Additive neuroprotective effects of 24(S)-hydroxycholesterol and allopregnanolone in an ex vivo rat glaucoma model

In a rat ex vivo acute glaucoma model, high pressure (75 mmHg) causes swelling of ganglion cell axons and elevates levels of the endogenous steroids 24(S)-hydroxycholesterol (24SH) and allopregnanolone (AlloP). Furthermore, 24SH (0.1 µM) alone elevates AlloP levels via NMDA receptors. With this model, we now investigate possible interactions between 24SH and AlloP. We found that inhibition of AlloP synthesis with dutasteride under high pressure results in severe excitotoxicity in addition to axonal swelling. The excitotoxicity is prevented by exogenous AlloP but not 24SH, indicating that endogenous AlloP is crucial for protection. However, inhibition of 24SH synthesis with voriconazole induces severe excitotoxicity under normal pressure. Paradoxically, the excitotoxicity by voriconazole is better prevented by AlloP than 24SH. These findings suggest that inhibition of 24SH synthesis becomes excitotoxic in the absence of AlloP. We also observed that co-administration of sub-micromolar 24SH (0.1 µM) and AlloP (0.1 µM), concentrations that are only partially effective when administered alone, prevents axonal swelling under high pressure. This apparent enhanced protection indicates strong interaction between the two neurosteroids to preserve neuronal integrity, with 24SH contributing to AlloP synthesis via NMDA receptors and with AlloP playing an essential role in neuroprotection via GABAA receptors.


Results
Combination effects of pressure elevation and inhibitors of neurosteroid synthesis. Effects of pressure elevation and picrotoxin. We initially confirmed our prior observation about the effects of pressure elevation on retinal morphology 18 . Retinas incubated at 10 mmHg for 24 hours exhibited no remarkable changes in morphology ( Fig. 2A). We next examined the effects of pressure loading. Consistent with our previous report 14 , pressure elevation (75 mmHg) induced axonal swelling in the nerve fiber layer (NFL) with the other retinal layers remaining relatively intact 17 (Fig. 2B). To determine whether GABA A receptors are involved in the maintenance of retinal morphology, 1 μM picrotoxin, a GABA A receptor antagonist, was administered at each pressure. Retinas incubated in the presence of picrotoxin showed no remarkable changes at 10 mmHg (Fig. 2C). By contrast, picrotoxin induced excitotoxic degeneration characterized by bull's eye formation in the inner nuclear layer (INL) and edematous changes in the inner plexiform layer (IPL) 19 along with axonal degeneration in the NFL under hyperbaric conditions (Fig. 2D). This excitotoxic damage was blocked by 50 µM APV, an NMDA receptor antagonist, although some nuclear degeneration in the ganglion cell layer (GCL) and axonal swelling in the NFL remained (Fig. 2E).
A quantitative assessment of structural changes induced by pressure elevation and administration of picrotoxin or APV is summarized in Supplementary Table 1 (also see Source data of Supplementary Table 1). The NFLT, NDS, and density of damaged cells in the GCL in retinas incubated with picrotoxin at 10 mmHg showed no significant changes compared to control retinas incubated at 10 mmHg. At 75 mmHg, NDS and the density of damaged cells in the GCL were greater in retinas incubated with picrotoxin compared to control retinas incubated with aCSF. The NFLT and the density of damaged cells in the GCL significantly decreased in retinas incubated with picrotoxin and APV compared to control retinas incubated with aCSF.
Effects of pressure elevation and dutasteride. Because of our interest in endogenous NSs, we examined the effects of 1 µM dutasteride, an effective inhibitor of AlloP synthesizing enzymes (types I and II 5α-reductase). Consistent with our previous report 14 , retinas incubated in the presence of dutasteride showed no remarkable changes at 10 mmHg (Fig. 2F). Administration of dutasteride induced severe excitotoxic retinal damage at 75 mmHg characterized by bull's eye formation in the INL and edematous changes in the IPL 19 along with axonal degeneration in the NFL (Fig. 2G). Administration of 50 μM APV substantially blocked excitotoxic retinal damage induced by dutasteride at 75 mmHg (Fig. 2H). Administration of 0.1 μM AlloP showed partial neuroprotection, but did not completely inhibit either dutasteride-associated retinal degeneration or axonal swelling (Fig. 2I), while 1 μM AlloP substantially inhibited the retinal damage at 75 mmHg (Fig. 2J). By contrast, we found that 1 µM 24SH failed to inhibit dutasteride-induced retinal degeneration or axonal swelling (Fig. 2K). Even at a concentration  of 30 μM, 24SH failed to prevent high pressure-induced axonal swelling in the presence of dutasteride, though excitotoxic damage was attenuated (Fig. 2L).
A quantitative assessment of structural changes induced by pressure elevation and administration of dutasteride, APV, or neurosteroids is summarized in Supplementary Table 2 (also see Source data of  Supplementary Table 2). The NFLT, NDS, and density of damaged cells in the GCL in retinas incubated with dutasteride at 10 mmHg showed no significant changes compared to control retinas incubated at 10 mmHg. At 75 mmHg, NDS and the density of damaged cells in the GCL significantly increased in retinas incubated with dutasteride compared to control retinas incubated with aCSF. The NFLT and the density of damaged cells in the GCL significantly decreased in retinas incubated with dutasteride and APV compared to control retinas incubated with aCSF. NDS and the density of damaged cells in the GCL significantly increased in retinas incubated with dutasteride and 0.1 μM AlloP compared to control retinas incubated with aCSF. The NFLT and the density of damaged cells in the GCL significantly decreased in retinas incubated with dutasteride and 1 μM AlloP compared to control retinas incubated with aCSF. NDS and the density of damaged cells in the GCL significantly increased in retinas incubated with dutasteride and 1 μM 24SH or with dutasteride and 30 μM 24SH compared to control retinas incubated with aCSF.
Effects of pressure elevation and voriconazole. Because we were interested in examining potential interactions between AlloP and 24SH, we also examined the effects of 10 μM voriconazole, an effective inhibitor of CYP46A1 (24SH synthesizing enzyme), on retinal morphology. We initially confirmed our prior observations that 10 μM voriconazole induced severe retinal degeneration characterized by bull's eye formation in the INL and edematous changes in the IPL along with axonal degeneration in the NFL at 10 mmHg (Fig. 3A). Because voriconazole-induced damage has histological features similar to excitotoxicity, we also examined the effects of glutamate receptor antagonists. In retinas incubated with voriconazole at 10 mmHg, administration of either 50 μM APV (Fig. 3B), or 30 μM CNQX alone (Fig. 3C) failed to eliminate the retinal damage. However, a combination of APV and CNQX resulted in substantial neuroprotection in the presence of voriconazole (Fig. 3D), supporting the idea that voriconazole-induced damage involves activation of both types of glutamate receptor.
A quantitative assessment of structural changes induced by pressure elevation and administration of voriconazole, glutamate receptor antagonists (APV, CNQX), neurosteroids (AlloP, 24SH), picrotoxin, and dutasteride is summarized in Supplementary Table 3 (also see Source data of Supplementary Table 3). The NFLT, NDS, and the density of damaged cells in the GCL in retinas incubated with voriconazole, voriconazole plus APV, or voriconazole plus CNQX at 10 mmHg significantly increased compared to control retinas incubated at 10 mmHg. The NFLT, NDS, and the density of damaged cells in the GCL in retinas incubated with a combination of voriconazole, APV, and CNQX at 10 mmHg showed no significant changes compared to control retinas incubated at 10 mmHg.
Consistent with our previous report 17 , administration of 30 μM 24SH showed nearly complete protection against voriconazole at 10 mmHg (Fig. 3E). However, in the presence of voriconazole and 30 µM 24SH at 10 mmHg, picrotoxin ( Fig. 3F) or dutasteride (Fig. 3G) induced severe excitotoxic degeneration, suggesting involvement of GABA A receptors and AlloP synthesis in the neuroprotective actions of 24SH against voriconazole-induced excitotoxicity. Interestingly, APV also induced similar excitotoxic degeneration in the presence of voriconazole and 30 μM 24SH at 10 mmHg (Fig. 3H), indicating that NMDA receptor activation is necessary for 24SH to prevent the excitotoxicity.
A quantitative assessment of structural changes induced by pressure elevation and administration of voriconazole, APV, picrotoxin, dutasteride, or neurosteroids is summarized in Supplementary Table 3 (also see Source data of Supplementary Table 3).
The NFLT, NDS, and the density of damaged cells in the GCL in retinas incubated with voriconazole, 30 μM 24SH and picrotoxin at 10 mmHg significantly increased compared to control retinas incubated with voriconazole and 30 μM 24SH at 10 mmHg. The NFLT, NDS, and density of damaged cells in the GCL in retinas incubated with voriconazole, 30 μM 24SH and dutasteride at 10 mmHg significantly increased compared to control retinas incubated with voriconazole and 30 μM 24SH at 10 mmHg. The NFLT, NDS, and density of damaged cells in the GCL in retinas incubated with voriconazole, 30 μM 24SH and APV at 10 mmHg significantly increased compared to control retinas incubated with voriconazole and 30 μM 24SH at 10 mmHg.
Next, we examined whether AlloP is effective against voriconazole-induced damage. At 10 mmHg, we found that either a combination of 5 μM 24SH plus 1 μM AlloP (Fig. 3I, also see Supporting data of Fig. 3I) or administration of a high concentration of AlloP (10 μM) alone (Fig. 3J) resulted in significant neuroprotection. Neuroprotection was not observed with 5 µM 24SH alone or 1 µM AlloP alone.
A quantitative assessment of structural changes induced by pressure elevation and administration of voriconazole, APV, CNQX, or neurosteroids is summarized in Supplementary Table 3 (also see Source data of Supplementary Table 3). The NFLT, NDS, and the density of damaged cells in the GCL in retinas incubated with voriconazole, 1 μM AlloP, and 5 μM 24SH at 10 mmHg significantly decreased compared to control retinas incubated with voriconazole at 10 mmHg. The NFLT, NDS, and the density of damaged cells in the GCL in retinas incubated with voriconazole and 10 μM AlloP at 10 mmHg significantly decreased compared to control retinas incubated with voriconazole at 10 mmHg.  At 75 mmHg, voriconazole was less neurotoxic compared to 10 mmHg (Fig. 4A). Similar to 10 mmHg, administration of 1 μM 24SH also resulted in substantial protection in the presence of voriconazole at 75 mmHg ( Fig. 4B). However, in the presence of voriconazole and 1 µM 24SH at 75 mmHg, picrotoxin ( Fig. 4C) or APV (Fig. 4D) induced severe excitotoxic degeneration. Next, we examined whether AlloP is effective against voriconazole-induced damage; we found that administration of AlloP inhibited voriconazole-induced retinal degeneration at 1 μM (Fig. 4E), indicating that the presence of AlloP is essential in the neuroprotection.
A quantitative assessment of structural changes induced by pressure elevation and/or administration of voriconazole, APV, picrotoxin, or neurosteroids is summarized in Supplementary Table 4 (also see Source data of Supplementary Table 4). At 75 mmHg, the NFLT significantly decreased in retinas incubated with voriconazole compared to control retinas incubated with aCSF. NDS and the density of damaged cells in the GCL significantly increased in retinas incubated with voriconazole compared to control retinas incubated with aCSF. The NFLT and the density of damaged cells in the GCL significantly decreased in retinas incubated with voriconazole plus 1 μM 24SH compared to control retinas incubated with aCSF. NDS and the density of damaged cells in the GCL significantly increased in retinas incubated with voriconazole and 1 μM 24SH, plus picrotoxin, or with voriconazole and 1 μM 24SH, plus APV compared to control retinas incubated with aCSF. The NFLT and the density of damaged cells in the GCL significantly decreased in retinas incubated with voriconazole plus 1 μM AlloP compared to control retinas incubated with aCSF.

Combination effects of AlloP and 24SH. We hypothesized that there might be an interaction between
AlloP and 24SH because both NSs are increased during high pressure loading and both are neuroprotective. To test this hypothesis, we applied AlloP alone, 24SH alone or a combination of AlloP and 24SH at moderate concentrations to ex vivo retinas at 75 mmHg. Consistent with our previous observation 14 , administration of 1 μM AlloP substantially protected the retina from axonal swelling induced by high pressure (Fig. 5A). Additionally, 0.2 μM AlloP inhibited axonal swelling induced by high pressure (Fig. 5B). However, AlloP at 0.1 μM failed to prevent axonal swelling (Fig. 5C). Also consistent with our previous observation 17 , administration of 1 μM 24SH was completely protective against high pressure (Fig. 5D). Similarly, 0.2 μM 24SH (Fig. 5E) almost completely inhibited axonal swelling, while 0.1 μM 24SH did not alter axonal swelling (Fig. 5F). Although exogenous AlloP alone or 24SH alone did not prevent axonal swelling at 0.1 μM, a combination of 0.1 μM AlloP and 0.1 μM 24SH also inhibited pressure-induced axonal swelling (Fig. 5G, also see Supporting data of Fig. 5G), indicating that these two NSs may induce additive protection at moderate concentrations. A combination of 0.05 μM AlloP and 0.05 μM 24SH failed to protect retinas from pressure-induced axonal swelling (Fig. 5H).
A quantitative assessment of nerve fiber layer thickness (NFLT), NDS, and the density of damaged cells in the GCL by a combination with AlloP and 24SH is summarized in Supplementary Table 5 (also see Source data of Supplementary Table 5). Administration of 1 μM AlloP, 0.2 μM AlloP, 1 μM 24SH, 0.2 μM AlloP, and a combination of 0.1 μM AlloP and 0.1 μM 24SH significantly decreased the NFLT, NDS, and the density of damaged cells in the GCL compared to control retinas incubated with aCSF at 75 mmHg. The NFLT and the density of damaged cells in the GCL significantly decreased in retinas incubated with 0.1 μM AlloP or with 0.1 μM 24SH compared to control retinas incubated with aCSF at 75 mmHg. The NFLT, NDS, and the density of damaged cells in the GCL showed no significant changes in retinas incubated with a combination of 0.05 μM AlloP and with 0.05 μM 24SH compared to control retinas incubated with aCSF at 75 mmHg.
Effects of pressure elevation and neurosteroids on optic nerve head morphology. As the optic nerve head is the region where RGC axons come together, we examined the effects of pressure elevation and neurosteroids on optic nerve heads. When incubated at 10 mmHg for 24 hours, optic discs exhibited no remarkable changes in morphology (Fig. 6A). Administration of 1 μM dutasteride also induced no remarkable changes at 10 mmHg (Fig. 6B). In contrast, administration of 10 μM voriconazole induced excitotoxic degeneration in the optic disc at 10 mmHg (Fig. 6C), while a combination of APV and CNQX protected the optic nerve head from excitotoxic degeneration induced by voriconazole (Fig. 6D). 24SH (30 μM) protected the optic nerve head from voriconazole-induced neuronal damage at 10 mmHg (Fig. 6E), while picrotoxin induced severe degeneration with voriconazole and 30 μM 24SH at 10 mmHg (Fig. 6F). AlloP (10 μM) protected the optic nerve head from voriconazole-induced degeneration at 10 mmHg (Fig. 6G).
A quantitative assessment of structural changes induced by pressure elevation and/or dutasteride, voriconazole, glutamate receptor antagonists, picrotoxin, and neurosteroids on the NDS in the optic nerve head is summarized in Supplementary Tables 1-6 (10 mmHg) and Supplementary Tables 2-6 (75 mmHg) (also see Source data  Tables 1-6 and Supplementary Tables 2-6). NDS showed no significant changes in retinas incubated with dutasteride or a combination of voriconazole, APV and CNQX at 10 mmHg compared to control retinas incubated with aCSF at 10 mmHg. NDS significantly increased in retinas incubated with voriconazole at 10 mmHg compared to control retinas incubated with aCSF at 10 mmHg. NDS significantly increased in retinas incubated with voriconazole, 30 μM 24SH, and picrotoxin at 10 mmHg compared to control retinas incubated with voriconazole and 30 μM 24SH at 10 mmHg. NDS showed no significant changes in retinas incubated with voriconazole and 1 μM AlloP at 10 mmHg compared to control retinas incubated with voriconazole and 30 μM 24SH at 10 mmHg.
At 75 mmHg, NDS significantly decreased in retinas incubated with dutasteride plus APV or dutasteride plus 1 μM AlloP compared to control retinas incubated with dutasteride. NDS significantly decreased in retinas incubated with voriconazole plus 1 μM 24SH or voriconazole and 1 μM AlloP at 75 mmHg compared to control retinas incubated with voriconazole at 75 mmHg. NDS significantly decreased in retinas incubated with a combination of 0.1 μM AlloP and 0.1 μM 24SH at 75 mmHg compared to control retinas incubated with 0.1 μM AlloP at 75 mmHg.
Effects of pressure elevation on RGC survival and neuroprotection with neurosteroids. In whole mounted retinas, RGCs were specifically labeled with anti-NeuN antibody (an RGC nuclear marker) at 10 mmHg (Fig. 7A). RGC damage induced by pressure elevation (75 mm Hg) was visualized as reduced numbers of cells that were positive for NeuN (Fig. 7B). An intermediate degree of protection was observed with 0.1 μM AlloP alone (Fig. 7C) or 0.1 μM 24SH alone (Fig. 7D). Compared to retinas treated with 0.1 μM AlloP alone or 0.1 μM 24SH alone, co-administration of 0.1 μM AlloP plus 0.1 μM 24SH significantly promoted survival of NeuN-positive RGCs at 75 mmHg (Fig. 7E). The graph in Fig. 7F shows the number of NeuN-positive RGCs in the retina in each condition (also see Source data of Fig. 7F).
Pressure-induced apoptosis and neuroprotection with neurosteroids. At 10 mmHg, a small number of TUNEL-positive cells are observed only in the ONL (Fig. 7G). Exposure to elevated pressure (75 mmHg) induced apoptosis that was apparent in the GCL, INL, and ONL (Fig. 7H). An intermediate degree of protection was observed with 0.1 μM AlloP alone (Fig. 7I) or 0.1 μM 24SH alone (Fig. 7J) at 75 mmHg. Co-administration of 0.1 μM AlloP and 0.1 μM 24SH substantially inhibited apoptosis in the retina at 75 mmHg (Fig. 7K). The graph in Fig. 7L shows the number of apoptotic cells in the retina in each condition (also see Source data of Fig. 7L). These data indicate again the additive effect of the two NSs.

LC-MS/MS analysis. Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis revealed
that pressure elevation (75 mmHg) increased AlloP levels compared to 10 mmHg, while pressure-induced AlloP production was inhibited by 50 μM APV (Fig. 8A, also see Source data of Fig. 8A). LC-MS/MS also revealed that levels of 24SH increased in hyperbaric conditions (75 mmHg) compared to 10 mmHg, while pressure-induced 24SH production was inhibited by APV (Fig. 8B, also see Source data of Fig. 8B). These results indicate that synthesis of both 24SH and AlloP under hyperbaric conditions occurs downstream of NMDA receptor activation.   As shown above, the damage induced by voriconazole is inhibited by AlloP, but that of dutasteride is not inhibited by 24SH, implying that AlloP is more critical for neuroprotection than 24SH. We hypothesized that 24SH mobilizes membrane cholesterol to promote AlloP synthesis. To test this hypothesis, we applied 24SH (0.1 μM) to the ex vivo retinal specimens and determined the levels of AlloP using LC-MS/MS. At 10 mmHg, 24SH significantly increased AlloP levels. Administration of 50 μM APV inhibited the 24SH-induced AlloP increase, suggesting that the effect of 24SH on AlloP synthesis is mediated by NMDA receptor activation (Fig. 8C, also see Source data of Fig. 8C).

Discussion
As previously reported, high pressure incubation in the ex vivo retina increases levels of both AlloP and 24SH. We also found that high pressure-induced axonal swelling is prevented by either of these NSs 14,15,17 . Thus, in this work, we investigated possible interactions between AlloP and 24SH to protect retinas from high pressure damage. Current findings along with a series of our previous observations are summarized in Fig. 8D.
With regard to pressure-induced axonal swelling, the present study revealed that the simultaneous administration of AlloP and 24SH is able to inhibit pressure-induced axonal swelling at sub-micromolar concentrations (Fig. 8D), resulting in apparent additive neuroprotection. Although we have already reported enhanced AlloP and 24SH synthesis in the hyperbaric condition 15,17 , the pressure-induced increase of these two NSs is not sufficient to inhibit axonal swelling in our ex vivo system. However, it is possible that if these two NSs are more efficiently synthesized, or if the pressure elevation is more slowly applied to the retina, then the endogenous increase in these two NSs may be sufficient for retinal protection 15 . While these two NSs are similar in terms of being induced by exposure to pressure elevation and having neuroprotective properties, these NSs appear to work in a complementary additive manner because their pharmacological actions are vastly different. While AlloP exerts neuroprotective effects, at least in part, via GABA A receptors 14 , 24SH is a positive allosteric modulator of NMDA receptors 16,20 .
In the present study, we found that a low concentration of 24SH is sufficient to enhance AlloP synthesis in the normobaric condition. To clarify the mechanisms of these two NSs, we examined the protective effects of 24SH and AlloP against excitotoxicity induced by dutasteride, an inhibitor of 5α-reductase, or voriconazole, an inhibitor of CYP46A1. Dutasteride is neurotoxic at 75 mmHg compared to 10 mmHg, and the neurotoxic mechanism involves excitotoxicity via activation of glutamate receptors 15 . Voriconazole also induces excitotoxicity, but is less neurotoxic at 75 mmHg compared to 10 mmHg. Whereas dutasteride-induced neurodegeneration is blocked by APV alone (Fig. 2H), voriconazole-induced excitotoxicity is not blocked by APV alone (Fig. 3B) but by a combination of APV and CNQX (Fig. 3D). These disparities strongly suggest different mechanisms of retinal degeneration induced by depletion of 24SH or AlloP.
Dutasteride-induced excitotoxcity can be prevented by exogenously administered AlloP at 1 µM but not at 0.1 µM. A high concentration of 24SH (30 µM) can also attenuate this excitotoxicity but failed to prevent axonal swelling, indicating that axonal swelling and excitotoxicity involve distinct mechanisms.
In contrast to dutasteride, voriconazole induced more severe excitotoxicity in the normobaric condition than in the hyperbaric condition, and high concentrations of exogenous 24SH are required to overcome this excitotoxicity. This finding likely reflects the fact that basal AlloP levels are low in the normobaric condition. The presence of AlloP, however, is pivotal even for protection against excitotoxicity induced by voriconazole. As we hypothesized, AlloP synthesis is enhanced by administration of 24SH in the normobaric condition (Fig. 8C). However, if AlloP is not synthesized in a timely fashion or in sufficient amounts, it is unable to protect against the toxicity induced by inhibition of 24SH synthesis. Conversely, a relatively low concentration of 24SH can inhibit voriconazole-induced excitotoxicity in the presence of AlloP, while AlloP alone at a higher concentration can prevent voriconazole-induced toxicity. Thus, it appears that AlloP production is an essential mechanism for protection against voriconazole-induced excitotoxicity.
Elevated pressure simultaneously enhances synthesis of both AlloP and 24SH 14,15,17 , while 24SH induces AlloP synthesis at normal pressure. Therefore, it is possible that 24SH may contribute to induction of pressure-induced AlloP synthesis. If this is the case, AlloP production should be reduced when voriconazole is administered to inhibit 24SH synthesis. However, voriconazole actually does not inhibit AlloP synthesis in the hyperbaric condition 17 . Based on these findings, it appears that 24SH is not required to drive synthesis of AlloP in hyperbaric conditions and that other mechanisms come into play. One possibility is that more cholesterol is mobilized to produce AlloP and possibly other NSs when 24SH synthesis is blocked (Fig. 8E). It is also possible that voriconazole itself induces AlloP synthesis.
The increase of 24SH induced by pressure elevation is completely inhibited by voriconazole, while the level of 24SH in the normobaric condition is also lowered 17 . Furthermore, voriconazole is severely excitotoxic at 10 mmHg 17 . Because the normobaric AlloP level is minimal, the normobaric level of 24SH does not appear to contribute significantly to AlloP synthesis. However, the normobaric level of 24SH helps to preserve the retina even if the AlloP level is minimal, whereas the level of 24SH in the presence of voriconazole results in excitotoxicity. These observations suggest that there may be two pools of 24SH: a basal pool that persists in the presence of voriconazole and another pool that is promptly affected by voriconazole and can be mobilized acutely to provide neuroprotection in response to stressors such as pressure elevation. Taken together, our observations indicate that both AlloP and 24SH help to preserve retinal integrity in the hyperbaric condition. However, their mechanisms of neuroprotection are distinct and their actions are complimentary rather than synergistic. Even in the normobaric condition voriconazole is not excitotoxic when AlloP is present (Fig. 3J); even without voriconazole the hyperbaric condition alone is excitotoxic when AlloP is not available (Fig. 2G). Thus, it appears that the presence of AlloP is more critical than the increase in 24SH for neuroprotection. In spite of this, dutasteride alone is not excitotoxic in the normobaric condition (Fig. 2F) and 24SH overcomes the excitotoxicity of voriconazole (Fig. 3E). However, 24SH fails to protect the retina in the presence of both voriconazole and dutasteride (Fig. 3G). This finding suggests again that AlloP is more critical for retinal protection than the increase in 24SH.
The action of AlloP is mediated via effects on GABA A receptors. Similar to our previous finding that picrotoxin prevents the neuroprotective action of AlloP against dutasteride in the hyperbaric condition 14 , picrotoxin promotes voriconazole-induced excitotoxicity in the hyperbaric condition even in the presence of 24SH (Fig. 4C). This finding suggests strongly that AlloP synthesized in the hyperbaric condition works via GABA A receptor activation. Furthermore, we now find that the hyperbaric condition alone is excitotoxic when GABA A receptors are blocked (Fig. 2D).
A key finding in this study is that AlloP synthesis by 24SH requires NMDA receptor activation (Fig. 8C), just like AlloP synthesis in the hyperbaric condition (Fig. 8A). Although 24SH is neuroprotective against voriconazole in the normobaric condition, blockade of NMDA receptors overcomes its neuroprotective action (Fig. 3H) because AlloP cannot be synthesized, even though NMDA receptor inhibition is often neuroprotective. Indeed, we found that in the hyperbaric condition APV is neuroprotective against dutasteride (Fig. 2H) and even picrotoxin (Fig. 2E). However, in the situation where AlloP synthesis is critical, blocking NMDA receptors may be paradoxically neurodegenerative. Figure 8E provides a scheme depicting possible interactions and mechanisms underlying the neuroprotection offered by 24SH and AlloP: 24SH potentiates NMDA receptors and activation of NMDA receptors triggers cholesterol metabolism 21 resulting in AlloP synthesis and enhanced GABA receptor activation to control retinal excitability, glutamate receptors and excitotoxicity. Taken together, our results suggest that use of a combination of NSs may represent an important and novel strategy to achieve neuroprotection under glaucomatous conditions.

Methods
Protocols for animal use were approved by the Akita University Animal Studies Committee in accordance with the guidelines of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.
Rat ex vivo eyecup preparation. Rat ex vivo eyecups were prepared from 28-32 day-old male Sprague-Dawley rats (Charles River Laboratories International Inc., Wilmington, MA) as previously described 18,22 . The anterior retina was carefully removed and the empty eyecup was placed on a flat cutting surface and immersed in ice-cold aCSF. The retina was not detached from the sclera. During experiments, several eyecups (3)(4)(5)(6)(7)(8) were placed at the bottom of a 100 ml glass beaker filled with aCSF containing (in mM): 124 NaCl, 5 KCl, 2 MgSO4, 2 CaCl2, 1.25 NaH2PO4, 22 NaHCO3, and 10 glucose, and incubated at 30 °C for 24 hours using a closed pressure-loading system. In this model, a glass beaker was carefully placed at the bottom of an acrylic pressure chamber (2,000 ml) with pH maintained at 7.35 to 7.40. A 95% O 2 -5% CO 2 gas mixture was delivered via plastic tubing and an air filter (Cat#SLGP033RS, Merck Millipore, Billerica, MA), with the tubing terminating 1 cm above the bottom of the beaker. Gas flow was regulated with an infusion valve and a control dial on the lid of the pressure chamber. Acutely prepared eyecups were incubated in gassed aCSF for at least 1 h at 30 °C before pressure loading. The 95% O 2 -5% CO 2 gas mixture was infused until the pressure reading from a manometer reached the desired level. The pressure was then locked in place by adjusting the control dial of an effusion valve, and monitored continuously for 24 h at 30 °C. After maintaining the chamber at the set pressure (10 and 75 mmHg) for the indicated time, the pressure inside the chamber was carefully decreased by opening the effusion valve. The loading pressure was adjusted to intraocular pressure (IOP) in the normal condition (10 mmHg) and IOP that can occur during an acute angle closure attack (75 mmHg). In some experiments, AlloP (0.05 μM, 0.1 μM, 0.2 μM, 1 μM), 24SH (0.05 μM, 0.1 μM, 0.2 μM, 1 μM), dutasteride (a type I and II 5α-reductase inhibitor, 1 μM), voriconazole (CYP46A1 inhibitor, 10 μM), D-2-amino-5-phosphonovalerate (APV; 50 μM), and 6-cyano-7 -nitroquinoxaline-2,3-dione (CNQX; 30 μM) were added to the aCSF.
The key advantages of this model include preservation of retinal morphology without ischemic degeneration induced by IOP elevation, making it possible to investigate direct effects of pressure-loading on histology and neurosteroidogenesis. Furthermore, the ex vivo model avoids interference from circulating steroids. Light Microscopy. At the end of each experiment, eyecup preparations were fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer overnight at 4 °C. The fixed eyecups were rinsed in 0.1 M phosphate buffer and placed in 1% buffered osmium tetroxide for 60 minutes. The eyecups were dehydrated with an ethanol dilution series, embedded in epoxy resin (Epon 812, TAAB Laboratories, Aldermaston, UK) and cut into 1 μm thick semi-thin sections. The tissue was then stained with toluidine blue and evaluated by light microscopy.

Data Analysis.
In histological studies, we examined the middle portion of the retina, greater than 1,200 μm from the center of the optic disc along the inner limiting membrane (ILM) according to previously described methods 22 . The nerve fiber layer thickness (NFLT) was measured by light microscopy along 5-6 lines perpendicular to the pigment epithelium at a distance of 15 μm from each other around 1,200 μm from the center of the optic disc. The average NFLT was determined in 9 or 10 different light micrographs taken from 3 to 5 eyecup samples in each condition, divided by total retinal thickness, and mean ± standard deviation (SD) was analyzed and compared with control.
The severity of neuronal damage was assessed by light microscopy in 9 or 10 retinal fields or 3 optic nerve head fields from each experiment using a neuronal damage score (NDS) as previously described 18,22 . The NDS was determined in 10 different light micrographs taken from 3 to 5 eyecup samples in each condition.
The density of degenerated cells in the GCL was determined by counting 9 or 10 fields of 250 μm length at 10 different locations in light micrographs taken from the block of the middle retinal part 950 to 1450 μm from the center of the optic disc.
These morphometrical parameters were assessed by three raters, who remained unaware of the experimental condition. Upon completion of data assessment, significance of individual differences among raters was evaluated using five randomly selected samples in each morphometric parameter by one-way analysis of variance (one-way ANOVA) followed by a post-hoc test. There were no significant differences among the raters in any of the morphometric measurements.