Natural Ceramide Reverses Fas Resistance of Acid Sphingomyelinase −/− Hepatocytes*

The role of the second messenger ceramide in Fas-mediated death requires clarification. To address this issue, we generated hepatocytes from paired acid sphingomyelinase(ASMase; asmase) +/+ andasmase − /− mice.asmase − /− hepatocytes, derived from 8-week-old mice, manifested normal sphingomyelin content and normal morphological, biochemical, and biologic features. Nonetheless, ASMase-deficient hepatocytes did not display rapid ceramide elevation or apoptosis in response to Jo2 anti-Fas antibody.asmase − /− hepatocytes were not inherently resistant to apoptosis because staurosporine, which did not induce early ceramide elevation, stimulated a normal apoptotic response. The addition of low nanomolar quantities of natural C16-ceramide, which by itself did not induce apoptosis, completely restored the apoptotic response to anti-Fas inasmase − /− hepatocytes. Other sphingolipids did not replace natural ceramide and restore Fas sensitivity. Overcoming resistance to Fas inasmase − /− hepatocytes by natural ceramide is evidence that it is the lack of ceramide and not ASMase which determines the apoptotic phenotype. The ability of natural ceramide to rescue the phenotype without reversing the genotype provides evidence that ceramide is obligate for Fas induction of apoptosis in hepatocytes.

Engagement of Fas/CD95/APO-1 receptors by Fas ligand or anti-Fas antibody results in formation of a death-inducing signaling complex comprised of the adaptor molecules FADD/ MORT-1 and caspase 8, resulting in release of active caspase 8 to initiate the apoptotic process (1). Apoptosis, however, proceeds only after an ensuing commitment step, which allows effector caspase activation. Recently, Peter and co-workers (2) provided evidence that the quality and quantity of the activated caspase 8 signal are regulated in a cell type-specific fashion and suggested that Fas initiates apoptosis via two different mechanisms. In type I cells, apoptosis occurs after high caspase 8 activation, which signals exclusively via a hierarchical caspase cascade, independent of mitochondrial dysfunction. In contrast, apoptosis in type II cells proceeds after minimal caspase 8 activation through an amplification cascade involving mitochondrial dysfunction, release of mitochondrial cytochrome c, and activation of Apaf-1 and caspase 9. Although many of the upstream elements in type II apoptosis are unknown, this event is inhibited by Bcl-2 and can be mimicked by exogenous addition of short chain ceramides. In both mechanisms, apoptosis eventually ensues after effector caspases, such as 3 and 7, are activated.
There are numerous reports of early ceramide elevation upon Fas activation (3,4). Ceramide is a second messenger in an evolutionarily conserved stress response pathway that, in different cells, signals events as diverse as differentiation, proliferation, and apoptosis (3). Ceramide is generated from sphingomyelin by the action of a neutral or acid sphingomyelinase (NSMase 1 or ASMase) or by de novo synthesis coordinated through the enzyme ceramide synthase (5). In response to Fas ligation, ceramide elevation is often biphasic, a rapid elevation with return to baseline is followed by a more prolonged elevation that is progressive. Evidence has been provided that both ASMase and NSMase might contribute to Fas-mediated ceramide elevation (6,7), with ASMase involvement in the acute event. Recent investigations titrated down the amount of FADD or procaspase 8 transfected into HeLa and 293T cells to the point where they no longer induced apoptosis (8). Under these conditions, FADD and procaspase 8 nonetheless induced maximal ceramide generation, which was inhibitable by the initiator caspase inhibitor CrmA. These experiments provided evidence that ceramide generation is initiator caspase-dependent and occurs prior to commitment to the effector phase of the apoptotic process.
Whether the ceramide generated in response to Fas ligation is involved in the apoptotic process or is an epiphenomenon is a matter of ongoing debate. Testi and co-workers (9) claimed that Epstein-Barr virus-transformed lymphoblasts from Niemann-Pick disease (NPD) patients, which have an inherited deficiency of ASMase activity, displayed deficits in ceramide generation and apoptosis, whereas Borst and co-workers (10) found no differences. Recently, Green and co-workers (7) reported partial resistance to Jo2 anti-Fas-induced liver failure and death in asmase Ϫ/Ϫ mice in vivo. Furthermore, the establishment of primary cultures of several cell types derived from asmase Ϫ/Ϫ mice showed that hepatocytes (11) but not thymocytes (7) displayed resistance to Fas-mediated death. Although these genetic models tended to support a possible cell typespecific role for ASMase, and indirectly for ceramide, in Fasinduced apoptosis, other interpretations are possible. For in-stance, the NPD lymphoblasts described above were transformed with Epstein-Barr virus and the transformation process itself, as well as known Epstein-Barr virus strainspecific phenotypic differences (12)(13)(14), might have affected the outcome. Further, it remains possible that the resistance of cells from asmase Ϫ/Ϫ mice to Fas-induced apoptotic death might result from subtle alterations in membrane structure/ function independent of ceramide itself.
To address directly the issue of the involvement of ceramide in Fas-mediated death, we examined the ability of primary cultures of hepatocytes derived from asmase ϩ/ϩ and asmase Ϫ/Ϫ mice to undergo apoptosis in response to anti-Fas antibody. Like embryonic fibroblasts and thymocytes from asmase Ϫ/Ϫ mice (15), asmase Ϫ/Ϫ hepatocytes displayed normal sphingomyelin content and were indistinguishable from asmase ϩ/ϩ hepatocytes by numerous morphological, biochemical, and biologic criteria. In contrast to wild type hepatocytes, however, AS-Mase-deficient hepatocytes failed to generate ceramide upon activation of Fas and were markedly resistant to Fas-mediated apoptotic death. Further, the addition of low nanomolar quantities of natural C 16 -ceramide, which by itself does not effect apoptosis, restored Fas-mediated apoptosis in asmase Ϫ/Ϫ hepatocytes. Overcoming apoptosis resistance by provision of natural ceramide is evidence that it is the lack of ceramide and not ASMase that determines the Fas sensitivity. The ability of natural ceramide to rescue the phenotype without reversing the genotype provides evidence that ceramide is obligate for efficient Fas induction of apoptosis in hepatocytes.

MATERIALS AND METHODS
Animals-ASMase knockout mice, maintained in a sv129 ϫ C57BL/6 background, are propagated using heterozygous breeding pairs and genotyped as described (16). Experimental mice were 8 -12 weeks old and sacrificed by carbon dioxide asphyxiation.
Ex Vivo Hepatocyte Culture Prepared by Mechanical Disruption-Harvested livers were washed three times with phosphate-buffered saline at 37°C, cut into small pieces, and separated into small clumps with a spatula. Individual hepatocytes were dispersed mechanically by passage through an 18-gauge needle, filtered through a 100-M cell strainer, washed once with RPMI 1640 complete medium, and resuspended into the same medium containing 10% fetal bovine serum, as described (11). Viability was always greater than 95% as defined by trypan blue exclusion analysis. After 1 h, cells (5 ϫ 10 6 /ml) were placed in 24-well plates pretreated with 10 mg/ml bovine serum albumin fraction V (Sigma) in an incubator at 37°C with 5% CO 2 atmosphere. Cells were treated with anti-Fas Jo2 antibody (Pharmingen) or staurosporine (Sigma) for the indicated times while under constant agitation using a nutator (Becton Dickinson).
Ex Vivo Hepatocyte Culture Prepared by Collagenase Perfusion-Hepatocytes were isolated by cannulation of the portal vein and retrograde in situ collagenase perfusion according the method described previously by Picardo et al. (17). Briefly, livers were perfused under sterile conditions at 40°C consecutively with Leffert's buffer (100 mM HEPES, pH 7.4, 30 mM KCl, 1.3 M NaCl, 10 mM NaH 2 PO 4 , 100 mM D-glucose) with 0.5 mM EGTA for 4 min, without EGTA for 2 min, and then with 33% (w/v) collagenase type V (Roche) and 5 mM CaCl 2 for 12 min using an Easy Load perfusion pump (Masterflex, Cole Parmer; flow rate of 16 ml/min). The liver was then minced in cold Leffert's buffer containing 5 mM CaCl 2 , and isolated hepatocytes were filtered through a 100-m cell strainer. Isolated hepatocytes were washed twice at 50 ϫ g and resuspended into RPMI 1640 complete medium containing 10% fetal bovine serum. Viability was always greater than 95% using this method.
Fluorescence-activated Cell Sorting (FACS)-A single cell suspension of hepatocytes was washed twice and resuspended into 100 l of H/S (132 mM NaCl, 20 mM HEPES, 5 mM KCl, 1 mM CaCl 2 , 0.7 mM MgCl 2 , 0.8 mM MgSO 4 ) containing 2% fetal calf serum and 0.2% NaN 3 supplemented with 1 g/ml anti-Fas Jo2 antibody or an isotype-specific con-trol immunoglobulin. After 45 min on ice, cells were washed with the same buffer and stained with a fluorescein isothiocyanate-coupled antihamster antibody (Pharmingen 12084D). Cell surface Fas was determined in labeled cells by flow cytometry, as described (11).
Terminal Nucleotidyl Transferase Assay-Hepatocytes were trypsinized after treatment with anti-Fas antibody or staurosporine and then permeabilized with a solution of 0.1% Triton X-100 and 0.1% sodium citrate at 4°C for 5 min. Apoptosis was assessed by terminal nucleotidyl transferase according to the manufacturer's instructions (Roche, Indianapolis, IN). At least 200 cells were counted for each point.
Ethidium Bromide/Acridine Orange Staining-Cells were stained with 4 g/ml each ethidium bromide and acridine orange for 10 min and analyzed by fluorescence microscopy as described (19).
DEVDase Assay-The fluorogenic substrate Ac-DEVD-AFC was used to measure caspase 3-like activity according to the manufacturer's instructions (Kamiya, Seattle, WA).
Statistical Analysis-Statistical analysis was performed by Student's t test and t test for correlation coefficient.

RESULTS
Characterization of asmase Ϫ/Ϫ Hepatocytes-We reported recently that 8 -10 week old asmase Ϫ/Ϫ mice derived from our colony display normal sphingomyelin content in multiple tissues (15,20). Consequently, these animals are physiologically normal and do not manifest NPD, as NPD requires sphingomyelin accumulation (21). In the present studies, we generated primary cultures of hepatocytes from livers of asmase ϩ/ϩ and asmase Ϫ/Ϫ mice. Livers from 8-week-old ASMase-deficient mice displayed normal architecture, and isolated hepatocytes manifested physiologic morphology. No measurable ASMase activity was detected in the asmase Ϫ/Ϫ hepatocytes (Table I); NSMase activity was unchanged. Further, asmase Ϫ/Ϫ hepatocytes manifested normal sphingomyelin and ceramide content (Table I), consistent with the observation that these animals develop normally into adulthood (15,22,23). The distribution of glycosphingolipid-enriched microdomains, a compartment putatively involved in ceramide signaling (24) and their content was similar in asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes (Table I), as was the expression of Fas on the cell surface, measured by FACS using the Jo2 anti-Fas antibody (Fig. 1). Similarly, the levels of factors known to participate in Fasmediated apoptosis in hepatocytes including FADD/MORT-1, Bid, Bcl-2, Apaf-1, and caspases 3, 8, and 9 were identical in asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes (Table I). Moreover, the activation patterns of p38, c-Jun terminal, and extracellular signal-regulated kinases in response epidermal growth factor, tumor necrosis factor-␣, heat shock, and hyperosmolarity were also similar in asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes (Table I). Thus, asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes appeared similar as assessed by numerous criteria.
asmase Ϫ/Ϫ Hepatocytes Are Defective in Fas-induced Ceramide Generation-To investigate whether ASMase-deficient cells displayed defects in Fas-mediated signaling through the sphingomyelin pathway, we treated primary cultures of asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes with increasing concentrations of Jo2 anti-Fas antibody. Hepatocytes were treated in suspension, as Kass and co-workers (25) showed that plating hepatocytes as monolayers conferred Fas resistance, i.e. the development of a requirement for cycloheximide for death induction (25). In contrast, hepatocytes in suspension retained sensitivity to Fasmediated death. Fig. 2A shows that asmase ϩ/ϩ hepatocytes displayed dose-dependent elevation in ceramide content when measured at 5 min of treatment. As little as 50 ng/ml Jo2 increased ceramide content from a baseline of 170 Ϯ 20 pmol/ 10 6 cells to 210 Ϯ 20 pmol/10 6 cells, and a maximal increase to 515 Ϯ 15 pmol/10 6 cells (p Ͻ 0.01) was achieved with 1,000 ng/ml. In contrast, asmase Ϫ/Ϫ hepatocytes failed to manifest increased ceramide levels at any Jo2 dose up to 8,000 ng/ml.
Studies were also performed to determine the time course of ceramide elevation after Jo2 treatment using a 500-ng/ml dose (Fig. 2B). In asmase ϩ/ϩ hepatocytes, ceramide levels increased to 1.6-fold of control by 30 s (p Ͻ 0.01), peaked at 5 min, and remained elevated for at least 20 min. asmase Ϫ/Ϫ hepatocytes, however, did not demonstrate elevation of ceramide levels at any time up to 20 min. Thus, asmase Ϫ/Ϫ hepatocytes display a complete defect in the early phase of ceramide generation upon Jo2 treatment.
asmase Ϫ/Ϫ Hepatocytes Are Resistant to Anti-Fas-mediated Apoptosis-To evaluate whether asmase Ϫ/Ϫ hepatocytes also displayed defects in Fas-mediated apoptosis, hepatocytes were rested for 1 h in RPMI 1640 complete medium containing 10% fetal bovine serum prior to treatment with Jo2 anti-Fas antibody. Prior investigations showed that Jo2 antibody induced time-and dose-dependent apoptosis in wild type hepatocytes in vivo and ex vivo (7, 11, 26 -30). Initial studies determined the time course of induction of apoptosis in asmase ϩ/ϩ hepatocytes as measured by terminal nucleotidyl transferase using 1,000 ng/ml Jo2 antibody, a maximally effective dose. Apoptosis was detected by 2 h and was maximal by 8 h (not shown). Subsequent investigations measured the dose dependence of Jo2induced apoptosis in asmase ϩ/ϩ hepatocytes after 8 h of treatment. As shown in Fig. 3A, apoptosis occurred over 2-3 logs of Jo2 doses; as little as 100 ng/ml Jo2 was effective; a maximal apoptotic effect was achieved using 2,000 ng/ml. Thus, the dose dependence for Jo2-induced ceramide generation is slightly lower than that for apoptosis. asmase Ϫ/Ϫ hepatocytes, which contain normal numbers of cell surface Fas receptors (Fig. 1), were nonetheless resistant to treatment with anti-Fas antibody. In contrast to the total resistance to ceramide generation, apoptosis was inducible by increasing the dose of Jo2 antibody about 1 log. Thus, apoptosis was elicited with a 2,000-ng/ml dose of Jo2 in asmase Ϫ/Ϫ hepatocytes and increased to a maximum at 8,000 ng/ml Jo2. Similar results were obtained using ethidium bromide and acridine orange staining to detect morphological changes of apoptosis (Fig. 3B). These studies indicate that asmase Ϫ/Ϫ hepatocytes can undergo ASMase-dependent and -independent apoptosis after Jo2 treatment.
Because the hepatocytes used in these studies were obtained by mechanical dispersion, we compared the effect of Jo2 to induce death using asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes obtained by an alternative technique, in vivo collagenase treatment. For these studies, collagenase was delivered to intact livers after cannulation of the inferior hepatic vein and isolated hepatocytes derived as described (17). Using this approach, we ϩ ϩ a ASMase and NSMase activities were measured as in (84). Sphingomyelin levels were determined by charring as described in (85). Ceramide levels were quantitated by the diacylglycerol kinase assay. The data (mean Ϯ S.D.) are compiled from two studies performed in duplicate. Note that the residual "ASMase" activity detected in the asmase Ϫ/Ϫ hepatocytes likely represents NSMase activity manifest at pH 5.0. b For morphology and structure studies, livers were fixed in 4% paraformaldehyde overnight, and 5-m sections, stained with hematoxylin/eosin, were observed by light microscopy.
FIG. 1. asmase ؉/؉ and asmase ؊/؊ hepatocytes express comparable levels of cell surface Fas. The level of Fas on the surface of hepatocytes was determined by flow cytometry as described (11). To detect Fas, asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes were stained with Jo2 and a fluorescein isothiocyanate-coupled anti-hamster secondary antibody (green and blue lines, respectively). For controls, asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes were stained with an isotype-specific control antibody followed by fluorescein isothiocyanate-coupled secondary antibody (orange and red lines, respectively).
again observed a 1-log deficit in apoptotic death in asmase Ϫ/Ϫ hepatocytes in suspension culture (not shown). Thus, by two separate techniques for hepatocyte isolation asmase Ϫ/Ϫ hepatocytes are defective in Fas-mediated apoptotic death. Unless otherwise indicated, subsequent studies used mechanical dispersion for hepatocyte isolation.
The apoptotic response was also assessed using a biochemical assay for caspase 3 activity after a dose of 1,000 ng/ml Jo2 (see Fig. 5B). Consistent with the results shown in Fig. 3A and 3B, asmase ϩ/ϩ hepatocytes showed a 4.1 Ϯ 0.1-fold increase in caspase activity at 8 h of Jo2 treatment while asmase Ϫ/Ϫ hepatocytes did not exhibit any increase in caspase 3 activity. Hence, multiple assays confirm that lack of ASMase renders cells resistant to Jo2-induced apoptosis.
asmase Ϫ/Ϫ Hepatocytes Are Sensitive to Staurosporine-induced Apoptosis-To evaluate whether the mutation leading to the asmase Ϫ/Ϫ phenotype resulted in a general inactivation of the apoptotic machinery, hepatocytes were exposed to a different proapoptotic stimulus. Staurosporine is a potent protein kinase inhibitor capable of inducing apoptosis by a mechanism that does not involve ASMase-mediated ceramide generation (20). As in other cell types, staurosporine failed to generate ceramide within the first 30 min in either asmase ϩ/ϩ or asmase Ϫ/Ϫ hepatocytes (not shown). Nonetheless, staurosporine induced apoptosis in asmase ϩ/ϩ hepatocytes within 6 -8 h (Fig.  4), which peaked by 12 h (not shown). A dose as low as 0.1 M staurosporine induced apoptosis in 30% of the population at 12 h, whereas 75% apoptosis was observed with 5-10 M staurosporine (Fig. 4). In contrast to the Fas experiments, the asmase Ϫ/Ϫ hepatocytes showed no resistance to staurosporineinduced cell death, and the level of apoptosis was essentially identical to that of the asmase ϩ/ϩ hepatocytes (Fig. 4). These studies indicate that asmase Ϫ/Ϫ hepatocytes are fully competent in generating an apoptotic response and suggest that the lack of apoptosis to anti-Fas may be specifically associated with ASMase deficiency.
Natural Ceramide Can Rescue the Apoptotic Response of asmase Ϫ/Ϫ Hepatocytes to Anti-Fas-To provide definitive evi-dence that the lack of ceramide is the cause of the resistance of asmase Ϫ/Ϫ cells to anti-Fas treatment, we pretreated asmase Ϫ/Ϫ hepatocytes with low doses of natural C 16 -ceramide for 10 min prior to the treatment with 1,000 ng/ml Jo2. This dose was selected because it provides the greatest discrimination in apoptotic responses between asmase ϩ/ϩ and asmase Ϫ/Ϫ hepatocytes. Fig. 5A shows that 5-25 nM C 16 -ceramide was without direct effect on apoptosis, nor did it alter the apoptotic response to 1,000 ng/ml Jo2 in asmase ϩ/ϩ hepatocytes. However, 5-25 nM C 16 -ceramide restored the sensitivity of asmase Ϫ/Ϫ hepatocytes to Jo2 antibody. In fact, at 25 nM C 16ceramide the apoptotic response of asmase Ϫ/Ϫ hepatocytes was FIG. 3. asmase ؊/؊ hepatocytes are relatively resistant to anti-Fas Jo2 antibody. Panel A, hepatocytes were isolated from asmase ϩ/ϩ and asmase Ϫ/Ϫ littermates, placed in 24-well plates precoated with bovine serum albumin, and treated with the indicated concentrations of anti-Fas monoclonal antibody. Cells were harvested after 8 h, and apoptosis assessed by terminal nucleotidyl transferase assay. 200 cells were evaluated per point. The data (mean Ϯ S.E.) are compiled from three studies performed in duplicate. Panel B, apoptosis was also assessed in hepatocytes from asmase ϩ/ϩ and asmase Ϫ/Ϫ littermates 8 h after the indicated doses of Jo2 by ethidium bromide/ acridine orange staining as described (19). The data (mean Ϯ S.E.) are compiled from six studies performed in triplicate.
indistinguishable from that of asmase ϩ/ϩ hepatocytes. Identical results were obtained using hepatocytes isolated by collagenase digestion (not shown). The capacity of natural ceramide to rescue the ASMase-deficient apoptotic phenotype was also evaluated using the caspase 3 biochemical assay. As shown in Fig. 5B, natural ceramide alone at 25 nM also conferred Jo2induced caspase activation onto asmase Ϫ/Ϫ hepatocytes. Thus, the addition of natural ceramide to the asmase Ϫ/Ϫ hepatocytes rescues the apoptotic phenotype without reverting the genotype, providing proof that the resistance of ASMase-deficient hepatocytes to anti-Fas antibody is mediated by the lack of ceramide generation and not by the lack of ASMase. Natural C 16 -ceramide but Not Other Sphingolipids Rescues ASMase-deficient Hepatocytes-To address further the specific role of ceramide in Fas-mediated apoptosis, we investigated the capacity of other sphingolipids to restore Jo2-mediated death to asmase Ϫ/Ϫ hepatocytes. Fig. 5C shows that although 25 nM C 16 -ceramide rescued the apoptotic phenotype of asmase Ϫ/Ϫ hepatocytes, as much as a 50 nM concentration of other sphingolipids including C 16 -dihydroceramide, sphingomyelin, sphingosine 1-phosphate, sphingosine, or sphinganine failed to rescue Jo2-induced apoptosis. It should be noted that these other sphingolipids also did not induce apoptosis in untreated asmase ϩ/ϩ or asmase Ϫ/Ϫ hepatocytes (not shown). Thus, natural ceramide specifically restores Jo2-induced apoptosis to asmase Ϫ/Ϫ hepatocytes. DISCUSSION The present studies show that primary cultures of hepatocytes from asmase Ϫ/Ϫ mice display a large deficit in anti-Fas induced death ex vivo compared with hepatocytes from asmase ϩ/ϩ littermates. In contrast to asmase ϩ/ϩ hepatocytes, asmase Ϫ/Ϫ hepatocytes failed to demonstrate an early ceramide elevation despite containing normal levels of sphingomyelin. Further, the apoptosis resistance of asmase Ϫ/Ϫ hepatocytes was overcome by natural ceramide. This provides proof that it is the lack of ceramide, and not ASMase, which determines the asmase Ϫ/Ϫ phenotype. Consistent with this notion, only natural ceramide and not other sphingolipids could restore Fas sensitivity. Apoptosis resistance could also be overcome by very high doses of Jo2 anti-Fas. The mechanism for this ceramide-independent apoptosis induction by Jo2 may involve superaggregation of trimerized Fas (31)(32)(33). However, it is not certain that this ceramide-independent apoptotic mechanism is operative under physiologic conditions. In this regard, two independent in vivo events, phytohemagglutinin-induced hepatitis, and anti-CD4-induced deletion of CD3 ϩ and CD4 ϩ cells, which are mediated by Fas ligand up-regulation on peripheral blood lymphocytes, are completely defective in our asmase Ϫ/Ϫ mice despite normal Fas ligand up-regulation (11).
Prior studies reported that ASMase-deficient mice were partially resistant to animal death upon intravenous injection of Jo2. In contrast to the 1-log defect in apoptosis observed in hepatocytes ex vivo, ASMase-deficient mice displayed only a moderate reduction in sensitivity to the lethal effect of Jo2. The reason for the difference in requirement for ASMase for these events is presently uncertain. However, animals expressing a Bcl-2 transgene only in their livers were defective in Fasinduced hepatitis yet displayed little resistance to death (34). Thus, it is not clear that rapid hepatic failure is causative in animal death after anti-Fas. Because the dose dependence for hepatocyte death in vivo has not been determined in animals treated with anti-Fas, a comparison of the effects of Jo2 in vivo and in vitro is not presently possible.
The present studies add to emerging data indicating that young ASMase-deficient animals display normal sphingomyelin levels and hence do not manifest evidence of NPD (35). NPD is caused by the progressive accumulation of sphingomyelin (35). It is therefore not surprising that hepatocytes, murine embryonic fibroblasts, and thymocytes from 8 -10-week-old asmase Ϫ/Ϫ mice appear normal as assessed by a battery of biochemical, histological, and biological tests (7,15,20,22,23). Consistent with this observation, the asmase Ϫ/Ϫ mice develop normally into adulthood. In our colony, the earliest clinical manifestation of NPD, a resting tremor, begins at 12-16 weeks of age. Routinely, our asmase Ϫ/Ϫ mice survive to 9.6 Ϯ 0.4 (mean Ϯ S.D.) months of age. Nevertheless, asmase Ϫ/Ϫ mice and primary cultures of cells from young asmase Ϫ/Ϫ mice display specific defects in apoptosis. asmase Ϫ/Ϫ hepatocytes display resistance to Fas-induced death ex vivo and, as mentioned above, to phytohemagglutinin-induced hepatitis in vivo (11). Similarly, murine embryonic fibroblasts from asmase Ϫ/Ϫ mice display complete resistance to ionizing radiation-induced apoptosis and partial resistance to tumor necrosis factor-␣/actinomycin D-and serum withdrawal-induced apoptosis (20). In both hepatocytes and murine embryonic fibroblasts, however, sensitivity to staurosporine is maintained, indicating that the use of the sphingomyelin pathway within any cell may be stress type-specific. asmase Ϫ/Ϫ mice also display defects in apoptosis in vivo. Endothelium in the lung and throughout the central nervous system of these mice is almost completely resistant to induction of apoptosis with doses as high as 100 Gray, whereas asmase ϩ/ϩ littermates display dose-dependent apoptosis beginning with as little as 5 Gray (22,23). ASMase-deficient mice also manifest a marked defect in the ovarian developmental program (36). A failure to delete oocytes normally in asmase mutant females during embryogenesis results in ovarian hyperplasia at birth (37). Further, oocytes isolated from adult ASMase-deficient females are markedly resistant to doxorubicin-induced apoptotic death (36). In contrast, thymic cells from asmase Ϫ/Ϫ mice display no resistance to apoptosis in vivo or ex vivo by any stimulus tested so far (7,23). These data show the use of the sphingomyelin pathway for induction of apoptosis is stress type-and cell type-specific.
In summary, the present investigation provides evidence that ceramide is required for optimal Fas signaling of the death response in hepatocytes. Preliminary data suggest that ceramide is required for capping of Fas, an event necessary for transmembrane signal transmission for other cell surface receptors. 2 Overcoming apoptosis resistance by natural ceramide is evidence that it is the lack of ceramide, and not ASMase, which determines the Fas-resistant phenotype in hepatocytes.