Human vascular endothelial cells are a rich and regulatable source of secretory sphingomyelinase. Implications for early atherogenesis and ceramide-mediated cell signaling.

We recently reported that macrophages and fibroblasts secrete a Zn2+-dependent sphingomyelinase (S-SMase), which, like lysosomal SMase, is a product of the acid SMase gene. S-SMase may cause subendothelial retention and aggregation of lipoproteins during atherogenesis, and the acid SMase gene has been implicated in ceramide-mediated cell signaling, especially involving apoptosis of endothelial cells. Because of the central importance of the endothelium in each of these processes, we now sought to examine the secretion and regulation of S-SMase by vascular endothelial cells. Herein we show that cultured human coronary artery and umbilical vein endothelial cells secrete massive amounts of S-SMase (up to 20-fold more than macrophages). Moreover, whereas S-SMase secreted by macrophages and fibroblasts is almost totally dependent on the addition of exogenous Zn2+, endothelium-derived S-SMase was partially active even in the absence of added Zn2+. Secretion of S-SMase by endothelial cells occurred both apically and basolaterally, suggesting an endothelial contribution to both serum and arterial wall SMase. When endothelial cells were incubated with inflammatory cytokines, such as interleukin-1beta and interferon-gamma, S-SMase secretion by endothelial cells was increased 2-3-fold above the already high level of basal secretion, whereas lysosomal SMase activity was decreased. The mechanism of interleukin-1beta-stimulated secretion appears to be through increased routing of a SMase precursor protein through the secretory pathway. In summary, endothelial cells are a rich and regulatable source of enzymatically active S-SMase, suggesting physiologic and pathophysiologic roles for this enzyme.

Experimental evidence suggests that products of the acid SMase (ASM) gene may have a prominent role in both atherogenesis and ceramide-mediated apoptosis. Hydrolysis of lipoprotein SM retained on subendothelial matrix would clearly be expected to be an extracellular event. In this context, our laboratory has reported that the ASM gene in macrophages and fibroblasts gives rise not only to lysosomal SMase (L-SMase) but also, via differential trafficking of the ASM precursor protein, 2 to a secretory SMase (S-SMase) (15). Importantly, S-SMase secreted by these cells, which is activated by physiologic levels of Zn 2ϩ (15), can hydrolyze and cause the aggregation of atherogenic lipoproteins, even at neutral pH (16). Several cell culture studies have also implicated a role for acid SMase activity in cytokine-induced, ceramide-mediated apoptosis (17)(18)(19), and mice in which the ASM gene has been disabled by homologous recombination show defective radia-tion-and endotoxin-induced apoptosis in vivo (20,21). The products of the ASM gene that mediate these cell signaling events have not been identified, but S-SMase might be an excellent candidate, since the outer leaflet of the plasma membrane is rich in SM (22).
The endothelium is central to these processes. Our previous work on S-SMase focused on macrophages (15), which enter the arterial wall in response to the initial retention of lipoproteins in the subendothelial matrix (5,23) and therefore may contribute to lipoprotein aggregation after lesion initiation. Lipoprotein aggregation also occurs, however, in prelesional susceptible segments of the subendothelium (10), and thus the endothelium would be a likely candidate for a source of S-SMase in the prelesional arterial wall. Likewise, recent in vivo studies have shown that the endothelium is the key tissue in cytokine-induced, ASM-mediated apoptosis (20,21). Moreover, cytokines are important regulators of endothelial function (24,25), atherogenesis, and apoptosis (26,27), which raises the possibility that cytokines influence endothelial secretion of S-SMase.
In this report, we demonstrate that cultured endothelial cells, including human coronary artery endothelial cells, are an even more abundant source of S-SMase than are macrophages. Furthermore, we show that the secretion of S-SMase by endothelial cells is strongly regulated by cytokines known to be present in atherosclerotic and inflammatory lesions. The mechanism of this regulation is primarily through an alteration in the cellular trafficking of the ASM precursor protein. Thus, the vascular endothelium is likely to be a major, regulatable source of S-SMase that may contribute to atherogenesis and inflammation.

EXPERIMENTAL PROCEDURES
Materials-The Falcon tissue culture plasticware used in these studies was purchased from Fisher. Millicell-CM 0.4-m culture plate inserts were from Millipore Corp. Tissue culture media and other tissue culture reagents were obtained from Life Technologies, Inc. Fetal bovine serum (FBS) was obtained from Hyclone Laboratories (Logan, UT) and was heat-inactivated for 1 h at 65°C (HI-FBS). Human recombinant cytokines were obtained as follows: interferon-␤ and interferon-␥ from Biogen (Cambridge, MA), interleukin-4 from Peprotech (Rocky Hill, NJ), and interleukin-1␤ from R & D Systems Inc. (Minneapolis, MN). [9,10-3 H]Palmitic acid (56 Ci/mmol) was purchased from DuPont NEN, and [N-palmitoyl-9 -10-3 H]Sphingomyelin was synthesized as described previously (15,28,29). N,N-Dimethylformamide; 1,3-dicyclohexylcarbodiimide; N-hydroxysuccinimide; and N,N-diisopropylethylamine were purchased from Aldrich. Precast 4 -20% gradient polyacrylamide gels were purchased from NOVEX (San Diego, CA). Nitrocellulose was from Schleicher and Schuell. FLAG-tagged S-SMase and rabbit anti-S-SMase were kindly provided by Dr. Henry Lichtenstein (Amgen, Boulder, CO); the FLAG-tagged S-SMase was purified by anti-FLAG immunoaffinity chromatography from the conditioned medium of cells transfected with a human ASM-FLAG cDNA. 2 For the immunoprecipitation protocol (see below), this antibody was further purified on an S-SMase-Affi-Gel affinity column (Dr. G. Andrew Keesler, Amgen, Boulder, CO). A murine monoclonal antibody against the FLAG epitope was purchased from Eastman Kodak Co., and peroxidase-conjugated sheep anti-mouse IgG was from Amersham Corp. Peroxidase-conjugated goat anti-rabbit IgG and concanavalin A-Sepharose was purchased from Pierce. DEAE-Sephacel was from Pharmacia Biotech Inc. Bovine liver 215-kDa mannose 6-phosphate receptor linked to Affi-Gel 15 was made as described by Varki and Kornfeld (30) and was kindly supplied by Dr. Peter Lobel (Center for Advanced Biotechnology and Medicine, Piscataway, NJ). All other chemicals and reagents were from Sigma, and all organic solvents were from Fisher.
Cells-Primary cultures of HUVEC were obtained from fresh umbilical cords and maintained in cell culture as described previously (31). Human coronary artery endothelial cells from a 38-year-old female donor were purchased from Clonetics Corp. (San Diego, CA; catalog number CC-2585, strain 3033). The cells were grown in endothelial basal media (Clonetics) supplemented with 10% HI-FBS, 10 g/ml endothelial cell growth supplement (Sigma; catalog number E0760), and 50 g/ml gentamicin. Primary cultures of bovine aortic endothelial cells were established and subcultured as described previously (32). For the experiment in Fig. 2, the cells were plated on Millicell-CM 0.4-m culture plate inserts placed in 35-mm wells in ␣-MEM containing 10% FBS; the inserts had been previously coated with 0.1% gelatin for 1 h and then with 5 g/ml collagen and 20 g/ml fibronectin for 1 h. Monolayer cultures of J774.A1 cells (from the American Type Culture Collection; see Ref. 33) were grown and maintained in spinner culture with DMEM/HI-FBS/PSG (penicillin, streptomycin, and glutamine) as described previously (33,34). Human peripheral blood monocytes were isolated from normal subjects as described previously (35) and induced to differentiate into macrophages by the addition of 1 ng of granulocytemacrophage colony-stimulating factor/ml of media on days 1, 4, and 11 of culture as described previously (35); by day 14, the cells were differentiated as assessed by morphological changes (e.g. increased spreading) and increased expression of scavenger receptor activity (cf. Ref. 36). Unless indicated otherwise, the cells were plated in 35-mm (6-well) dishes in media containing HI-FBS for 48 h. The cells were then washed three times with PBS and incubated for 24 h in 1 ml of fresh serum-free media containing 0.2% BSA. This 24-h conditioned medium was collected for SMase assays.
Harvesting of Cells and Conditioned Media-Following the incubations described above and in the figure legends, cells were placed on ice, and the serum-free conditioned medium was removed. The cells were washed two times with ice-cold 0.25 M sucrose and scraped into 0.3 and 3.0 ml of this sucrose solution per 35-and 100-mm dish, respectively. Unless indicated otherwise, the scraped cells were disrupted by sonication on ice using three 5-s bursts (Branson model 450 sonifier), and the cellular homogenates were assayed for total protein by the method of Lowry et al. (22) and for SMase activity as described below. The conditioned media were spun at 800 ϫ g for 5 min to pellet any contaminating cells and concentrated 6-fold using a Centriprep 30 (Amicon; Beverly, MA) concentrator (molecular weight cut-off ϭ 30,000).
SMase Assay-As described previously (15), the standard 200-l assay mixture consisted of up to 40 l of sample (conditioned media or homogenized cells; see above) and a volume of assay buffer (0.1 M sodium acetate, pH 5.0) to bring the volume to 160 l. The reaction was initiated by the addition of 40 l of substrate (50 pmol of [ 3 H]sphingomyelin) in 0.25 M sucrose containing 3% Triton X-100 (final concentration of Triton X-100 in the 200-l assay mix was 0.6%). When added, the final concentrations of EDTA and Zn 2ϩ were 5 and 0.1 mM, respectively, unless indicated otherwise. The assay mixtures were incubated at 37°C for no longer than 3 h and then extracted by the method of Bligh and Dyer (37); the lower, organic phase was harvested, evaporated under N 2 , and fractionated by TLC using chloroform/methanol (95:5). The ceramide spots were scraped and directly counted to quantify [ 3 H]ceramide. Typically, our assay reactions contained approximately 20 g of cellular homogenate protein and a volume of conditioned media derived from a quantity of cells equivalent to approximately 50 g of cellular protein.
SMase Immunohistochemistry of Murine Aorta-Hearts from chowfed, 4-week-old female ASM knockout mice and wild-type mice of the same genetic background (SV129/C57BL6) were perfused, embedded in optimum-cutting-temperature (OCT) compound (Sakura Finetek, Torrance, CA), and snap-frozen as described previously (38). 8-m-thick sections of the proximal aorta were cut on a cryostat and fixed in 10% buffered formalin for 5 min at room temperature. The sections were blocked in 2% murine serum in PBS for 2 h at room temperature and then incubated with 20 g of anti-SMase antibody/ml of PBS containing 0.1% Triton X-100 for 2-4 h at room temperature. Bound antibody was detected with a biotinylated secondary antibody followed by streptavidin peroxidase (Vectastain Elite ABC-peroxidase kit; Vector Laboratories Inc., Burlingame, CA) and 3,3Ј-diaminobenzidine. The specimens were counterstained with hematoxylin and then viewed with an Olympus IX 70 inverted microscope using a ϫ 100 objective.
Partial Purification of L-and S-SMase from HUVECs-A modification of previously published procedures (39) was used. Cells were scraped in ice-cold 40 mM Tris-HCl, 0.1 mM ZnCl 2 , 0.1% Nonidet P-40, pH 7.2 (buffer A), disrupted by sonication on ice using three 5-s bursts (Branson model 450 sonifier), and centrifuged at 100,000 ϫ g for 1 h. The supernatant (Ͻ5 ml) was applied to a 5-ml column of DEAE-Sephacel, and the column was capped and incubated end-over-end for 90 min at room temperature. The column was then uncapped, and the flow-through fraction was collected; the column was washed with another 1-2 ml of buffer A, and this flow-through fraction was added to the first. The combined flow-through fractions were loaded onto another DEAE-Sephacel column, and the procedure was repeated. The flowthrough fractions from this second column were pooled, concentrated using a Centriprep 30 (see above), assayed for protein concentration by the method of Lowry et al. (40), and subjected to SDS-PAGE and immunoblotting.
The conditioned medium was concentrated, dialyzed against buffer A, and subjected to the same two rounds of DEAE-Sephacel chromatography as above. The combined flow-through fractions from the second DEAE-Sephacel step were then concentrated and dialyzed against 10 mM Tris-HCl, 500 mM NaCl, 1 mM MgCl 2 , 1 mM MnCl 2 , 1 mM CaCl 2 , 0.1% Nonidet P-40, and 0.02% NaN 3 , pH 7.2 (buffer B). This solution (Ͻ1 ml) was added to a 1-ml concanavalin A-Sepharose column and incubated with the column exactly as described for the DEAE-Sephacel step. The column was then washed with two bed volumes of buffer B containing 10 mM methylglucopyranoside, followed by two bed volumes of buffer B containing 1 M methylglucopyranoside. This last eluate was concentrated, assayed for protein concentration by the method of Lowry (40), and subjected to SDS-PAGE and immunoblotting.
Northern Blot of SMase mRNA-Total cellular RNA was isolated from HUVECs using RNAzol B (Tel-Test, Inc., Friendswood, TX). Approximately 10 g of RNA was separated on a 1% agarose gel and blotted onto a nylon membrane. The membrane was hybridized in a QuikHyb hybridization solution (Stratagene) with a 1.6-kilobase pair EcoRI-XhoI fragment of human ASM cDNA (41) that was labeled with 32 P by the random-priming procedure (Life Technologies, Inc.). The amount of RNA loaded for each condition was normalized using a probe to glyceraldehyde-3-phosphate dehydrogenase. The relative intensities of the mRNA bands were determined by densitometric scanning of the autoradiograms using a Molecular Dynamics computing densitometer (model 300A) with ImageQuant software or a Bio-Rad molecular imager (model GS525).
Lactate Dehydrogenase Assay-Lactate dehydrogenase activity was assayed in medium and cells as reported previously (42) using a kit purchased from Sigma (catalogue number 500); the assay measures the unreacted pyruvic acid using a colorometric assay.
SDS-PAGE and Immunoblotting-Protein samples were boiled in buffer containing 1% SDS and 10 mM dithiothreitol for 10 min, loaded onto 4 -20% gradient polyacrylamide gels, and electrophoresed for 50 min at 35 milliamps in buffer containing 0.1% SDS. Following electrophoresis, the proteins on the gels were electrotransferred (100 V for 1.5 h) to nitrocellulose for immunoblotting. Next, the nitrocellulose membranes were incubated with 5% Carnation nonfat dry milk in buffer C (24 mM Tris, pH 7.4, containing 0.5 M NaCl) for 3 h at room temperature. The membranes were then incubated with mouse anti-FLAG monoclonal antibody (1:1000) or rabbit anti-L-SMase antibody (1:1000) in buffer D (buffer C containing 0.1% Tween 20, 3% nonfat dry milk, and 0.1% bovine serum albumin) for 1 h at room temperature. After washing four times with buffer C containing 0.1% Tween 20, the blots were incubated with horseradish peroxidase-conjugated sheep anti-mouse IgG (1:20,000) or goat anti-rabbit IgG (1:20,000) for 1 h in buffer D at room temperature. The membranes were subsequently washed twice with 0.3% Tween 20 in buffer C and twice with 0.1% Tween 20 in buffer C. Finally, the blots were soaked in the enhanced chemiluminescence reagent (Pierce "Super Signal" kit) for 2 min and exposed to x-ray film for 1 min.
Immunoprecipitation of S-SMase-20 l of HUVEC conditioned medium, 20 l of affinity-purified anti-SMase antibody, and 160 l of PBS were incubated for 1 h at 4°C. 100 l of Protein A-Sepharose (50 mg/ml in 50 mM Tris buffer, pH 7.0) was then added to the incubation mixture, and the slurry was mixed end-over-end for 18 h at 4°C. The suspension was centrifuged for 1 min in a microcentrifuge. The supernatant was harvested and assayed for SMase activity.
Statistics-Unless otherwise indicated, results are given as means Ϯ S.D. (n ϭ 3); absent error bars in the figures signify S.D. values smaller than the symbols.

Cultured Human Vascular Endothelial Cells Secrete Abundant Amounts of S-SMase-Previous
work from our laboratories revealed that human and murine macrophages were a relatively abundant source of S-SMase compared with other cell types, such as COS-7 and murine migroglial cells (15). As shown in Fig. 1A, however, both human coronary and umbilical vein endothelial cells (HUVECs) are a much more abundant source of S-SMase; for example, HUVECs secrete almost 20fold more S-SMase than human macrophages. Interestingly, endothelium-derived S-SMase was partially active in the absence of exogenously added Zn 2ϩ (Fig. 1A), whereas S-SMase secreted by macrophages and other cell types previously examined by us was almost entirely Zn 2ϩ -dependent (15). 3 This observation may be important in the regulation of endothelium-derived S-SMase activity (see "Discussion"). The data in Fig. 1B show that SMase activity in the cell homogenate, which is not stimulated by exogenous Zn 2ϩ and represents L-SMase activity (15), is also very abundant in endothelial cells (recall that S-SMase and L-SMase originate from the same gene (15), mRNA (15), and protein precursor 2 ). In summary, the data in Fig. 1 demonstrate that human endothelial cells, including those derived from coronary arteries, are an abundant source of S-SMase, much of which is active in the absence of exogenously added Zn 2ϩ .
We next addressed the polarity of secretion of S-SMase from cultured endothelial cells. Cultures of endothelial cells were 3 95% of S-SMase activity from HUVECs was immunoprecipitated by an affinity-purified anti-S-SMase antibody, and 97% of HUVEC S-SMase activity was inactivated by chelation of Zn 2ϩ with 10 mM EDTA plus 10 mM 1,10-phenanthroline (see Footnote 2). These data indicate that a single enzyme accounts for SMase activity in the conditioned medium of HUVECs. established on Millicell-CM inserts in 35-mm dishes so that S-SMase from the upper (apical) and lower (basolateral) chambers could be assayed separately. As a control for apically secreted enzyme that either "leaked" or was transcytosed into the lower chamber, we set up parallel dishes in which immunoaffinity-purified FLAG-tagged S-SMase was added to the upper chamber and then monitored using an anti-FLAG immunoblot assay. As shown in Fig. 2, substantial amounts of S-SMase activity were found in both the upper and lower chambers of the dishes. In contrast, FLAG-tagged S-SMase added to the upper chamber did not appear in the lower chamber during this 1-h experiment (Fig. 2, inset). From these data, we conclude that cultured endothelial cells secrete S-SMase both apically and basolaterally. If the endothelium acts similarly in vivo, these findings implicate the endothelium as a possible source of both subendothelial and serum S-SMase (cf. Ref. 43).
SMase Is Present in the Endothelium of Murine Aorta-To complement the cell culture studies described above, we used immunohistochemistry to detect SMase in slices of murine proximal aorta. Because L-and S-SMase are derived from the same protein precursor and differ only by post-translational modifications (15), 2 the anti-SMase antibodies currently available to us recognize both forms of the enzyme. To ensure the specificity of the signal for SMase, proximal aorta from ASM knockout mice (44), which lack both L-and S-SMase (15), were used as a negative control (Fig. 3A). Fig. 3B demonstrates substantial SMase staining in the endothelium; much of the stain in this image appears to be intracellular (open arrow), which is most likely L-SMase and possibly some S-SMase in the secretory pathway. By comparison, medial staining was much weaker, although still specific. These data demonstrate that the high levels of intracellular SMase in cultured endothelial cells (Fig. 1B) reflect the situation in an actual vessel wall. In addition, we also noticed areas of dark staining on the lumenal edge of some of the endothelial cells (closed arrow). This staining is specific (i.e., not in panel A) and does not appear lysosomal, and so it is possible that this signal represents S-SMase that has been secreted and perhaps retained on the cell surface.
Secretion of S-SMase from Human Endothelial Cells Is Regulated by Cytokines-Inflammatory cytokines are important constituents of atherosclerotic lesions and may contribute to various aspects of atherogenesis (26). A major target of these cytokines is the endothelium (24,25). To determine if cytokines known to be present in atherosclerotic lesions affect endothelium-derived S-SMase, cultured human endothelial cells were exposed to IL-1␤, interferon-␥, interferon-␤, and IL-4. Each of the first three cytokines substantially increased the accumulation of S-SMase in the conditioned media of these cells, although IL-4 had no effect (Fig. 4A). In particular, IL-1␤ and interferon-␥ increased S-SMase activity ϳ3-fold. Interestingly, the stimulatory cytokines led to a decrease in L-SMase activity (Fig. 4B). This pattern is distinct from that observed during monocyte-to-macrophage differentiation, in which both S-and L-SMase activities are increased (15).

IL-1␤ Increases S-SMase Secretion by HUVECs via Alteration in the Trafficking of the ASM Precursor
Protein-We next sought to determine the mechanism of cytokine-mediated induction of S-SMase secretion by HUVECs. For these studies, we focused on IL-1␤, and we first determined whether this cytokine increased SMase mRNA levels in HUVECs. As shown in Fig. 5, there was an ϳ40% increase in SMase mRNA in IL-1␤-treated HUVECs, when normalized for glyceraldehyde-3-phosphate dehydrogenase mRNA and quantified by either densitometry or molecular imaging. Because differences less than 2-fold in Northern blot assays may not be significant and because the IL-1␤-induced increase in S-SMase activity was greater than 3 times control, we conclude that most, if not all, of the induction by this cytokine was post-transcriptional. This conclusion is consistent with the finding that the cytokinemediated increase in S-SMase is accompanied by a decrease in L-SMase (above).
Next, we asked whether the mechanism of IL-1␤-induced S-SMase secretion could be the result of cellular lysis. For example, cellular release of certain molecules involved in inflammation may occur physiologically by this mechanism (45). To test this idea, the release of the cytosolic protein lactate dehydrogenase was assayed. Under conditions in which S-SMase was induced by IL-1␤ to 3.6 times control (Fig. 6A), lactate dehydrogenase release increased only 60% (Fig. 6B). Thus, at most, cellular lysis can account for only a small proportion (i.e. ϳ20%) of the IL-1␤-induced increase in S-SMase.
We recently reported that S-SMase is secreted through a Golgi pathway that bypasses the lysosomal pathway of L-SMase. 2 Thus, we considered three general mechanisms to explain how IL-1␤ increases the secretion of SMase: secretion of mannose-phosphorylated SMase during trafficking of the ASM precursor to lysosomes (e.g., due to saturation of the mannose 6-phosphate receptor; cf. Ref. 46 ing of the ASM precursor protein through the normal Golgi secretory pathway. To begin, we assessed the mannose phosphorylation state of S-SMase from control and IL-1␤-treated HUVECs by passing conditioned media from these cells over a mannose-phosphate receptor column (cf. Ref. 48). By Western blot analysis quantified by densitometry, only 7.9% of control S-SMase and 4.9% of S-SMase from the cytokine-treated cells bound to the column and could be eluted with mannose 6-phosphate. As a control for this experiment, we have previously shown that the massive secretion of SMase by ASM-transfected Chinese hamster ovary cells can be partly explained by secretion of mannose-phosphorylated SMase due to saturation of the mannose 6-phosphate receptor 2 ; when conditioned medium from these cells was passed over the receptor column, 40.3% bound and could be eluted with mannose 6-phosphate. Based on this set of experiments, we conclude that IL-1␤-mediated induction of SMase secretion by HUVECs is not due to increased secretion of mannose-phosphorylated SMase during trafficking of the ASM precursor to lysosomes.
Next, we compared the zinc dependence of S-SMase and L-SMase activities from control and IL-1␤-treated HUVECs. Zn 2ϩ increased the enzymatic activity of S-SMase 1.77 Ϯ 0.03fold and 1.78 Ϯ 0.17-fold from IL-1␤-treated and untreated cells, respectively, whereas Zn 2ϩ increased the enzymatic activity of L-SMase 1.13 Ϯ 0.08-and 1.09 Ϯ 0.04-fold from IL-1␤-treated and untreated cells, respectively. Thus, S-SMase from IL-1␤-treated HUVECs had a zinc dependence similar to that of S-SMase from untreated cells, not to that of L-SMase. These data support the idea that IL-1␤ treatment results in increased trafficking of the ASM precursor protein through the normal Golgi secretory pathway, not in increased secretion of L-SMase.
To further test this conclusion, we took advantage of our previous observation that L-SMase has a lower apparent M r FIG. 3. SMase immunohistochemistry of proximal aorta from wild-type and SMase knockout mice. 8-m sections of proximal aorta from an ASM (SMase) knockout mouse (A) and from a wild-type mouse of the same genetic background (B-D) were stained with an anti-SMase antibody, counterstained with hematoxylin, and viewed by microscopy as described under "Experimental Procedures." The open arrows show areas of intracellular staining, and the closed areas show staining that appears to be on the lumenal surface of the endothelial cells.

FIG. 5. Northern blot analysis of HUVECs incubated in the absence or presence of IL-1␤.
HUVECs were incubated for 18 h in DMEM, 0.2% BSA alone (Con) or containing human recombinant interleukin-1␤ (5 ng/ml) (IL1␤). Total RNA was extracted from the cells, electrophoresed, blotted, and probed for ASM (SMase) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA as described under "Experimental Procedures." Quantification by molecular imaging or densitometry indicated ϳa 40% increase in the ratio of ASM to glyceraldehyde-3-phosphate dehydrogenase mRNA signals with IL-1␤ treatment. than S-SMase on SDS-PAGE, 2 presumably because of different post-translational modifications. Thus, we sought to determine if conditioned medium from IL-1␤-treated HUVECs had increased amounts of higher M r SMase, indicating increased flux through the normal secretory pathway or normal levels of the higher M r form plus the appearance of the lower M r form, indicating secretion or release of L-SMase from lysosomes (Fig.  7). It should be noted that the HUVEC enzyme, particularly HUVEC L-SMase, is not well recognized by our antibody, and so the immunoblot signal, particularly for the intracellular enzyme (L-SMase), is weak. Nonetheless, the data in Fig. 7 confirm the differences in M r between L-and S-SMase in this cell type (compare lanes 1 and 3) and show that only the higher M r S-SMase was detectably increased in the IL-1␤-treated cells (compare lanes 3 and 4). By densitometry, this increase was approximately 2-fold. Even when the immunoblot was overexposed, the conditioned medium from the IL-1␤-treated cells contained no evidence of a band at the M r of L-SMase (data not shown). These data further support the conclusion that the increase in S-SMase from HUVECs induced by IL-1␤ is primarily due to an increased flux of the common SMase precursor through the Golgi-secretory pathway. DISCUSSION The data in this report demonstrate that endothelial cells are a rich source of active S-SMase, particularly in the presence of inflammatory cytokines. The potential physiologic relevance of these findings is related to the postulated role of the endothelium in extracellular SMase-induced lipoprotein aggregation and retention and to the demonstrated role of this tissue in cytokine-induced, acid SMase-mediated cell signaling (see Introduction and below). In addition, our results reveal several novel mechanisms for the regulation of S-SMase, particularly by alterations in intracellular protein trafficking.
We have previously shown that retained and aggregated lipoproteins in atherosclerotic lesions are hydrolyzed by an extracellular arterial wall SMase (4), and we found that S-SMase was the only SMase secreted by arterial wall cells (15). Mohan Das et al. (49) have reported that a magnesium-dependent SMase is externally oriented on neurons. We have not yet investigated whether this enzyme, which has not yet been cloned, is present on the surface of arterial wall cells. Nonethe-less, we imagine that a secreted enzyme would have more access to lipoproteins retained on subendothelial matrix than a cell surface-bound enzyme. These points, together with the findings reported herein, lead us to propose the following model: basal levels of endothelium-derived S-SMase help initiate the atherosclerotic lesion by promoting lipoprotein retention and aggregation (5). Then, as T-cells and macrophages enter lesions (23,26) and secrete cytokines and (at least in the case of macrophages) additional S-SMase, lipoprotein-SM hydrolysis and lipoprotein retention and aggregation would be amplified.
We have also speculated that endothelium-derived S-SMase may play a role in ceramide-mediated apoptosis. This speculation is based upon several pieces of information. First, a product of the acid SMase gene plays an important role in endothelial apoptosis in vivo (20,21). Specifically, the pulmonary endothelium of ASM knockout mice was the major tissue demonstrating a defect in ceramide generation and apoptosis in response to total body irradiation; radiation-induced apoptosis FIG. 7. Immunoblot analysis of L-and S-SMase from control and IL-1␤-treated HUVECs. HUVECs were incubated for 18 h in DMEM, 0.2% BSA alone (Con) or containing human recombinant interleukin-1␤ (5 ng/ml) (IL1␤). SMase was partially purified from the cell homogenates and conditioned media of these cells, and then aliquots derived from equivalent numbers of cells were subjected to immunoblot analysis. Specifically, conditioned medium from control and IL-1␤treated cells yielded 256.0 and 262.5 g of concanavalin A eluate protein, of which 60.8-and 60.0-g aliquots were loaded onto the SDSpolyacrylamide gel, respectively. Cell homogenates from control and IL-1␤-treated cells yielded 168.1 and 160.6 g of DEAE flow-through protein, of which 59.5-and 60.0-g aliquots were loaded onto the SDS-polyacrylamide gel, respectively. No band at the molecular weight of L-SMase was visible in conditioned media (right panels), even when overexposed (not shown).
FIG. 6. Effect of IL-1␤ on the release of lactate dehydrogenase from HUVECs. HUVECs were incubated for 18 h in DMEM, 0.2% BSA alone (Con) or containing human recombinant interleukin-1␤ (5 ng/ml) (IL1). Conditioned media from these cells were then assayed for SMase activity (A) and lactate dehydrogenase activity (B). in thymocytes and splenocytes was much less diminished in the ASM knockout versus wild-type mice (20). Moreover, very recent studies in mice have shown that the endothelium is the target of tumor necrosis factor-␣-mediated apoptosis after injection of lipopolysaccharide. This response is associated with an increase in endothelial ceramide content and is greatly diminished in ASM knockout mice, indicating involvement of a SMase arising from the ASM gene (21). Second, the data in this report show that endothelial cells are a rich and cytokineregulatable source of S-SMase, a product of the ASM gene. Third, S-SMase may have more access to the most abundant pools of cellular SM (22). In contrast, since lysosomal membranes have little SM (50), signaling by L-SMase would have to occur during trafficking of the nascent enzyme to lysosomes or would require delivery of SM into lysosomes. Thus, we propose that apoptotic stimuli, such as radiation and cytokines, increase the amount of endothelium-derived S-SMase from a subthreshold basal level to a level capable of generating enough cellular ceramide to trigger or enhance cell-signaling events. It is important to note, however, that this hypothesis is based upon several controversial assumptions. For example, despite the compelling data obtained using ASM knockout mice (20,21), there are some cell culture systems that have shown a role for the neutral, magnesium-dependent SMase in ceramidemediated apoptosis (6). In addition, there are conflicting data regarding the role of cell surface versus intracellular pools of SM in ceramide-mediated signaling (cf. Ref. 51, and see "Discussion" therein). Nonetheless, the data presented in this report give impetus for further in vitro and in vivo studies on the role of endothelium-derived S-SMase in atherogenesis and ceramide-mediated apoptosis.
Regarding the regulation of S-SMase, the data in this report provide evidence to support at least three separate mechanisms for control of S-SMase. Two of these mechanisms, alterations in protein trafficking and accessibility to cellular zinc, are based upon the following model of how the ASM gene gives rise to both L-and S-SMase 2 ; when a common precursor protein derived from the ASM gene is mannose-phosphorylated and thus is targeted to lysosomes, it becomes L-SMase. During this targeting, the enzyme acquires cellular Zn 2ϩ and so does not require exogenous Zn 2ϩ for enzymatic activity. In contrast, when the common precursor is not mannose-phosphorylated, and thus is targeted to the Golgi-secretory pathway, it gives rise to S-SMase. In the secretory pathway, the enzyme does not acquire cellular Zn 2ϩ , so S-SMase secreted by macrophages requires exogenous Zn 2ϩ for enzymatic activity. Based on this model, we predicted that one level of control of S-SMase secretion might be the proportion of the common ASM precursor trafficked into the lysosomal versus secretory pathways. The current data indicate that this mechanism is indeed primarily responsible for the increase in S-SMase in response to IL-1␤. How might a cytokine influence protein trafficking?
The key regulatory step in the trafficking of the SMase precursor is mannose phosphorylation of the precursor by N-acetylglucosaminyl-1-phosphotransferase (52, 53), 2 and we speculate that cytokines may affect (e.g. by protein phosphorylation) either the activity of this phosphotransferase or the suitability of the SMase precursor as its substrate. Further work will be needed to test this and other possible mechanisms.
A second level of regulation of S-SMase is via Zn 2ϩ -induced activation (15). The zinc requirement of S-SMase is similar to that of matrix metalloproteinases (15,43,54,55), and so under conditions in which these proteinases are active, such as in atherosclerotic lesions (56), one would expect S-SMase to be active as well. In addition, Zn 2ϩ levels have been reported to be elevated in atherosclerotic (57) and inflammatory (58) lesions. Nonetheless, the accessibility of extracellular zinc, perhaps modulated by zinc-binding proteins such as metallothionein (59), may represent a regulatory mechanism. Our current results indicate that accessibility of cell-derived zinc is also relevant; S-SMase from endothelial cells was partially activated in the absence of exogenously added Zn 2ϩ (Fig. 1). In contrast, S-SMase from macrophages, fibroblasts, and Chinese hamster ovary cells is almost entirely inactive in the absence of added Zn 2ϩ (Fig. 1 and Ref. 15). Based upon our model (above), we propose that SMase in the secretory pathway of endothelial cells, unlike SMase in the secretory pathway of the other cells examined, has partial access to cellular pools of Zn 2ϩ . 3 The findings that endothelial cells secrete abundant amounts of S-SMase and that this S-SMase is partially active in the absence of added Zn 2ϩ suggest that endothelium-derived S-SMase has unique physiologic roles.
A third point of S-SMase regulation is extracellular pH. S-SMase, like L-SMase, has an acidic pH optimum when assayed in vitro using sphingomyelin in detergent micelles as substrate (15). Thus, S-SMase may be particularly active in environments in which the pH is relatively low, such as in advanced atherosclerotic lesions (60 -63), in certain types of inflammatory processes (62,64), and possibly after reuptake into acidic endosomes. Calahan (65) noted, however, that pH affects only the K m , not the V max , of L-SMase. This finding suggests that access to SM is the issue, and we showed recently that S-SMase can extensively hydrolyze the SM of atherogenic lipoproteins (e.g., oxidized LDL) and lesional LDL at neutral pH (16). This point is of particular importance regarding our hypothesis that endothelium-derived S-SMase plays a role in early lesional events, where the arterial wall pH would be expected to be neutral. These observations may also be relevant to the hydrolysis of cellular sphingomyelin by endotheliumderived S-SMase in neutral pH environments.
In summary, endothelial cells, which we have postulated are important in subendothelial, extracellular lipoprotein SM hydrolysis and which others have shown are important in cytokine-induced, ASM-mediated cell signaling, are a rich and regulatable source of the ASM gene product, S-SMase. These findings have formed the basis of ongoing work directed at further testing the physiologic and pathophysiologic roles of endothelium-derived S-SMase.