Secondary Retinal Ganglion Cell Death and the Neuroprotective Effects of the Calcium Channel Blocker Lomerizine

Abstract

Purpose.: After partial optic nerve (ON) injury, intact retinal ganglion cells (RGCs) undergo secondary death, but the topographic distribution of this death is unknown, and it is unclear which cell death pathways are involved. Although the calcium channel blocker lomerizine reduces RGC death after partial ON injury, it is unknown whether this drug alleviates necrotic or apoptotic death.

Methods.: The dorsal ON was transected in adult Piebald-Virol-Glaxo (PVG) rats, and the site of secondary RGC death was determined using anterograde and retrograde DiI tracing. RGC death was assessed at 2 and 3 weeks. Retrograde tracing with fluorogold injected into the superior colliculus 3 days before euthanatization was used to identify RGCs undergoing secondary death. Overall cell loss was quantified using βIII-tubulin immunohistochemistry. Lomerizine (30 mg/kg, oral) or vehicle was given twice daily, and retinal wholemounts were analyzed for necrotic morphology (nucleic acid stain) or anticleaved caspase-3 expression at 2 and 3 weeks.

Results.: Ventral retina was identified as the site of secondary RGC death, and central and dorsal retinae were defined as sites of both primary and secondary death. Overall RGC loss occurred by 2 weeks in central and ventral retina (P < 0.05) and by 3 weeks in dorsal retina (P < 0.05). Secondary RGC death was characterized mainly by necrotic morphology, with caspase-3 expression in some RGCs. Lomerizine reduced secondary necrosis at 2 weeks and secondary caspase-3 expression at 3 weeks.

Conclusions.: Lomerizine had differential effects on necrotic and apoptotic death with time, but its inability to completely prevent secondary death suggests that full neuroprotection will require combinatorial treatments.

After traumatic injury, some neurons are axotomized by the lesion (the primary injury) and die rapidly over several weeks. ^1,2 Other neurons escape axotomy, and, though initially undamaged, are vulnerable to secondary degeneration. ^3–5 Similarly, in glaucoma, the transport of neurotrophic factors to retinal ganglion cell (RGC) somata is reduced, leading to primary apoptotic RGC death ^2,6 and an ensuing wave of secondary death among neighboring RGCs. ^3,4,6,7 Therapeutic strategies that aim to replace ⁸ or regenerate ^9–11 axotomized neurons have had limited success. However, it is now recognized that the rescue of tissue affected by secondary degeneration is not only feasible but is critical to improving long-term function after CNS injury or glaucoma. ^2,12  

Partial transection of the optic nerve (ON) is emerging as a useful CNS model in which to differentiate damage resulting from primary (direct) compared with secondary (indirect) degeneration. ^3,4,13 Transection of the dorsal ON allows spatial separation of RGCs axotomized in the primary injury from initially uninjured RGC axons in ventral ON and their respective parent somata in the eye. RGCs in dorsal retina project their optic axons strictly along the dorsal ON in primates, ^3,14 marsupials, ¹⁵ and the Wistar rat. ⁴ In contrast, ON topography in the Dark Agouti rat is reported to be inverted; optic axons in dorsal ON have corresponding parent somata in ventral retina. ¹⁶ Differences between rat strains are therefore apparent, ¹⁷ but the topography of dorsal optic nerve axons in the strain used in this study, the hooded Piebald-Virol-Glaxo (PVG) rat, is unknown. 

Programmed cell death after injury to the CNS is associated with at least three morphologically and molecularly distinct processes—apoptosis, autophagic cell death, and necrosis—though autophagy is considered an accompaniment rather than a cause of cell death. ¹⁸ Apoptosis is an energy-dependent, genetic process that has been detected using TUNEL immunohistochemistry after ON axotomy ^19–22 and ON crush. ^23,24 However, TUNEL immunohistochemistry lacks specificity and contrasts with detection of caspase-3, which is collectively activated by initiating caspases 8, 9, and 10 and is responsible for the breakdown of RGC nuclear contents into discrete vesicles bound with an impermeable membrane. ^25,26 Caspase-independent apoptosis may also occur through the activation of apoptosis inducing factor by calpain. ²⁷ Unregulated RGC death (often called necrosis) has been reported to occur immediately after high intraocular pressure-induced ischemic injury, ²⁸ in intense excitotoxic conditions, ²⁹ and after partial ON crush injury. ²³ Depletion of energy resources causes failure of adenosine triphosphate (ATP)–dependent sodium and potassium pumps, resulting in ionic and osmotic influxes that lead to cell swelling and the loss of membrane permeability. Semicondensed chromatin haphazardly breaks down into heterogeneous-sized fragments, and the nucleus slowly begins dissolution. ^26,30 However, it is increasingly apparent that CNS injury results in a spectrum of cell death processes encompassing various combinations of apoptotic and necrotic features (termed aponecrosis or necroptosis), with apoptosis and necrosis perhaps at either ends of a continuum. ^27,31,32 The death processes of RGCs undergoing secondary degeneration are yet to be conclusively identified. 

Loss of intracellular calcium homeostasis is a final common pathway for many events leading to secondary RGC death. High cytoplasmic calcium concentrations occur as a consequence of chronic membrane depolarization, allowing calcium influx largely through voltage sensitive L-type calcium channels. ³³ Depletion of ATP stores prevents calcium extrusion; high intracellular calcium concentrations in turn trigger the release of further calcium stored in the smooth endoplasmic reticulum. ^2,34 The resultant increase in cytoplasmic calcium concentrations activates intracellular calcium-dependent signaling processes ³⁵ that may lead to secondary death by apoptotic or necrotic processes. ^36–38  

The relatively newly synthesized calcium channel blocker lomerizine, also known as KB-2796, is a diphenylmethylpiperazine compound currently marketed as an antimigraine drug ³⁹ and is in phase II clinical trials for glaucoma. ^40,41 Lomerizine is highly selective for L-type channels in the CNS and cerebral arteries and therefore has few peripheral side effects. ^39–42 In addition, lomerizine rescues neurons from glutamate excitotoxicity–induced death through the blockade of calcium influx by L-type voltage-sensitive calcium channels ⁴¹ and rescues RGCs from secondary death after partial ON crush injury and partial transection. ^42,43 In the present study, we define the location and proportion of RGC somata undergoing secondary cell death in the retina, after partial transection of the ON in PVG rat. We quantify RGCs undergoing secondary cell death involving necrotic morphology or caspase 3 expression and determine whether lomerizine alleviates these types of cell death. 

Materials and Methods

Animals and Surgical Procedures

The authors confirm adherence to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and procedures were approved and monitored by The University of Western Australia Animal Ethics Committee. Female PVG Hooded (Rattus norvegicus) adult rats weighing 150 to 200 g each were obtained from the Animal Resources Centre (Murdoch, WA) and maintained on a 12-hour light/12-hour dark cycle in a temperature-controlled room, with ad libitum access to standard rat chow and water. From the day of surgery, animals used for assessment of nucleic acid staining (Styox Green; Invitrogen, Carlsbad, CA) or caspase-3 expression, were treated twice daily with 30 mg/kg lomerizine hydrochloride (LKT Laboratories, St. Paul, MN) or vehicle for 2 or 3 weeks. The dose of lomerizine chosen was based on a dose-response study. ⁴² Each dose of lomerizine was mixed with butter into a paste and administered using a spatula, with minimal stress to the animal. Lomerizine was administered by carefully placing the spatula tip holding lomerizine in butter on the left side of the protruding front teeth of the rat and wiping the spatula tip behind the teeth such that the lomerizine remained in the rat's mouth. Each rat was held and inspected for several seconds to ensure that the treatment was swallowed. Anesthesia, the partial ON transection procedure, and subsequent retrograde labeling of RGCs with fluorogold were conducted as previously described. ^3,4,43 Partial transection (200-μm cut) of the dorsal nerve was undertaken using a diamond radial keratotomy knife. The depth of the cut was determined by the protrusion of the blade beyond a surrounding precision-calibrated guard, allowing precise and reproducible injury. ^3,4,13,43 Fluorogold labeling was conducted 3 days prior to perfusion in animals used for assessment of nucleic acid staining (Sytox Green; Invitrogen) or caspase-3 expression (normal animals: Sytox staining, n = 3; caspase-3 expression, n = 3; perfusion at 2 weeks: Sytox staining, n = 12, caspase-3 expression, n = 13; perfusion at 3 weeks: Sytox staining, n = 11; caspase-3 assessment, n = 13). Animals were perfused transcardially with 0.9% saline followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.2). 

Retrograde DiI Tracing In Vivo after Partial Optic Nerve Transection

Immediately after partial transection of the ON, a few crystals of DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate; Molecular Probes, Inc., Eugene, OR) were placed directly into the injury site, with care taken to confine dye placement to the partial transection site (n = 4). Use of DiI for retrograde labeling of RGCs is a technique established in the literature, with the dye remaining within the axotomized cells it enters. ^44,45 As such, DiI placed into the partial transection site precisely and reproducibly labels only RGCs axotomized by the primary injury. Four days after surgery, animals were perfused as described, and the right eye and attached ON were removed and postfixed for 30 minutes. The cornea, lens, and sclera were dissected away, and the retina was disconnected from the ON head, mounted onto a slide, and coverslipped (Fluoromount-G; Southern Biotechnology, Birmingham, AL) without the removal of the vitreous. Numbers of DiI-labeled RGC somata were counted in 33 fields of view (400 × 400 μm) distributed in 11 regions across dorsal, central, and ventral retina. Mean RGCs per square millimeter for each region were grouped according to similar densities and portrayed schematically. The highest SEM for regions in each group is shown in the legend to Figure 1. 

Anterograde DiI Tracing Ex Vivo

Normal animals were terminally anesthetized and perfused, and both eyes were orientated with a dorsal suture (n = 5). The eye and attached ON were removed, and the cornea, lens, and vitreous were dissected to expose the retina. The surface of the retinal cup was thoroughly dried with hardened Whatman's 50 filter paper, and a crystal of DiI was applied ventrally. The eyecup and attached ON were immersed in fixative and incubated in the dark at 37°C for 23 days to allow for DiI transport. After 6 weeks, the fixed ON and attached eyecup were immersed in 15% sucrose in PBS overnight for cryopreservation. The eyecup was positioned for sagittal sectioning, with the nasotemporal retinal axis perpendicular to the block face so as to expose the dorsoventral retinal axis, embedded in OCT mounting medium (Tissue-Tek; Sakura Finetek, Torrance, CA), and frozen at −20°C to −24°C. Sections (20 μm) were cut using a cryostat (CM1900; Leica, Wetzlar, Germany) and collected onto glass slides (Superfrost Plus; Selby Biolabs, Australia). After 3 × 5-minute washes with PBS, slides were mounted (Fluoromount-G; Southern Biotechnology) and viewed under a fluorescent microscope. 

Immunohistochemistry

Immunohistochemical preparation of retinal tissue was conducted as previously described ⁴³ using primary antibodies to βIII-tubulin (TUJ1; 1:500; Covance, Princeton, NJ) diluted in 0.2% Triton X-100 in PBS or cleaved caspase-3 (1:500; Molecular Probes, Eugene, OR) ⁴⁶ diluted in 3% bovine serum albumin and 1% Triton X-100 in PBS. Previous studies have shown that βIII tubulin can be used as a marker for RGCs and distinguishes them from the smaller displaced amacrine cell population. ^47–50 Specific immunohistochemical identification of caspase-dependent apoptotic cells using antibodies to caspase-3 has also been described. ⁴⁶ After thorough washing, secondary antibodies goat anti-mouse (Alexa Fluor 488, 1:500 dilution; Molecular Probes, Inc, Eugene, OR) or anti-rabbit (Alexa Fluor 488, 1:400; Molecular Probes) and Hoechst nuclear stain (1:1000, Molecular Probes, Inc, Eugene, OR) diluted with 0.2% Triton X-100 in PBS were applied. 

Necrotic Morphology: Nucleic Acid Staining

After retrograde fluorogold labeling, animals were euthanatized and the right eye of each was immediately enucleated in semisterile conditions. The cornea and lens were dissected away in Hanks balanced salt solution (Invitrogen), and the eyecup was immersed in nucleic acid stain (Sytox Green; 1:4000; Molecular Probes, Eugene, OR) diluted in HBSS, and gently agitated for 15 minutes. After 5× 2-minute washes with HBSS, the eyecup was mounted, and small radial incisions were made with a new scalpel. The wholemount was coverslipped with a few drops of HBSS and counted immediately with a fluorescence microscope. The entire preparation was completed strictly within 15 minutes. Loss of stain because of immunohistochemical processing for βIII-tubulin precluded the use of this marker. 

Quantification of RGCs

βIII-Tubulin–Positive RGCs.

βIII-tubulin immunohistochemistry can be used as a marker for RGCs in the ganglion cell layer. ^10,43,47,48 The density of βIII-tubulin–positive RGCs after partial transection was counted in the 12 retinal sections closest to the ON/animal (n = 3 animals) with the use of a 100× objective to analyze three fields of view (100 × 100 μm) in each of far dorsal, central and far ventral retina. The optical fractionator counting technique ⁵¹ was used to avoid double counting of RGCs in adjacent sections. 

Necrotic Morphology.

The density of nucleic acid stain and fluorogold-positive RGCs was determined immediately after completion of the wholemount procedure using a 25× objective to analyze three fields of view (400 × 400 μm) each in the dorsal, central, and ventral retina, strictly within a 15-minute time limit. 

Caspase 3-Dependent Apoptosis.

RGCs positive for caspase-3, βIII tubulin, and fluorogold were counted in a serpentine manner across the entire wholemounted retina using a 25× objective (field of view, 400 × 400μm). Locations of positive RGCs were recorded onto a schematic retinal map and classified as located in the dorsal, central, or ventral retina. Densities of RGCs were expressed as mean ± SEM. RGCs per square millimeter were derived from a total retinal area of 65.2 ± 1.0 mm² (similar to Ref. 52). 

Densities of RGCs at different time points after partial transection were compared with RGC densities in normal animals for dorsal, central, or ventral retinal regions using one-way ANOVA and Bonferroni/Dunn post hoc tests, requiring a significance of P ≤ 0.05. Similarly, densities of RGCs in different retinal regions at each time point after partial transection were compared with each other using one-way ANOVA and Bonferroni/Dunn post hoc tests, requiring a significance of P ≤ 0.05. Comparisons between vehicle- and lomerizine-treated groups at each time point were conducted using Student's t-tests requiring a significance of P ≤ 0.05. 

Results

Secondary Death of RGCs with Somata in the Ventral Retina after Partial ON Transection

The location of parent somata of RGCs axotomized by partial dorsal ON transection was determined by precise application of DiI to the partial ON injury site and quantification of DiI-labeled RGCs in retinal wholemounts 4 days later. Ninety-five percent of DiI-labeled RGCs were located in dorsal or central retinal regions and were virtually absent in ventral and nasal retinal regions (Figs. 1A–D). The highest density of DiI-labeled somata (150 RGCs/mm²) lay in a region in central retina, surrounded by lower densities (51–150 RGCs/mm²). The lowest densities (50 DiI-labeled RGCs/mm²) were spread across peripheral dorsal and peripheral temporal retinal regions (Fig. 1A). In PVG rats, dorsal optic axons axotomized by the primary injury of partial ON transection therefore have parent somata predominantly in central and, to a lesser extent, dorsal retina. 

Figure 1.

 

Schematic diagram (A) and representative images (B–D) show parent somata of dorsal optic nerve axons labeled in vivo with DiI by placing DiI crystals precisely into the injury site immediately after partial optic nerve transection. (A) After 4 days of retrograde transport, retinal wholemounts were prepared (n = 4), and DiI-positive RGC somata were quantified in 33 fields of view (400 × 400 μm) distributed in 11 regions across central and peripheral retina. Mean RGCs/mm² for each region were grouped according to similar densities and portrayed schematically. The highest SEM for regions in each group is shown. (B–D) Representative images of DiI-positive somata in dorsal (B), central (C), and ventral (D) retina are shown. (E) Montaged representative images of ex vivo anterograde DiI tracing illustrate topography of optic axons arising from ventral RGC somata. Crystals of DiI taken up by RGC somata in ventral retina projected ventrally in the ON. Scale bars: (B–D) 50 μm; (E) 100 μm.

Conversely, DiI applied to a small area in ventral retina ex vivo labeled ventral RGC somata and their optic axons, which projected exclusively ventrally along the length of the ON past the site of partial transection (Fig. 1E). These data are thus consistent with data from other rat strains but differ from results previously reported for the Dark-Agouti strain. ¹⁶ RGCs with their somata in ventral retina are not axotomized by partial dorsal ON transection and are therefore vulnerable only to secondary, and not primary, cell death. 

RGC Densities at 2 and 3 Weeks after Partial ON Transection

In normal animals, RGC densities visualized by βIII tubulin immunohistochemistry are almost twice as high in central retina (2751 ± 61 RGCs/mm²) as in far dorsal (1572 ± 44 RGCs/mm²) and far ventral (1602 ± 64 RGCs/mm²) retina (P ≤ 0.05). A similar relative distribution of RGC densities is maintained across the retina after partial transection but with decreased densities in central (2197 ± 54 RGCs/mm²; 80% of normal value; P ≤ 0.05) and far ventral (1277 ± 80 RGCs/mm²; 80% of normal value; P ≤ 0.05) retina at 2 weeks. There is no further RGC loss in these regions by 3 weeks (central: 2071 ± 73 RGCs/mm², 74% of normal value, P > 0.05; far ventral: 1167 ± 30 RGCs/mm², 71%, P > 0.05). The density of RGCs in far dorsal retina at 2 weeks (1261 ± 72 RGCs/mm²) was not significantly different from normal (P > 0.05). Significant RGC loss in far dorsal retina (1193 ± 80.91 RGCs/mm², 76% of normal value, P ≤ 0.05) is, however, seen by 3 weeks. 

Effects of Lomerizine on the Density of RGCs Undergoing Secondary Cell Death with Necrotic Morphology

Optic axons of RGC somata retrogradely labeled with fluorogold are connected to the superior colliculus and are capable of functional axonal transport. ⁵³ Fluorogold was applied after partial optic nerve transection, 3 days before analysis, therefore labeling RGCs that were still intact but vulnerable to secondary degeneration while excluding the labeling of axotomized RGCs (those axotomized by the primary injury). Nucleic acid stain enters cells with permeable plasma membranes and binds to intracellular fragmented DNA with high affinity, but it does not stain apoptotic dying cells or apoptotic bodies because cell membrane integrity is maintained during true apoptosis; as such, nucleic acid stain indicates necrotic morphology. ⁵⁴ Combining fluorogold labeling with nucleic acid staining enabled definitive identification of RGCs undergoing secondary cell death with necrotic morphology at particular snapshots in time. 

Detectable, but very low, densities of RGCs undergoing secondary death with necrotic morphology were found in the dorsal retina of normal animals (Fig. 2). A previous study did not detect necrosis in uninjured retina. ²³ Compared with normal animals, densities of RGCs undergoing secondary cell death with necrotic morphology in central retina are increased 2 weeks after partial transection (P ≤ 0.05; Fig. 2). Secondary cell death of RGCs with necrotic morphology is not significantly different from normal rats 3 weeks after partial transection (Fig. 2; P > 0.05). Compared with vehicle, lomerizine reduces the density of RGCs undergoing secondary death with necrotic morphology in central retina at 2 weeks (2.5 ± 0.6 RGCs/mm²; 72% reduction; P ≤ 0.05), to levels similar to those seen in normal animals (Fig. 2). 

Figure 2.

 

Representative images (A, B) and quantification of fluorogold^+ve RGCs with necrotic morphology (C). (A) RGCs were stained with nucleic acid stain (green) and (B) retrogradely labeled with fluorogold ^53,54 (same cell labeled →). Mean ± SEM densities of nucleic acid stain^+ve, fluorogold^+ve RGCs were expressed graphically for vehicle and lomerizine treatment groups 2 or 3 weeks after partial transection. Normal untreated animals acted as a control. Statistically significant differences between treatment groups for each retinal region are indicated by asterisks (P ≤ 0.05). Scale bars, (A, B) 50 μm.

Effects of Lomerizine on the Density of RGCs Undergoing Caspase-3–Dependent Secondary Cell Death

As expected, in normal animals, densities of RGCs undergoing secondary death and expressing caspase-3 are negligible (0.005 ± 0.005 RGCs/mm²). After partial transection, overall densities of RGCs undergoing secondary death and expressing activated caspase-3 across the retina are approximately 50 to 100 times lower than densities of RGCs undergoing secondary death with necrotic characteristics. Compared with normal animals, after partial transection, the density of RGCs undergoing secondary cell death with caspase-3 expression was significantly higher in central retina at 2 weeks and 3 weeks (0.08 ± 0.009 RGCs/mm² and 0.08 ± 0.007 RGCs/mm², respectively; P ≤ 0.05; Fig. 3). ^55,56 There is no significant secondary cell death with caspase-3 expression in dorsal retina at either time point (Fig. 3). However, the density of RGCs undergoing secondary cell death with caspase-3 expression in ventral retina becomes significant 3 weeks after partial transection (0.04 ± 0.01 RGCs/mm²; P ≤ 0.05; Fig. 3). The density of RGCs undergoing secondary cell death with caspase-3 expression is significantly higher in central retina than in dorsal or ventral retina at each time point after partial transection (P ≤ 0.05). 

Figure 3.

 

Representative images (A, B) and quantification of fluorogold^+ve RGCs expressing caspase-3 (C). (A) RGCs were positive for βIII-tubulin (green) and caspase-3 and (B) retrogradely labeled with fluorogold ^55,56 (same cell labeled →). Mean ± SEM densities of caspase-3^+ve, βIII-tubulin^+ve, fluorogold^+ve RGCs were expressed graphically for vehicle and lomerizine treatment groups 2 or 3 weeks after partial transection. Normal untreated animals acted as a control. Statistically significant differences between treatment groups for each retinal region are indicated by asterisks (P ≤ 0.05). Scale bars, (A, B) 50 μm.

Compared with vehicle, lomerizine does not reduce the density of RGCs undergoing caspase-3–dependent secondary death in central retina at 2 weeks (P > 0.05), but it does so at 3 weeks (0.031 ± 0.01 RGCs/mm²; 61% reduction; P ≤ 0.05; Fig 3). Lomerizine treatment also significantly reduces the density of RGCs undergoing caspase-3–dependent secondary death in ventral retina at 3 weeks (0.004 ± 0.002 RGCs/mm²; 88% reduction; P ≤ 0.05; Fig 3). 

Discussion

The present study demonstrates that partial transection of dorsal optic axons in the PVG rat results in the primary and secondary death of RGCs in central and dorsal retina; in contrast, RGCs in ventral retina are unaffected by the primary injury and exclusively undergo secondary death. Combined with retrograde fluorogold labeling after partial ON transection, the model provides a unique opportunity to spatially separate and quantitatively analyze secondary cell death. Secondary RGC death occurs predominantly by necrosis in central retina, with a 50× to 100× greater incidence than that occurring by caspase-3–dependent processes in central and later ventral retina. Compared with vehicle, lomerizine results in early (2 weeks) rescue of RGCs undergoing secondary death with necrotic morphology, followed by later (3 weeks) rescue of the much smaller number of RGCs undergoing caspase-3–dependent death. 

Progressive Loss of RGCs after Partial Optic Nerve Transection

After partial dorsal optic nerve transection, we observed 20% loss of RGCs, visualized by βIII-tubulin immunohistochemistry, in central and ventral retina by 2 weeks but no further loss in these regions by 3 weeks; significant loss of RGCs in far dorsal retina was seen by 3 weeks, possibly because of the low number of RGCs in this region. βIII-tubulin immunohistochemistry labels RGCs ^10,47–50 regardless of whether they have undergone axotomy and are experiencing primary cell death or have escaped the initial injury and are succumbing to secondary cell death. Results of the present study support our previous observations on the loss of RGCs exclusively undergoing secondary cell death, identified by retrograde fluorogold labeling after partial optic nerve transection ⁴³ ; loss was significant in central retina at 7 days and had spread to ventral retina by 28 days. The similar spatiotemporal spread of RGC loss from central to peripheral retina in the two studies mirrors the high central to lowest ventral topography of axotomized RGCs in the PVG rat, which we report here using retrograde DiI labeling (Fig 1), and suggests that secondary RGC death spreads relatively quickly from central to ventral retina. In agreement with other studies, ^4,22 it is likely that the primary death of most RGCs in the partial transection model in PVG rats had already taken place in central and dorsal retina by approximately 2 weeks after injury, accompanied by concurrent but also more prolonged secondary RGC death. Total numbers of RGCs in the PVG rat retina range from a maximum of 178,000 centrally to 104,000 peripherally (average, 128,000 across the retina), values that are in line with the range of densities of RGCs observed in other rat species. ^47,57,58  

Predominance of Necrotic Morphology May Be Related to the Nature of the Injury

It has been proposed that predominantly necrotic-like cell death or bax-independent, calpain-dependent programmed cell death is a result of high oxidative stress and energy depletion, whereas caspase-dependent apoptosis is associated with neurotrophic factor withdrawal. ^27,59 Our demonstration that secondary RGC death after partial transection is characterized predominantly by necrotic morphology and, to a far lesser extent, by caspase-3 expression may indicate an oxidative stress-dependent insult. Alternatively, the predominance of secondary necrotic processes rather than apoptotic processes may be a result of an active process involving crosslinks between death pathways, ensuring that one mode of death is dominant over the other. ⁶⁰ The presence of necrotic morphology after injury to the ON has not been as well documented as that of caspase-dependent apoptosis. ⁶¹ However, after partial optic nerve crush ²³ or ischemic injury, ²⁸ necrosis accounted for most of the early phase of RGC loss and was followed by apoptosis. It is possible that the delayed onset of necrosis after partial ON transection is a consequence of the relatively confined nature of the primary injury, producing milder secondary effects compared with optic nerve crush, which is a more global injury in terms of both primary and secondary effects. The low but detectable number of dying cells with necrotic morphology in normal animals is consistent with the age-dependent decline in RGC numbers reported in humans and rodents. ^62–64  

Caspase-3 expression in a small proportion of RGCs undergoing secondary RGC death began in central retina, where primary RGC death is highest. By 3 weeks, caspase-3 expression has spread to ventral retina, the site of exclusively secondary RGC death. The effects of direct (primary) injury in central retina might have been spread by glutamate and calcium released into the extracellular space or by interconnected glial networks, resulting in increased glutamate and calcium uptake by adjacent RGCs or their axons in the optic nerve. ^28,34,65 Furthermore, RGC death is likely to have resulted in the release of proapoptotic death signaling molecules, such as bax and bim, ⁶⁶ into the extracellular space, where they could induce secondary caspase-3 expression. 

Lomerizine Rescues RGCs Undergoing Secondary Death by Both Necrosis and Caspase-3–Dependent Mechanisms

Lomerizine reduced the density of RGCs undergoing secondary cell death with both necrotic morphology and caspase-3 expression, extending the results of previous studies showing that lomerizine significantly reduced secondary death, 28 days after injury ^42,43 and indicating efficacy against cell death activation regardless of its nature or plasticity. ³² Lomerizine is known to specifically inhibit L-type calcium channels present on RGC somata ⁶⁷ and subsequently to reduce high concentrations of intrasomal calcium. ⁴⁰ Lomerizine may reduce secondary necrotic morphology by inhibition of L-type calcium channels on RGC somata, thereby reducing the influx of calcium responsible for RGC somata swelling and permeabilization. Lower intrasomal calcium concentrations also result in reductions in activity of the proapoptotic molecule Bax, calpains, and caspase-dependent apoptosis–inducing molecules, such as cytochrome c and endonucleases. ^25,26  

Lomerizine exhibited differential effects on caspase-3 expression in central retina (where primary death is highest), dependent on the time after partial transection. Glutamate-induced calcium influx is blocked by the calcium channel blocker nilvadipine in only approximately 50% of RGCs. ⁶⁸ It is feasible that a small subpopulation of RGCs vulnerable to apoptotic death processes is “insensitive” or has low affinity to lomerizine. The presence of specific subpopulations of RGCs with variable vulnerabilities or death “timetables” may explain the differential effects of lomerizine on caspase-3 expression at 2 weeks compared with 3 weeks. The presence of subpopulations of RGCs resistant to injury is thought to be dependent on the intrinsic properties of an RGC. ⁸ Such resistance to injury by specific subpopulations of RGCs has been demonstrated after complete optic nerve transection and in knockout mice lacking the antiapoptotic signaling molecule bcl-2. ⁶⁹ Alternatively, differences in the environments within the retina and the optic nerve at these time points may explain the differential effects lomerizine has on caspase-3 expression. 

Given the inability of lomerizine to completely protect RGCs from caspase-3 expression, it is likely that combinatorial approaches to prevent secondary degeneration will be required. Indeed, effects other than death of RGC somata have been described after partial optic nerve injury, including transient loss of axonal transport, demyelination, and synaptic plasticity within target tissue. ^{2,42,43,70,71} Combining lomerizine with alternative therapies known to target any of these factors could be a useful approach to preventing secondary death. We conclude, therefore, that lomerizine has clinical implications for the treatment of traumatic CNS injuries and eye diseases such as glaucoma, though full protection will require combinatorial treatments. 

Footnotes

 Supported by the Neurotrauma Research Program, Western Australia (MF, SAD), and by the National Health and Medical Research Council, Australia (MF, SAD). SAD is a Professorial Fellow (Research) and Senior Research Fellow of the National Health and Medical Research Council (Grant 254670).

Footnotes

 Disclosure: M. Fitzgerald, None; S.C. Payne, None; C.A. Bartlett, None; L. Evill, None; A.R. Harvey, None; S.A. Dunlop, None

Footnotes

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

The authors thank Woodside Australia for supporting their research by naming them beneficiaries of the Woodside Sustainability Award 2007, and Vince Clark and Michael Archer for assistance with technical procedures.