Overexpression of acid ceramidase (ASAH1) protects retinal cells (ARPE19) from oxidative stress[S]

Over 11 million people in the United States alone have some form of age-related macular degeneration (AMD). Oxidative stress, cell death, and the degeneration of retinal pigment epithelial (RPE) cells contribute to AMD pathology. Recent evidence suggests that ceramide (Cer), a cellular sphingolipid mediator that acts as a second messenger to induce apoptosis, might play a role in RPE cell death. The lysosomal breakdown of Cer by acid ceramidase [N-acylsphingosine amidohydrolase (ASAH)1] into sphingosine (Sph) is the major source for Sph 1-phosphate production, which has an opposing role to Cer and provides cytoprotection. Here, we investigated the role of Cer in human RPE-derived ARPE19 cells under hydrogen peroxide-induced oxidative stress, and show that Cer and hexosyl-Cer levels increase in the oxidatively stressed ARPE19 cells, which can be prevented by overexpression of lysosomal ASAH1. This study demonstrates that oxidative stress generates sphingolipid death mediators in retinal cells and that induction of ASAH1 could rescue retinal cells from oxidative stress by hydrolyzing excess Cers.

The generation of endogenous Cer is induced by noxious stimuli, such as UV A light (22) and proinflammatory cytokines (e.g., TNF-) (23), both of which promote the release of Cer from SM and/or by the de novo synthesis pathway via palmitate (24). Cer induces apoptosis via multiple mechanisms, such as through the inhibition of AKT-1 and activation of extrinsic and intrinsic apoptotic pathways (15). These pathways converge on the cleavage of caspase-3 and result in DNA fragmentation, degradation of the cytoskeleton and nuclear proteins, and the formation of apoptotic bodies, including DNA fragments (e.g., DNA ladder) (25). Apoptotic cell death is the potential driver for retinal neuron degeneration (e.g., photoreceptor cells, ganglion cells, and RPE cells), and this is what occurs in human retinitis pigmentosa, glaucoma, and AMD, respectively. In recent studies utilizing animal and cell culture models, Cer accumulation has been established as one of the underlying causes of retinal cell death (26)(27)(28). Previously, using FTY720, a Cer synthase (CerS) inhibitor that blocks de novo light-induced Cer production (11,(29)(30)(31), we have shown that Cer plays a second-messenger role in the induction of retinal cell death, and that its inhibition can prevent and/or delay cell death (29). Therefore, targeting Cer could be a potential strategy for developing neuroprotective therapies for retinal diseases.
The enzyme group that hydrolyzes accumulated Cer into sphingosine (Sph) and releases free fatty acids is called ceramidase, also known as N-acylsphingosine amidohydrolase (ASAH). The lysosomal form of this enzyme is acid ceramidase, or ASAH1, and has very high expression in retinal and RPE cells [N. Mandal, unpublished observations; (6)]. The lysosomal breakdown of Cer into Sph is the major source for intracellular Sph 1-phosphate (S1P) production, which has an opposing role to Cer and provides cytoprotection (32,33). We hypothesize that elevated ceramidase activity will reduce endogenous Cer levels, thereby protecting the cells from Cer-induced apoptosis.
In this study, we show that Cer and hexosyl-Cer (Hex-Cer) levels increase in ARPE19 cells that are oxidatively stressed by hydrogen peroxide (H 2 O 2 ), and that overexpression of lysosomal ASAH1 can reduce the levels of both of these bioactive lipids. ARPE19 cells are immortalized human RPE cells that maintain their native epithelial properties and are extensively used in eye research. This study demonstrates, for the first time, that the overexpression of ASAH1 could help to protect human retinal cells from oxidative stress-induced cell death.

Plasmid construction and introduction
cDNA of the known lysosomal transcript of human ASAH1 (GenBank accession number BC016481.1) was cloned from ARPE19 cells. In order to construct expression vectors, cDNA encoding lysosomal ASAH1 (Ly-ASAH1) was fused to the upstream DNA segment encoding the fluorescent protein, Venus (supplemental Fig. S1). Recombinant plasmids were linearized by restriction digestion and electroporated into ARPE19 cells (CUY21Pro-Vitro; NEPA GENE, Tokyo, Japan). Successful transgene expression of ASAH1 was confirmed by assessing the expression of Venus fluorescence by microscopy (Axiovert 40; Carl Zeiss, Oberkochen, Germany), real-time PCR, and Western blotting. The localization of ASAH1 in the cells was also investigated using a lysosomal marker, LysoTracker Red DND-99 (Invitrogen, Life Technologies Corp., Eugene, OR). Lysosomal ASAH1 (Ly-ASAH1)-transduced cells were sorted with the excitation wavelength of 488 nm by Cell Sorter SH800 (Sony, Tokyo, Japan), and stably transfected ARPE19-ASAH1 cells were selected.

Confocal laser scanning microscopy
For confocal laser scanning microscopy experiments, expression of ASAH1 was confirmed by assessing the expression of Venus fluorescence using a Zeiss 710 confocal laser scanning microscope (Carl Zeiss). Localization of ASAH1 in the cells was also investigated using a lysosomal marker, LysoTracker Red DND-99 (Invitrogen, Life Technologies Corp.). Live cells were seeded on glass coverslips and incubated with 500 nM of LysoTracker Red DND-99 for 2 h according to the manufacturer's instructions. Cells were then washed with 1× PBS, fixed with 4% paraformaldehyde for 15 min at room temperature, mounted in ProLong Diamond Antifade mountant with DAPI (Invitrogen, Life Technologies Corp.), and examined. Images were captured using ZEN 2012 imaging software (Carl Zeiss).

Western blots
Total cell lysates were used for Western blot analysis. Preparation of lysates was done utilizing T-PER tissue protein extraction reagent (Thermo Scientific, Rockford, IL) containing protease inhibitor (Roche, Indianapolis, IN) followed by brief sonication, centrifugation at 16,000 g for 10 min at 4°C, and supernatants stored at 20°C until needed. Protein estimation was performed utilizing a BCA assay (Thermo Scientific) for total protein concentration. Aliquots (15 g) of protein samples were loaded on a 4-20% Tris-glycine gel (Novex; Life Technologies) for gel electrophoresis, transferred with a Bio-Rad Trans-Blot turbo transfer system to PVDF membrane (Bio-Rad, Hercules, CA), and membranes were incubated with Odyssey blocking buffer (TBS) (Licor, Lincoln, NE) for 1 h. Membranes were incubated with primary antibodies for anti-ASAH1 (MilliporeSigma, Burlington, MA; 2 g/ml), anti-GFP (for the presence of Venus) (MilliporeSigma;,1:1,000), and -actin (Abcam, Cambridge, MA; 1:5,000) overnight at 4°C and then incubated with appropriate secondary antibodies for 1 h at 4°C. Proteins were visualized by ECL detection (Pierce; Thermo Scientific) and data were normalized against -actin.

Cell proliferation assay
In order to measure rate of growth, ARPE19 and ARPE19-ASAH1 cells were plated in 12-well plates at a concentration of 2 × 10 5 cells/plate in replication, and grown at 37°C for up to 48 h.

Oxidative stress and short-chain Cer cell viability assays
ARPE19 cells (1 × 10 4 ) were plated in 96-well plates and grown at 37°C for 24 h. They were cultured in serum-free medium containing 100-1,000 M H 2 O 2 (Santoku Chemical Industries Co., Tokyo, Japan). After 24 h of incubation, cell viability was assessed by using a CellTiter 96 aqueous nonradioactive cell proliferation assay kit (Promega, Madison, WI). In this kit, viable and living cells are quantified by measuring the absorbance of the assay plates at 490 nm following the addition of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS), which is reduced by metabolically active cells to yield a soluble formazan product. From preliminary and previous studies, we estimated that at least 800 M of H 2 O 2 were needed to induce measurable cell death within 24 h in ARPE19 cells (35). We used an application of 800-1,000 M of H 2 O 2 for 24 h to induce oxidative stress in ARPE19-ASAH1 cells, and ARPE19 cells served as controls. For biochemical and molecular analyses, we treated the cells with 1,000 M of H 2 O 2 for 3 and 24 h, and harvested and froze the cells. Deionized water in place of H 2 O 2 served as the vehicle (Veh) for Veh-treated groups. Additionally, ARPE19 cells were grown per the methods stated above and were treated with exogenous C2-Cer dissolved in ethanol (Avanti Polar Lipids, Alabaster, AL) at concentrations of 10-50 M, and an equivalent amount of ethanol-treated (100%) cells served as controls. Estimations based on preliminary data suggested that 20 M of C2-Cer was sufficient to induce measurable cell death within 24 h in ARPE19 cells. ARPE19 and ARPE19-ASAH1 cells were then treated with 10-20 M of C2-Cer for 24 h and cell viability was determined per the methods stated above. To supplement the CellTiter viability assay, cell viability following H 2 O 2 -induced oxidative stress, as described above, was determined indirectly in a repeated experiment by measuring released lactate dehydrogenase (LDH) into the medium using a commercial kit from Promega (CytoTox-One homogenous membrane integrity assay kit) following the manufacturer's instructions. For the LDH release assay, fluorescent measure of the release of LDH from cells with a damaged membrane was calculated by subtraction of culture medium background and the cell viability (percent) was calculated from 100% viability (mean value of untreated cells) to 0% viability (mean value of 2% Triton X-100-treated cells).

Quantitative analysis of sphingolipids
Sphingolipids were analyzed in the Lipidomics Core facility at Virginia Commonwealth University, Richmond, VA following previously established procedures (36). All internal standards were procured from Avanti Polar Lipids and added to samples at a ratio of ethanol:methanol:water (7:2:1) as a cocktail of 500 pmol each. Standards for sphingoid bases and sphingoid base 1-phosphates included the 17-carbon chain length analogs: C17-Sph, (2S,3R,4E)-2-aminoheptadec-4-ene-1,3-diol (d17:1-So); C17-sphinganine, (2S,3R)-2-aminoheptadecane-1,3-diol (d17:0-Sa); C17-S1P, heptadecasphing-4-enine-1-phosphate (d17:1-So1P); and C17-sphinganine 1-phosphate, heptadecasphinganine-1-phosphate (d17:0- Sphingolipid extraction was performed according to previously established procedures (11,29,36). Briefly, samples were added to 1 ml of CH 3 OH and 0.5 ml of CHCl 3 , along with internal standards, into 13 × 100 mm borosilicate tubes with a Teflon-lined cap (60827-453; VWR). Samples were dispersed using an ultra sonicator for 30 s at room temperature. The single phase mixture was then incubated overnight at 48°C. After cooling, 75 l of 1 M KOH in CH 3 OH were added to the mixture, briefly sonicated, and then incubated in a shaking water bath at 37°C for 2 h to cleave potentially interfering glycerolipids. The extract was then brought to a neutral pH with 6 l of glacial acetic acid, centrifuged, and the supernatant was collected and transferred into a new tube, where a SpeedVac was used to dry the extract. The dried extract was reconstituted in 0.5 ml of starting mobile phase solvent for LC/MS, sonicated for 15 s, and centrifuged. The supernatant was collected and transferred to the autoinjector vial for analysis.
Complex sphingolipids, sphingoid bases, and sphingoid base 1-phosphates were separated by reverse-phase LC with a Supelco 2.1 (internal diameter) × 50 mm Ascentis Express C18 column (Sigma-Aldrich) and a binary solvent system at a flow rate of 0.5 ml/min with the column oven set at 35°C. The column was equilibrated with a solvent mixture of 95% mobile phase A1 (CH 3 OH/ water/HCOOH, 58:41:1, v/v/v, with 5 mM ammonium formate) and 5% mobile phase B1 (CH 3 OH/HCOOH, 99:1, v/v, with 5 mM ammonium formate) for 0.5 min, and after a typical sample injection of 40 l, the A1:B1 ratio was maintained at 95:5 for 2.25 min, followed by a linear gradient to 100% B1 over 1.5 min, which was held at 100% B1 for 5.5 min, followed by a 0.5 min gradient return to 95:5 A1:B1. Re-equilibration of the column with 95:5 A1/B1 for 0.5 min was done before the next run.
Confirmation and quantification of individual species of sphingolipids was performed using a Shimadzu LC-20 AD binary pump system coupled to a SIL-20AC autoinjector and a DGU20A3 degasser coupled to an ABI 4000 quadrupole/linear ion trap (QTrap) (Applied Biosystems) operating in triple quadrupole mode. Sphingolipid species, Cer, Hex-Cer, lactosyl-Cer (Lac-Cer), SM, and sphingoid lipids, such as Sph, dihydro-Sph (dh-Sph), S1P, and dihydro-S1P, were identified based on their retention time and m/z ratio and quantified, as described previously (37,38), by comparing the target lipid ion of interest with the normalization of quantitated ion abundances. Quantitative lipid analyses for each sample were performed after normalizing against their levels of SM, which was selected due to its consistency as a structural sphingolipid.

RT-PCR
Total RNA extraction was carried out on harvested cells using the Ambion RNA mini extraction kit (Ambion TRIzol® Plus RNA purification kit; Life Technologies, Carlsbad, CA). The cDNA synthesis was followed by SuperScript™ III First-Strand Synthesis SuperMix (Invitrogen, Carlsbad, CA), according to the manufacturer's protocol. Primers for quantitative RT-PCR were designed in such a way that they spanned at least one intron, eliminating the chance of amplification from residual DNA contamination. The sequence of the primers is available from the corresponding author upon request. Expression of ASAH1, other sphingolipid metabolic genes, and oxidative and apoptotic marker genes were assayed by SYBR Green methods, and quantitation was calculated by the comparative Ct (threshold cycle) value method, as previously described (11,(39)(40)(41)(42)(43).

Data analysis and statistics
Data analysis and statistics for the sample sets were done by Student's t-test and two-way ANOVA, as appropriate, using GraphPad Prism 7 software (GraphPad Software Inc., La Jolla, CA). Statistical significance was considered P < 0.05.

Generation of stably transfected ARPE19-ASAH1 cells
We generated an ARPE19 cell line stably transfected with the recombinant human ASAH1 gene, henceforth referred to as ARPE19-ASAH1. Expression of the transgenic ASAH1 was confirmed by immunocytochemical assays (Fig. 1A-D). The transgenic ASAH1 protein colocalized with lysosomal marker in punctate structures, which is characteristic of lysosomal localization, confirming its cellular location ( Fig. 1A-D). This localization pattern of ASAH1 protein is consistent with previous findings (44). Stably transfected ARPE19-ASAH1 cells were selected based on Venus expression, propagated, and used in further experiments. The transgene expression of ASAH1 was confirmed via fluorescence microscopy and RT-PCR and a significant increase of ASAH1 mRNA was seen in ARPE19-ASAH1 cells compared with ARPE19 cells ( Fig. 1E; P < 0.001). Endogenous ASAH1 protein expression was noted in both transfected and untransfected ARPE19 cells (Fig.  1F). ASAH1 protein expression was more than doubled in ARPE19-ASAH1 cells relative to ARPE19 cells ( Fig. 1G; P < 0.001).

ASAH1 activity tested by its effect on C2-Cer
In order to determine the activity and function of transgenic ASAH1 in ARPE19 cells, we treated both ARPE19-ASAH1 and ARPE19 control cells with exogenous short-chain C2-Cer. C2-Cer is known to cause cell death and our data showed a trend of decreasing cell viability with an increasing concentration of C2-Cer ( Fig. 2A). A significant decrease Fig. 1. Generation of ASAH1-expressed ARPE19 cells. Immunocytochemistry of ASAH1-expressed cells. ASAH1 expression vector was electroporated into ARPE19 cells, and ASAH1-Venus-expressing cells were selected after sorting to develop stable pure culture. A: Cells were stained with a lysosomal marker, LysoTracker Red. B: Venus (green) expression was observed in ASAH1 cells. C: Venus-ASAH1 expression was majorly colocalized with the lysosomal marker. D: Enlarged image of ASAH1 expression with LysoTracker Red; ASAH1 shows cytoplasmic expression and ultimately moving to the punctate lysosomes. Scale bars: 20 m (A-C) and 10 m (D). E: Transgene expression was confirmed by RT-PCR, and a 10-fold increase in ASAH1 was seen compared with ARPE19. F: Protein expression indicates endogenous ASAH1 expression in both transfected and untransfected ARPE19 cells. G: ASAH1 protein levels were normalized against -actin, showing a >2-fold increase in ASAH1 expressionQ 21. ###P < 0.001 between cell lines.Q 31 in cell viability and an increase in cell death was noted beginning at a concentration of 10 M C2-Cer (P < 0.001; Fig. 2A, B). ARPE19-ASAH1 cells exhibited an apparent resistance to C2-Cer treatments at every dose. ARPE19-ASAH1 viability was significantly higher than ARPE19 at all concentrations of C2-Cer (P < 0.001; Fig. 2B). This result suggests that overexpression of ASAH1 can actively degrade C2-Cer and protect the cells from its lethal effect.

Characterization of ARPE19-ASAH1 cells
ASAH1 is the major Cer-hydrolyzing enzyme in cells; it hydrolyzes Cer to Sph, which serves as the major source of cellular S1P. Overexpression of ASAH1 is expected to shift the cellular Cer/S1P balance toward increasing S1P, which is a well-established factor for cell growth and proliferation. Therefore, we analyzed whether the growth and proliferation of the ARPE19-ASAH1 cells are different from the native ARPE19 cells. We seeded an equal number of both ARPE19 and ARPE19-ASAH1 cells, counted the cell numbers after 24 and 48 h of culture, and found that the numbers of ARPE19-ASAH1 are 35-40% higher than ARPE19 cells (Fig. 3A). To evaluate the sphingolipid profiles in these cells, we harvested the cells after 48 h and assessed sphingolipid levels (picomoles per 10 6 cells; Fig. 3B, C). In the ARPE19-ASAH1 cells, there was no significant increase in Cer 48 h after plating. However, a significant decrease was observed in Hex-Cer (P < 0.001) and SM (P < 0.01) at this time point (Fig. 3B), indicating that the synthesis of the Hex-Cer and SM was not favored in order to maintain the cellular pool of total Cer. Another reason for this difference may have been the catabolism of Hex-Cer and SM to maintain optimum cellular levels of Cer.
When we analyzed changes in the major intracellular lyso-sphingolipid species 48 h after plating, we found a significant increase in Sph (>2-fold; P < 0.001) and dh-Sph (P < 0.001) in ARPE19-ASAH1 cells compared with ARPE19 cells (Fig. 3C). This increase likely resulted from the breakdown of Cer to Sph in the lysosome by the overexpressed ASAH1 enzyme, and it suggests an active and functionally higher level of ASAH1 in ARPE19-ASAH1 cells. However, we did not see any changes in the S1P levels in these cells (Fig. 3C). In order to determine whether S1P was being secreted from the cells into the media, we harvested and analyzed the serum-free culture media of these cells and found moderate decreases in Cer, Hex-Cer, and S1P levels, as well as an increase in Sph levels (supplemental Fig. S2; P < 0.05). These data suggest that overexpression of ASAH1 increases Sph levels and reduces both Hex-Cer and SM levels. However, we did not note changes in S1P levels, which we believe to be due to higher levels of S1P present in the serum-containing media in which the cells were cultured.

ASAH1 overexpression protects ARPE19 cells from oxidative stress
In subsequent experiments, ARPE19-ASAH1 and wildtype ARPE19 cells were exposed to H 2 O 2 for 24 h to induce oxidative stress in serum-free media. After incubation, cell viability in ARPE19 cells was assessed using a cell proliferation assay. ARPE19 cells were resistant to low levels of oxidative stress, but viability decreased significantly starting at 800 M H 2 O 2 (Fig. 4A). Next, cell viability between ARPE19 and stably transfected ARPE19-ASAH1 cells was compared (Fig. 4B). Cells were treated for 24 h with either 800, 900, or 1,000 M of H 2 O 2 . In the ARPE19 cells, there was a significant increase in cell death with increasing H 2 O 2 concentration (800 M, P < 0.001; 900 M, P < 0.001; 1,000 M, P < 0.001; Fig. 4B). Likewise, the ARPE19-ASAH1 cells treated with H 2 O 2 showed a dose-dependent increase in cell death (800 M, P < 0.001; 900 M, P < 0.001; 1,000 M, P < 0.001). However, cell viability was significantly greater in ARPE19-ASAH1 cells compared with the control ARPE19 cells at each dose of H 2 O 2 ( Fig. 4B; 800 M, P < 0.001; 900 M, P < 0.001; 1,000 M, P < 0.001). Viability was reassessed using a LDH assay, which returned similar results. Specifically, ARPE19-ASAH1 cytotoxicity was significantly lower at all concentrations below 800 M of H 2 O 2 (P < 0.05; supplemental Fig. S3). These results clearly indicate that ASAH1 overexpression has a protective effect in ARPE19 cells from oxidative stress.

Analysis of cellular sphingolipids
We evaluated the absolute and relative composition of the major cellular sphingolipids (e.g., Cer, Hex-Cer, SM, and their subspecies), as well as cellular lyso-sphingolipids (e.g., Sph, dh-Sph, and S1P), in ARPE19 and ARPE19-ASAH1 cells with and without H 2 O 2 treatment. As mentioned earlier, cellular bioactive sphingolipids serve as signaling mediators. Hence, we analyzed the cells after 3 h of H 2 O 2 treatment, when cell death was not yet evident.
Quantitative analysis of major sphingolipids. To understand how H 2 O 2 -induced oxidative stress affects sphingolipids in RPE cells, we measured total Cer, Hex-Cer, Lac-Cer, SM, and lyso-sphingolipid levels after 3 h of oxidative stress, before cell death became apparent. We normalized all sphingolipid data using SM, based on the fact that SM levels normalized to milligrams of protein remained unchanged between the two cell lines and with H 2 O 2 treatment for 3 h (Fig. 5A). We found a significant increase in total Cer after H 2 O 2 treatment in both ARPE19 (P < 0.001) and  ARPE19-ASAH1 cells (P < 0.01; Fig. 5B). However, overexpression of ASAH1 significantly decreased the total Cer in ARPE19-ASAH1 cells compared with ARPE19 cells in untreated cells (P < 0.05) and in oxidatively stressed cells (P < 0.001, Fig. 5B). Both protein and SM normalization resulted in similar patterns of changes in all other lipids, and because SM was an internal control for each sample, SM was used for normalizing all the lipids henceforth. We did not see any significant differences in Lac-Cer levels between the ARPE19 cells and the ARPE19-ASAH1 cells with or without oxidative stress (Fig. 5C). However, we found that, in untreated cells, total Hex-Cer levels were significantly decreased in ARPE19-ASAH1 versus ARPE19 cells (P < 0.001; Fig. 5D). Hex-Cer was increased after H 2 O 2 treatment in both cell types, but was significantly less elevated in ARPE19-ASAH1 cells (P < 0.001; Fig. 5D). In summary, overexpression of ASAH1 appears to significantly reduce the levels of both Cer and Hex-Cer in untreated ARPE19 cells, while H 2 O 2 treatment increased the levels of these sphingolipids, supporting our hypothesis that cellular Cers increase during oxidative stress and serve as second messengers to induce cell death, as reduction of these Cers reduces cell death.

Quantification of bioactive lyso-sphingolipid species in ARPE19
and ARPE19-ASAH1 cells. Among the lyso-sphingolipids, we found Sph levels to be significantly higher in Vehtreated ARPE19-ASAH1 compared with ARPE19 cells (P < 0.001; Fig. 6). There was no difference in the Sph levels between the two cell lines after oxidative stress, but H 2 O 2 treatment significantly decreased Sph to roughly the same levels in both cell types (P < 0.001; Fig. 6). We detected higher levels of dh-Sph in ARPE19-ASAH1 compared with ARPE19 in Veh-treated cells (P < 0.01; Fig. 6). Following oxidative stress, there was no significant difference between the ARPE19 and ARPE19-ASAH1 cell lines, but there was a significant increase in dh-Sph levels in the ARPE19 cells (P < 0.05; Fig. 6). We then measured the S1P levels and found that H 2 O 2 treatment significantly increased S1P levels in both cell types (P < 0.001; Fig. 6); however, S1P levels were significantly reduced in ARPE19-ASAH1 cells compared with ARPE19 cells for both treatments (P < 0.001 and P < 0.01, respectively; Fig. 6). The lyso-sphingolipid, Sph, can induce cell senescence and S1P signals for cell survival. Finding major changes in cellular S1P levels following oxidative stress suggests its association with cell death and survival.
In summary, ASAH1 overexpression resulted in significant changes in the relative composition of Cer and Hex-Cer species in ARPE19 cells, both of which have been found to be altered due to exposure to oxidative stress. However, this effect was less pronounced among the Hex-Cer species.

Expression of sphingolipid, apoptotic, and inflammatory genes
We utilized RT-PCR to analyze sphingolipid, inflammatory, and apoptotic genes in ARPE19 and ARPE19-ASAH1 cells. As shown in Fig. 1E, expression of ASAH1 in the stably transfected cells was significantly higher than in the wildtype ARPE19 cells (P < 0.001). A 2-fold increase in expression of ASAH1 protein in ARPE19 cells transfected with ASAH1 was confirmed through normalization with -actin (P < 0.001; Fig. 1G). Treatment with H 2 O 2 did not make a significant difference in the expression of ASAH1 in these cells (supplemental Fig. S6). We further tested the expression of other sphingolipid metabolic genes. After 3 h of H 2 O 2 treatment, there was a significant increase in serine palmitoyl transferase 2 (SPT2; P < 0.05) in ARPE19-ASAH1 cells compared with ARPE19 cells (Fig. 7); yet, when comparing to ARPE19-ASAH1 controls, oxidative stress resulted in a significant decrease in SPT2 (P < 0.05; Fig. 7).

DISCUSSION
In the present study, we investigated how overexpression of lysosomal ceramidase (Ly-ASAH1) affects retinal cell death in vitro via oxidative stress. Oxidative stress is implicated in the etiology and pathogenesis of many neurodegenerative disorders, including the RD that occurs in retinitis pigmentosa, diabetic retinopathy, and AMD (45)(46)(47)(48)(49). In   RT-PCR analysis of sphingolipid metabolic genes with oxidative stress. Each sample included in a condition that consisted of three independent RT-PCR experiments that were averaged and normalized against housekeeping genes. Bar graphs are represented as mean ± SEM. Change in sphingolipid genes as a result of overexpressed ASAH1 and stress condition. Significant difference between cell types depicted: #P < 0.05; n = 4. SMPD1, SM phosphodiesterase 1; SGPL1, S1P lyase 1; Cers2, Cer synthase 2.
AMD, for example, UV sunlight exposure and smoking generate reactive oxygen species in the RPE cells, which causes oxidative stress and is considered to be the major pathological cause of the disease (45,46). However, the mechanism by which reactive oxygen species induce RPE cell death is not fully understood. It is well-documented that oxidative stress (e.g., light, chemical, and physical) can induce excess Cer biosynthesis, and we, along with other laboratories, recently established that Cer accumulation causes retinal cell death (11,12,29,50). In this study, we found that ARPE19 cells overexpressing lysosomal ceramidase (ARPE19-ASAH1) were protected from H 2 O 2 -induced oxidative stress by reducing the levels of bioactive Cer and Hex-Cer and S1P levels. The S1P increase due to oxidative stress (Fig. 6) could very well be a consequence of higher levels of Cer that counteract the Cer action toward cell death. Cer inhibits cell growth and promotes apoptosis and senescence, whereas S1P induces cell proliferation and migration by signaling through one of five S1P receptors (51)(52)(53). This delicate balance, or rheostat, between these two molecules ultimately determines whether the cell undergoes apoptosis or proliferation (54). Consequently, targeting pathways that decrease Cer accumulation could be a promising therapeutic approach for treating or preventing RD, as we demonstrated in earlier studies (12,29). This is the first report that uses a genetic means to reduce excess cellular Cer by stably overexpressing the ceramidase enzyme in mammalian cells, ultimately preventing oxidative stressinduced cell death.
In this study, we cloned the human ASAH1 gene from ARPE19 cells, fused it with a fluorescent marker (Venus), and transfected it into the same cells, thereby generating stable ARPE19 cells that overexpressed the ASAH1 gene (ARPE19-ASAH1; Fig. 1). We verified the expression of the transgene, the proper localization of the protein (lysosomes), and the function of the transgenic protein by lipid product analysis and its ability to protect cells from exogenous Cer-induced cell death (Figs. 1, 2). ASAH1 overexpression resulted in significantly higher levels of Sph in ARPE19 cells, and it reduced the levels of higher-order Cer anabolic products, such as Hex-Cers and SM (Fig. 3B, C). When we introduced oxidative stress to these cells, we observed significant protection of ARPE19-ASAH1 cells from H 2 O 2 compared with the ARPE19 cells (Fig. 4B). In ARPE19 cells, we further demonstrated that oxidative stress via H 2 O 2 induces an increase in Cer and Hex-Cer species, which is significantly reduced by overexpression of ASAH1, thereby protecting the cells (Fig. 5).
The mammalian ASAH1 is a lysosomal enzyme, which is active in the acidic environment of the lysosome. The acidic subcellular compartments, the late endosomes, and the lysosomes are the sites for constitutive degradation of sphingolipids and glycosphingolipids, with the end goal being Sph production. We detected Venus-fused ASAH1 expression in cytoplasm and late endosomes (not colocalized with LysoTracker) and in lysosomes (Fig. 1B-D). In the case of glycosphingolipids, the stepwise release of monosaccharide units by exohydrolases occurs at an acidic pH, leaving just the Sph portion of the molecule (i.e., Cer). Cer can be further hydrolyzed by acid ceramidase (ASAH1) to form Sph and a free fatty acid, both of which are able to Fig. 8. RT-PCR analysis of inflammatory and apoptotic genes with oxidative stress. Each sample included in a condition that consisted of three independent RT-PCR experiments that were averaged and normalized against housekeeping genes. Bar graphs are represented as fold over control mean ± SEM from n = 6 biological replications. Change in inflammatory and apoptotic markers as a result of cell type and oxidative stress. Significant difference between treatment groups: *P < 0.05, **P < 0.01. Significant difference between cell types depicted: #P < 0.05. FOSL1, FOS-like 1; BAX, BCL2-associated X protein; FAS, Fas cell surface death receptor; IL6, interleukin 6. leave the lysosome, unlike Cer. The salvage pathway reutilizes long-chain sphingoid bases to form Cer through the action of CerS at the surface of the ER or in ER-associated membranes. The salvage pathway has been estimated to contribute between 50% and 90% of sphingolipid biosynthesis, suggesting major associations between lysosomal degradation and recycling of sphingolipids in cells (60). ASAH1, being the ultimate enzyme in the lysosomal degradation pathway, plays an important role in mammalian cells. A mutation in the ASAH1 gene causes Farber disease in humans (Farber's lipogranulomatosis), which is an inborn lipid storage disease characterized by Cer accumulation in tissue that results in early death (61). ASAH1 plays an important role in balancing the levels of cellular Cer and S1P, and it was expected that an increase of ASAH1 activity would facilitate cell survival, while a decrease in ASAH1 activity would result in cell apoptosis. ASAH1 deficiency in the mouse retina causes early cell death (N. Mandal, unpublished observations), whereas overexpression of ASAH1 is associated with various human cancers, suggesting a role in cell proliferation (62). For the first time, we report in this study that overexpressing ASAH1 decreases levels of Hex-Cer and increases levels of Sph, which possibly maintains the steady-state levels of Cer that are needed to maintain healthy cell growth.
This study also underscores the involvement of the Cer lipids downstream of H 2 O 2 signaling. H 2 O 2 inhibits cell growth and proliferation via the PI3K/AKT/mTOR and Erk signaling pathways. AKT/protein kinase B, a serine/ threonine kinase, is a versatile pro-proliferative signaling protein activated by phosphorylation at Ser 473 by phosphoinositide-dependent kinase 1 (PDK1). Akt activation, in turn, signals to a variety of key downstream targets, including GSK3 and mTOR, which consequently inhibits apoptosis and promotes cell survival in a regulated manner (63). Cer, a multifunctional central molecule in the sphingolipid biosynthetic pathway, exerts potent apoptotic effects in a variety of cell types (64)(65)(66)(67)(68)(69)(70), and it is one of the potent mediators of RD. Cer biosynthesis mainly occurs via: 1) the de novo pathway, 2) SMase-mediated hydrolysis of SM, and/or 3) the inhibition of ceramidase or glucosylceramide synthase-mediated conversion of Cer into Sph and glucosylceramide into Cer, respectively. Cer-activated protein phosphatases, such as protein phosphatase 1 and protein phosphatase 2A, were shown to dephosphorylate Akt at Ser 473 , which ultimately led to the inactivation of AKT and the induction of apoptosis (71)(72)(73)(74)(75)(76).
Significant differences in gene expression between the ARPE19 and ARPE19-ASAH1 cells under oxidative stress serve to further demonstrate the importance of Cer metabolism in cell survival. In ARPE19-ASAH1 cells under oxidative stress, PARP expression is significantly reduced (Fig. 8). This is notable, as PARP is involved in DNA excision repair, and elevations in PARP activation have been reported in apoptosis resulting from cerebral ischemia, inflammation, and oxidative stress injuries. PARP is also implicated in Cer-induced apoptosis (77). cFOS, a member of the AP-1 transcription factor complex associated with apoptotic cell death (78), was also significantly reduced in ASAH1-overexpressing cells. Bcl2 family proteins involved with pro-and anti-apoptotic pathways (79) are also decreased in ARPE19-ASAH1 cells compared with controls (Fig. 8). Inflammatory gene expression is also substantially influenced by oxidative stress and the activity of ASAH1, as evidenced by significant differences between ARPE19 and ARPE19-ASAH1 expression of CCL2, ICAM, HO-1, and TNFa (Fig. 8), which are all involved in inflammation and/or apoptotic pathways (80)(81)(82)(83). Clearly, ASAH1 overexpression and the resultant disruption of sphingolipid metabolic processes has wide-ranging effects that include increased cell survival in oxidative stress through modulation of apoptotic and inflammatory pathways, though the mechanisms of these actions and Cer's exact role in them are yet to be fully established. Maintaining steady-state levels of Cer is essential for cell survival and growth, and the salvage pathway plays an important role in this balance. We found higher levels of Sph in ASAH1-ARPE19 cells, which suggests activation of the salvage pathway to maintain the cellular Cer level, as seen in our results (Figs. 3, 6). The salvage pathway, which generates Cer from Sph, is prosurvival, whereas ER stress is increased by an overly activated de novo pathway (84), which might have played a major role in protecting ARPE19-ASAH1 cells from oxidative stress.
In conclusion, in our in vitro study using human immortal RPE cell lines, we have established that bio-active Cer plays a role in oxidative stress-induced cell death, which could ultimately be targeted to develop novel therapeutic treatments for human AMD and other degenerative diseases. These experiments suggest that overexpression of the ASAH1 gene can help to protect RPE cells from cell death by metabolizing excess Cer (Fig. 9), which helps to justify further investigation into using or targeting Cermodifying enzymes to protect retinal cells. The transgenic ASAH1 overexpression strategy has been shown to be effective to prevent retinal photoreceptor degeneration in Drosophila (85), and we expect that this strategy could result in similar findings in higher organisms. Our future studies focusing on ASAH1 gene therapy in the mammalian system will generate further valuable information on how this intracellular target (Cer) is associated with human retinal diseases, and how we can ultimately develop novel therapeutic strategies.