De novo ceramide accumulation due to inhibition of its conversion to complex sphingolipids in apoptotic photosensitized cells.

The oxidative stress induced by photodynamic therapy (PDT) with the photosensitizer phthalocyanine 4 is accompanied by increases in ceramide mass. To assess the regulation of de novo sphingolipid metabolism during PDT-induced apoptosis, Jurkat human T lymphoma and Chinese hamster ovary cells were labeled with [14C]serine, a substrate of serine palmitoyltransferase (SPT), the enzyme catalyzing the initial step in the sphingolipid biosynthesis. A substantial elevation in [14C]ceramide with a concomitant decrease in [14C]sphingomyelin was detected. The labeling of [14C]ceramide was completely abrogated by the SPT inhibitor ISP-1. In addition, ISP-1 partly suppressed PDT-induced apoptosis. Pulse-chase experiments showed that the contribution of sphingomyelin degradation to PDT-initiated increase in de novo ceramide was absent or minor. PDT had no effect on either mRNA amounts of the SPT subunits LCB1 and LCB2, LCB1 protein expression, or SPT activity in Jurkat cells. Moreover in Chinese hamster ovary cells LCB1 protein underwent substantial photodestruction, and SPT activity was profoundly inhibited after treatment. We next examined whether PDT affects conversion of ceramide to complex sphingolipids. Sphingomyelin synthase, as well as glucosylceramide synthase, was inactivated by PDT in both cell lines in a dose-dependent manner. These results are the first to show that in the absence of SPT up-regulation PDT induces accumulation of de novo ceramide by inhibiting its conversion to complex sphingolipids.

Sphingolipids are widely present in eukaryotic cells. Their biosynthesis begins with the condensation of L-serine with palmitoyl-CoA to give rise to 3-ketosphinganine. The reaction is catalyzed by serine palmitoyltransferase (SPT) 1 (1). At least two genes, LCB1 and LCB2, encoding two respective SPT subunits, are essential for expression of mammalian SPT activity (1). 3-Ketosphinganine is reduced to form sphinganine (dihydrosphingosine). In mammalian cells sphinganine is acylated to produce dihydroceramide, which is desaturated to give rise to ceramide (2). In the reactions catalyzed by phosphatidylcholine:ceramide phosphocholine transferase (sphingomyelin synthase) and UDP-glucose:ceramide glucosyltransferase (glucosylceramide synthase) ceramide is converted to sphingomyelin and glucosylceramide, respectively (3). Glucosylceramide is a building block for more complex glycosphingolipids (4). Although ceramide is the predominant hydrophobic backbone for numerous complex sphingolipids, some sphingolipids are derived from dihydroceramide and phytoceramide (3).
The aim of the present study was to address de novo ceramide production in apoptosis after the novel oxidative stress inducer photodynamic therapy (PDT). In PDT, after the uptake of a photosensitive dye, the dye is activated by red light with subsequent formation of reactive oxygen species to induce cell death (14). Apoptosis is an important mechanism in tumor ablation by PDT (15). We have demonstrated in a variety of cell types that apoptosis induced by photosensitization with the silicon phthalocyanine 4 (Pc 4) is accompanied by increases in ceramide mass (16 -22), as well as sphinganine (23). We have also shown that de novo sphingolipids are involved in initiation of mitochondrial as well as extramitochondrial apoptosis after Pc 4-PDT (23,24).
Here we provide the first evidence from Pc 4-photosentized Jurkat and CHO cells that (i) in the absence of SPT up-regulation accumulation of de novo ceramide is a result of inhibition of its conversion to complex sphingolipids and that (ii) de novo ceramide may play an important role in PDT-induced apoptosis.
from Invitrogen and Hyclone, respectively. Mouse monoclonal anti-LCB1 and anti-KDEL (10C3) antibodies were from BD Biosciences and Stressgen, respectively. Anti-rabbit IgG (Bio-Rad) and anti-mouse IgG (Amersham Biosciences) horseradish peroxidase-labeled secondary antibodies were used. Thin layer chromatography (TLC) plates (aluminum sheets of silica gel 60) were from EM Science.
Cell Culture and Treatments-Jurkat human lymphoma T cells were purchased from American Type Culture Collection. Jurkat cells were cultured in RPMI 1640 medium (supplemented with 10% fetal bovine serum). CHO cells were grown in Ham's F-12 medium (supplemented with 10% newborn calf serum). All cell cultures were grown in medium supplemented with 100 units/ml penicillin and 100 g/ml streptomycin and were maintained at 37°C in a 5% CO 2 atmosphere. For experiments, an aliquot of a stock solution of Pc 4 (0.5 mM in dimethyl formamide) was added to the cells in the culture medium to give a final concentration of 200 and 500 nM for Jurkat and CHO cells, respectively. After overnight incubation, the cells were irradiated with red light (2 milliwatts/cm 2 ; max ϳ670 nm) using a light-emitting diode array light source (EFOS) at various fluences at room temperature. Following PDT, cells were incubated at 37°C for desired periods of time. CHO cells were harvested using trypsin (0.25%) and gentle scraping in modified Hanks' solution (supplemented with 1.2 mM EDTA and 25 mM HEPES, pH 7.4).
Metabolic Radiolabeling of Cellular Lipids-In pulse experiments, following PDT cells in the culture medium were labeled with [ 14 C]serine (28 kBq) for 2 h. In some experiments, ISP-1 (1 M) was added 1 h prior to PDT. In pulse-chase experiments Pc 4-treated cells were labeled with [ 14 C]serine (28 kBq) for 2 h, incubated with 10 mM L-serine for 30 min, irradiated, and then incubated for the desired periods of time. Following post-treatment incubations, cells were harvested by centrifugation at 4°C. Total extracted cellular lipids (25) were separated by TLC (methyl acetate, n-propyl alcohol, chloroform, methanol, 0.25% potassium chloride; 25:25:25:10:9, v/v). After chromatography, the TLC plates were exposed to PhosphorImager screens (Amersham Biosciences) for 48 h. The samples were analyzed by a STORM 860 (Amersham Biosciences) imaging system. Data were normalized per milligram of protein. Individual lipid bands were visualized by staining with Coomassie Brilliant Blue (26) and identified by comparison to a concomitantly run standard curve comprised of known amounts of lipids. The R F values for ceramide, glucosylceramide, and sphingomyelin were 0.83, 0.65, and 0.04, respectively. The R F values for phosphatidylethanolamine and phosphatidylserine were 0.28 and 0.14, respectively.
TLC Detection of Intracellular [ 14 C]Serine-Untreated or PDTtreated cells in the culture medium were labeled with [ 14 C]serine (28 kBq) for desired periods of time. After incubations, cells were harvested rapidly on ice and washed three times with ice-cold PBS to remove exogenous [ 14 C]serine. After extraction (25), aqueous upper phases were dried down and separated by TLC (0.6% NaCl, methanol, 28% ammonia; 10:10:1, v/v) as described elsewhere (27). After chromatography, the TLC plates were exposed to PhosphorImager screens (Amersham Biosciences) for 24 h. The samples were analyzed by the STORM 860 (Amersham Biosciences) imaging system. Serine and phosphoserine bands were identified by comparison to concomitantly run respective standards, which were visualized by staining with 0.2% ninhydrin in acetone and heating to 100°C. 2 RNA Isolation-Following PDT treatment, Jurkat cells (5 ϫ 10 6 ) were harvested and washed in PBS. Total cellular RNA was isolated from cell lysates using an RNeasy minikit (Qiagen) according to the manufacturer's instructions. Following RNA purification, DNase treatment (Ambion) was performed following the manufacturer's directions to ensure that there was no contaminating genomic DNA.
Reverse Transcription-Total RNA was reverse transcribed into cDNA using Superscript III (Invitrogen) according to the manufacturer's instructions. Briefly 2 g of RNA, 10 mM dNTP mixture, and 0.5 g of oligo(dT) [12][13][14][15][16][17][18] were mixed in a total volume of 13 l and incubated at 65°C for 5 min. Then the following components were added together on ice: 4 l of 5ϫ first strand buffer, 1 l of 0.1 M dithiothreitol, 1 l of the RNase inhibitor RNaseOUT (40 units/l), and 1 l of Superscript III (200 units/ml). The reaction was incubated at 50°C for 45 min and at 70°C for 15 min. Following addition of 1 l of RNase H (2 units/l, Invitrogen) the samples were incubated at 37°C for 20 min. RNase-free H 2 0 (80 l) was then added, and samples were frozen (Ϫ20°C) until use.
Quantitative Reverse Transcription-PCR-The quantitation of gene expression of the SPT subunits hLCB1 and hLCB2 was performed as described elsewhere (28). The following primers were designed using the Primer Express software (ABI) and synthesized by Sigma-Genosys: human SPT subunit 1 (hLCB1), 5Ј-CAGCTTCGTTACCTCCCCTG-3Ј, 5Ј-AGAAGTGTCTCCCTCCTCCCAG-3Ј, and 5Ј-TATCAGTGCCAACA-TGGAGAATG-3Ј; and human SPT subunit 2 (hLCB2), 5Ј-TTGGACG-GGAGATGCTGAA-3Ј, 5Ј-AGCCAGACTGTCAGGAGCAAC-3Ј, and 5Ј-AGCTGAAGTATTCCCGTCATCG-3Ј. The GenBank TM accession numbers for these genes are as follows: hLCB1, Y08685; and hLCB2, Y08686. The ABI 5700 sequence detection system was used for the real time PCR. All methods for reactions and quantitation were performed as recommended by the manufacturer. Primers were tested against control cDNA (from a normal fibroblast cell line) using the following ABI 5700 PCR parameters: 2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C, 1 min at 60°C with a final hold at 4°C. PCR was performed with the ABI Sybr Green PCR Core reagents (29). The PCR components (25 l) were as follows: 0.5 l of cDNA template, 2.5 l of 10ϫ buffer, 3.0 l of 10 mM MgCl 2 , 2.0 l of 10 mM dNTP mixture, 0.125 l of Amplitaq Gold, 0.25 l of AmpErase uracil N-glycosylase, 16.13 l of H 2 0, 0.25 l of each primer (at 50 nM final concentration for each primer). Real time PCR was performed in 96-well optical plates (ABI). All samples were done in triplicate. The calibrator primers used for each sample were designed to amplify glyceraldehyde-3-phosphate dehydrogenase. A master mixture was made containing all reagents (described above) except primers and template. The mixture was aliquoted into microcentrifuge tubes (four reactions/tube), and then the appropriate template and primers were added. This was then aliquoted in triplicate into the 96-well optical plate. Wells were also allocated for no template controls. PCR was performed in the ABI 5700 with the cycling parameters described above. Data were analyzed on a Microsoft Windows NT work station with the ABI 5700 software. The ⌬C T values were then exported into Microsoft Excel for quantitation. ⌬C T values were calculated by subtracting the average glyceraldehyde-3-phosphate dehydrogenase ⌬C T from the average experimental ⌬C T values for each cDNA template. ⌬⌬C T values were calculated by subtraction of ⌬C T values of a control cDNA template from the experimental cDNA templates for each set of primers. The -fold change of expression of a mRNA was calculated by the following formula: 2 Ϫ⌬⌬CT .
SPT Activity Assay-The enzyme activity was assayed as described previously (34). Enzyme activity in 100 g of microsomal membranes was measured in 50 mM HEPES (pH 7.5), 5 mM dithiothreitol, 5 mM EDTA, and 50 M pyridoxal 5Ј-phosphate. The reaction was initiated by the addition of 200 M palmitoyl-CoA and 740 kBq of L-[ 3 H]serine (1 mM, final concentration). A control containing all of the components except palmitoyl-CoA was included as well. Following incubation for 10 min at 37°C, the reaction was terminated with 0.2 ml of 0.5 N NH 4 OH. The 3 H-labeled lipid product 3-ketosphinganine was extracted, and the radioactivity was measured by scintillation counting. SPT activity was expressed as pmol of 3-ketosphinganine generated/10 min/mg of protein after subtracting the background radioactivity (i.e. the minus palmitoyl-CoA control).
Sphingomyelin and Glucosylceramide Synthase Activity Assays-The sphingomyelin synthase activity assay using C 6 -NBD-ceramide was performed as described previously (35). Enzyme activity in 50 -100 g of microsomal membranes was measured in 50 mM HEPES (pH 7.5), 5 mM EDTA, and 10 M C 6 -NBD-ceramide complexed with fatty acid-free bovine serum albumin (0.1 mM). To determine glucosylceramide synthase activity, 500 M UDP-glucose was included in the assay mixture (35). Following incubation for 5 min at 37°C, the reaction was terminated with chloroform/methanol (1:2, v/v). C 6 -NBD-ceramide-labeled lipid products were extracted and separated by TLC using chloroform/ methanol/water (65:25:4, v/v), and their fluorescence was detected and quantified by the STORM 860 imaging system. Sphingomyelin synthase and glucosylceramide synthase activities were expressed as pmol of sphingomyelin and glucosylceramide, respectively, generated/5 min/mg of protein after subtracting the background fluorescence.
Hoechst Staining-Nuclear apoptotic changes were determined by staining with the DNA-binding Hoechst 33342 dye as described previously (37). After a 30-min fixation at room temperature with formalin, the cell pellets were washed with PBS and stained with Hoechst dye (24 g/ml, overnight at 4°C), and 200 -300 cells were counted for the incidence of apoptotic chromatin condensation under a fluorescence microscope (Zeiss).
Statistical Analysis-Results were expressed as mean Ϯ S.E. Statistical analyses were performed by Student's t test. Significance was defined as a two-tailed p Ͻ 0.05.

PDT-induced Increased Labeling of Ceramide with [ 14 C]Serine
Is ISP-1-sensitive in Jurkat Cells-To assess the effect of PDT on the de novo sphingolipid biosynthesis in cells, [ 14 C]serine was used for labeling of cellular lipids. A dose-dependent increase in [ 14 C]ceramide was observed at 2 h after PDT (Fig. 1, A and C). Specifically exposure of Pc 4 (200 nM)-treated Jurkat cells to light fluences of 135, 270, or 400 mJ/cm 2 increased [ 14 C]ceramide labeling by 4.4-, 5.2-, and 11.0fold, respectively. The effect was abolished when cells were co-exposed to PDT with ISP-1, a potent inhibitor of SPT (38) (Fig. 1, B and C). ISP-1 was used in these and all subsequent experiments at the nontoxic concentration of 1 M, which effectively suppresses apoptosis in Jurkat and other cell lines (23). ISP-1 alone also inhibited incorporation of [ 14 C]serine into ceramide at rest (Fig. 1, B and C). Thus, PDT induced increased, ISP-1-sensitive, 14 C labeling of ceramide.
PDT and Degradation of de Novo Complex Sphingolipids in Jurkat Cells-To test whether PDT-induced increases in the amount of [ 14 C]ceramide in the pulse labeling experiments were due to degradation of de novo complex sphingolipids, we performed pulse-chase experiments. Following prelabeling of Jurkat cells with [ 14 C]serine for 2 h, the cells were treated with L-serine (10 mM) for 30 min, irradiated, and then incubated for 5, 15, or 60 min. No changes in labeling of lipids were detected under these conditions (Fig. 2, A, B, and C). Therefore, PDTinduced increases in pulse-labeled ceramide amount did not result from accelerated degradation of de novo synthesized complex sphingolipids.
Effect of PDT on [ 14  SPT Is Not Up-regulated after PDT in Jurkat Cells-We have shown that sphinganine generation after PDT is ISP-1sensitive (23). Since SPT can be up-regulated by inducers, such as UV radiation (39) or endotoxin (40), and photo-oxidative stress can up-regulate certain genes (41-45), we first tested whether PDT can induce up-regulation of mRNA for SPT subunits LCB1 and LCB2. Based on the findings that rapid PDT apoptosis is not associated with either transcription or new protein synthesis (46), we hypothesized that during delayed rather than rapid apoptosis SPT mRNA may be up-regulated. Exposure of Jurkat cells to various PDT doses (200 nM Pc 4 ϩ 135, 270, or 400 mJ/cm 2 ) leads to apoptosis at 1.5 h in 17, 40, and 46% cells, respectively (23). The same PDT doses did not up-regulate mRNA of either LCB1 or LCB2 at 2 or 4 h (Table I). At the lower PDT doses (200 nM Pc 4 ϩ 20 or 40 mJ/cm 2 ), apoptosis was not observed until 24 h after PDT (not shown). The low PDT doses did not up-regulate LCB1 or LCB2 transcript up to 24 h (not shown). These results demonstrate the absence of SPT mRNA up-regulation after Pc 4-PDT irrespective of the rapidity of apoptosis.
The possibility of SPT up-regulation was further tested at the protein level. Pc 4-treated Jurkat cells were exposed to different light fluences (135, 270, or 400 mJ/cm 2 ), and microsomal LCB1 protein amounts were analyzed by Western blot. As depicted in Fig. 4A, there was no change in LCB1 protein expression. Similarly LCB1 protein levels remained unchanged up to 16 h post-PDT (200 nM Pc 4 ϩ 135 mJ/cm 2 ; Fig. 4B).
Since SPT can be activated without transcriptional up-regulation and increased protein expression (8, 10, 47), in the next series of experiments we measured SPT activity in microsomes isolated from either control or PDT-treated Jurkat cells. The enzyme activity did not change when cells were exposed to various PDT doses (200 nM Pc 4 ϩ 135, 270, or 400 mJ/cm 2 ; Fig.  4C). Similarly SPT activity was not affected by PDT (200 nM Pc 4 ϩ 135 mJ/cm 2 ) up to 16 h (Fig. 4D). At 24 h SPT activity was attenuated by 23% (p Ͻ 0.05) compared with Pc 4-treated samples. Since at that time point ϳ90% of Jurkat cells were apoptotic (not shown), the inactivation of SPT is probably a consequence of cell death. Overall our data support no upregulation of SPT by PDT. . Cells were exposed to [ 14 C]serine immediately after PDT. At the indicated times (A) or at 2 h (B), cells were rapidly harvested, washed, and processed for TLC analysis and phosphorimaging as described under "Experimental Procedures." A, the data (pixels/mg of protein) are shown as means Ϯ S.E. of two to five independent determinations. B, representative results from three to five independent determinations are shown. Con, untreated control; Pc 4, Pc 4-treated control; P-Serine, phosphoserine. mJ/cm 2 )-triggered inhibition of glucosylceramide synthase was detected over a period of 1-16 h with the peak reduction (48%) in enzyme activity at 8 h (Fig. 5D). Hence both sphingomyelin synthase and glucosylceramide synthase are inactivated after PDT. The data strongly suggest that de novo ceramide accumulation is a consequence of inhibition of its conversion into sphingomyelin and glucosylceramide.
PDT Also Induces Increase in the Amount of de Novo Ceramide in CHO Cells-To test whether the observations obtained from Jurkat cells are reproducible in other cell types, we examined the effects of PDT on metabolic labeling of ceramide in CHO cells. PDT doses (500 nM Pc 4 ϩ 100 or 200 mJ/cm 2 ) increased 14 C labeling of ceramide at 2 h by 2.7-and 8.4-fold, respectively (Fig. 6, A and C). The effect was abolished when cells were co-exposed to PDT ϩ ISP-1 (Fig. 6, B and C). ISP-1 alone also completely inhibited incorporation of [ 14 C]serine into ceramide at rest (Fig. 6, B and C). Hence, in CHO cells, similar to Jurkat cells, PDT initiates substantial increases in ISP-1sensitive [ 14 C]ceramide.
Increases in [ 14 C]ceramide were paralleled by a dose-dependent inhibition of the [ 14 C]serine incorporation into sphingomyelin post-PDT (Fig. 6, A and D). PDT ϩ ISP-1 abolished the incorporation of [ 14 C]serine into sphingomyelin (Fig. 6, B and  D). ISP-1 alone also abrogated labeling of sphingomyelin with [ 14 C]serine at rest (Fig. 6, B and D).
The incorporation of [ 14 C]serine into phosphatidylserine and phosphatidylethanolamine was inhibited after PDT ( Fig. 6A and  B). Neither ISP-1 alone nor PDT ϩ ISP-1 had any effect on 14 C labeling of phosphatidylserine (Fig. 6B). While resting [ 14 C]phosphatidylethanolamine was unaffected by ISP-1, PDT ϩ ISP-1 inhibited 14 C labeling of phosphatidylethanolamine (Fig. 6B). Because sphingoid bases, like sphinganine that is produced in the de novo sphingolipid biosynthesis, can be catabolized to phosphoethanolamine, which is then used for biosynthesis of phosphatidylethanolamine (51), this may explain why PDT ϩ ISP-1 further inhibits 14 C labeling of phosphatidylethanolamine.

Effects of PDT on Catabolism of Complex Sphingolipids and [ 14 C]Serine Uptake in CHO Cells-
To test whether catabolism of sphingolipids is affected by PDT in CHO cells, pulse-chase experiments were carried out. While no significant changes in either [ 14 C]ceramide or glucosylceramide were detected (Fig. 7,  A, B, and D), 14 C labeling of sphingomyelin was attenuated (Fig. 7C). Specifically, following PDT [ 14 C]sphingomyelin amounts were reduced by 35,40, and 43% at 5, 15, and 60 min, respectively. These findings suggest that PDT-induced increases in [ 14 C]ceramide in CHO cells is, at least in part, due to the degradation of sphingomyelin. In addition, [ 14 C]phosphatidylethanolamine and [ 14 C]phosphatidylserine amounts were also inhibited post-PDT (Fig. 7A), indicating a rapid turnover of the two lipids.
The effect of PDT on the [ 14 C]serine pool was also assessed in CHO cells. Exposure of Pc 4 (500 nM)-treated CHO cells to light fluences of 100 or 200 mJ/cm 2 did not substantially affect intracellular [ 14 C]serine amounts (Fig. 7E). However, a higher PDT dose (500 nM Pc 4 ϩ 400 mJ/cm 2 ) induced substantial loss of [ 14 C]serine (not shown). There was no difference in cellular [ 14 C]serine between untreated and PDT (500 nM Pc 4 ϩ 200 mJ/cm 2 )-treated CHO cells at either 0 or 120 min (not shown). The data support that PDT (500 nM Pc 4 ϩ 100 or 200 mJ/cm 2 ) had no effect on [ 14 C]serine uptake in CHO cells.

Sphingomyelin Synthase and Glucosylceramide Synthase
Are Inhibited after PDT in CHO Cells-We next tested whether PDT affects the activities of sphingomyelin synthase and glucosylceramide synthase in CHO cells. Similar to our findings in Jurkat cells, PDT profoundly inhibited sphingomyelin synthase in CHO cells (Fig. 8A). PDT doses (200 nM Pc 4 ϩ 100, 200, or 400 mJ/cm 2 ) inhibited the enzyme activity at 1 h by 50, 84, and 92%, respectively. Glucosylceramide synthase was also inhibited in Pc 4-photosensitized CHO cells. PDT doses (200 nM Pc 4 ϩ 100, 200, or 400 mJ/cm 2 ) attenuated the enzyme activity at 1 h by 24, 60, and 81%, respectively (Fig. 8B). Therefore, both sphingomyelin synthase and glucosylceramide synthase were also inhibited in CHO cells post-PDT.
PDT Induces Loss of LCB1 in CHO Cells-In the next series of studies we addressed the regulation of LCB1 by PDT directly. To determine the expression amounts of LCB1, Pc 4 (500 nM)-treated CHO cells were exposed to various light fluences (100, 200, or 400 mJ/cm 2 ) and then incubated for 1 h. Microsomal LCB1 protein amounts were analyzed and quantified by Western blot and chemifluorescence image analyzer. PDT doses reduced LCB1 protein amounts by 41, 63, and 90% (Fig.  9A). At the highest PDT dose (500 nM Pc 4 ϩ 400 mJ/cm 2 ), the expression amount of LCB1 protein was reduced by ϳ90% immediately after PDT (Fig. 9B).
We next measured SPT activity in microsomes isolated from either control or PDT-treated CHO cells. PDT caused a dosedependent inhibition of SPT at 1 h, since co-exposure of CHO  (Fig. 9C). The inhibition of the enzyme activity (74%) was detected as early as 0 min after PDT and was maintained for at least 1 h (Fig. 9D). ation is paralleled by partial inhibition of apoptosis in various cell types, including Jurkat cells (23). To test whether ISP-1induced inhibition in [ 14 C]ceramide labeling translates into changes in apoptosis, we determined DEVDase activity and nuclear apoptosis in photosensitized cells. Exposure of CHO cells to PDT led to a dose-dependent activation of DEVDase, since PDT (500 nM ϩ 100, 200, or 400 mJ/cm 2 ) increased the activity of enzyme by 2.4-, 13.8-, and 16.1-fold, respectively (Fig. 10A). Co-exposure of CHO cells to ISP-1 and the same PDT doses attenuated DEVDase activity by 55, 47, and 44%, respectively (Fig. 10A).

ISP-1 Inhibits DEVDase Activation and Apoptosis after Pc 4-PDT in CHO Cells
Similarly, treatment of CHO cells with PDT resulted in a dose-dependent induction of nuclear apoptotic changes, because PDT (500 nM Pc 4 ϩ 100, 200, or 400 mJ/cm 2 ) increased apoptotic cell population by 6.7-, 13.1-, and 20.1-fold, respectively (Fig. 10B). Treatment of CHO cells with ISP-1 reduced the induction of apoptosis in response to the same PDT doses from 10.1, 19.6, and 30.2% to 3.9, 9.3, and 19.2%, respectively (Fig. 10B). Thus, there is a correlation between the inhibition of [ 14 C]ceramide and apoptosis by ISP-1, suggesting the involvement of the de novo ceramide in the process. DISCUSSION This is the first report showing that in the absence of SPT up-regulation PDT initiates de novo ceramide generation via inhibition of its conversion to complex sphingolipids. PDTinduced increases in [ 14 C]ceramide reflect de novo ceramide accumulation, since (i) the incorporation of the SPT substrate [ 14 C]serine into ceramide was stimulated post-PDT, (ii) the SPT inhibitor ISP-1 abolished labeling of ceramide with [ 14 C]serine, and (iii) PDT did not accelerate degradation of de novo synthesized complex sphingolipids in Jurkat cells (although accelerated degradation might occur in CHO cells). In addition, the pulse-chase data showed that [ 14 C]ceramide amounts were not changed following PDT, ruling out the possibility that the increases in the amount of [ 14 C]ceramide resulted from slower degradation of de novo synthesized ceramide in PDT-treated cells. The present data are consistent with the notion that previously observed PDT-induced increases in ceramide mass in numerous cell lines (16 -22), including CHO (17) and Jurkat cells, 3 are a result of de novo ceramide accumulation. Treatment of cells with ISP-1 suppresses not only PDT-induced accumulation of de novo ceramide but also PDTinduced apoptosis ( Fig. 10 and Ref. 23). Hence, we propose that the lack of ceramide accumulation may cause apoptotic resistance to PDT. Collectively our findings further support the role of ceramide in apoptosis of photosensitized cells.
The following evidence from both Jurkat and CHO cells is consistent with sphingomyelin synthase as a molecular PDT target. (i) The enzyme was inactivated in a dose-dependent manner after treatment. (ii) [ 14 C]Serine pulse labeling of sphingomyelin was inhibited post-PDT. Similarly, tumor necrosis factor and ceramide inhibit sphingomyelin synthase (48). Our findings strongly support that sphingomyelin synthase controls cellular ceramide amounts and functions, as suggested by Hannun and co-workers (52).
Glucosylceramide synthase has been shown to be activated (49) as well as inhibited (50) in response to some stimuli. However, the enzyme is not involved in apoptosis after etoposide and ionizing radiation in Jurkat cells (53). We have shown that in both Jurkat and CHO cells PDT-induced inhibition of glucosylceramide synthase was not as effective as the inactivation of sphingomyelin synthase (e.g. in Fig. 5 compare A and B with C and D). Despite glucosylceramide synthase inactivation by PDT, 14 C labeling of glucosylceramide increased in photosensitized CHO cells. The pulse-chase data showed no significant changes in [ 14 C]glucosylceramide amounts, suggesting the absence of PDT-induced up-regulation of glucosylceramide catabolism. Perhaps accumulated de novo ceramide due to inhibition of sphingomyelin synthase can translocate to the site of glucosylceramide synthesis to increase 14 C-labeled glucosylceramide amounts despite partial inhibition of glucosylceramide synthase.
The biochemical evidence from CHO cells supports that LCB1 is a Pc 4-PDT target, since (i) the native 52-kDa LCB1 protein is lost on Western blots, and (ii) the SPT activity is inhibited. Both effects are dose-dependent. The observations that the inactivation and the disappearance of LCB1 are immediate (0 min post-PDT) indicate that PDT induces direct photodamage of the LCB1 subunit of SPT. Rapid photodamage is predicted, since the major PDT damaging reactive oxygen 3 D. Separovic, unpublished observations.

FIG. 10. ISP-1 inhibits PDT-induced activation of DEVDase and apoptosis in CHO cells.
After overnight preincubation with Pc 4 (500 nM), cells were treated with ISP-1 (1 M) for 1 h prior to irradiation at the indicated light fluences and then incubated for 2 h. The data (pixels/mg of protein) are shown as means Ϯ S.E. from two to five independent determinations. A, DEVDase activity was measured in cell lysates spectrofluorometrically using DEVD-7-amino-4-methyl-coumarin as the substrate. B, nuclear apoptotic changes were detected by staining with Hoechst 33342 using fluorescence microscopy. I, ISP-1; Pc 4, Pc 4-treated control. species, singlet oxygen, which is formed where the photosensitizer is localized, has a short lifetime (Ͻ0.04 s) and a short radius of action (Ͻ0.02 m) (54). Hence it is expected that most of the singlet oxygen will react very near to its site of production. Pc 4 is localized to intracellular membranes, including the endoplasmic reticulum (55), where SPT is localized as well (32). Of note, Bcl-2, which can also be found in the endoplasmic reticulum, is directly damaged by Pc 4-PDT (56,57). Moreover, we have shown that PDT affects metabolic labeling of phosphatidylserine, indicating that PDT also damages phosphatidylserine synthases, which are localized to the endoplasmic reticulum (58). In contrast to CHO cells, in Jurkat cells SPT is resistant to photosensitization with Pc 4. Although the reasons for the differential susceptibility of SPT to PDT remain unclear, rapid PDT apoptosis, as shown in our present and previous studies (23,24), is not associated with RNA up-regulation or de novo protein synthesis (46). In pulse-chase experiments, PDT significantly reduced [ 14 C]sphingomyelin amounts during chase in CHO cells but not in Jurkat cells (Fig. 7C versus Fig. 2C), suggesting that the effect of PDT on catabolism of sphingomyelin is also cell type-dependent.
In summary, our data strongly support that the observed increases in de novo ceramide in photosensitized apoptotic cells in the absence of SPT up-regulation are a result of inhibition of ceramide conversion to complex sphingolipids. Our novel findings imply that PDT-induced inhibition of conversion of ceramide to complex sphingolipids plays a critical role in regulating de novo ceramide amounts. Our data also indicate that sphingomyelin synthase and glucosylceramide synthase are key regulators of both de novo ceramide generation and its putative apoptotic function in photosensitized cells.