Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

An important role for mitochondria is frequently observed during apoptosis or drug-induced cell death. Although the exact contribution of this organelle in APAP-induced liver injury and cell death is unclear, alterations to mitochondrial respiration with APAP treatments have been demonstrated (Burcham and Harman, 1991). Mechanistically, the arylation of free thiols (Qiu et al., 1998b) as well as oxidative stress (Adamson and Harman, 1993) have been proposed to play roles in the toxicity of APAP, but the consequences of these changes are still uncertain (Pumford and Halmes, 1997). More recent studies have identified proteins covalently modified by APAP or its less hepatotoxic isomer N-acetyl-m-aminophenol (AMAP), and indicate that an important component of the differential toxicity of these compounds is mediated at the level of the mitochondrion (Qiu et al., 1998a,b).

The hepatotoxicity of APAP has been traditionally thought of as a centrilobular necrotic event (Pumford and Halmes, 1997). Other studies have attempted to define an apoptotic component to APAP-mediated cell death, but these investigations have not been entirely definitive (Ray et al., 1996; Lawson et al., 1999). Consequently, we have addressed the possibility that the BCL-2 family of proteins may have a functional role in the progression of liver damage after APAP overdose. The BCL-2 protein family consists of both proapoptotic (e.g., BAX, BAK) and antiapoptotic (e.g., BCL-2, BCL-X_L) members, which play an important role in the determination of apoptosis in response to many physiological and pathological effectors (Wei et al., 2001). This pro- versus antiapoptotic balance has been suggested to be controlled through dimerization of BCL-2 family members (Yin et al., 1994; Rosse et al., 1998) and/or by phosphorylation (Haldar et al., 1998). Although there is an extensive literature on the antiapoptotic properties of BCL-2, its mechanism of cytoprotection is still unknown. It has been proposed that the cytoprotective action of BCL-2 may lie in its ability to act as an antioxidant (Hockenbery et al., 1993), to block cytochrome c release (Cai and Jones, 1998), or to inhibit caspase activity after cytochrome c release (Rosse et al., 1998). Further studies are consequently required to address this uncertainty in BCL-2 function, particularly with recent reports indicating that BAX translocation to mitochondria is not prevented in apoptotic neurons despite overexpression of BCL-2 (Putcha et al., 1999).

Based on observations that APAP-induced cell death is partly apoptotic (Ray et al., 1996; Lawson et al., 1999), we examined APAP-mediated hepatotoxicity for alterations to the subcellular distribution of BAX, a BCL-2 family member implicated as a central effector of mitochondrially-mediated apoptotic cell death (Wei et al., 2001). We report here that BAX is redistributed with APAP treatment and, consequently, may play an early role in APAP-mediated hepatotoxicity. In addition, we hypothesized that overexpression of BCL-2 protein should offer protection to liver tissue exposed to doses of APAP that would otherwise cause centrilobular necrosis. Our results indicate a more pronounced liver injury produced by APAP in BCL-2-overexpressing animals with morphological and biochemical evidence of increased damage shifting from centrilobular to throughout the entire lobule (panlobular). These unexpected findings provide a basis for the further elucidation of the role of mitochondria in APAP-induced liver injury and cell death.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animal Care and Dosing Protocols. Human bcl-2 transgenic mice (C57/B6C3H background) were kindly provided by Dr. S. J. Korsmeyer (Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA). Animals were housed in a temperature- and humidity-controlled specific pathogen free facility maintained on a 12-h light/dark cycle with free access to food and water. Animals were genotyped using tail DNA in a polymerase chain reaction assay. Briefly, tail snips were digested in proteinase K overnight at 55°C. Proteinase K was then heat inactivated at 95°C for 10 min. Polymerase chain reaction was performed using 1 µl of the crude extract as a template and primers spanning an intron in bcl-2 to yield a 300-base-pair product from the human gene. BCL-2 expression is under the inducible control of a metallothionein promoter. Male mice 16 weeks old (22-28 g) were given 25 mM ZnSO₄ in drinking water for 5 to 7 days prior to dosing and BCL-2 levels were evaluated by immunoblot. Animals were fasted for 16 h before dosing with APAP or AMAP (500 mg/kg i.p.) or vehicle (saline). To exclude nonspecific metal effects, ZnSO₄-supplemented water was freely available to all animals during the fast and after dosing. Liver tissue and plasma samples were taken 6 and 24 h after injection for histopathological, immunoblotting, and other biochemical determinations.

Isolation of Subcellular Fractions and Immunoblotting Procedures. BAX subcellular localization studies were with fractions prepared by standard differential centrifugation procedures from wild-type B6C3F₁ (Fig. 1) or Swiss-Webster mice (data not presented) receiving 600 mg/kg (i.p.) APAP, AMAP, or vehicle control (saline) at 2 and/or 4 h after dosing as described previously (Tonge et al., 1998). The relative purity of subcellular fractions was assessed by monitoring lactate dehydrogenase (cytosol) and succinate cytochrome c reductase (mitochondria) marker enzyme activities. Subcellular fractions or total homogenized liver tissue (50 µg per lane) were resolved on SDS-polyacrylamide minigels (10 to 15%, Mini-Protean II; Bio-Rad, Hercules, CA) and transferred to nitrocellulose (1 h, 15 V, Trans-Blot SD SemiDry Transfer Cell; BioRad). Immunodetection using antibodies to BAX (B9; Santa Cruz Biotechnology, Santa Cruz, CA), BCL-2 (N19; Santa Cruz Biotechnology), catalase (Calbiochem, San Diego, CA), cytochrome c (Pharmingen, San Diego, CA), mouse albumin (ICN Pharmaceuticals, Costa Mesa, CA), GAPDH (Dietze et al., 1997), or anti-APAP protein adducts (Tonge et al., 1998) was preformed by standard procedures and detected by chemiluminescence (SuperSignal ULTRA; Pierce, Rockford, IL).

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Mitochondrial alterations induced by APAP and the nonhepatotoxic positional isomer, AMAP. Subcellular fractions from B6C3F1 mice were isolated at various times after control (vehicle), APAP, or AMAP treatment (600 mg/kg, i.p) and immunoblotted using standard procedures. Top two blots, mitochondrial fractions immunoblotted for BAX and cytochrome c content indicated early and sustained BAX- relocalization to mitochondria in all treatments with transient cytochrome c depletion in APAP treatments. Bottom two blots, cytosolic fractions immunoblotted for BAX and cytochrome c release indicated release of cytochrome c from mitochondria of APAP-treated animals and either an absence of BAX- immunoreactivity in APAP treatments or a relative depletion of BAX- in AMAP-treated animals. Molecular mass of immunoreactive proteins, in kilodaltons, is indicated on the right of each blot.

Transaminase Activities. Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were assessed spectrophotometrically using either the Dimension Clinical Chemistry System (Dade Behring, Deerfield, IL) analyzer on plasma samples diluted with DADE diluent to within the linear range of the analyzer (typically 1:40-1:50) or standard commercial methods (Sigma Diagnostics Infinity, Procedure 122-UV or 152-UV) exactly as described by the manufacturer. In both cases aminotransferase activities are presented as units per liter.

Assessment of Hepatotoxicity by Histopathology. Liver tissue sections from animals dosed with APAP, AMAP, or saline were fixed in formalin and embedded in paraffin. Sections of 5-µm thickness were stained with hematoxylin and eosin using standard procedures. Sections were independently evaluated without prior knowledge of treatment regime and digital photomicrographs were taken at 100× or 200× magnification.

Assessment of Cellular Morphology by Electron Microscopy. Liver sections from animals dosed with APAP, AMAP, or saline were fixed in Karnovsky's fixative (one-half strength glutaraldehyde-formaldehyde), embedded and stained with uranyl acetate and Reynold's lead citrate. Specimens were examined using a Philips 410 Transmission Electron Microscope at 12,000× or 6,000× magnification.

Determination of Reduced (GSH) and Oxidized (GSSG) Liver Glutathione Content. Hepatic glutathione levels were determined as described previously (Luderer et al., 2001). Briefly, liver tissue was homogenized in 5% (w/v) sulfosalicylic acid then immediately processed independently for either GSH or GSSG. For GSSG, free GSH was derivatized with 2-vinylpyridine at room temperature and excess 2-vinylpyridine extracted into the organic phase with chloroform. GSSG was then reduced by the addition of 50 µl of 1 mM NADPH and 20 units/ml GSH reductase for 1 h at room temperature, followed by derivatization with 20 µl of 12.5 mM monobromobimane in the dark for 30 min before HPLC analysis (Shimadzu LC-6A HPLC equipped with an Alltech 15 × 0.5-cm C18 reversed-phase column using a binary gradient [solvent A, 1.0 mM tetrabutylammonium phosphate, pH 3.0; solvent B, methanol] at a flow rate of 1.5 ml/min with starting conditions of 95% A/5% B). Eluted peaks were monitored fluorometrically at ex = 375 nm and em = 475 nm. Reduced GSH was assessed by further dilution (1:10) with 5% (w/v) sulfosalicylic acid. A 100-µl volume of the diluted sample was then combined with 50 µl 10% (v/v) triethanolamine, mixed, and 100 µl of the mixture was added to 200 µl of buffer (100 mM NaH₂PO₄, 1 mM EDTA) before derivatization and HPLC as described above for GSSG.

Caspase Activation. Caspase activation was determined as described previously (Franklin et al., 1998). Briefly, 50 µg of liver tissue homogenate were incubated for 60 min at 37°C in 100 µl of caspase assay buffer [50 mM HEPES, pH 7.4, 100 mM sodium chloride, 2 mM EDTA, 20% sucrose (w/v), 0.2% CHAPS (w/v), 10 mM DTT] with 20 µM Ac-DEVD-amc (Alexis Biochemicals, San Diego, CA) and fluorescence monitored on a Packard Fluorocount (Packard Instrument Company, Meriden, CT) microplate fluorometer with an excitation wavelength of 360 nm and an emission wavelength of 460 nm. Data are presented as fold activation over extracts from untreated cells.

Determination of Mitochondrial Respiration. Oxidative phosphorylation activity, via complex III, ubiquinol:ferricytochrome c oxidoreductase (E. C. 1.10.2.2), was assayed as described previously (Gudz et al., 1997). In brief, the rate of reduction of cytochrome c in isolated mitochondria was monitored spectrophotometrically (Cary 3E UV-Vis Spectrophotometer, Varian, Australia) with or without antimycin A, a specific complex III inhibitor, for 1 min at 550 nm after the addition of decylubiquinol substrate. The antimycin A sensitive rate is then calculated using the extinction coefficient of 19.1 mM¹ cm¹ (Gudz et al., 1997) and reported as micromoles of cytochrome c reduced per minute per milligram of mitochondrial protein.

Statistical Analysis. Results are calculated as mean ± S.E. Statistical significance was determined for transaminase and caspase activities using two-tailed, unpaired t test analysis and for GSH/GSSG contents using a paired, two-tailed t test. The statistical significance of complex III activities was determined by Mann-Whitney U test. Values for p  0.05 were considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Altered Processing and Subcellular Levels of BAX after Drug Treatment. Increased immunoreactivity of proapoptotic BAX was found in the mitochondrial fraction after APAP or AMAP treatment of wild-type B6C3F₁ mice (Fig. 1, top). Molecular mass calculations performed on the data indicated that in these nontransgenic animals, the  splice variant of BAX represented the active isoform. Similarly, a loss of BAX immunoreactivity from the cytosol, consistent with the BAX- isoform, was observed (Fig. 1, bottom). Moreover, at these early time points the extent of BAX- loss seemed to correlate well with the relative hepatotoxic potential of the two compounds being completely absent in the cytosol of APAP treatments only. Drug-induced increases to BAX- mitochondrial levels, however, were associated only with release of cytochrome c into the cytosol in animals treated with APAP for 2 h and not observed with AMAP treatment (Fig. 1).

Assessment of Liver Damage. Liver damage was assessed histologically and by determining plasma ALT and AST activities. WT (/) and bcl-2 (/+) transgenic animals induced with zinc demonstrated no alterations to hepatic morphologies (compare Fig. 2, C and D, with Fig. 2A) and no significant differences in plasma ALT and AST levels (data not shown). Consequently, WT (/) animals were used as the basis for comparison with treatment groups. At 6 h after injection, all treatment transaminase levels were elevated compared with vehicle-treated control animals (Table 1). The 24-h ALT and AST enzyme levels were significantly raised in APAP-treated bcl-2 (/+) transgenic mice (55.8- and 18-fold higher, respectively) versus control (vehicle-treated) WT mice. In comparison, APAP-treatment in WT mice resulted in considerably smaller elevations of plasma ALT and AST activities versus control (vehicle-treated) WT mice (22- and 6.6-fold, respectively; Table 1). These 24-h data indicate a previously unreported difference in the biological response to APAP between WT and bcl-2 (/+) animals with a more severe form of liver damage found in BCL-2 overexpressing mice.

View larger version (110K): [in this window] [in a new window]   Fig. 2.   Assessment of hepatotoxicity by morphology. Hematoxylin- and eosin-stained liver sections 6 and 24 h after dosing with APAP or AMAP (500 mg/kg, i.p.). All animals were induced with 25 mM ZnSO4 in the drinking water for 5 to 7 days before dosing. A, control WT (/) treated with saline showing only normal hepatic morphology. B, WT (/) animals 24 h after dosing with APAP showing distinct areas of centrilobular necrosis. C-D, zinc-induced bcl-2 (/+) animals dosed with saline only at either 100× (C) or 200× (D) magnification showing typical hepatic lobule structure and no evidence of damage. E-F, bcl-2 (/+) animals 6 h (E) and 24 h (F) after dosing with APAP. At 6 h, microvesicular steatosis (hollow arrowheads, E) is observed, but by 24 h, a massive confluent necrosis is evident (solid arrowheads, F) with considerable hemorrhage. G-H, AMAP-treated animals developed a qualitatively more pronounced microvesicular steatosis at 6 h (hollow arrowheads, G). In contrast to APAP, AMAP-treated bcl-2 (/+) animals did not progress to extensive damage (H). All photographs are 100× magnification, except D is 200×. CV, central vein.

View this table: [in this window] [in a new window]   TABLE 1 Plasma Transaminase Activities in wild-type (/) and bcl-2 (/+) transgenic mice treated with 500 mg/kg APAP or AMAP Values are reported as Units per liter. [mean (S.E.)]. n = 3 to 6 for each treatment group.

Histopathological Analyses. In comparison with vehicle-treated control animals (Fig. 2A), APAP treatment in WT mice resulted in a clear centrilobular (zone 3) necrosis with areas of focal bridging necrosis and hemorrhage (Fig. 2B). In places where the hepatocytes were less obscured by blood (i.e., at the edges of the lesion), microvesicular steatosis was also identified in intact centrilobular hepatocytes (data not shown).

View larger version (185K): [in this window] [in a new window]   Fig. 3.   Assessment of cellular morphology changes by electron microscopy. Liver sections were fixed in Karnovsky's fixative and stained with Reynold's lead citrate 6 h after dosing with APAP and AMAP 500 mg/kg i.p. A, control bcl-2 (/+) treated only with ZnSO4 showing normal hepatocyte morphology. B, WT (/) animals 6 h after dosing with APAP showing lipid deposition (solid arrowheads), mitochondrial proliferation, and nuclear condensation. C-D, bcl-2 (/+) animals 6 h after dosing with APAP and AMAP, respectively. APAP treatment (C) shows association of the endoplasmic reticulum with the mitochondria and ring-shaped mitochondria (hollow arrowhead), whereas AMAP-treatment (D) shows almost normal morphology with the exception of electron dense aggregates, which are seen in all bcl-2 (/+) animals (*). All magnifications are ~12,000X except (D) at ~6,000×. Scale bar, 5 µm.

Further Characterizations of BCL-2 Overexpression in Control and APAP-Treated Animals. To further assess the effects of BCL-2 overexpression per se, several biochemical parameters were determined in addition to liver morphology. Early, APAP-specific increases in mitochondrial BAX content (and most likely BAX-) were reproducibly observed in bcl-2 (/+) transgenic mice by chemiluminescent immunoblot analysis but required extended, overnight, exposures to be detected (Fig. 4A, top blot). In comparison, no differences in saline-treated mitochondrial BAX content were observed between bcl-2 (/) and bcl-2 (/+) animals (Fig. 4A, top blot). However, APAP-specific BAX mitochondrial translocations in bcl-2 (/+) mice failed to induce cytochrome c release into the cytosol (Fig. 4A, middle blot) despite apparently excellent subcellular fractionations of mitochondria from the cytosol (cytochrome c versus GAPDH; Fig. 4A, middle and bottom blots).

View larger version (45K): [in this window] [in a new window]   Fig. 4.   Biochemical characterizations of control, APAP-treated bcl-2 (/+), and WT (/) transgenic animals. A, subcellular BAX content from mitochondrial and cytosolic fractions prepared from bcl-2 (/) and bcl-2 (/+) animals 2 h after treatment with vehicle only (saline/control) or APAP (500 mg/kg, i.p.) as described under Materials and Methods. APAP-specific BAX translocation to the mitochondrial fraction was observed in BCL-2 overexpressing animals (top blot). Cytochrome c and GAPDH remained in the mitochondria and cytosol, respectively, in all treatment groups (lower 2 blots). B, hepatic GSH and GSSG contents of bcl-2 (/) and bcl-2 (/+) mice induced with ZnSO4 in the drinking water for 5 to 7 days and dosed with saline vehicle only as described under Materials and Methods. a, total GSH or GSSG content reported as mean ± S.E. (n = 4); b, significant increase (p < 0.05) in GSSG in bcl-2 (/+) transgenic animals. C, APAP-specific elevations of hepatic catalase content in "heavy membrane" (mitochondria + peroxisomes) fractions of bcl-2 (/) and bcl-2 (/+) animals isolated 6 h after vehicle or drug treatment.

Immunoblot Detection of BCL-2. Expression of human bcl-2 transgene was confirmed by immunoblotting mouse liver tissue homogenates with commercially available antisera directed to the N-terminal region of BCL-2 (N19; Materials and Methods), which recognizes both human and endogenous mouse BCL-2. The data demonstrated clearly an overexpression of the human bcl-2 transgene in transgenic animals, whereas little or no endogenous murine BCL-2 was detected in WT mice (Fig. 5A, top). Immunoblotting also revealed apparent decreases in BCL-2 protein content in APAP- but not AMAP-treated transgenic mice relative to GAPDH loading controls (Fig. 5A, compare top and bottom blots). These slight decreases in BCL-2 protein seemed to correlate with an increased breakdown of mouse hepatic albumin in APAP-treated groups (Fig. 5A, compare top and middle blots) and could not be attributed to general sample proteolysis as the GAPDH loading standards remained unaffected (Fig. 5A, bottom blot).

View larger version (34K): [in this window] [in a new window]   Fig. 5.   BCL-2 transgene overexpression and caspase activities in control and APAP-dosed WT (/) and bcl-2 (/+) animals. A, confirmation by immunoblot analysis of liver BCL-2 content in tissue homogenates from control, APAP or AMAP (500 mg/kg, i.p.) treatments prepared as described under Materials and Methods. The blot was probed with antisera to BCL-2 (N19) (Santa Cruz), GAPDH, and albumin (ICN) using standard procedures. Molecular weight markers (kilodaltons) are indicated to the left of the panel. Samples were taken at 6 h (lanes 1-4) and 24 h (lanes 5-8) after treatment. Endogenous murine BCL-2 was not detected in saline-injected WT animals (lanes 1 and 5) or APAP-treated WT animals (lanes 3 and 7) using BCL-2 (N19) antiserum. High levels of BCL-2 expression were confirmed in bcl-2 (/+) animals treated with ZnSO4 and dosed with AMAP (lanes 2 and 6) or APAP (lanes 4 and 8). Degradation of albumin was observed predominantly in the APAP-treated animals. Equal loading of samples was demonstrated by GAPDH content after membrane stripping, blocking and reprobing. B, the activation of caspase-3-like enzymes was evaluated as described under Materials and Methods. Activity is represented as fold change in Ac-DEVD-amc cleavage versus untreated animals after 60 min of incubation in caspase assay buffer. Error bars represent S.E. of mean expressed as a percentage of absolute activity and range from 1.1 to 6.2% of total activity. n = 3 to 6 animals for each treatment group. *p < 0.05 (compared with control values).

Caspase Activation. BCL-2 degradation is generally associated with a proapoptotic state and increased activity of the effector caspase, caspase-3. To further examine the possibility of BCL-2 cleavage to its proapoptotic form by caspase-3, the activation of caspase-3-like enzymes in liver tissue homogenates was evaluated. Only modest increases in Ac-DEVD-amc cleavage were observed after either APAP- or AMAP-treatment of bcl-2 (/+) transgenic or WT mice (Fig. 5B). For example, at 6 h after treatment, AMAP-treated transgenic mice had the highest change in caspase-3-like activity (1.28-fold increase versus control, p  0.05), whereas at 24 h after treatment, APAP-treated transgenic mice were highest with a 1.47-fold increase in activity versus control (p  0.05). These changes are in agreement with our previous studies, indicating only slight caspase-3-like activation after APAP-treatment in vitro and in vivo (Pierce et al., submitted). Although statistically significant the biological relevance of low-level Ac-DEVD-amc cleavage activity in APAP- and AMAP-treated mice observed in these studies is unclear.

Complex III Activities. To evaluate overall mitochondrial respiratory chain activity, ubiquinol:ferricytochrome c oxidoreductase (complex III) activity was determined in control and treatment groups. At 6 h and before overt liver damage, all three treatment groups demonstrated a statistically significant decrease in complex III activity compared with control levels (Fig. 6). The extent of inhibition correlated well with eventual liver damage as APAP-treated transgenic mice were most adversely affected (with the exception of AMAP-treatment; see next paragraph). The same trend was also seen in APAP-treated transgenic animals at 24 h when liver damage was well advanced but the absolute level of complex III inhibition was less. This may be a consequence of the differential centrifugation procedures used to isolate mitochondria from damaged tissue with the selection of intact, and relatively functional, organelles.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   Complex III activity. Complex III (ubiquinol:ferricytochrome c oxidoreductase) activities for each treatment group (APAP or AMAP, 500 mg/kg, i.p.) was determined as described under Materials and Methods. Values are reported as percentage of mean control rate (micromoles of cytochrome c reduced per minute per milligram of mitochondrial protein). Errors, expressed as percentage of S.E., ranged from 1.4 to 7% for control groups and 3.5 to 10.5% for treatment groups. n = 3 to 6 for all treatment groups except APAP (/+) 24 h and APAP (/) 24 h for which n = 2. *p = 0.05, **p < 0.05 (compared with control values).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The cell death and liver injury produced by APAP has been studied extensively for more than 20 years. Necrosis has usually been associated with APAP-induced hepatotoxicity; more recently, a role for apoptosis has been proposed but not supported by extensive data (Ray et al., 1996; Lawson et al., 1999). For example, studies have concluded that APAP-associated hepatotoxicity is a mixed necrotic and apoptotic event, with approximately 40% of cells undergoing apoptosis based on oligosomal DNA laddering, whereas 60% undergo necrosis as determined by plasma transaminase elevations (Ray et al., 1996). The data presented here suggest an early role for BAX in APAP-induced liver injury, indicating that necrotic cell death in these circumstances may be initiated at the mitochondrion in a proapoptotic manner. Nonetheless, the precise relationship between APAP-mediated protein modifications, mitochondrial BAX content, and cell death has yet to be fully determined.

Mitochondria play an important role in APAP-mediated cell death. This has been deduced from many studies that indicate preferential mitochondrial glutathione depletion, calcium deregulation (Tirmenstein and Nelson, 1989, 1991), and selective protein modifications within the mitochondrion (Qiu et al., 1998a,b) after APAPbut not AMAPtreatment. Consistent with general mitochondrial dysfunction, both decreased oxygen utilization and ATP levels have been associated with APAP overdosing (Tirmenstein and Nelson, 1990; Burcham and Harman, 1991). In a related observation, it has been proposed previously that ATP levels are critical in determining the path to cell death (i.e., apoptosis versus necrosis) (Leist et al., 1997).

Most studies have used protective agents to elucidate the mechanism by which APAP elicits its toxicity (e.g., antioxidants, calcium chelators). Our studies have focused on the events downstream of APAP-protein adduction and examined the cellular responses to such modifications that ultimately lead to cell death. BAX has been shown to regulate the release of mitochondrial cytochrome c via components of the permeability transition pore (Shimizu et al., 1999). Consequently, APAP- and AMAP-mediated cell death and liver injury may be attributable, at least in part, to this mechanism. Other contributing factors seem to be required, however, because mitochondrial BAX localization does not always correlate with cytochrome c loss into the cytosol (Figs. 1 and 4A; Putcha et al., 1999). In this regard, it should be noted that BAX-dependent mitochondrial cytochrome c release is insufficient to account for the capacity to undergo apoptosis in a trophic factor withdrawal model of sympathetic neuron cell death (Deshmukh and Johnson, 1998). An alternative explanation for the absence of cytochrome c in the cytosol, despite mitochondrial BAX localization, can be found with a recent report, which indicates that the absolute levels of intracellular BAX may determine the extent and even reversibility of mitochondrial cytochrome c release (Pastorino et al., 1999). Finally, BAX-induced apoptosis requires a functional F₀F₁-ATPase proton pump in mammalian cells (Matsuyama et al., 1998), which may explain recent findings indicating that F₀F₁-ATPase inhibition protects against APAP-mediated damage (Banerjee et al., 1998).

It has also been recognized that BCL-2 itself, as well as other members, such as BCL-X_L (the BCL-2 functional homolog expressed in hepatocytes; Tzung et al., 1997), heterodimerize with BAX (e.g., Yin et al., 1994). BCL-2 dimerization with BAX is thought to mediate its antiapoptotic action (Yin et al., 1994; Rosse et al., 1998). Recognizing the function of BAX translocation as an initiating mitochondrial event in many apoptotic systems (Wei et al., 2001), and the potential for regulatory control by BCL-2, we have attempted to determine whether BCL-2 overexpression may protect against APAP-induced hepatotoxicity as reported for other stimuli (Rosse et al., 1998; Putcha et al., 1999). In this manner, perturbing the balance between BAX and BCL-2 in favor of BCL-2, should be hepatoprotective.

The liver damage we have observed in APAP-treated bcl-2 (/+) transgenic animals was considerably greater than that found with WT animals, contrary to our expectations (Fig. 2, Table 1). Moreover, we were able to detect APAP-specific localization of BAX to mitochondria in bcl-2 (/+) animals (Fig. 4A). However, these levels seemed to be considerably lower than those of comparably treated nontransgenic mice as determined by immunoblot sensitivity. The only measurable difference we have observed between control BCL-2-overexpressing and WT animals is an elevation of hepatic oxidized glutathione (GSSG) in bcl-2 (/+) animals (Fig. 4B). Although an antioxidant function has been attributed to BCL-2, our findings are in agreement with a previous report proposing a prooxidant capacity for BCL-2 (Steinman, 1995). In addition, it has also been observed that BCL-2 overexpression fails to prevent the action of classical inducers of permeability transition in isolated mouse liver mitochondria (Yang et al., 2000). Moreover, our caspase-3-like activities (Ac-DEVD-amc cleavage) were much lower than that observed with most other apoptosis models (Fig. 5B). This low-level caspase activation was expected based on previous reports of no caspase activation (Lawson et al., 1999) and our observations of low caspase activation in APAP- and AMAP-treated mice (Pierce et al., submitted). These observations are consistent with a proapoptotic initiation of liver damage during APAP overdose and subsequent advance to death via necrosis.

Disruption of oxidative phosphorylation, specifically at complex I and complex II, after direct exposure to NAPQI, the APAP reactive intermediate, has been previously reported (Burcham and Harman, 1991) but an inhibition of complex III in vivo has not been observed. However, in vitro studies have demonstrated decreased complex III activity after NAPQI exposure to inverted membrane particles (Ramsay et al., 1989). From the studies presented here we conclude that alterations to complex III activities can be a sensitive and early marker of the eventual extent of liver damage.

The generation of a proapoptotic BCL-2 cleavage fragment by caspase-3 or other proteolytic cleavage activity may offer an explanation for the enhanced APAP-induced hepatotoxicity we have observed (Cheng et al., 1997; Kirsch et al., 1999). Although our studies suggest that BCL-2 protein levels were decreased 24 h after APAP treatments, we have not detected BCL-2 fragments consistent with cleavage by caspase-3 to the proapoptotic form (Fig. 5A). Our data (i.e., low caspase activation and no evidence for a specific cleavage product) suggest that BCL-2 disappearance in animals treated with APAP for 24 h may be attributable to another cellular proteolytic activity (e.g., proteasome; Breitschopf et al., 2000; Pierce et al., submitted). Further experiments will be required to determine the mechanism of BCL-2 loss in APAP-treated animals. An alternative explanation for enhanced APAP-induced hepatotoxicity in BCL-2-overexpressing animals is the observation that APAP transiently inhibits proteasomal activity (Qiu et al., 1998b; Pierce et al. submitted). Finally, our data suggest the possibility that BCL-2 exacerbation of liver injury by APAP does not occur by a mitochondrially mediated mechanism for two reasons. BAX translocation in APAP-treated bcl-2 (/+) is not associated with cytochrome c release into the cytosol and peroxisomal catalase content was induced with APAP treatment adding to a basal, but apparently nonpathological, oxidative stress in untreated bcl-2 (/+) animals. Further studies will be required to determine the role of other organelles, and especially the ER, in this form of liver injury.

In conclusion, our studies indicate that the overexpression of human BCL-2 in C57/B6C3H mice significantly enhances APAP-induced hepatotoxicity. Biochemical and morphological evidence also confirms a key role for mitochondria in the progression of this damage. These novel findings allow for new avenues to approach and define the mechanism of APAP-mediated hepatotoxicity.