Amplification of CD95 Activation by Caspase 8-induced Endosomal Acidification in Rat Hepatocytes*

Although in rat hepatocytes CD95 is predominantly located inside the cell with almost undetectable immunostaining at the plasma membrane, the addition of CD95-ligand (CD95L) induces hepatocyte apoptosis, which is preceded by a targeting and activation of intracellularly localized CD95 to the plasma membrane including formation of the death-inducing signaling complex. This process involves an NADPH oxidase-dependent generation of reactive oxygen species (ROS) through a ceramide- and protein kinase Cζ-dependent pathway, which leads to an activating phosphorylation of p47phox. The mechanisms underlying CD95L-induced ceramide formation were addressed in the present study. It was found that CD95L lowered within seconds the apparent vesicular pH from 6.0 to 5.7 in a fluorescein isothiocyanate-dextran-accessible endosomal compartment, which was previously shown to contain acidic sphingomyelinase, and decreased N-(ethoxycarbonylmethyl)-6-methoxyquinolinium bromide fluorescence, suggestive for an increase of cytosolic [Cl-]. Bafilomycin or 4,4′-diisothiocyanostilbene-2,2′-disulfonic acid disodium salt largely abolished the CD95L-induced endosomal acidification, ceramide formation, and downstream events, such as p47phox phosphorylation, ROS formation, CD95 activation, and apoptosis. These responses were also abolished after knock-down of acidic sphingomyelinase in rat hepatocytes. Interestingly, caspase 8 inhibitors abolished these CD95L-induced signaling events, including the increase in cytosolic [Cl-], endosomal acidification, ceramide formation, and ROS generation as well as CD95 targeting to the plasma membrane and CD95 activation. The data suggest that CD95L initiates a rapid caspase 8-dependent endosomal acidification, which triggers ceramide-dependent ROS formation as an upstream event of trafficking of intracellularly stored CD95 to the plasma membrane. It is concluded that a rapid caspase 8 activation in response to CD95L signals to intracellularly stored CD95, which becomes activated and targeted to the plasma membrane. This autoamplification of CD95-activation is required for apoptosis induction.

CD95 2 (Apo-1/Fas) belongs to the death receptor family and plays an important role in apoptosis induction in many cell types. In hepatocytes, CD95 is mainly located within the cellular interior and can either be activated after ligation with its natural ligand (CD95L) or in a ligand-independent way by hydrophobic bile salts or hyperosmotic cell shrinkage (Refs. 1-4; for review, see Ref. 5). Ligand-dependent and -independent CD95 activation in hepatocytes is a complex process that finally results in CD95 trafficking from the cellular interior to the plasma membrane, subsequent formation of the death-inducing signaling complex (DISC), and eventually apoptosis (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12).
In hepatocytes, proapoptotic stimuli such as CD95 ligand (CD95L), hydrophobic bile salts, or hyperosmotic cell shrinkage induce a rapid oxidative stress (ROS) response (2,4), which triggers a Yes-dependent but ligand-independent activation of the epidermal growth factor receptor (EGFR) (6,7) and within 30 min a c-Jun N-terminal kinase (JNK)-dependent association of the EGFR with CD95 (2,4). Subsequent EGFR-catalyzed CD95-tyrosine phosphorylation allows for CD95 oligomerization and targeting of the EGFR/CD95-protein complex to the plasma membrane in a microtubule-dependent way, where formation of the DISC occurs (8,12). Evidence has been presented that the ROS response, which is elicited by CD95L, hyperosmolarity, or hydrophobic bile salts is due to NADPH oxidase activation (9 -11). This is achieved by an activating phosphorylation of the regulatory subunit p47 phox in response to a ceramide-dependent activation of protein kinase C (9 -11). Hyperosmotic or bile salt-triggered ceramide formation results from an endosomal acidification with subsequent activation of acidic sphingomyelinase (11,13). The mechanisms underlying the rapid ROS response and NADPH oxidase activation after the addition of CD95L, however, remained unclear. Furthermore, in resting hepatocytes CD95 is located inside the hepatocyte, whereas CD95 immunostaining of the plasma membrane is, if at all barely detectable. This raises the question of how externally added CD95L activates intracellularly stored CD95 in hepatocytes. These questions were addressed in the present study, which shows that CD95L induces rapid endosomal acidification and ceramide formation leading to NADPH oxidase activation. The resulting ROS response in turn triggers the targeting of intracellular CD95 to the plasma membrane and DISC formation. Interestingly, all these processes were sensitive to caspase 8, but not caspase 3 inhibition, suggestive for a caspase 8-dependent autoamplification of death receptor signaling. This amplification is apparently initiated by a ligand-dependent activation of small, by conventional immunostaining undetectable amounts of CD95 receptors present in the plasma membrane whereby the resulting activation of caspase 8 triggers via endosomal acidification and ceramide formation the activation of intracellular CD95 and its recruitment to the plasma membrane. Thus, caspase 8 may not only play a role in initiating proapoptotic cascades through activation of downstream caspases but also by amplifying the amount of activated CD95 receptors.
Cell Preparation and Culture-As described previously (1), hepatocytes were isolated from livers of male Wistar rats fed ad libitum with a standard diet by a collagenase perfusion technique. Aliquots of 1.5 ϫ 10 6 hepatocytes were plated on collagen-coated 6-well culture plates (Falcon, Heidelberg, Germany) and cultured as published recently (1) for 24 h unless indicated otherwise before the experiments were started. Osmolarity changes were performed by appropriate addition or removal of NaCl from the medium. The viability of the hepatocytes was more than 95% as assessed by trypan blue exclusion.
ASM Protein Knock-down-Antisense oligonucleotides directed against ASM and the corresponding controls (i.e. nonsense oligonucleotides) were designed and manufactured by Biognostik (Göttingen, Germany). Immediately after hepatocytes were plated on collagen-coated culture plates (diameter, 6 cm; Falcon) at a density of 4 ϫ 10 6 cells/ plate, 4 mol/liter of the respective oligonucleotides (nonsense or ASM antisense) supplemented with Lipofectamine 2000 were added according to the manufacturer's recommendations. Thereafter, cells were kept in culture for up to 4 days. In control experiments, uptake of FITC-labeled nonsense oligonucleotides (2 mol/liter) was already visible after 1 h of incubation and lasted for up to 4 days.
Immunoprecipitation-Hepatocytes were harvested in lysis buffer as recently published (1). Equal protein amounts (200 g) of each sample were incubated for 2 h at 4°C with a polyclonal rabbit anti-CD95, rabbit anti-EGFR, rabbit anti-p47 phox , or rabbit anti-Yes antibody (dilution 1:100; Santa Cruz, CA) to immunoprecipitate CD95, EGFR, p47 phox , or Yes. Then 10 l of protein A-and 10 l of protein-G-agarose (Santa Cruz) were added and incubated at 4°C overnight. Immunoprecipitates were washed 3 times as published recently (2) and then transferred to Western blot analysis as described above. Activation of p47 phox by serine phosphorylation was detected using an anti-phosphoserine antibody (15). The anti-phospho-Src family-Tyr 418 antibody was used to detect activating phosphorylation of Yes (16), and the anti-phosphotyrosine antibody was used to detect tyrosine phosphorylation of EGFR in the respective immunoprecipitates. FADD and caspase 8 association or tyrosine phosphorylation of the immunoprecipitated CD95 samples were detected by Western blot analysis using the respective antibodies (anti-FADD, anti-caspase 8, and anti-phosphotyrosine).
Detection of ROS-Hepatocytes were seeded on collagencoated 6-well culture plates (Falcon) and cultured for 24 h. Cells were incubated with PBS containing 5 mol/liter CM-H 2 DCFDA for 30 min at 37°C and 5% CO 2 . To detect ROS generation, CM-H 2 DCFDA-loaded cells were supplemented again with culture medium and then exposed to CD95L for the indicated time period. Then cells were washed briefly using ice-cold PBS, and cells were lysed in 0.1% Triton X 100 (v/v) dissolved in aqua bidest. Lysates were centrifuged immediately (10,000 ϫ g, 4°C, 1 min), and fluorescence of the supernatant was measured at 515-565 nm using a luminescence spectrometer LS-5B (PerkinElmer Life Sciences) at a 488-nm excitation wavelength. Fluorescence of untreated control cells was arbitrarily set to 1.

Determination of Apparent Endosomal pH (pH ves ) and Intracellular Cl Ϫ Concentration by Single Cell Fluorescence
Recording-To determine changes in apparent vesicular pH (pH ves ), primary hepatocytes grown for 24 h on glass coverslips (diameter, 30 mm) were washed with PBS and then incubated in the same solution with FITC-dextran (5 mg/ml) for another 60 min at 37°C. The intracellular vesicular compartment reached by FITC-dextran under these conditions is seen to reflect an endosomal compartment (17)(18)(19). For fluorescence recording, the coverslips were mounted with PBS at 37°C equilibrated with room atmosphere resulting in a pH of 7.4. Measurements of apparent pH ves in single cells were performed with an inverted fluorescence microscope (Zeiss, Axiovert) combined with the QuantiCell 2000-calcium imaging setup (VisiTech, Sunderland, UK). This apparatus allows FITC-dextran fluorescence measurements at the single-cell levels at the excitation wavelengths of 488/440 nm with a time resolution of 10 Hz by a monochromator, and emission was measured at 515-565 nm using a CCD camera as provided by the QuantiCell 2000-calcium imaging setup. By use of regions of interest using the QuantiCell 2000 software, the field of measurement was chosen to be within one single cell. The raw fluorescence signals were corrected for autofluorescence as published previously (17). Values of apparent pH ves were obtained from the corrected ratios of 488/440 nm after appropriate calibration according to Thomas et al. (20).
To determine changes of the intracellular Cl Ϫ -concentration, primary hepatocytes grown for 24 h on glass coverslips (diameter, 30 mm) were loaded for 6 -10 h with the chloridesensitive dye MQAE (1 mmol/liter) in phenol red-free culture medium in a humidified 5% CO 2 atmosphere at 37°C (21). Then coverslips were transferred to single cell fluorescence recording using an inverted fluorescence microscope (Zeiss, Axiovert) combined with the QuantiCell 2000 calcium imaging setup (VisiTech). Cells were excited at 350 nm with a time resolution of 10 Hz by a monochromator, and emission was measured at 480 -520 nm using a CCD camera as provided by the QuantiCell 2000 calcium imaging setup. By use of regions of interest using the QuantiCell 2000 software, the field of measurement was chosen to be within one single cell. A decrease in intracellular MQAE fluorescence at an excitation wavelength of 350 nm thereby reflects an increase in the intracellular Cl Ϫ concentration (21). Fluorescence is given as relative fluorescence compared with the fluorescence at the beginning of the respective experiment (first second of the recording).
Lipid Extraction and High Performance Thin Layer Chromatography-Cells were harvested at the indicated time points by scrapping the cells off the plate on ice. Pellets were washed and sonicated. Quantification of lipids was done using 500 g of protein for Folch extraction (22). Analysis of ceramides included a mild alkaline hydrolysis. The lower phase of Folch extraction was evaporated under nitrogen. The lipids were dissolved in chloroform/methanol (2:1, v/v).
Samples and standards were separated on silica gel high performance thin layer chromatography plates (20 ϫ 10 cm) Merck 60F 254s (Merck) prewashed for 60 min in 2-propanol and dried for 30 min at 120°C. Samples and standards were applied to the TLC plates using a CAMAG Linomat IV (CAMAG, Berlin, Germany). For determination of ceramides, samples were separated using an automated multiple development (AMD) procedure on an AMD2 device (CAMAG). This procedure consisted of 7 repeated developments of the chromatogram using a stepwise elution gradient with methanol, dichloromethane, and n-hexane (methanol/dichloromethane/ n-hexane: 100/0/0; 10/90/0; 9/91/0; 8/92/0; 3/97/0; 2/98/0; 0/0/ 100) (23) on a CAMAG AMD2 device as described earlier (24). Visualization of separated bands was done by postchromatic derivatization after dipping in a manganese chloride solution according to Grether-Beck et al. (24) in an automated dipping device (CAMAG). After heating the plate for 10 min at 120°C in a temperature-controlled oven, the plate was dried and scanned using a CAMAG TLC Scanner II and CATS software. Quantification was done by absorption at 550 nm with a plot of peak area versus weight spotted for a series of standards using a second-order polynomial calibration with 4 standard mixes in the range 50 -1000 ng.
Subcellular Fractionation-Hepatocytes were cultured on collagen-coated culture plates (diameter, 10 cm; Falcon) at a density of 8 ϫ 10 6 cells/plate. Huh7 were cultured on culture plates (diameter, 6 cm; Falcon) and transfected with CD95-YFP as described above. Cells were lysed in a buffer containing 10 mmol/liter Tris, 30 mmol/liter mannitol, and 10 mmol/liter CaCl 2 , pH 7.5. After centrifugation of the samples (5 min, 1200 ϫ g), the supernatants were subjected to ultracentrifuga-tion (35 min, 40,000 g) to separate the plasma membrane fraction (pellet) from the cytosolic compartment (supernatant). The latter fractions then underwent Western blotting as described above for CD95, GAPDH, and annexin II. GAPDH and annexin II served as markers for the cytosolic and the plasma membrane fraction, respectively. GAPDH was not detectable in the plasma membrane fraction, and annexin II was not detectable in the cytosolic fraction, indicating a high efficacy of separation.
CD95 Translocation to the Plasma Membrane-For determination of membrane surface trafficking of CD95 in primary rat hepatocytes, cells were cultured for 24 h on collagen-coated glass coverslips (diameter, 30 mm) in 6-well culture plates (Falcon). Permeabilized and non-permeabilized cells were stained as published recently (1, 2, 4) using a polyclonal rabbit anti-CD95 antibody (dilution 1:500 in PBS) and a secondary antirabbit Cy3-conjugated antibody. Cells were visualized using an Axioskop (Zeiss, Oberkochen, Germany), and pictures were taken with a 3CCD Camera (Intas, Göttingen, Germany).
Receptor membrane trafficking was defined as the appearance of fluorescent spotting on the surface of the non-permeabilized cells compared with the non-permeabilized control cells (1,2,4). For each condition, at least 100 cells per independent experiment from at least three different cell preparations were scored for CD95 membrane trafficking.
CD95 translocation to the plasma membrane of Huh7 cells and primary rat hepatocytes, respectively, was also studied by detecting total CD95 amount in cytosolic and membrane fractions obtained by ultracentrifugation as described above by use of Western blotting. For determination of membrane surface trafficking of CD95-YFP in Huh7, cells were transfected as described recently (25) and plated on glass-bottom dishes (Mattek, Ashland, MA). 24 h after transfection CD95-YFP was detected in living cells. YFP was excited with 488 nm, and confocal pictures were taken in about 10-min intervals for up to 180 min using the LSM 510 META (Zeiss) at 37°C (25).
Detection of Apoptosis-Terminal deoxynucleotidyltransferase-mediated X-dUTP nick-end labeling of FITC-conjugated deoxyuridine triphosphate (TUNEL) technique was performed as described recently (1). The number of apoptotic cells was determined by counting the percentage of fluorescein-positive cells. At least 100 cells from three different cell preparations were counted for each condition. Cells were visualized on an Axioskop (Zeiss).
Statistics-Results from at least three independent experiments are expressed as means Ϯ S.E. n refers to the number of independent experiments. Results were analyzed using Student's t test; p Ͻ 0,05 was considered statistically significant.
Endosomal acidification involves both the activity of vacuolar type H ϩ ATPase and a chloride conductance in the endosomal membrane, which is required for maintaining electroneutrality during electrogenic proton pumping into the vesicular interior. Evidence has been presented for a direct activation of the H ϩ -ATPase by cytosolic chloride (32,33). To study the effect of CD95L on the cytosolic chloride concentration, rat hepatocytes were loaded for 6 -10 h with the Cl Ϫ -sensitive fluorescent dye MQAE (21). Upon excitation at 350 nm, this dye yields a fluorescence signal, which underlies a diffusion-limited

CD95L-induced changes of apparent pH ves in FITC-dextran-accessible vesicles
For determination of apparent pH ves , 24-h-cultured primary rat hepatocytes were allowed to endocytose FITC-dextran (5 mg/ml) for 60 min. Cells were then exposed to CD95L (100 ng/ml), TLCS (100 mol/liter), or hyperosmolarity (405 mosmol/ liter), and changes in apparent pH ves were detected as described under "Experimental Procedures." When indicated, cells were preincubated for 30 min with DIDS (500 mol/liter), bafilomycin A1 (100 nmol/liter), pan-caspase inhibitor (50 mol/liter Z-VAD-FMK), caspase 8 inhibitor (50 mol/liter IETD-CHO), or caspase-3 inhibitor (50 mol/liter Z-DQMD-FMK). Data represent steady state values after the addition of the respective agents are given as the means Ϯ S.E. and are from at least nine independent measurements for each condition. Cells were obtained from at least three independent preparations. collisional quenching in presence of Cl Ϫ , whereas the affinity of this dye to other physiologically relevant anions (especially HCO 3 Ϫ ) was shown to be minor (21). However, MQAE fluorescence only gives qualitative information on changes of the intracellular chloride concentration, because no ratio technique can be applied (21).

Condition
Under control conditions MQAE-loaded hepatocytes gave a stable fluorescence signal, indicating negligible dye loss during the experimental period. The addition of CD95L resulted in a reversible and dose-dependent decrease of MQAE fluorescence (Fig. 1, A and B), suggestive for an increase in the cytosolic chloride concentration. This effect was bafilomycin-insensitive (Fig. 1C), indicating that endosomal acidification is not the trigger for the presumed CD95L-induced increase of cytosolic chloride concentration. However, after caspase 8 (Fig. 1E) or pan-caspase inhibition (Fig. 1D), but not after caspase 3-inhibition (Fig. 1F), the CD95L-induced decrease of MQAE fluorescence was no longer observed. This suggests that CD95L-induced caspase 8 activation is located upstream of the CD95L-induced increase of the cytosolic chloride concentration.
ASM and Caspase 8 Are Required for CD95 Ligand (CD95L)induced NADPH Oxidase Activation-A CD95L-induced increase of ceramide levels was recently identified as an important upstream event of CD95L-induced NADPH oxidase activation in rat hepatocytes (9). CD95L-induced ceramide formation was no longer found in rat hepatocytes, which underwent protein knock-down of ASM by antisense oligonucleotides, whereas CD95L increased ceramide levels in control hepatocytes treated with nonsense oligonucleotides ( Fig. 2A). As shown in Fig. 2A, ASM knock-down had little effect on basal ceramide levels. One likely explanation is that basal ceramide levels are primarily determined by the activity of sphingomyelinase isoenzymes distinct from ASM. Alternatively, compensatory mechanisms may be induced within the 4 days of ASM knock-down induction, which maintain basal ceramide levels. ASM knock-down also abolished the recently demonstrated (34,35) ceramide-and protein kinase C-dependent CD95L-induced serine phosphorylation of p47 phox , an activating subunit of NADPH oxidase (Fig. 2B), as well as ROS generation (Fig. 2C). These findings indicate that CD95L-induced ceramide formation is due to ASM, whereas other sphingomyelinase isoenzymes probably play a minor role.
Acidic sphingomyelinase is found not only at the plasma membrane but also in intracellular vesicles such as lysosomes (36) and endosomes, which are accessible within 60 min to endocytosed FITC-dextran (9). Therefore, the question was addressed of whether CD95L-induced endosomal acidification accounts for ASM-dependent ceramide formation. ASM has a pH optimum of about 5 (37), and a CD95L-induced lowering of apparent pH ves from 6.0 to 5.7 is accordingly expected to activate the enzyme. In line with this, bafilomycin and DIDS, which prevented CD95L-induced endosomal acidification (Table 1, supplemental Fig. 1), largely abolished the CD95L-induced increase of ceramide levels (Fig. 3A) and p47 phox -serine phosphorylation (Fig. 3B) as well as the CD95L-induced ROS response (Fig. 3C). These data suggest that a CD95L-induced vesicular acidification largely accounts for the stimulation of ceramide formation by ASM (38), which in turn activates NADPH oxidase through a recently reported protein kinase C-dependent phosphorylation of its activating subunit p47 phox (39).
Inhibition of caspase 8 or pan-caspase inhibition, which abolish CD95L-induced endosomal acidification (Table 1), also markedly inhibited the CD95L-induced increase of the intracellular ceramide (Fig. 3A), p47 phox -serine phosphorylation (Fig. 3B), and ROS generation (Fig. 3C), whereas a caspase 3 inhibitor was ineffective (Figs. 3, B and C). On the other hand, induction of p47 phox -serine phosphorylation by the proapoptotic bile acid TLCS or hyperosmolarity was insensitive toward caspase 8 or pan-caspase inhibition (Fig. 3B). These findings indicate that caspase 8 is required for NADPH oxidase activation in response to CD95L but not in response to hydrophobic bile salts or hyperosmolarity.
Apoptosis Induction by CD95L Requires Caspase 8-dependent Endosomal Acidification-CD95L triggers a complex series of events that leads to an activation of the CD95 system and finally results in hepatocyte apoptosis (2,8,9). This sequence of events has been described and characterized in detail recently (for review, see Ref. 5). In rat hepatocytes CD95L induces a ROS response through NADPH oxidase activation, ROS-dependent activation of the Src-kinase family member Yes, and a Yes-dependent activation of the EGFR (2,8,9). These events take place within 1 min and are followed by a JNK-dependent EGFR/CD95 association, subsequent CD95-tyrosine phosphorylation, intracellular CD95 oligomerization, and translocation of the protein complex to the plasma membrane, where formation of the DISC, i.e. recruitment of FADD and caspase 8, occurs (2,8,9). As shown in Figs. 2 and 4, knock-down of ASM in rat hepatocytes not only abolished CD95L-induced p47 phox -serine phos-  3). a.s., antisense. B, CD95Linduced p47 phox -serine phosphorylation. After 96 h significant ASM protein knock-down was achieved in hepatocytes treated with ASM antisense oligonucleotides compared with cells treated with nonsense oligonucleotides. GAPDH expression served as the loading control. p47 phox was immunoprecipitated 1 min after CD95L addition and detected for serine phosphorylation by Western blotting. Total p47 phox served as a loading control. ASM knockdown prevents the CD95L (100 ng/ml)-induced increase in p47 phox -serine phosphorylation (n ϭ 3). C, CD95L-induced ROS-response. In all experiments the CD95L-induced increase of DCFDA fluorescence was measured 1 min after CD95L addition (100 ng/ml). Fluorescence obtained in untreated control cells without CD95L addition was arbitrarily set to 1. In cells cultured in control medium or medium containing nonsense oligonucleotides CD95L induced a rapid ROS response compared with untreated control cells that was significantly inhibited after ASM protein knock-down (*, p Ͻ 0.05) (n ϭ 3). phorylation (Fig. 2B) but also Yes, EGFR and JNK activation, EGFR/ CD95 association, CD95-tyrosine phosphorylation, and DISC formation in response to CD95L (Fig. 4A). In addition, ASM knock-down significantly blunted CD95L-induced hepatocyte apoptosis, as detected by TUNEL reactivity (Fig. 4B). CD95Linduced Yes and EGFR activation, EGFR/CD95 association, CD95-tyrosine phosphorylation, and DISC formation as well as CD95L-induced ROS-formation were also strongly blunted in presence of bafilomycin or DIDS (Fig. 3C and  5A). These findings indicate that endosomal acidification is required for CD95L-induced activation of the CD95 system and apoptosis induction. In line with this, DIDS and bafilomycin strongly blunted CD95L-induced hepatocyte apoptosis, as detected by TUNEL staining (Fig. 5B). However, both DIDS and bafilomycin by themselves exerted some proapoptotic effects (Fig. 5B), as also reported by others (40 -42).
Caspase 8 and pan-caspase, but not caspase 3 inhibitors, not only prevented CD95L-induced ROS formation (Fig. 3C) but also the downstream events leading to CD95 activation and DISC formation, i.e. Yes, EGFR, and JNK activation, EGFR/CD95 association, CD95-tyrosine phosphorylation, and DISC formation (Fig. 5A). These data suggest that an initial caspase 8-activation is required for the full activation of the CD95 system by CD95L in the sense of autoamplification. In line with this, caspase 8-or pancaspase-, but not caspase 3 inhibition, also prevented the translocation of intracellularly stored CD95 to the plasma membrane (Table 2) and the CD95L-induced trafficking of CD95-YFP in transfected Huh7 human hepatoma cells (Fig. 6, A and  B; see also the supplemental film). Caspase 8 dependence of CD95-induced CD95 translocation to the plasma membrane was also confirmed in a biochemical approach, i.e. in subfractionation studies on Huh7 cells (Fig. 6B) and primary rat  Ͻ 0.05; n ϭ 4). B, CD95L-induced p47 phox -serine phosphorylation was sensitive to DIDS, bafilomycin A1, and pan-caspase and caspase 8 inhibition, whereas inhibition of caspase 3 had no effect on CD95L-induced p47 phox -serine phosphorylation (n ϭ 3). In contrast, hyperosmotic and TLCS-induced p47 phoxserine phosphorylation was unaffected by caspase 8 or 3 or pan-caspase inhibition (n ϭ 3). C, the CD95Linduced increase of DCFDA-fluorescence was measured after 1 min of CD95L addition. Fluorescence obtained in cells that were exposed to control medium was arbitrarily set to 1. CD95L induced a rapid ROS response compared with control cells that was significantly inhibited by DIDS and bafilomycin A1 (n ϭ 3) as well as pan-caspase and caspase 8 inhibition (*, p Ͻ 0.05) (n ϭ 6), whereas inhibition of caspase 3 had no effect (n.s., not significant) (n ϭ 3). JANUARY 25, 2008 • VOLUME 283 • NUMBER 4 hepatocytes (Fig. 6C). On the other hand, pan-caspase inhibition was without effect on the TLCS-or hyperosmolarity-induced CD95 trafficking to the plasma membrane (Table 2, Fig. 6).

DISCUSSION
CD95L-induced Endosomal Acidification-As shown in the present study, CD95L induces a rapid endosomal acidification in rat hepatocytes that is also observed in response to proapoptotic bile salts (13) or hyperosmolarity (11). Endosomal acidification is triggered by the vacuolar type H ϩ -ATPase and requires an anion conductance to maintain electroneutrality during electrogenic proton pumping into the vesicular interior (for review, see Ref. 38). Evidence is presented here that CD95L may interfere with the activity of H ϩ -ATPase and/or chloride transport since DIDS and bafilomycin largely abolished the CD95L-mediated effect on apparent pH ves , i.e. vesicular acidification. However, it should be kept in mind that bafilomycin may also affect the pH in vesicular compartments distinct from those accessible to FITC-dextran. As shown in the present study, like proapoptotic bile acids, CD95L also induced a rapid decrease of MQAE fluorescence. This suggests an increase in the cytosolic chloride concentration, which is known to activate the vesicular proton pump and may thereby trigger endosomal acidification (32,33). Chloride channels in the plasma membrane play a major role in hepatocyte cell volume regulation and have been characterized at the functional, pharmacological, and molecular level (for review, see Ref. 43). In hepatocytes, under physiological conditions, intracellular chloride is at electrochemical equilibrium due to the high chloride conductance of the hepatocellular plasma membrane (44). Increases of the intracellular chloride concentration are expected to occur in response to a depolarization of the plasma membrane potential. Indeed, activation of CD95 receptors in myeloid U-937 cells by agonistic CD95 antibodies induces membrane depolarization (45) as do bile salts through electrogenic Na ϩ /bile salt cotransport (46) and/or activation of anion and cation channels (47,48). Thus, it is likely that a depolarization-induced increase in the intracellular chloride concentration could directly activate the vesicular proton pump and thereby trigger endosomal acidification in response to CD95L (32,33). Interestingly, both the CD95L-induced increase of [Cl Ϫ ] and endosomal acidification were abolished in presence of a caspase 8 inhibitor, whereas caspase 3-inhibition was ineffective. This suggests that caspase 8 activation is upstream of chloride fluxes and endosomal acidification. In line with this, CD95 activation-induced depolarization of the plasma membrane was reported to be caspase-dependent in myeloid U-937 cells (45). The mechanisms linking caspase 8-activation to chloride fluxes and endosomal acidification remain unclear; however, multiple protein substrates for caspases have been described (49,50). It is an interesting speculation whether CD95L could also target intracellular chloride channels of the ClC-type chloride channel family (ClC3-7), which have been reported to be expressed in intracellular vesicles, including endosomes (for review, see Refs. 43 and 51) and play a role in hepatocellular endosomal acidification (52).
Endosomal Acidification and CD95 Activation-The present data show that CD95L-induced endosomal acidification is a  Fig. 3B). Cells were then exposed to control medium or CD95L (100 ng/ml). A, CD95 activation. Yes, EGFR, and CD95 were immunoprecipitated as described under "Experimental Procedures" and analyzed by Western blotting. Activating Yes-Tyr 418 -phosphorylation and EGFR-tyrosine phosphorylation (EGFR-Tyr-P) were detected 1 min after CD95L addition. EGFR/CD95 association and CD95-tyrosine phosphorylation (CD95-Tyr-P) were detected after 60 min of CD95L exposure and caspase 8/CD95 and FADD/CD95 association (i.e. DISC formation) 3 h after CD95L addition. Total Yes, EGFR, and CD95 served as respective loading controls. CD95L-induced JNK1 activation was measured 30 min after CD95L exposure by use of phosphospecific antibodies. Total JNK1 served as loading control. ASM knockdown largely prevents CD95L-induced Yes and EGFR phosphorylation (stimulation for 1min), JNK activation (stimulation for 30 min), EGFR/CD95 association (stimulation for 60 min), subsequent CD95-tyrosine phosphorylation (detected after 60 min of CD95L exposure), and DISC formation (detected after 3 h of CD95L stimulation) (n ϭ 3). B, apoptosis. The percentage of CD95L-induced apoptotic cells was determined using the TUNEL staining as described under "Experimental Procedures." CD95L exposure for 12 h induced a significant increase in the percentage of apoptotic hepatocytes (p Ͻ 0.05), which is significantly blunted after ASM knock-down (*, p Ͻ 0.05) (n ϭ 3). major upstream event for CD95L-induced ceramide formation, NADPH oxidase activation, and activation of the CD95 system. All these processes were sensitive to bafilomycin and DIDS, which prevented CD95L-induced endosomal acidification and ceramide formation. The latter is most likely due to an activation of endosomal ASM (11) and apparently does not involve other sphingomyelinase isoenzymes. This is in line with the reported role of ASM in CD95-mediated apoptosis in vivo (53) and of ceramide and sphingolipids in apoptosis regulation (for review, see Ref. 54). In view of a pH optimum of ASM around 5, a CD95L-induced shift of the apparent pH ves from 6.0 to 5.7 is expected to increase ASM activity and thereby to augment ceramide formation. As shown recently (9), downstream consequences of CD95Linduced ceramide formation are CD95L-induced NADPH oxidase activation and ROS formation, which mediate an activation of JNK and Src-family kinase Yes (2,9). Activated Yes triggers EGFR activation, which subsequently associates with CD95 and triggers CD95 activation (9). The JNK signal is required for association of the activated EGFR with CD95 (2), which then allows for CD95-tyrosine phosphorylation by the EGFR-tyrosine kinase activity as a prerequisite for the translocation of intracellularly stored CD95 to the plasma membrane, where DISC formation occurs (2,8). All these events were strongly blunted by DIDS, bafilomycin, or ASM FIGURE 5. CD95L-induced activation of the CD95 system and apoptosis are DIDS-, bafilomycin A1-, and caspase 8 inhibitor-sensitive. Hepatocytes were cultured for 24 h and then exposed to control medium or CD95L (100 ng/ml). When indicated, DIDS (500 mol/liter), bafilomycin A1 (100 nmol/liter), pan-caspase (50 mol/liter Z-VAD-FMK), caspase 8 (50 mol/liter IETD-CHO), or caspase 3 inhibitor (50 mol/liter Z-DQMD-FMK) were added 30 min before CD95L addition. A, CD95L-induced activation of the CD95 system and DISC formation. Yes, EGFR, and CD95 were immunoprecipitated as described under "Experimental Procedures" and analyzed by Western blotting. Activating Yes-Tyr 418 -phosphorylation, Yes/EGFR association, and EGFR-tyrosine phosphorylation (EGFR-Tyr-P) were detected 1 min after CD95L exposure, and EGFR/CD95 association and CD95-tyrosine phosphorylation (CD95-Tyr-P) were detected 60 min after CD95L addition and DISC formation, i.e. caspase 8/CD95 and FADD/CD95 association after 3 h of CD95L exposure. Total Yes, EGFR, and CD95 served as respective loading controls. Within 1 min CD95L induced a Yes activation and Yes/EGFR association followed by a Yesmediated EGFR-tyrosine phosphorylation which was sensitive to inhibition by DIDS or bafilomycin A1. Also, EGFR/CD95 association and CD95-tyrosine phosphorylation as well as DISC formation were largely prevented by DIDS and bafilomycin A1 (n ϭ 3). Also, inhibition of caspase 8 or pan-caspase inhibition blunted the otherwise observed CD95L-induced activation of the CD95-system (n ϭ 6), whereas inhibition of caspase 3 was ineffective (n ϭ 3). B, CD95L-induced apoptosis. The percentage of CD95L-induced apoptotic cells was determined using the TUNEL staining as described under "Experimental Procedures." CD95L exposure for 12 h induced a significant increase in the percentage of apoptotic hepatocytes (p Ͻ 0.05). CD95L induced hepatocyte apoptosis, which was sensitive to inhibition by DIDS or bafilomycin A1 (*, p Ͻ 0.05) (n ϭ 5).

Condition
Cells with CD95 membrane staining knock-down and also resulted in an inhibition of CD95L-induced apoptosis, as assessed by the TUNEL-reaction (this study; see also Ref. 13).
Caspase 8-dependent Autoamplification of CD95 Activation-Interestingly, as shown in the present study, caspase 8, but not caspase 3 inhibitors, blocked all events leading to intracellular CD95 activation and translocation to the plasma membrane including DISC formation. These findings suggest that the role of initiator caspase 8 may not only reside in a signaling toward downstream effector caspases but also in an amplification of CD95 activation. Recent immunocytochemical studies on rat hepatocytes under permeabilized and non-permeabilized conditions (1,2) showed that CD95 is located inside the hepatocyte with no detectable CD95 staining at the plasma membrane. This raised the question of how externally added CD95L may get access to its receptor. However, small amounts of CD95 in the plasma membrane may escape immunocytochemical detection. In line with this, in cyan/yellow fluorescent protein-CD95-transfected Huh7 hepatoma cells, small amounts of partly oligomerized CD95 were detectable at the plasma membrane (12). Also in this experimental systems, the addition of CD95L resulted in a strong CD95 targeting to the plasma membrane. The most likely explanation for the present findings is that in resting hepatocytes small, by conventional immunocytochemistry, hardly detectable amounts of CD95 are present at the cell surface, whereas the bulk of CD95 is stored inside the hepatocyte. These "sentinel" receptors in the plasma membrane become activated by externally added CD95L, and the resulting caspase 8 activation may then trigger chloride fluxes, endosomal acidification, and the complex downstream signaling events leading to the CD95 translocation to the plasma membrane and activation. In line with this hypothesis that caspase 8  A and B) and primary rat hepatocytes (C). A, Huh7 cells were transfected with CD95-YFP as described under "Experimental Procedures" resulting in a homogeneous intracellular distribution of the fluorescent CD95-YFP fusion protein when YFP was excited with 488 nm. Huh7 cells were subsequently exposed for at least up to 180 min to CD95L (100 ng/ml), and confocal pictures were taken roughly every 10 min in living cells at 37°C (25). If indicated, pan-caspase (50 mol/liter Z-VAD-FMK), caspase 8-(50 mol/liter IETD-CHO), or caspase 3 inhibitor (50 mol/liter Z-DQMD-FMK) were added 30 min before CD95L addition. Representative samples at time t ϭ 0 min (i.e. immediately before CD95L addition) (left side) and t ϭ 180 min (CD95L) or t ϭ 210 min (CD95L ϩ respective inhibitor) (right side) from at least three independent experiments are shown. To view the entire video, see the supplemental material. The addition of CD95L results within 180 min in a translocation of the CD95-YFP construct to the plasma membrane, which was sensitive to pan-caspase or caspase 8 inhibition, whereas inhibition of caspase 3 was ineffective. B and C, Huh7 cells (B) or primary rat hepatocytes (C), respectively, were exposed for 3 h to CD95L (100 ng/ml). When indicated, pan-caspase inhibitor (50 mol/liter Z-VAD-FMK), caspase 8 inhibitor (50 mol/liter IETD-CHO), or caspase-3 inhibitor (50 mol/ liter Z-DQMD-FMK) were added 30 min before CD95L addition. In another set of experiments the respective inhibitors were given alone for 3.5 h. To detect CD95 translocation to the plasma membrane, the respective samples were transferred to ultracentrifugation as described under "Experimental Procedures." The total amount of CD95-YFP (B) or CD95 (C), respectively, in the membrane and cytosolic fraction was analyzed by Western blotting. GAPDH and annexin II served as markers for the cytosolic and the plasma membrane fraction, respectively. GAPDH was not detectable in the plasma membrane fraction, and annexin II was not detectable in the cytosolic fraction, indicating a high efficacy of separation. The addition of CD95L results within 180 min in a translocation of the CD95-YFP (B) or CD95 (C), respectively, to the plasma membrane in both Huh 7 cells (B) and primary rat hepatocytes (C), which was sensitive to pan-caspase or caspase 8 inhibition, whereas inhibition of caspase 3 was ineffective.
amplifies CD95 activation would be that activation of only 2% of total caspase 8 by CD95-activating antibodies was sufficient to induce a maximal ASM activation (55). Furthermore, CD95-dependent ceramide formation in HeLa and 293T cells was reported to require active caspase 8 (56). Also in Jurkat T (55,57) and glioma cells (58), CD95-mediated caspase activation was located upstream of ASM activation and ROS generation, whereas CD95-independent ASM activation did not require caspase 8 activity (55). The abovementioned hypothesis is also corroborated by the finding that CD95-independent endosomal acidification and subsequent NADPH oxidase activation, when induced by proapoptotic bile salts or hyperosmolarity, were not affected by caspase 8 nor pan-caspase inhibitors.
Taken together, the present results provide a mechanistic link between chloride fluxes, endosomal acidification, ceramide formation, and the recruitment and activation of intracellularly stored CD95 in response to external CD95L. All these processes are dependent on caspase 8 activity. This suggests a crucial role of caspase 8 as an upstream event in an autoamplifying CD95 activation. However, the mechanisms linking caspase 8 activation to cellular ion homeostasis remain to be elucidated.