BLOOD, 1 JULY 2003 ⅐ VOLUME 102, NUMBER 1 VITAMIN C INHIBITS FAS-INDUCED APOPTOSIS 337RPMI containing 100 U penicillin, 100 U streptomycin (Gemini Bioprod- cell samples were incubated with several concentrations of DHA from aucts, Woodland, CA), L-glutamate, and 10% fetal calf serum (Omega solution containing 0.5 Ci (0.0185 MBq) L-14C-AA (speciﬁc activity 8.0Scientiﬁc, CA). Human monocytes were isolated from peripheral blood mCi/mmol (296 MBq/mmol); Perkin-Elmer Life Sciences, MA), 1.0 to 4.0obtained from the New York Blood Center (NY). The following day, mM AA (Sigma), and 86 U/mL ascorbate oxidase (Sigma) in incubationperipheral blood mononuclear cells were separated by Ficoll gradient buffer (15.0 mM HEPES [N-2-hydroxyethylpiperazine-NЈ-2-ethanesul-(Accu-Prep; Accurate Chemicals and Scientiﬁc, Westbury, NY). Monocytes fonic acid], 135.0 mM NaCl, 5.0 mM KCl, 1.8 mM CaCl2, and 0.8 mMwere immediately puriﬁed by negative selection using the magnetic- MgCl2 [pH 7.4]). Cells were then washed twice with cold PBS and lysedactivated cell separation monocyte isolation kit (Miltenyi Biotec, Auburn, with 0.5 mL sodium dodecyl sulfate (SDS)–lysis solution (10.0 mM TrisCA) following the manufacturer’s instructions. Monocytes were incubated [tris(hydroxymethyl)aminomethane]–HCl [pH 8.0] and 0.2% SDS). Cell-in RPMI containing 100 U penicillin, 100 U streptomycin, L-glutamate, and associated radioactivity was quantiﬁed by scintillation spectrometry. To20% human AB serum (Irvine Scientiﬁc, Santa Ana, CA). Purity was determine the accumulation of vitamin C supplied as AA, cells wereassessed by staining with anti-CD3 (RPE-Cy5; DAKO, Glostrup, Den- incubated with varying concentrations of AA from a solution prepared withmark), anti-CD56 (phycoerythrin [PE]; Becton Dickinson, San Diego, CA), 0.5 Ci (0.0185 MBq) L-14C-AA, 1.0 to 4.0 mM L-AA, and 0.1 mManti-CD20 (peridinin chlorophyll protein [PerCP]; Becton Dickinson), and 1,4-dithiothreitol in incubation buffer and cell-associated radioactivityanti-CD14 (ﬂuorescein isothiocyanate [FITC]; Becton Dickinson) ﬂuores- measured by scintillation spectrometry. Additionally, accumulation ofcent antibodies. The cells were suspended in staining buffer consisting of vitamin C supplied as DHA was determined by high-performance liquid1% heat-inactivated fetal calf serum and 0.1% (wt/vol) sodium azide chromatography (HPLC) electrochemical detection (ESA, Chelmsford,(Sigma, St Louis, MO) in phosphate-buffered saline (PBS), pH 7.4, MA).27 Cells incubated with DHA were washed twice in cold PBS and thencontaining the ﬂuorescent antibody (Ab) for 30 minutes at 4°C. After lysed in 60% methanol. Lysates were centrifuged, the supernatants ﬁlteredstaining, the cells were washed with buffer, ﬁxed in 1% paraformaldehyde through 0.2-m nylon membrane ﬁlters (Nalgene, Rochester, NY) andand analyzed by ﬂow cytometry (FACScalibur; Becton Dickinson) within equal amounts of each sample injected onto a modiﬁed C18 column (no.24 hours of staining. The results were analyzed using the CELLQuest 70-4160; ESA). The column was equilibrated at 30°C with a mobile phasesoftware (Becton Dickinson). Monocyte preparations were always higher buffer consisting of 50 M sodium phosphate, 50 M sodium acetate, 189than 95% purity. M dodecyltrimethylammonium chloride, and 3.66 M tetraoctylammo- nium bromide in 30:70 methanol-water (vol/vol, HPLC grade; J. T. Baker,Detection of CD95 Phillipsburg, NJ), pH 4.8. The ﬂow rate was 0.3 mL/min. Under these conditions, the retention time of AA was 4.9 minutes. Peak areas for AAThe antihuman CD95-RPE Ab (DAKO) was used to detect the CD95 were determined using a dominant potential for AA of 100 mV, and AA wasantigen in cell lines and monocytes by ﬂow cytometry. Cell-surface staining quantiﬁed from a calibration curve.and analysis were performed as described in the previous section. Cell volume determinationInduction of apoptosis Estimation of cell volume was performed as previously described.28 Brieﬂy,An anti-FAS Ab and soluble FAS ligand (FASL) were used to induce 5 ϫ 106 cells were incubated for 60 minutes at RT in 200 L incubationapoptosis. For the induction of apoptosis in monocytes, an anti-FAS Ab buffer containing 1.0 mM 3-oxy-methyl-D-glucose and 5 Ci (0.185 MBq)(immunoglobulin G1 [IgG1], clone DX2; R and D Systems, Minneapolis, 3H-3-oxy-methyl-D-glucose. Uptake was stopped by adding 2 L of 2.0MN) or an isotype control Ab (R and D Systems) were plate-immobilized in mM cytochalasin B (Sigma), which blocks glucose transport and preventsa 12–ﬂat well plate overnight. The plate was washed twice in serum-free the efﬂux of trapped glucose. The cells were then washed twice in ice-coldRPMI and blocked with RPMI containing 10% human type AB serum for PBS containing 20 M cytochalasin B and cell-associated radioactivity30 minutes at 37°C. The plate was then washed and the cells were added determined by scintillation spectrometry. Glucose reached equilibrium afterimmediately. For induction of apoptosis in U937 cells, 10 L protein G (2.5 60 minutes of incubation, therefore the amount of radioactivity accumu-mg/mL suspension in agarose beads; Boehringer Mannheim, Mannheim, lated inside the cells is in direct proportion to the intracellular volume. TheGermany) per g anti-FAS Ab or control Ab was mixed in cold serum-free cell volume estimated for monocytes was 0.15 L per 106 cells and 1.0 LRPMI for 5 minutes at room temperature (RT) before addition to the cell per 106 U937 cells.suspension. To induce apoptosis with soluble FASL, Super FASL (AlexisBiochemicals, Lausen, Switzerland) was added to U937 cells or to freshly Vitamin C loadingisolated monocytes. Cells treated with anti-FAS Ab, FASL, or control cellswere incubated at 37°C in an environment of 5% CO2 in air. Cells were washed twice in incubation buffer at pH 7.4, and DHA (Aldrich, Milwaukee, WI) was added. After incubation with DHA for up to 60Detection of apoptosis minutes at 37°C, the cells were washed twice in serum-free RPMI.The annexin V–FITC/propidium iodine (PI) apoptosis detection kit (Pharm- Detection of ROSingen, San Diego, CA) was used to determine the frequency of apoptosis inU937 cells, following the manufacturer’s instructions. Alternatively, PI Intracellular ROS in U937 were estimated by 2Ј7Ј dichloroﬂuoresceinstaining of DNA was used to analyze the sub-G1 region by ﬂow cytometry. diacetate (DCFH-DA) ﬂuorescence.8 Cells were washed twice in Krebs-For PI staining, cells were washed twice with PBS and ﬁxed with 70% Ringer buffer (20.0 mM HEPES, 10.0 mM dextrose, 127.0 mM NaCl, 5.5ethanol for at least 30 minutes at 4°C. Cells were then washed twice with mM KCl, 1.0 mM CaCl2, and 2.0 mM MgSO4 [pH 7.4]) and stained withPBS and resuspended in 50 L of a 100 U/mL RNAse-A solution 0.1 g/mL DCFH-DA (Molecular Probes, Eugene, OR) and 1.0 mM maleic(Boehringer Mannheim) and 20 L PI (0.1 mg/mL; Sigma) per 1 ϫ 105 acid diethyl ester (Sigma) in Krebs-Ringer buffer. Fluorescence wascells for 30 minutes at RT in the dark and kept at 4°C until ﬂow cytometry determined by ﬂow cytometry after incubation at 37°C, and the data wereacquisition. Apoptosis in monocytes was determined by analysis of the analyzed using CELLQuest software.sub-G1 region by DNA staining and ﬂow cytometry. For PI staining, themonocytes were incubated with trypsin for 5 minutes at 37°C, washed twice Assessment of mitochondrial membrane potential (⌬)with PBS, and then stained as described in the previous section. U937 cells were washed twice with PBS, stained with 40 nM 3,3Ј- dihexyloxacarbocyanine iodide (DiOC(6)(3); Molecular Probes) with orVitamin C uptake without 100 M carbonyl cyanide m-chlorophenylhydrazone (Sigma) as aVitamin C uptake was measured as intracellular accumulation after control for membrane potential disruption, and incubated at 37°C in theincubation of cells with DHA or AA at 37°C, as previously described.23,24 dark for 15 minutes before determination of ﬂuorescence by ﬂow cytom-To determine the accumulation of vitamin C supplied as DHA, triplicate etry. These experiments and the ROS quenching studies were not performed
338 PEREZ-CRUZ et al BLOOD, 1 JULY 2003 ⅐ VOLUME 102, NUMBER 1in monocytes because cell handling and the required trypsinization led toerratic results in these assays. ResultsCytochrome C release Vitamin C inhibits FAS-induced apoptosisRelease of CytC into the cytosol was detected with a CytC enzyme-linked FAS-R (CD95) was expressed in U937 cells and monocytes, andimmunosorbent assay (ELISA) kit (MLB, Nagoya, Japan), following the both anti-FAS Ab and FASL induced apoptosis in these cells in amanufacturer’s instructions. To obtain cytosolic fractions, cells were dose-dependent manner (Figure 1). Incubation of U937 cells for 3washed twice with ice-cold PBS, resuspended in cold buffer (1.28 M NaCl, hours with FASL induced up to a 3.5-fold increase in apoptosis,50.0 mM KCl, 50.0 mM MgSO4, 13.0 mM CaCl2, 0.5 M HEPES, 1.0 mM whereas incubation with anti-FAS Ab caused a 2-fold increasephenylmethylsulfonyl ﬂuoride, 10% vol 2.5 M sucrose, 10.0 mM 1,4- (Figure 1B-C). In monocytes, 3-hour treatment with FASL resulteddithiothreitol, 1.0 mM reduced glutathione, and 1% glycerol) and homoge-nized mechanically. This total cellular extract was centrifuged at 2000 rpm in up to a 7-fold increase in apoptosis, whereas incubation withfor 3 minutes at 4°C and the cytosolic supernatant recovered and analyzed anti-FAS Ab resulted in a 4-fold increase (Figure 1E-F). An isotypefor the presence of cytochrome C. control Ab did not cause apoptosis in either cell type (data not shown). FASL was more efﬁcient than anti-FAS Ab in inducing apoptosis in both U937 cells and monocytes (Figure 1).Caspase assays We investigated the effect of loading cells with vitamin C onCaspase activity was measured using the caspase-3 (Sigma), caspase-8, or FAS-R–dependent apoptosis. U937 cells and monocytes transportcaspase-10 (R and D Systems) colorimetric assay kits, following the vitamin C in the form of DHA through facilitative glucosemanufacturer’s instructions. Brieﬂy, 1 ϫ 107 cells were lysed and equal transporters.29,30 As measured by 14C-substrate uptake, there wasquantities of protein loaded in MaxiSorp 96-well plates (Nunc, Roskilde, little evidence for direct transport of AA, whereas DHA was readilyDenmark). The caspase-3 inhibitor Ac-Asp-Glu-Val-Asp-aldehyde (Sigma) taken up by the cells (Figure 2A-D). We estimated the internalor caspase-8 inhibitor benzylloxycarbonyl-Ileu-Glu-Thr-Asp-ﬂuoromethyl volume of U937 cells at 1.0 L/106 cells and monocytes atketone (Z-IEDT-FMK; BioVision Research Products, CA) was added as acontrol at 20-M ﬁnal concentration. The plate was incubated at 37°C 0.15 L/106 cells from tritiated methylglucose equilibrium stud-and color development quantiﬁed by reading with an ELISA plate reader ies.23,28 Based on this internal volume and cell-associated radioac-at 405 nm. tivity, the intracellular accumulation of AA in U937 cells and monocytes was estimated. The intracellular AA concentration calculated using this method was higher than values determined byRecombinant caspase-8 in vitro assay HPLC electrochemical detection (Figure 2A,D). The small accumu-Recombinant caspase-8 (Alexis Biochemicals) was incubated with different lation of vitamin C in cells incubated with AA was probably due toconcentrations of AA and activity measured using the caspase-8 assay kit (R oxidation of AA to DHA in air.and D Systems). Background was determined from triplicates of each The frequency of apoptosis induced by anti-FAS Ab or FASL inconcentration of vitamin C without the enzyme and subtracted from the U937 cells and monocytes was prominently reduced by cellularexperimental results. The results were expressed as the percentage of loading with vitamin C (Figure 2). Data regarding the intracellularcaspase activity in the absence of vitamin C. AA concentrations reported in these experiments were based on HPLC electrochemical detection. The frequency of apoptosisCaspase-8 Western blotCells were lysed in a buffer containing 30 mM Tris-HCl (pH 7.5), 150mM NaCl, 10% glycerol, and 1% triton and protease inhibitors. Celllysates were separated by SDS–polyacrylamide gel electrophoresis onprecast Tris-HEPES-SDS-polyacrylamide gradient minigels (4%-20%;Gradipore, Frenchs Forest, Australia) according to the manufacturer’sprotocols. The gels were then transferred onto a nitrocellulose mem-brane (Bio-rad Laboratories, Hercules, CA) in a buffer containing 25mM Tris (pH 8.3), 192 mM glycine, and 20% methanol in a semidrysystem (Bio-rad) for 15 minutes (15 V) at 25°C. Residual binding siteson the membrane were blocked by incubating the membrane inTris-buffered saline (TBS; 20 mM Tris-HCl [pH 7.4] and 0.15 M NaCl),containing 5% nonfat dry milk and 0.1% Tween-20, for 2 hours at 25°C(blocking buffer). Membranes were incubated with monoclonal anti–caspase-8 (12F5; Alexis Biochemicals) or anti-␤tubulin (H-235; SantaCruz Biotechnology, Santa Cruz, CA) Abs (1:1000 dilution) for 16 hoursat 4°C. The primary Ab was removed, and the blots were washed 3 timesin TBS. To detect Ab reaction, the blots were incubated for one hourwith anti–mouse or anti–rabbit IgG horseradish peroxidase (HRP)–conjugated Abs (Bio-rad), diluted 1:3000 in blocking buffer, washedwith TBS, and the protein bands were revealed using the enhanced Figure 1. FAS-R ligation induces apoptosis in U937 cells and in monocytes. (A)chemiluminescence (ECLϩplus) Western blotting detection system U937 cells express the FAS-R (CD95) as determined by ﬂow cytometry. (B-C) FASL(Amersham Pharmacia Biotech, Piscataway, NJ) before exposure to and anti-FAS Ab induce apoptosis in U937 cells in a dose-dependent manner. (D)BioMax ﬁlm (Kodak, Rochester, NY). Fresh human monocytes express CD95 as determined by ﬂow cytometry. (E-F) FASL and anti-FAS Ab induce apoptosis in monocytes in a dose-dependent manner. The frequency of apoptosis for U937 cells was determined by annexin Vϩ/PIϪ staining;Statistical analysis apoptosis in monocytes was determined by the frequency of events in the sub-G1 region. The results are presented as the fold increase in the frequency of apoptosisThe paired 1-tail distribution Student t test was applied to compare results over control cells left untreated for 3 hours and are representative of at least 5with a criterion for statistical signiﬁcance of a P value of .05 or less. independent experiments for each cell type.
BLOOD, 1 JULY 2003 ⅐ VOLUME 102, NUMBER 1 VITAMIN C INHIBITS FAS-INDUCED APOPTOSIS 339Figure 2. Vitamin C decreases FAS-induced apopto-sis in U937 cells and monocytes. (A) Intracellularaccumulation of vitamin C in U937 cells after exposure to1.0 mM DHA or AA. Results are presented as theintracellular concentration of accumulated AA, deter-mined by cell-associated radioactivity. The standard de-viation is smaller than the symbols in the graphs, and wasalways less than 10% of the values for each point. (B-C)U937 cells loaded with vitamin C were treated for 3 hourswith 200 ng/mL FASL or 1 g/mL anti-FAS Ab, andapoptosis was measured by annexin Vϩ/PIϪ staining orby analysis of the sub-G1 region. The intracellular accu-mulation of AA was estimated by HPLC. (D) Accumula-tion of AA in monocytes after exposure to 0.1 mM DHA orAA, calculated by cell-associated radioactivity. (E-F)Monocytes were loaded with different amounts of vitaminC estimated by HPLC analysis before treatment with 100ng/mL FASL or 0.1 g/mL anti-FAS Ab for 3 hours tomeasure apoptosis by analysis of the sub-G1 region.Asterisks indicate statistically signiﬁcant differences(P Յ .05) using the Student t test between control andcells loaded with vitamin C before challenge with FASL.These experiments were repeated 6 times for each celltype with similar results. A representative experimentis shown. Error bars represent the SD of triplicate values.induced by FASL (200 ng/mL) in U937 cells was 12% and in the levels of ROS, which were quenched by loading cells withdecreased to 7.8% in cells loaded with approximately 13 mM vitamin C (Figure 3B). The production of ROS correlated with thevitamin C (P ϭ .01) (Figure 2B). U937 cells incubated 3 hours frequency of apoptosis in cells incubated with FASL in thesewith 1 g/mL anti-FAS Ab had a frequency of apoptosis of 5.8%, experiments (data not shown).which decreased to 4.0% when the cells were loaded with Vitamin C reduces mitochondrial damage induced byapproximately 13 mM vitamin C (P ϭ .03) (Figure 2C). The FAS-R ligationpercentage of reduction in apoptosis for U937 cells loaded with 13mM vitamin C compared with control ranged from 22% to 27% In apoptotic cells, the decrease in mitochondrial membrane differ-(mean, 25%) when the cells were challenged with anti-Fas Ab and ential potential (⌬) occurs before chromatin condensation andfrom 26% to 50% (mean 35%) when the cells were treated with DNA fragmentation,36 and precedes the enhanced release of ROSFASL. Vitamin C loading did not affect the expression of CD95 in by the mitochondria.7 Since vitamin C prevented the increase inthese cells (data not shown). Monocytes loaded with vitamin C also cellular ROS levels induced by FAS ligation, we studied the effecthad a reduced frequency of apoptosis after treatment with anti-FAS of vitamin C on mitochondrial apoptosis in U937 cells culturedAb or FASL. FASL (100 ng/mL) induced 37.1% apoptosis in with FASL by using the dye DiOC(6)(3). Cells with healthymonocytes, and when these cells were loaded with approximately mitochondria have a high uptake of DiOC(6)(3) and therefore a high14 mM vitamin C the frequency was reduced to 20.0% (P ϭ .04) ﬂuorescence signal, whereas cells with damaged mitochondria(Figure 2E). Monocytes treated with 0.1 g/mL anti-FAS Ab for 3 show reduced ﬂuorescence. Only 1.1% of control U937 cells hadhours showed 7.2% apoptosis, which was reduced to 5.1% in cells dull DiOC(6)(3) ﬂuorescence (Figure 4A, upper left panel), whereasloaded with approximately 14 mM vitamin C (P ϭ .047) (Figure2F). This represents a mean reduction in apoptosis of 31% to44%. The lowest intracellular concentration of AA used in theseexperiments was 4 mM, which approximates the concentrationof AA reported in monocytes obtained from blood of healthydonors (3 mM).31,32 AA (4 mM) conferred signiﬁcant protectionfrom FAS-induced apoptosis in monocytes (P ϭ .02)(Figure 2E-F).Vitamin C quenches ROS produced by FAS-R ligationROS levels are elevated in cells stimulated via the FAS pathway,and it has been suggested that ROS are an important component ofFAS signaling.8,33,34 Since vitamin C is a potent antioxidant and Figure 3. Vitamin C inhibits ROS generated by FAS-R ligation. (A) U937 cellsquenches ROS in cells under oxidative stress,35 we examined the were loaded with approximately 13 mM vitamin C prior to treatment with 1 g/mLeffect of vitamin C on ROS accumulation after FAS ligation. anti-FAS Ab for 3.5 hours and then stained with DCFH-DA to detect intracellular ROSIncubation of U937 cells with anti-FAS Ab for 3.5 hours resulted in by detection of dye ﬂuorescence within 90 minutes by ﬂow cytometry. (B) U937 cells were loaded with vitamin C, treated with 200 ng/mL FASL for 3 hours, and stainedapproximately a 2.5-fold increase in cellular ROS, whereas loading with DCFH-DA for 30 minutes for ﬂuorescence measurement. The results arethe cells with vitamin C reduced ROS accumulation in the cells presented as the mean ﬂuorescence intensity of DCFH-DA in arbitrary ﬂuorescence(Figure 3A). FASL incubation for 3 hours induced a 2-fold increase units and are representative of 5 independent experiments.
340 PEREZ-CRUZ et al BLOOD, 1 JULY 2003 ⅐ VOLUME 102, NUMBER 1 vitamin C, however, prevented FASL-induced apoptosis in similar proportions (Figure 4B). Apoptotic cells release cytochrome C (Cyt C) from the mitochon- dria into the cytosol,38 and we sought to determine the effect of vitamin C on this process. The amount of Cyt C found in the cytosolic fractions of cells incubated with FASL was nearly twice that found in control cells (Figure 4C). In cells loaded with vitamin C, there was a marked reduction of Cyt C in the cytoplasmic fractions, even below the control levels. These results indicate that vitamin C loading results in inhibition of the release of Cyt C from mitochondria after FAS-R ligation. Vitamin C reduces FAS-induced caspase-3 and caspase-10 activation A consequence of mitochondrial damage during apoptosis is activation of caspase-3,39 and we studied the effect of vitamin C on the activation of this caspase. Activation of caspase-3 was induced by FASL incubation in U937 cells and in monocytes (Figure 5A-B). The activity of caspase-3 induced by FASL in U937 cells loaded with vitamin C was modestly reduced (30%), although the difference in caspase activity of cells with or without vitamin C was statistically signiﬁcant (Figure 5A, P ϭ .039). In contrast, caspase-3 activity induced by FASL in monocytes loaded with vitamin C was markedly reduced (Figure 5B, P ϭ .02). The results suggested that caspase-3 in U937 could be activated by pathways independent of mitochondrial signaling, and we therefore analyzed the effect of vitamin C on alternative mecha- nisms for caspase-3 activation in U937 cells. Caspase-10 isFigure 4. Vitamin C prevents mitochondrial damage induced by FAS-R ligation. homologous to caspase-8 and is activated upstream in the apoptosis(A) U937 cells were loaded with 13 mM vitamin C prior to 3 hours of treatment withFASL (200 ng/mL) and were stained with 40 nM DiOC(6)(3) for 15 minutes at 37°C. cascade and can activate caspase-3.40 We explored the effect ofControl cells were treated with 0.2 M Z-IEDT-FMK. DiOC(6)(3) ﬂuorescence was vitamin C on caspase-10 activity in U937 cells treated with FASL.analyzed by ﬂow cytometry. The intensity ﬂuorescence of DiOC(6)(3) is shown on the The activity of caspase-10 increased 36% after incubation withx-axis, and the forward scatter is represented on the y-axis. The numbers in the ﬁgureindicate the frequency of the population with low intensity of DiOC(6)(3) ﬂuorescence FASL and was reduced by about 50% when the cells were loadedfor each treatment and are representative of 7 independent experiments (except for with vitamin C (Figure 6). We did not detect caspase-10 activity inthe experiment with Z-IEDT-FMK, which was performed twice). (B) Cells cultured monocytes after treatment with FASL (data not shown), and to ourunder the conditions described in panel A were stained with annexin-PI to assess knowledge caspase-10 activity in monocytes has not been reported.frequency of apoptosis. These results are the mean of 3 independent experiments,performed in duplicates, and are presented as the fold increase in apoptosis over Vitamin C inhibits caspase-8 activation induced via FASLuntreated control cells. (C) U937 cells were loaded with approximately 13 mM vitaminC prior to treatment with 200 ng/mL FASL. Cytosolic cell extracts were obtained andanalyzed by ELISA for the presence of Cyt C. The results show the normalized Cyt C ROS production, mitochondrial damage, and Cyt C release arecontent for each experimental condition in arbitrary units based on the Cyt C cytosolic downstream consequences of initiator caspases such as caspase-content of control cells and represent 2 independent experiments. Asterisk indicates 8.37 We therefore addressed the hypothesis that vitamin C inhibitedstatistically signiﬁcant differences (P ϭ .039) using the Student t test between control FAS-dependent caspase-8 activation. In U937 cells incubated withand cells loaded with vitamin C before challenge with FASL. Error bars represent theSD of triplicate values.42% of U937 cells cultured with FASL for 3 hours incorporatedlittle DiOC(6)(3) (Figure 4A, upper right panel). We found thatloading the cells with vitamin C conferred protection fromFASL-induced mitochondrial damage, and these cells maintainedhigh DiOC(6)(3) ﬂuorescence (Figure 4A, lower left panel). Controlcells treated with cyanide m-chlorophenylhydrazone showed amarked decrease in mitochondrial ⌬ (data not shown). Caspase-8activation precedes the reduction in mitochondrial ⌬ and thechange in mitochondrial membrane permeability.37 Thus, themitochondrial damage and the mitochondrial collapse should be Figure 5. Effect of vitamin C on caspase-3 activity. (A) U937 cells loaded withprevented by incubation with the speciﬁc caspase-8 inhibitor approximately 13 mM vitamin C prior to treatment with FASL (200 ng/mL) were lysed, and the activity of caspase-3 in the cell lysates was determined as described inZ-IETD-FMK. The addition of 0.2 M Z-IETD-FMK inhibited “Materials and methods.” (B) Monocytes loaded with approximately 14 mM vitamin Capproximately 60% of FASL-induced mitochondrial ⌬ loss prior to treatment with FASL (100 ng/mL) were lysed, and the activity of caspase-3 incompared with nearly 100% inhibited with vitamin C (Figure 4A, the cell lysates analyzed. The results are presented as the percentage increase in caspase activity compared with cells without treatment and are the mean of triplicatelower panels). We therefore reasoned that vitamin C was inhibiting samples. These results are representative of 3 independent experiments for each cellevents at both the mitochondrial level and upstream sites in the type. Asterisks indicate statistically signiﬁcant differences (in A, P ϭ .03 and in B,apoptosis cascade. The caspase-8 inhibitor Z-IETD-FMK and P ϭ .02) in caspase activity between cells loaded with vitamin C and control.
BLOOD, 1 JULY 2003 ⅐ VOLUME 102, NUMBER 1 VITAMIN C INHIBITS FAS-INDUCED APOPTOSIS 341 activation of caspase-8. The molecular mechanisms of inhibition of the activation of caspase-8 are not known. We postulate that the activity of caspase-8 is dependent on production of ROS at the level of the FAS-R and that the antioxidant properties of vitamin C could inhibit this early step of FAS-initiated apoptosis signaling. There is support for this notion in the recent ﬁnding that preloading HL-60 cells with vitamin C inhibits the processing of pro– caspase-8 induced by hydrogen peroxide.43 Vitamin C also main- tains the integrity of the mitochondrial membrane, which is disrupted during FAS signaling, and inhibits further downstream effects. Thus, the results suggest that vitamin C inhibits FAS-R–induced apoptosis in mononuclear phagocytes because of its antioxidant capabilities.Figure 6. Effect of vitamin C on caspase-10 activity. U937 cells were loaded with Discussion13 mM vitamin C prior to FASL treatment (200 ng/mL) and lysed to determine activityof caspase-10. The results are presented as the percentage increase in caspase Dietary vitamin C is critical for normal host defense, and vitamin Cactivity compared with cells without treatment and are the mean of triplicate samples. is widely believed to enhance immune function when usedThese results are representative of 2 independent experiments. Asterisk indicates astatistically signiﬁcant difference (P ϭ .033) in caspase-10 activity of cells loaded with pharmacologically.18-20 Vitamin C may also have anti-inﬂamma-vitamin C compared with control. tory properties.21,44-46 There is now a substantial body of evidence deﬁning a central role for oxidative reactions in cell signal-FASL there was a 65% increase in caspase-8 activity, while vitamin ing12-15,47,48 and showing that antioxidants can inhibit importantC loading resulted in activity reduced to control levels (Figure 7A). signaling pathways in immune cells.16,17 For example, granulocyte-A caspase-8 Western blot analysis indicated that cellular loading macrophage colony-stimulating factor (GM-CSF) and interleu-with vitamin C did not modify the expression of pro–caspase-8 in kin-3 (IL-3) signaling involves ROS,17 and our evidence suggestsU937 cells (Figure 7B). In monocytes, FASL induced caspase-8 that vitamin C inhibition of GM-CSF signaling occurs at the levelactivity by 130%, whereas vitamin C–loaded cells showed little of JAK-2 kinase activation.46 In this study, we undertook toactivation (Figure 7C). understand the role of vitamin C in FAS-mediated signaling for The inhibitory effect of vitamin C on the activity of caspase-8 in apoptosis in human monocytes and the monocytic cell line U937.cells cultured with FASL could be due to a direct effect of vitamin Vitamin C circulates in the plasma as AA; however, it isC on the enzyme or to inhibition of enzyme activation. It has been generally transported into cells as DHA through the facilitativereported that vitamin C can directly inhibit certain enzymes.41,42 As glucose transporters.23,24 Inside the cell, DHA is rapidly reduced tovitamin C is accumulated intracellulary in its reduced form, AA, we AA.24,28 In this way cells can accumulate high intracellularinvestigated if AA had a direct inhibitory effect on caspase-8 concentrations of vitamin C. For example, the intracellular concen-activity. The activity of recombinant caspase-8 was assayed in the tration of AA in mononuclear cells is reported to be about 3 mM,presence of AA (1.0-20.0 mM). To inhibit the conversion of AA to with circulating concentrations of AA in the range of 50 M.31,32DHA, 1,4-dithiothreitol was added to the reaction mixture. There Specialized cells can transport AA directly through sodium/was no signiﬁcant effect of AA on the observed activity of ascorbate cotransporters.25 AA is often used in in vitro experiments,recombinant caspase-8 (Figure 7D), while 20 M Z-IEDT-FMK leading to confounding results due to the lack of direct transport ofcompletely inhibited the activity of caspase-8 (not shown). There- ascorbate and because AA acts as a pro-oxidant in the presence offore, the data indicate that vitamin C is a potent inhibitor of the the free transition metals ubiquitously found in tissue culture.26 Weactivation of caspase-8 without a direct effect on enzymatic activity. therefore used DHA to load cells with vitamin C24 and analyzed its These results indicate that vitamin C inhibits apoptosis induced effect on the intracellular events mediating FAS-induced apoptosisby FAS-R in monocytes and U937 cells largely by preventing the in monocytic cells.Figure 7. Effect of vitamin C on caspase-8 activity. (A) U937 cells were loaded with 13 mM vitamin C prior to FASL treatment. The cell lysates were analyzed for caspase-8activity. Error bars represent the SD of triplicate values. (B) A Western blot was performed to analyze the expression of pro–caspase-8 in U937 cells loaded with 13 mM vitaminC (arrow, upper panel). Detection of ␤-tubulin was used as control for protein loading (arrow, lower panel). (C) Monocytes were loaded with 14 mM vitamin C prior to FASLtreatment and the cell lysates were analyzed for caspase-8 activity. The results are the mean of triplicate samples expressed as the percentage increase in caspase activitycompared with untreated cells and represent 3 (U937) or 2 (monocyte) experiments. (D) The activity of recombinant caspase-8 (1.25 U/well) with and without vitamin C wasdetermined as detailed in “Materials and methods.” These results are expressed as the percentage of activity in relation to control wells (without vitamin C) and arerepresentative of 3 independent experiments. Asterisks indicate statistically signiﬁcant differences (P ϭ .0005) in caspase activity of cells loaded with vitamin C beforechallenge with FASL.
342 PEREZ-CRUZ et al BLOOD, 1 JULY 2003 ⅐ VOLUME 102, NUMBER 1 The central observation of this work was a signiﬁcant inhibition did not result in more protection from apoptosis than the selectiveof FAS-induced apoptosis by vitamin C loading in monocytes and inhibition of caspase-8. This result is consistent with the notion thatU937 cells. We used pharmacologic concentrations of vitamin C to the major mechanism by which vitamin C prevents apoptosis isascertain the effect of vitamin C loading on the apoptotic process. through the inactivation of caspase-8; however, the data also pointThe lowest intracellular concentration of AA used in these experi- to an independent effect of vitamin C on mitochondrial membranements was higher than the physiologic intracellular AA reported for stabilization.normal human monocytes (3 mM).31,32 Loading monocytes with 4 We studied the downstream consequences of inhibition ofmM AA conferred signiﬁcant protection from FAS-R–induced caspase-8 activation by vitamin C by measuring the liberation ofapoptosis. Viability of U937 cells and monocytes was not impaired Cyt C into the cytosol and the subsequent activation of caspase-3.by pharmacologic concentrations of vitamin C. The general cellular Cyt C was not detected in the cytosol of FAS-stimulated U937 cellslevel of ROS increased after FAS-R ligation, and that increase was loaded with vitamin C, conﬁrming that mitochondrial integrity wasabrogated by loading cells with vitamin C. The liberation of ROS preserved in cells protected with vitamin C. Also, a FAS-inducedduring the apoptotic process is largely due to uncoupling of increase in caspase-3 activity in monocytes was not detected inoxidative metabolism in the mitochondria.7 Others have noted that cells loaded with vitamin C. Vitamin C reduced basal levels ofN-acetyl-L-cysteine reduces ROS in monocytes after FAS liga- caspase-3 activity in these cells likely because there is some degreetion,8 possibly through an increase in glutathione. The release of of oxidatively-induced apoptosis occurring in vitro.52 In U937ROS in apoptosis is preceded by a decrease in mitochondrial cells, however, vitamin C did not greatly inhibit caspase-3,membrane potential (⌬),7 and we found that the mitochondrial ⌬ maintaining 70% of FAS-induced activity. This phenomenon maywas maintained in U937 cells loaded with vitamin C and stimulated be due to the activity of caspase-10, which can activate caspase-353with FAS, suggesting that vitamin C inhibited events occurring and which was active in U937 cells incubated with FAS (38%before injury to the mitochondria. above basal). FASL-induced aggregation of FAS-R is believed to cause Beneﬁcial effects of vitamin C on immune function arecaspase-8 autoactivation, presumably related to proximity. Vitamin frequently cited, and it is widely assumed that vitamin C enhancesC loading, however, prominently inhibited FASL-mediated activa- immunity via its antioxidant function. Blood monocytes migrate totion of caspase-8 in U937 cells and in monocytes. In a cell-free tissues where some mature into macrophages, which participatesystem AA did not inhibit the enzymatic activity of recombinant importantly in the modulation of the adaptive immune system ascaspase-8, and therefore these results suggest that vitamin C phagocytes and antigen-presenting cells and by producing cyto-inhibits the activation of caspase-8, rather than its enzymatic kines. Upon maturation into macrophages, monocytes down-activity. We postulate that local or “micro-oxidation” is required for regulate the expression of caspase-8 and up-regulate the expressioncaspase-8 activation and that vitamin C inhibits that process by of FAS-associated death domain–like IL-1 beta-converting enzyme-quenching ROS. This thesis implies that FAS ligation leads to the inhibitory protein, a natural caspase-8 inhibitor. Macrophageslocal generation of ROS, which is required for caspase-8 activa- therefore are resistant to FAS-induced apoptosis.54,55 Before matu-tion. Indeed, a rapid and transient synthesis of oxygen radicals ration, however, monocytes are highly susceptible to FAS-inducedfollowing FAS ligation has been reported that is independent of apoptosis while resting and during bacterial phagocytosis.52,56 Herecaspase activation,49,50 indicating that the initiation of FAS- we present evidence that the intracellular accumulation of vitaminmediated signaling is temporally associated with local generation C in monocytes causes resistance to FAS-induced apoptosis byof ROS. ROS production after receptor ligation has been noted for inhibiting caspase-8 activation. A strong antioxidant such asthe IL-1R, TNF␣-R, and other cytokine receptor systems,12 and vitamin C could prolong the active life of monocytes before theythere is abundant evidence for ROS-mediated signaling for growth reach a FAS-insensitive stage. Maintenance of monocyte viabilityfactor receptors.13-15,17,51 by vitamin C may be a physiologic mechanism of vitamin C in host Both vitamin C and the caspase-8 inhibitor Z-IETD-FMK defense. Further, the pharmacologic use of vitamin C administeredinhibited apoptosis to a similar degree. The loss of mitochondrial as DHA57 could be a means to favorably affect host defense in⌬ as a consequence of FAS-R ligation was inhibited by both pathologic states.vitamin C and Z-IETD-FMK, although to a different extent, withalmost complete protection provided by vitamin C. The greaterstabilization of the mitochondrial membrane resulting from high Acknowledgmentsconcentrations of vitamin C in cells challenged with FAS may bedue to vitamin C’s capability to quench ROS generated at the We appreciate the technical help of O. Borquez, C. Tat, andreceptor level and elsewhere in the cell. Nevertheless, this effect A. Pedraza.References 1. Rathmell J, Thompson C. The central effectors of and apoptosis initiation in the absence of intermediate dependent pathway. J Immunol. cell death in the immune system. Annu Rev Im- caspase-8. J Biol Chem. 2001;276:46639-46646. 1996;156:3469-3477. munol. 1999;17:781-828. 6. Li H, Zhu H, Xu C, Yuan J. Cleavage of BID by 9. Laochumroonvorapong P, Paul S, Elkon K, 2. Rudin C, Thompson C. Apoptosis and disease: caspase 8 mediates the mitochondrial damage in Kaplan G. H2O2 induces monocyte apoptosis regulation and clinical relevance of programmed the Fas pathway of apoptosis. Cell. 1998;94:491- and reduces viability of Mycobacterium avium-M. cell death. Annu Rev Med. 1997;48:267-281. 501. intracellulare within cultured human monocytes. 3. Strasser A, O’Connor L, Dixit VM. Apoptosis sig- Infect Immun. 1996;64:452-459. naling. Annu Rev Biochem. 2000;69:217-245. 7. Zamzami N, Marchetti P, Castedo M, et al. Se- quential reduction of mitochondrial transmem- 10. Sugiyama H, Kashihara N, Makino H, Yamasaki 4. Boldin M, Goncharov T, Goltsev Y, Wallach D. Involvement of MACH, a novel MORT1/FADD- brane potential and generation of reactive oxygen Y, Ota Z. Reactive oxygen species induce apo- interacting protease, in Fas/APO-1- and TNF re- species in early programmed cell death. J Exp ptosis in cultured human mesangial cells. J Am ceptor-induced cell death. Cell. 1996;85:803-815. Med. 1995;182:367-377. Soc Nephrol. 1996;7:2357-2363. 5. Kischkel F, Lawrence D, Tinel A, et al. Death re- 8. Um H, Orenstein J, Wahl S. Fas mediates apo- 11. Li P, Nijhawan D, Budihardjo I, et al. Cytochrome c ceptor recruitment of endogenous caspase-10 ptosis in human monocytes by a reactive oxygen and dATP-dependent formation of Apaf-1/caspase-9
BLOOD, 1 JULY 2003 ⅐ VOLUME 102, NUMBER 1 VITAMIN C INHIBITS FAS-INDUCED APOPTOSIS 343 complex initiates an apoptotic protease cascade. 27. Washko PW, Hartzell WO, Levine M. Ascorbic c, alteration of mitochondrial membrane potential Cell. 1997;91:479-489. acid analysis using high-performance liquid chro- and activation of multiple caspases induced by12. Nakamura H, Nakamura K, Yodoi J. Redox regu- matography with coulometric electrochemical de- H(2)O(2), in human leukemia cells. Biochem lation of cellular activation. Annu Rev Immunol. tection. Anal Biochem. 1989;181:276-282. Pharmacol. 2002;63:1325-1335. 1997;15:351-369. 28. Vera JC, Rivas CI, Zhang RH, Farber CM, Golde 44. Williams R, Paterson C, Eakins K, Bhattacherjee13. Bae YS, Kang SW, Seo MS, et al. Epidermal DW. Human HL-60 myeloid leukemia cells trans- P. Ascorbic acid inhibits the activity of polymor- growth factor (EGF)-induced generation of hydro- port dehydroascorbic acid via the glucose trans- phonuclear leukocytes in inﬂamed ocular tissues. gen peroxide: role in EGF receptor-mediated ty- porters and accumulate reduced ascorbic acid. Exp Eye Res. 1984;39:261-265. rosine phosphorylation. J Biol Chem. 1997;272: Blood. 1994;84:1628-1634. 45. Nowak D, Ruta U, Piasecka G. Ascorbic acid in- 217-221. 29. Vera J, Rivas C, Zhang R, Golde D. Colony- hibits polymorphonuclear leukocytes inﬂux to the14. Suzukawa K, Miura K, Mitsushita J, et al. Nerve stimulating factors signal for increased transport place of inﬂammation-possible protection of lung growth factor-induced neuronal differentiation of vitamin C in human host defense cells. Blood. from phagocyte-mediated injury. Arch Immunol requires generation of Rac1-regulated reactive 1998;91:2536-2546. Ther Exp. 1989;37:213-218. oxygen species. J Biol Chem. 2000;275:13175- 30. May J, Mendiratta S, Qu Z, Loggins E. Vitamin C 46. Carcamo JM, Borquez-Ojeda O, Golde DW. Vita- 13178. recycling and function in human monocytic U-937 min C inhibits granulocyte macrophage-colony-15. Droge W. Free radicals in the physiological con- cells. Free Radic Biol Med. 1999;26:1513-1523. stimulating factor- induced signaling pathways. trol of cell function. Physiol Rev. 2002;82:47-95. 31. Levine M, Conry-Cantilena C, Wang Y, et al. Vita- Blood. 2002;99:3205-3212.16. Weber C, Erl W, Pietsch A, Strobel M, Ziegler- min C pharmacokinetics in healthy volunteers: 47. Schoonbroodt S, Piette J. Oxidative stress inter- Heitbrock HW, Weber PC. Antioxidants inhibit evidence for a recommended dietary allowance. ference with the nuclear factor-kappa B activation monocyte adhesion by suppressing nuclear fac- Proc Natl Acad Sci U S A. 1996;93:3704-3709. pathways. Biochem Pharmacol. 2000;60:1075- tor-kappa B mobilization and induction of vascu- 32. Levine M, Wang Y, Padayatty S, Morrow J. A new 1083. lar cell adhesion molecule-1 in endothelial cells recommended dietary allowance of vitamin C for 48. Herrlich P, Bohmer F. Redox regulation of signal stimulated to generate radicals. Arterioscler healthy young women. Proc Natl Acad Sci U S A. transduction in mammalian cells. Biochem Phar- Thromb. 1994;14:1665-1673. 2001;98:9842-9846. macol. 2000;59:35-41.17. Sattler M, Winkler T, Verma S, et al. Hematopoi- 33. Kasahara Y, Iwai K, Yachie A, et al. Involvement of reactive oxygen intermediates in spontaneous 49. Gulbins E, Brenner B, Schlottmann K, et al. Fas- etic growth factors signal through the formation of and CD95 (Fas/APO-1)-mediated apoptosis of induced programmed cell death is mediated by a reactive oxygen species. Blood. 1999;93:2928- neutrophils. Blood. 1997;89:1748-1753. Ras-regulated O2- synthesis. Immunology. 1996; 2935. 89:205-212.18. Hodges R, Hood J, Canham J, Sauberlich H, 34. Suzuki Y, Ono Y, Hirabayashi Y. Rapid and spe- Baker E. Clinical manifestations of ascorbic acid ciﬁc reactive oxygen species generation via 50. Banki K, Hutter E, Gonchoroff NJ, Perl A. Eleva- deﬁciency in man. Am J Clin Nutr. 1971;24:432- NADPH oxidase activation during Fas-mediated tion of mitochondrial transmembrane potential 443. apoptosis. FEBS Lett. 1998;425:209-212. and reactive oxygen intermediate levels are early events and occur independently from activation of19. Heuser G, Vojdani A. Enhancement of natural 35. Guaiquil V, Vera J, Golde D. Mechanism of vita- caspases in Fas signaling. J Immunol. 1999;162: killer cell activity and T and B cell function by buff- min C inhibition of cell death induced by oxidative 1466-1479. ered vitamin C in patients exposed to toxic chemi- stress in glutathione-depleted HL-60 cells. J Biol cals: the role of protein kinase-C. Immunophar- Chem. 2001;276:40955-40961. 51. Simon A, Rai U, Fanburg B, Cochran B. Activa- macol Immunotoxicol. 1997;19:291-312. 36. Marchetti P, Hirsch T, Zamzami N, et al. Mito- tion of the JAK-STAT pathway by reactive oxygen chondrial permeability transition triggers lympho- species. Am J Physiol. 1998;275:C1640-C1652.20. Jacob R, Kelley D, Pianalto F, et al. Immunocom- petence and oxidant defense during ascorbate cyte apoptosis. J Immunol. 1996;157:4830-4836. 52. Kiener P, Davis P, Starling G, et al. Differential depletion of healthy men. Am J Clin Nutr. 1991; 37. Hatano E, Bradham C, Stark A, Iimuro Y, Lemas- induction of apoptosis by Fas-Fas ligand interac- 54:1302S-1309S. ters J, Brenner D. The mitochondrial permeability tions in human monocytes and macrophages. J transition augments Fas-induced apoptosis in Exp Med. 1997;185:1511-1516.21. Bowie A, O’Neill L. Vitamin C inhibits NF-kappa B activation by TNF via the activation of p38 mito- mouse hepatocytes. J Biol Chem. 2000;275: 53. Fernandes-Alnemri T, Armstrong R, Krebs J, et al. gen-activated protein kinase. J Immunol. 2000; 11814-11823. In vitro activation of CPP32 and Mch3 by Mch4, a 165:7180-7188. 38. Yang J, Liu X, Bhalla K, et al. Prevention of apo- novel human apoptotic cysteine protease con-22. Carcamo JM, Pedraza A, Borquez-Ojeda O, ptosis by Bcl-2: release of cytochrome c from mi- taining two FADD-like domains. Proc Natl Acad Golde DW. Vitamin C suppresses TNFalpha-in- tochondria blocked. Science. 1997;275:1129- Sci U S A. 1996;93:7464-7469. duced NFkappaB activation by inhibiting Ikappa- 1132. 54. Perera L, Waldmann T. Activation of human Balpha phosphorylation. biochemistry. 2002;41: 39. Budihardjo I, Oliver H, Lutter M, Luo X, Wang X. monocytes induces differential resistance to apo- 12995-13002. Biochemical pathways of caspase activation dur- ptosis with rapid down regulation of caspase-8/23. Vera J, Rivas C, Fischbarg J, Golde D. Mamma- ing apoptosis. Annu Rev Cell Dev Biol. 1999;15: FLICE. Proc Natl Acad Sci U S A. 1998;95:14308- lian facilitative hexose transporters mediate the 269-290. 14313. transport of dehydroascorbic acid. Nature. 1993; 40. Srinivasula S, Ahmad M, Fernandes-Alnemri T, 55. Perlman H, Pagliari LJ, Georganas C, Mano T, 364:79-82. Litwack G, Alnemri E. Molecular ordering of the Walsh K, Pope RM. FLICE-inhibitory protein ex-24. Vera J, Rivas C, Velasquez F, Zhang R, Concha I, Fas-apoptotic pathway: the Fas/APO-1 protease pression during macrophage differentiation con- Golde D. Resolution of the facilitated transport of Mch5 is a CrmA-inhibitable protease that acti- fers resistance to fas-mediated apoptosis. J Exp dehydroascorbic acid from its intracellular accu- vates multiple Ced-3/ICE-like cysteine proteases. Med. 1999;190:1679-1688. mulation as ascorbic acid. J Biol Chem. 1995; Proc Natl Acad Sci U S A. 1996;93:14486-14491. 56. Baran J, Weglarczyk K, Mysiak M, et al. Fas 270:23706-23712. 41. Russell P, Williams A, Gapuz D. Inhibition of rab- (CD95)-Fas ligand interactions are responsible25. Tsukaguchi H, Tokui T, Mackenzie B, et al. A fam- bit muscle adenylate kinase by vitamin C. Bio- for monocyte apoptosis occurring as a result of ily of mammalian Naϩ-dependent L-ascorbic acid chem Biophys Res Commun. 1997;233:386-388. phagocytosis and killing of Staphylococcus au- transporters. Nature. 1999;399:70-75. 42. Schrammel A, Koesling D, Schmidt K, Mayer B. reus. Infect Immun. 2001;69:1287-1297.26. Clement M, Ramalingam J, Long L, Halliwell B. Inhibition of puriﬁed soluble guanylyl cyclase by 57. Huang J, Agus D, Winfree C, et al. Dehydroascor- The in vitro cytotoxicity of ascorbate depends on L-ascorbic acid. Cardiovasc Res. 2000;47:602- bic acid, a blood-brain barrier transportable form the culture medium used to perform the assay 608. of vitamin C, mediates potent cerebroprotection and involves hydrogen peroxide. Antioxid Redox 43. Gruss-Fischer T, Fabian I. Protection by ascorbic in experimental stroke. Proc Natl Acad Sci U S A. Signal. 2001;3:157-163. acid from denaturation and release of cytochrome 2001;98:11720-11724.