Mitochondrial permeability transition in acetaminophen-induced necrosis and apoptosis of cultured mouse hepatocytes



Acetaminophen overdose causes massive hepatic failure via mechanisms involving glutathione depletion, oxidative stress, and mitochondrial dysfunction. The ultimate target of acetaminophen causing cell death remains uncertain, and the role of apoptosis in acetaminophen-induced cell killing is still controversial. Our aim was to evaluate the mitochondrial permeability transition (MPT) as a key factor in acetaminophen-induced necrotic and apoptotic killing of primary cultured mouse hepatocytes. After administration of 10 mmol/L acetaminophen, necrotic killing increased to more than 49% and 74%, respectively, after 6 and 16 hours. MPT inhibitors, cyclosporin A (CsA), and NIM811 temporarily decreased necrotic killing after 6 hours to 26%, but cytoprotection was lost after 16 hours. Confocal microscopy revealed mitochondrial depolarization and inner membrane permeabilization approximately 4.5 hours after acetaminophen administration. CsA delayed these changes, indicative of the MPT, to approximately 11 hours after acetaminophen administration. Apoptosis indicated by nuclear changes, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling, and caspase-3 activation also increased after acetaminophen administration. Fructose (20 mmol/L, an adenosine triphosphate–generating glycolytic substrate) plus glycine (5 mmol/L, a membrane stabilizing amino acid) prevented nearly all necrotic cell killing but paradoxically increased apoptosis from 37% to 59% after 16 hours. In the presence of fructose plus glycine, CsA decreased apoptosis and delayed but did not prevent the MPT. In conclusion, after acetaminophen a CsA-sensitive MPT occurred after 3 to 6 hours followed by a CsA-insensitive MPT 9 to 16 hours after acetaminophen. The MPT then induces ATP depletion–dependent necrosis or caspase-dependent apoptosis as determined, in part, by ATP availability from glycolysis. (HEPATOLOGY 2004;40:1170–1179.)

Acetaminophen overdose causes severe hepatotoxicity that leads to liver failure in experimental animals and humans.1 Although much acetaminophen is metabolized via conjugation with glucuronic acid and sulfate and then excreted, a portion of it is metabolized by cytochrome P-450.2 Cytochrome P-450 oxidation of acetaminophen forms a chemically reactive metabolite, N-acetyl-p-benzoquinone imine (NAPQI), which reacts with glutathione (GSH) to form an acetaminophen–GSH conjugate. Once GSH is exhausted, NAPQI binds to cellular proteins, including a number of mitochondrial proteins,3, 4 which leads to hepatocellular death. In mouse hepatocytes, acetaminophen inhibits mitochondrial oxidative phosphorylation, resulting in depletion of adenosine triphosphate (ATP).5–9 Acetaminophen also depletes both cytosolic and mitochondrial GSH, the potent scavenger of reactive oxygen and peroxynitrite species.10, 11 Although mitochondria are sensitive to acetaminophen toxicity, the precise role of mitochondria, if any, in acetaminophen-induced cell death remains unclear.

Previous reports indicate that the immunosuppressive drug cyclosporin A (CsA) is protective in acetaminophen toxicity both in vivo and in vitro.12, 13 In addition to its immunosuppressive action, CsA inhibits the mitochondrial permeability transition (MPT), a phenomenon characterized by mitochondrial swelling, uncoupling, and inner membrane permeabilization to solutes of molecular mass up to 1500 Da.14 The MPT is implicated in lethal cell injury from anoxia, ischemia/reperfusion, and oxidative stress to liver, heart, and other cell types.15 Increased Ca2+ and reactive oxygen and nitrogen species activate the MPT, whereas CsA inhibits it. The MPT plays an important role initiating both apoptotic and necrotic cell death. Specifically, ATP depletion caused by uncoupling of oxidative phosphorylation after the MPT leads to necrotic cell killing, whereas cytochrome c release caused by mitochondrial swelling and outer membrane rupture after the MPT initiates apoptosis.16, 17 However, the role of acetaminophen in inducing MPT-dependent necrosis and apoptosis to hepatocytes has not been characterized. Studies also indicate that with increasing strength of induction, the onset of the MPT is no longer inhibited by CsA and occurs in the absence of Ca2+. Namely, the Ca2+-dependent, CsA-sensitive MPT progresses to a Ca2+-independent, CsA-insensitive form of mitochondrial inner membrane permeabilization and large amplitude swelling.18

The relative importance of necrosis and apoptosis in acetaminophen-induced liver damage is controversial. Some reports indicate apoptosis via terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) and DNA laddering after exposure of mice and mouse hepatocytes to acetaminophen.19, 20 Other studies indicate that the percentage of apoptosis is low in acetaminophen-induced liver injury and that necrosis is the principal mechanism of acetaminophen-induced liver cell death.21

Because the MPT can initiate both necrotic and apoptotic cell killing, we investigated its role in acetaminophen-induced cytotoxicity of cultured hepatocytes. The data suggest that acetaminophen first induces a CsA-sensitive regulated MPT and then a CsA-insensitive unregulated MPT. Onset of the MPT then initiates necrosis or apoptosis, depending on ATP availability.


MPT, mitochondrial permeability transition; CsA, cyclosporin A; NAPQI, N-acetyl-p-benzoquinonimine; GSH, glutathione; ATP, adenosine triphosphate; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling; HDM, hormonally defined medium; PI, propidium iodide; TMRM, tetramethylrhodamine methylester; PBS, phosphate-buffered saline; FLICA, fluorochrome-labeled inhibitor of caspase.

Materials and Methods

Isolation and Culture of Mouse Hepatocytes.

Animal protocols were approved by the Institutional Animal Care and Use Committee of the University of North Carolina at Chapel Hill. Hepatocytes were isolated from 25- to 30-g overnight-fasted male C3Heb/FeJ mice (Jackson Laboratory, Bar Harbor, ME) via collagenase perfusion through the inferior vena cava, as described previously.22 Hepatocytes were resuspended in Waymouth's medium MB-752/1 containing 2 mmol/L L-glutamine, 10% fetal calf serum, 100 nmol/L insulin, 100 nmol/L dexamethasone, 100 U/mL penicillin, and 100 μg/mL streptomycin. Cell viability was greater than 90%, as determined via trypan blue exclusion. Hepatocytes were plated in 24-well microtiter plates (1.5 × 105 cells per well), 60-mm culture dishes (9 × 105 cells) (Falcon, Lincoln Park, NJ), or 35-mm Petri dishes (6 × 105 cells) with 14-mm microwell cover glasses (MatTek Corporation, Ashland, MA). Plates and coverslips were coated with 0.1% type 1 rat-tail collagen. Hepatocytes were allowed to attach for 4 hours in humidified 5% CO2, 95% air at 37°C. Subsequently, hepatocytes were washed once and then incubated in hormonally defined medium (HDM) containing 240 nmol/L insulin, 2 mmol/L L-glutamine, 1 μg/mL transferrin, 0.3 nmol/L selenium, 1.5 μmol/L free fatty acids, 100 U/mL penicillin, and 100 μg/mL streptomycin in RPMI 1640 medium (Gibco, Rockville, MD) at pH 7.4.

Fluorometric Assay of Cell Viability.

After attachment to 24-well plates, hepatocytes were washed once and replaced with HDM containing 30 μmol/L propidium iodide (PI). Fluorescence was measured using a multiwell fluorescence reader (BMG Lab Technologies, Offenburg, Germany), as described previously.23 Cell killing assessed by PI fluorometry correlates closely with trypan blue exclusion and enzyme release as indicators of oncotic necrosis.

Laser Scanning Confocal Microscopy.

Confocal microscopy after loading of hepatocytes with calcein and tetramethylrhodamine methylester (TMRM) was performed essentially as described previously.22, 24 Briefly, hepatocytes plated on cover glasses were incubated in HDM with 50 mmol/L hydroxyethylpiperazine-N-2 ethanesulfonic acid (pH 7.4) to stabilize pH. After 2.5 hours of treatment with acetaminophen or no addition, cells were loaded with 100 nmol/L TMRM, 1 μmol/L acetoxymethyl ester of calcein, and 3 μmol/L PI in HDM. Images were collected using Zeiss LSM 410 and LSM 510 laser scanning confocal microscopes (Zeiss, Germany) at 37°C. TMRM fluorescence was excited at 543 nm and emission was imaged at more than 560 nm. Calcein fluorescence was excited at 488 nm and emission was collected at 500 to 530 nm using a band pass emission filter.


To measure glutathione, hepatocytes plated on 60-mm culture dishes were washed with phosphate-buffered saline (PBS) and scraped with 10 mmol/L hydroxyethylpiperazine-N-2 ethanesulfonic acid buffer solution. Scraped hepatocytes were frozen at −70°C to break cell membranes. After thawing, total glutathione (GSH + oxidized GSH) in cell lysates was measured with a commercial kit (OXIS International, Portland, OR) according to the manufacturer's instructions.


Hepatocytes plated on 60-mm culture dishes were washed with PBS, and 1 mL of cold perchloric acid (0.6 mol/L) was added. After the cells were scraped, then centrifuged at 9,000g for 1 minute, 0.8 mL of supernatant was taken to which 0.2 mL of 5 mol/L KOH and 0.4 mol/L imidazole were added to neutralize pH to 7.75. After centrifugation at 9,000g for 1 minute, the supernatant was diluted 200-fold in H2O, and ATP was determined with a luciferin–luciferase kit (Promega Enliten, Madison, WI), as described previously.22

Nuclear Morphology.

Hepatocytes plated on cover glasses were fixed with cold 1% paraformaldehyde in PBS. After washing twice with cold PBS, the fixed cells were incubated with HDM containing 10 μmol/L PI for 30 minutes and then treated with 375 μmol/L digitonin to permeabilize plasma membranes. Red nuclear fluorescence was imaged by a laser scanning confocal microscopy using excitation at 568 nm and emission at more than 590 nm.

Detection of Caspase-3 Activation.

FAM-DEVD-fmk, a fluorochrome-labeled inhibitor of caspase (FLICA) (Immunochemistry Technologies, LLC, Bloomington, MN) was used to detect caspase-3 activation. Cell-permeable FLICA inhibitors bind covalently to specific caspases after activation with enhancement of fluorescence.25 After staining with FAM-DEVD-fmk according to the manufacturer's instructions, more than 250 cells were counted for each coverslip; the percentage of staining was expressed as the number of positively stained cells divided by the total number of cells.

TUNEL Assay.

To detect nuclear DNA fragmentation, TUNEL assay was performed using a commercial kit according to the manufacturer's instructions (In Situ Cell Death Detection Kit, Fluorescein, Roche, Indianapolis, IN). Hepatocytes were plated on 40-mm round glass cover slips in 60-mm culture dishes. TUNEL was performed after fixation with 4% paraformaldehyde in PBS. Nuclear staining was determined via green nuclear fluorescence observed through laser scanning confocal microscopy in more than 250 cells. TUNEL staining was expressed as the number of positively stained nuclei divided by the total number of nuclei assessed via bright field microscopy.


Calcein-AM and TMRM were purchased from Molecular Probes (Eugene, OR). Tacrolimus was purchased from Fujisawa Healthcare (Deerfield, IL). NIM811 was a kind gift from Dr. Peter C. Waldmeier of Novartis Pharma AG (Basel, Switzerland). Other chemicals were of analytical grade obtained from the usual commercial sources. When used, acetaminophen was dissolved in absolute ethanol, and CsA, NIM811, and FAM-DEVD-fmk were dissolved in dimethyl sulfoxide.


Differences between means were compared via ANOVA followed by Tukey's multiple comparison procedure using a P value of less than .05 as the criterion for significance. P values of less than .01 were also noted. Data were expressed as the mean ± SE.


Delay by MPT Blockers of Acetaminophen-Induced Necrotic Cell Killing.

Acetaminophen-induced necrotic cell killing was measured using a PI fluorescence assay. Acetaminophen (10 mmol/L) caused time-dependent necrotic cell killing. After acetaminophen, necrotic killing increased to 49% ± 2% after 6 hours and 74% ± 2% after 16 hours (Fig. 1A). CsA (2 μmol/L, a specific MPT blocker) temporarily decreased cell killing to 26% ± 1% after 6 hours. However, protection by CsA was lost after 16 hours. NIM811 (2 μmol/L), a non-immunosuppressive CsA analog,26 similarly decreased acetaminophen-induced hepatocyte killing after 6 hours but did not protect after 16 hours (Fig. 1B). By contrast, tacrolimus, an immunosuppressant that does not block the MPT,27 did not prevent acetaminophen-induced cell killing (see Fig. 1A).

Figure 1.

Acetaminophen-induced necrotic cell killing and GSH depletion in mouse hepatocytes: delay by CsA and NIM811. Isolated mouse hepatocytes were exposed to acetaminophen (10 mmol/L). Cell viability was determined via PI fluorometry. (A) Hepatocytes were treated with CsA (2 μmol/L) and tacrolimus (FK506) (2 μmol/L) beginning 20 minutes prior to acetaminophen administration. (B) Hepatocytes were treated with NIM811 (2 μmol/L) or no addition prior to acetaminophen administration. (C) Mouse hepatocytes were exposed to acetaminophen in the presence or absence of CsA. After 0 to 2 hours, cell extracts were prepared, and GSH levels were measured as described in Materials and Methods. In all panels, control represents hepatocytes unexposed to acetaminophen. Values are the mean ± SE from 3 or more hepatocyte isolations. *P < .01 versus acetaminophen; †P < .05 versus control. AAP, acetaminophen; CsA, cyclosporin A; Tac, tacrolimus.

To investigate whether CsA affected the metabolism of acetaminophen, GSH in hepatocytes was measured (Fig. 1C). NAPQI, the toxic metabolite of acetaminophen, binds to GSH covalently and causes GSH depletion.3, 4 Thus, after exposure of hepatocytes to acetaminophen the rate of GSH depletion parallels the rate of NAPQI formation.7, 28, 29 After acetaminophen, GSH decreased by approximately 50% after 1 hour and 80% after 2 hours (see Fig. 1C). GSH in untreated hepatocytes did not change significantly. When hepatocytes were treated with acetaminophen in the presence of CsA, GSH depletion was virtually identical (Fig. 1C). These findings are consistent with the conclusion that CsA does not inhibit NAPQI formation.

Protection by Fructose Plus Glycine Against Acetaminophen-Induced Necrotic Cell Death.

Hepatocytes were exposed to acetaminophen in the presence of fructose (20 mmol/L) and glycine (5 mmol/L) beginning 20 minutes prior to acetaminophen administration and then continuously afterwards. Fructose is a glycolytic substrate that protects against ATP depletion–dependent necrotic cell killing.17, 30–32 Glycine is a membrane-stabilizing amino acid that prevents plasma membrane failure and onset of necrotic cell killing.22, 33, 34 Fructose plus glycine prevented acetaminophen-induced necrotic cell killing almost completely 16 hours after acetaminophen administration (Fig. 2A). The action of fructose plus glycine was synergistic, because the protection by each compound alone was weak or absent (data not shown). In additional experiments, fructose plus glycine prevented acetaminophen-induced ATP depletion over the first 6 hours of exposure (Fig. 2B).

Figure 2.

Protection by fructose plus glycine against acetaminophen-induced necrotic cell killing and ATP depletion. Hepatocytes were exposed to acetaminophen as described in Fig. 1. Some hepatocytes were treated with fructose (20 mmol/L) plus glycine (5 mmol/L) at 20 minutes before acetaminophen. (A) Cell killing was measured via PI fluorometry. (B) ATP was measured via luciferase assay after 0 to 6 hours. ATP is expressed as the percentage of initial ATP in each group prior to acetaminophen addition. Controls are hepatocytes unexposed to acetaminophen. Values are the mean ± SE from 3 or more hepatocyte isolations. *P < .01 versus acetaminophen; **P < .05 versus acetaminophen. AAP, acetaminophen; F, fructose; G, glycine; ATP, adenosine triphosphate.

The MPT After Acetaminophen Administration.

To assess the ability of acetaminophen to induce the MPT, mouse hepatocytes were loaded with TMRM and calcein. After 3.5 hours, mitochondria were bright red-fluorescing spheroids in confocal images (Fig. 3A, upper row), which indicated these mitochondria were polarized. Additionally, the red-fluorescing mitochondria excluded the green fluorescence of calcein, which showed impermeability of the mitochondria to this 623-Da solute (see Fig. 3A, lower row). After 4 hours and 20 minutes, mitochondria began to lose TMRM fluorescence and fill with calcein, which indicated mitochondrial depolarization and inner membrane permeabilization, respectively. Cell-surface blebbing, which is an early indication of cell injury, also became evident. After a few more minutes, mitochondria lost TMRM fluorescence completely and some of the cells lost viability, as indicated by nuclear staining with PI and release of cytosolic calcein into the medium. By contrast, in the presence of CsA, mitochondria remained polarized and continued to exclude the cytosolic calcein for up to 10 hours after acetaminophen (Fig. 3B). Subsequently after 11 hours, mitochondria of the acetaminophen- and CsA-treated hepatocytes depolarized and became permeable to calcein, and the cells began to bleb. These results show that CsA delayed but did not prevent the onset of the MPT in acetaminophen-treated mouse hepatocytes.

Figure 3.

Acetaminophen induction of mitochondrial depolarization and inner membrane permeabilization: delay by CsA. (A, B) Hepatocytes were exposed to acetaminophen (10 mmol/L) in the absence (A) or presence (B) of CsA (2 μmol/L), as described in Fig. 1. After 2.5 hours, hepatocytes were loaded with TMRM (100 nmol/L), PI (3 μmol/L), and acetoxymethyl ester of calcein (1 μmol/L). The red fluorescence of TMRM and PI and the green fluorescence of calcein were imaged using laser scanning confocal microscopy, as described in Materials and Methods. Each experiment is typical of 3 or more replicates. TMRM, tetramethylrhodamine methylester; PI, propidium iodide; AAP, acetaminophen; CsA, cyclosporin A.

Onset of the MPT in the presence of fructose plus glycine was also assessed with confocal microscopy. After acetaminophen was added in the presence of fructose plus glycine, mitochondria released TMRM fluorescence and took up calcein fluorescence within 6 hours of acetaminophen addition. However, cell killing did not follow, as indicated by the absence of nuclear staining with PI after 7 hours (Fig. 4A). When hepatocytes were treated additionally with CsA in the presence of fructose plus glycine, mitochondria retained TMRM and excluded calcein after acetaminophen for up to 13 hours, at which time mitochondria released TMRM and took up calcein (Fig. 4B). However, hepatocytes retained viability, as assessed by the absence of nuclear PI staining after 15 hours.

Figure 4.

Onset of the mitochondrial permeability transition after acetaminophen and fructose plus glycine. Mouse hepatocytes were treated with acetaminophen; treated with fructose plus glycine without (A) and with (B) CsA and loaded with TMRM, PI, and calcein; and imaged, as described in Fig. 2. Each experiment is typical of 3 or more replicates. TMRM, tetramethylrhodamine methylester; PI, propidium iodide; AAP, acetaminophen; F, fructose; G, glycine; CsA, cyclosporin A.

Apoptosis in Acetaminophen-Treated Hepatocytes.

Apoptosis of acetaminophen-treated hepatocytes was evaluated via several end points. To examine nuclear morphological changes, hepatocytes were fixed with paraformaldehyde after 30 hours, digitonin-permeabilized, and labeled with PI. After acetaminophen alone, 27% of hepatocytes showed chromatin condensation or nuclear fragmentation, both of which are characteristics of apoptosis (Fig. 5A). After administration of acetaminophen and fructose plus glycine to prevent necrotic cell killing, apoptotic nuclear changes developed in 55% of hepatocytes despite the protection against necrotic cell death (see Fig. 5B).

Figure 5.

Nuclear fragmentation and chromatin condensation after acetaminophen exposure. Mouse hepatocytes were incubated with 10 mmol/L acetaminophen without (A) and with (B) fructose plus glycine, as described in Fig. 2. After 30 hours, hepatocytes were fixed, permeabilized, and stained with PI (10 μmol/L), as described in Materials and Methods. Red nuclear PI fluorescence was imaged by laser scanning confocal microscopy. Arrows identify apoptotic chromatin condensation and/or nuclear fragmentation. AAP, acetaminophen; F, fructose; G, glycine.

Caspase-3 is a key executioner of apoptosis.35 To investigate caspase-3 activation in hepatocytes exposed to acetaminophen, FAM-DEVD-fmk fluorescence was monitored via confocal microscopy because it binds covalently to activated caspase-3 with enhancement of fluorescence. Hepatocytes were incubated for 9 hours and then exposed to FAM-DEVD-fmk for 1 hour prior to fixation. In the absence of acetaminophen, almost no FAM-DEVD-fmk fluorescence was observed (Fig. 6). By contrast, after exposure to acetaminophen, 30% ± 6% of hepatocytes became FAM-DEVD-fmk positive. Treatment with acetaminophen in combination with fructose plus glycine increased to 61% ± 3% the number of FLICA-positive hepatocytes. CsA decreased FAM-DEVD-fmk staining after acetaminophen and fructose plus glycine treatment to 28% ± 3%, comparable to apoptosis after acetaminophen alone (see Fig. 6).

Figure 6.

Caspase-3 activation after acetaminophen exposure. Mouse hepatocytes were incubated with no addition, acetaminophen alone, acetaminophen with fructose plus glycine, and acetaminophen with fructose plus glycine and CsA, as described in Figs. 3 and 4. After 9 hours, hepatocytes were incubated with FAM-DEVD-fmk for 1 hour, and green fluorescence was imaged using laser scanning confocal microscopy as described in Materials and Methods and superimposed with bright field images. In panel A, representative FLICA fluorescence is shown for untreated hepatocytes (control) and hepatocytes treated with acetaminophen, acetaminophen and fructose plus glycine, and acetaminophen, fructose plus glycine and CsA. In B, the percentages of FLICA-stained cells are plotted for the experimental groups. Results are the mean ± SE from 3 or more hepatocyte isolations. *P < .01 versus control; **P < .01 versus acetaminophen; ***P < .01 versus acetaminophen and fructose plus glycine. AAP, acetaminophen; F, fructose; G, glycine; CsA, cyclosporin A; FLICA, fluorochrome-labeled inhibitor of caspase.

TUNEL was used to detect DNA fragmentation in situ. In untreated hepatocytes, virtually no TUNEL staining occurred (Fig. 7). After exposure to acetaminophen for 16 hours, 31% ± 4% of cells became TUNEL-positive. DEVD-fmk, a caspase-3 inhibitor, prevented TUNEL almost completely (see Fig. 7), although it did not prevent necrotic cell killing evaluated via PI fluorometry (data not shown). After fructose plus glycine treatment of acetaminophen-exposed hepatocytes, TUNEL of nuclei increased to 59% ± 5%. CsA decreased TUNEL of hepatocytes treated with acetaminophen and fructose plus glycine to 26% ± 6%, whereas DEVD-fmk prevented TUNEL almost completely (see Fig. 7). Thus, different independent indices of apoptosis evaluated 9, 16, and 30 hours after acetaminophen exposure showed 27% to 31% apoptosis. Thus, acetaminophen-induced apoptosis was enhanced by fructose plus glycine and was partially prevented by CsA.

Figure 7.

TUNEL assay of acetaminophen-treated hepatocytes. Mouse hepatocytes were incubated with acetaminophen, as described in Figs. 3 and 4. Some hepatocytes were exposed to FAM-DEVD-fmk (100 μmol/L) and/or fructose plus glycine with and without CsA. After 16 hours, TUNEL staining was performed, as described in Materials and Methods. Results are the mean ± SE from 3 or more hepatocyte isolations. *P < .01 versus control; §P < .01 versus acetaminophen; **P < .05 versus acetaminophen; †P < .01 versus acetaminophen and fructose plus glycine. AAP, acetaminophen; F, fructose; G, glycine; TUNEL, terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling; CsA, cyclosporin A.


Acetaminophen caused time- and dose-dependent killing of cultured mouse hepatocytes. CsA delayed cell killing after acetaminophen in agreement with earlier reports of cytoprotection against acetaminophen in rat liver slices with the combination of CsA, trifluoperazine, and fructose and in human hepatoma cells with CsA alone.12, 13 A recent report also showed that CsA decreases lethality in acetaminophen-treated mice.13 In the present work, protection by CsA against acetaminophen was lost after more than 9 hours of exposure (see Fig. 1A). CsA is an inhibitor of the MPT,14 which suggests that CsA exerts protection by blocking the MPT, at least at early time points. However, CsA also inhibits the protein phosphatase, calcineurin, which mediates the immunosuppressive action of CsA.36 Accordingly, we evaluated NIM811, a non-immunosuppressive CsA analog that inhibits the MPT similarly to CsA.26 NIM811, like CsA, was protective at early time points but not late time points (see Fig. 1B). Tacrolimus is an immunosuppressant and calcineurin inhibitor that does not block the MPT.27 Tacrolimus, unlike CsA and NIM811, did not delay cell killing after acetaminophen (see Fig. 1A).

To confirm that acetaminophen was inducing the MPT in situ, we used confocal microscopy to monitor mitochondrial membrane potential and inner membrane permeability. After acetaminophen treatment, the MPT evidenced by mitochondrial depolarization and inner membrane permeabilization occurred after an average of approximately 4.5 hours. CsA delayed onset of the MPT after acetaminophen administration to 10 to 11 hours on average (see Fig. 3). With or without CsA, cell death quickly followed mitochondrial depolarization and inner membrane permeabilization, as indicated by nuclear labeling with PI. Taken together, these findings are consistent with the conclusion that in mouse hepatocytes, acetaminophen induces the MPT, which is delayed but not prevented by CsA.

The delay of the MPT and cell death by CsA could not be ascribed to an inhibition of the metabolic activation of acetaminophen. NAPQI formed by acetaminophen metabolism reacts with GSH, causing GSH depletion. Thus, the rate of GSH depletion parallels the rate of NAPQI formation under these conditions.7, 28, 29 Because the time course and extent of GSH depletion in hepatocytes after acetaminophen exposure was nearly identical with and without CsA (see Fig. 1C), we can conclude that CsA likely does not inhibit the bioactivation of acetaminophen. Thus, the delay of the MPT and cell death by CsA appears to be due to an effect on MPT pores rather than prevention of the initiating events of acetaminophen hepatotoxicity.

Although CsA inhibition is often regarded as an essential and identifying feature of the MPT, numerous reports show that a CsA-insensitive MPT can occur under a variety of circumstances.18, 37–39 Such observations have led to the proposal that MPT pores have two open conductance modes: regulated and unregulated.18 Regulated MPT pore opening occurs after low level chemical induction, requires mitochondrial Ca2+ uptake, and is inhibited by CsA. Unregulated MPT pore opening occurs after high level chemical induction. The unregulated MPT occurs in the absence of Ca2+ and is not inhibited by CsA.

The findings of the present study are consistent with the proposal that acetaminophen initially induces a regulated MPT, but after prolonged exposure an unregulated MPT occurs. Thus, CsA initially prevents mitochondrial depolarization, inner membrane permeabilization, and cell death after treatment with acetaminophen. After longer exposure to acetaminophen, however, CsA-insensitive depolarization and permeabilization occur, representing late onset of an unregulated MPT. Unfortunately, a specific inhibitor of the unregulated MPT has not been identified that could more directly test the hypothesis that an unregulated MPT underlies CsA-insensitive mitochondrial depolarization and inner membrane permeabilization after acetaminophen treatment. Future studies will also be needed to determine the specific intracellular events precipitating both the regulated and unregulated MPT, such as mitochondrial protein adduct formation with NAPQI, calcium dysregulation, and enhanced formation of reactive oxygen and nitrogen species.

The role of apoptotic cell death in acetaminophen hepatotoxicity is controversial.19–21 To evaluate apoptosis, several assays were employed, including caspase-3 activation (see Fig. 6), TUNEL (see Fig. 7), and nuclear morphological changes (see Fig. 5). After treatment with 10 mmol/L acetaminophen, apoptosis developed in 30%, 31%, and 27% of cells after 9, 16, and 30 hours as accessed, respectively, through the three different assays.

The interaction between apoptosis and necrosis was accessed by treating hepatocytes with fructose plus glycine. Fructose is an ATP-generating glycolytic substrate that protects hepatocytes against cell death after mitochondrial injury.17, 30–32 Glycine is a plasma membrane–stabilizing amino acid that blocks opening of nonspecific anion channels that lead to colloid osmotic swelling, plasma membrane failure, and onset of oncotic necrosis.33, 34 Fructose plus glycine almost completely prevented acetaminophen-induced necrotic cell killing (see Fig. 2A), but the MPT still occurred, as assessed by mitochondrial depolarization and inner membrane permeabilization (see Fig. 4). In fructose plus glycine-treated hepatocytes, CsA delayed but did not prevent the MPT after acetaminophen, just as CsA had delayed but not prevented the MPT in the absence of fructose plus glycine. These observations indicate that fructose plus glycine blocks acetaminophen-induced necrotic cell death downstream of regulated and unregulated MPT pore opening.

Previously, fructose was shown to protect against lactate dehydrogenase release induced by acetaminophen in rat liver slices.32 Fructose protection is generally attributed to glycolytic ATP formation and can be overcome through the activation of mitochondrial ATPases with an uncoupler.30, 31 Because a role for ATP in fructose protection against acetaminophen toxicity has been questioned,40 we measured ATP in acetaminophen-exposed hepatocytes. Fructose plus glycine stabilized ATP levels and prevented acetaminophen-induced ATP depletion (see Fig. 2B). However, in distinct contrast to this protection against necrotic cell killing, fructose plus glycine increased acetaminophen-induced apoptosis, as assessed by nuclear morphological changes (see Fig. 5), caspase-3 activation (see Fig. 6), and TUNEL (see Fig. 7). This apoptosis was nearly completely inhibited by the caspase-3 inhibitor FAM-DEVD-fmk and over 50% inhibited by CsA (see Fig. 7). FAM-DEVD-fmk did not prevent acetaminophen-induced necrotic killing, as measured with PI fluorescence (data not shown). Inhibition of apoptosis by CsA is consistent with the conclusion that the MPT also promotes caspase-dependent apoptosis in acetaminophen-treated hepatocytes through MPT-dependent mitochondrial swelling, outer membrane rupture, and cytochrome c release. Previously in models of ischemia/reperfusion and calcium ionophore toxicity to cultured hepatocytes, onset of the MPT induced both apoptosis and necrosis as regulated by ATP availability.16, 17 When MPT onset induced profound ATP depletion, rapid caspase-independent necrotic cell death ensued. However, when necrosis was prevented by glycine and fructose, apoptosis developed instead—provided that ATP depletion was at least partially prevented, because ATP (or dATP) is required for cytochrome c–dependent activation of caspase-3.41–44

A similarly operating ATP switch between necrosis and apoptosis appeared to occur during acetaminophen hepatotoxicity (Fig. 8). In the absence of fructose plus glycine, MPT onset precipitated ATP depletion (see Fig. 2B) and oncotic necrosis (see Fig. 1A, B). By contrast, in the presence of fructose plus glycine, ATP levels were sustained (see Fig. 2B), which prevented necrosis (see Fig. 2A) but promoted caspase-dependent apoptosis (see Figs. 5–7). Thus, the balance of ATP hydrolysis after the MPT and ATP synthesis by glycolysis seems to operate a switch between necrosis and apoptosis during acetaminophen cytotoxicity. However, unlike ischemia/reperfusion and calcium ionophore-induced killing, CsA did not completely prevent apoptosis (see Figs. 6, 7), which is consistent with involvement of an unregulated MPT in acetaminophen cytotoxicity. Moreover, unlike ischemia/reperfusion and calcium ionophore cytotoxicity, glycine and fructose added individually were poorly protective against acetaminophen-induced killing. These differences in ischemia/reperfusion and ionophore-induced injury may reflect the greater overall toxic stress of acetaminophen.

Figure 8.

Schematic of MPT-dependent acetaminophen-induced hepatocyte killing. Acetaminophen induces NAPQI formation, GSH depletion, and onset of the MPT. CsA blocks the MPT early in the injury, but the MPT becomes insensitive to CsA later. Onset of the MPT leads to either necrotic or apoptotic cell killing. If ATP depletion occurs, glycine-sensitive membrane failure and necrotic cell killing occur. If ATP levels are partially maintained by glycolysis, cytochrome c released by swollen mitochondria causes ATP-dependent caspase-3 activation, leading to apoptosis. NAPQI, N-acetyl-p-benzoquinonimine; GSH, glutathione; MPT, mitochondrial permeability transition; CsA, cyclosporin A; ATPase, adenosine triphosphatase; ATP, adenosine triphosphate.

In cultured hepatocytes, the conversion of cell death from necrosis to apoptosis might seem of little overall benefit, because the hepatocytes die in either event. In vivo, however, the conversion to apoptosis may be beneficial. In necrosis, intracellular contents are released that induce a strong inflammatory response from resident hepatic macrophages and circulating leukocytes. Such inflammation extends and amplifies acetaminophen-induced injury.45, 46 By contrast, in apoptosis the inflammatory response may be blunted. As a consequence, overall injury is likely ameliorated, perhaps substantially. Our observation that fructose, a glycolytic substrate, contributes to protection against acetaminophen-induced toxicity is also relevant to the clinical observation that fasted individuals are more susceptible to acetaminophen hepatotoxicity. Although modest decreases in hepatic GSH may contribute to this increased susceptibility,47, 48 the depletion of hepatic glycogen as an ATP-generating glycolytic substrate likely also contributes to vulnerability after fasting. Glycogen acts similarly to fructose in preventing anoxia and drug-induced necrotic cell killing of hepatocytes.30In vivo, glycogen may do the same by preventing necrotic cell killing or by converting oncotic necrosis to apoptosis, which is better tolerated.

In conclusion, acetaminophen induces the MPT in cultured mouse hepatocytes, which is regulated early and unregulated late after addition of the toxicant. The MPT then leads to both necrosis and apoptosis (see Fig. 8). When the MPT causes rapid ATP depletion, the plasma membrane ruptures to precipitate oncotic necrosis. However, when necrosis is prevented by fructose plus glycine, caspase-dependent apoptosis occurs. Thus, pathways leading to apoptosis and necrosis share the MPT. Such sharing of these pathways represents and likely explains observations of both apoptosis and necrosis in acetaminophen hepatotoxicity.21, 49, 50


The authors thank Elizabeth A. Doyal for technical assistance.