily dose (<1mg/kg 4 times daily; Bush et al., 2000). Retinal wholemounts and slices were labeled with an antibody against rhodopsin and outcomes were measured indirectly via the scotopic ERG b-wave which showed a remaining b-wave (~20 μV) in 4 out of 10 animals treated at P36. Though the amplitude was small (compared to ~800 μV in wild type eyes), this result was highly significant as the rd1 scotopic b-wave is basically flat around P21 (Pawlyk et al. 2002). Additionally, the diltiazem-treated cohort at P36 had ~18600 surviving rods, a 2-3 fold increase compared to 7500 rods in non-treated animals; a modest protective effect on cone survival was also observed. These early results prompted a clinical study which found that diltiazem, supplemented with vitamin E, improves vision in human RP patients (Pasantes-Morales et al. 2002).
The immediate concern voiced by Tso was that the rhodopsin antibody used by Frasson et al. might have depicted ongoing rod degeneration rather than healthy, “protected”, cells (Edward and Tso, 2000). Several groups, which attempted to replicate the original study using D-cis-diltiazem in different models of RP degeneration, reported mostly discouraging results. A histological & ERG study by Sieving's group found no significant rescue in the P23H rat that carries a dominant negative opsin mutation (Bush et al. 2000). A follow-up study using the same methodology in the rcd1 dog that carries a stop codon in its PDE6b gene found no neuroprotection (Pearce-Kelling et al. 2001). Yamazaki et al. (2002), Fox et al. (2003) and Takano et al. (2004) observed no effect of D-cis diltiazem on Pde6brd1 rods after treating P9 mice for 7 days with 1 mg/ml diltiazem (i.e., at ~50 times lower dose than the Frasson study). Finally, Pawlyk et al (2002), using a different rd1 strain, found no significant differences in a double blind study that reproduced the original study conditions. There were no changes in rhodopsin and cone (S and M) opsin immunoreactivity nor were changes observed in the ERG, leading the authors to conclude “there is no compelling rationale for testing D-cis-diltiazem (Cardizem, Hoechst-Marion Roussel) as a possible treatment for RP.”
As D-cis diltiazem in the Frasson study only partially (~50%) blocked L-type channels, a stronger effect would be expected with full channel block or in vitro. A subsequent study found that elimination of the β2 subunit of the L-type calcium channel results in degeneration at a (delayed) rate similar to that in the Frasson study (Read et al., 2002). A 40% increase in the number of surviving cells at P21 was observed in Pde6brd1 - β2 KO animals, directly and unambiguously confirming involvement of L-type channels in rd1 degeneration. Subsequently, Sanges et al. (2006) showed that the D-cis stereoisomer blocked Ca2+ overload, resulting in a decreased number of TUNEL-positive rd1 rods. Translocation of AIF and caspase-12 to the nucleus of remaining rods was inhibited, confirming the effectiveness of D-cis diltiazem in retinal rd1 explants.
Nilvadipine, an L-type antagonist often used in Japan, was effective in decreasing photoreceptor cell loss in the RCS rat (Yamazaki et al. 2002) and Pde6brd1 (Takano et al. 2004) & rds mice (Takeuchi, etal. 2008). The effect of nilvadipine vs. the lack of effect seen with diltiazem treatment was ascribed to greater permeability of dihydropyridines vs. benzothiazepins (Takano et al. 2004). Surprisingly, while nilvadipine appeared to preserve photoreceptor structure, it had no effect on ERG b-wave amplitudes. Microarray analysis after nilvadipine treatment showed downregulation of proapoptotic genes coding for caspases-3, -9 and -14 as well as and ADP-ribosyl cyclase, the product of which acts as an allosteric modulator of the ryanodine receptor (RyR). RyR itself, a major cellular mediator of calcium-induced calcium release in the ER, was upregulated (Takano et al. 2004), implicating Ca2+ stores in rd1 degeneration.
D-cis-diltiazem inhibits not only L-type calcium channels, but also CNG channels and vice versa, the L-cis isomer inhibits L-type channels with an IC50 ~ 75 μM (Koch and Kaupp, 1985; Baumann et al., 2004). A slight advantage of L-cis over the D-cis isomer was suggested for retinas treated with zaprinast (Vallazza-Deschamps et al. 2005) and lead (Fox et al., 2003). In conclusion, while elimination of L-type-mediated Ca2+ entry delays, but does not prevent, rd1 rod degeneration, the accumulated evidence supports the notion that Ca2+ overload mediated by some type of Ca2+ channel potentiates degeneration.
The ineffectiveness of L-type channel blockade in rat and dog might have been due to drug dosing and accessibility issues (some dihydropyridines are notoriously difficult to get into solution let alone across the blood-brain barrier). Alternatively, the rhodopsin mutation in the P23H mutant and the Mertk mutation in the RCS rat may involve different degeneration mechanisms unrelated to Ca2+ influx. A final possibility is that Ca2+ regulation in degenerating photoreceptors involves players that were not directly affected by L-type and/or CNG channel antagonists.
Edward et al. (1991) hinted at the contribution of IP3-sensitive intracellular Ca2+ stores localized in the ER. Indeed, suppression of Ca2+ sequestration into the ER kills differentiating photoreceptors (Linden, 1999; Chiarini et al. 2003), presumably due to global suppression of mRNA translation that occurs in cells with depleted Ca2+ stores (Brostrom and Brostrom, 2003). Conversely, at P10-P12, Ca2+ overload in rd1 cells coincides with peak expression of pro-apoptotic ER markers such as caspase-12, phospho-pancreatic ER kinase (p-PERK), phospho-eukaryotic initiation factor 2α (p-eIF2α) and glucose-regulated protein-78 (GRP78/BiP) (Yang et al., 2007, Lin et al., 2007), possibly resulting in their amplification. This suggests that a critical junction for preventing Ca2+-mediated apoptosis might be Ca2+ stores and/or store-operated channels rather than L-type/CNG channels.
Store-operated channels belonging to different canonical TRPC classes are prominently expressed in rods and cones (Szikra et al., 2008; Krizaj et al., 2008a) together with their activators, STIM1 proteins (Szikra et al., 2009). TRPCs play an increasingly recognized role relative to Ca2+ overloads that drive calpains and apoptosis (Marasa et al., 2006; Shan et al., 2008) and neuroprotection (Bollimuntha et al., 2006; Yamamoto et al., 2007). In normal retinas, Ca2+ influx through these channels is likely to serve a neuroprotective function by preventing pathological decreases in [Ca2+]i in cells exposed to saturating light, which when prolonged was shown to trigger photoreceptor degeneration (Leconte and Barnstable, 2000; Woodruff et al., 2003). TRPC transcription and expression is dramatically affected in Pde6brd1 degeneration together with upregulated expression of the major protective PMCA1 calcium transporters (Krizaj et al., 2008b). This suggests that rd1 rods are battling Ca2+ overload by upregulation of endogenous Ca2+ clearance mechanisms. It remains to be seen whether upregulation of these ATPases prolongs rod survival. In summary, the Pde6brd1 mouse model has greatly facilitated studies of retinal degeneration caused by defects in expression and function of PDE6. This degeneration is intimately associated with chronic elevation in [Ca2+]i which has a causal role in activating pro-apoptotic pathways within the inner segment. The conflicting interpretations of pharmacological effects may stem partly from lack of drug specificity and partly from our lack of a comprehensive understanding of Ca2+ regulation and trafficking in the photoreceptor IS. Successful therapeutic interventions will depend on our understanding of signaling pathways involved in the degeneration process.
This work was supported by the Knights Templar Eye Foundation (PB and CCP), NIH (EY13870), Foundation Fighting Blindness and an unrestricted grant from Research to Prevent Blindness to the Moran Eye Institute. We thank Dr. Wolfgang Baehr for helpful comments.