Deoxyribonuclease 1 aggravates acetaminophen-induced liver necrosis in male CD-1 mice


  • Potential conflict of interest: Nothing to report.


An overdose of acetaminophen (APAP) (N-acetyl-p-aminophenol) leads to hepatocellular necrosis induced by its metabolite N-acetyl-p-benzoquinone-imine, which is generated during the metabolic phase of liver intoxication. It has been reported that DNA damage occurs during the toxic phase; however, the nucleases responsible for this effect are unknown. In this study, we analyzed the participation of the hepatic endonuclease deoxyribonuclease 1 (DNASE1) during APAP-induced hepatotoxicity by employing a Dnase1 knockout (KO) mouse model. Male CD-1 Dnase1 wild-type (WT) (Dnase1+/+) and KO (Dnase1−/−) mice were treated with 2 different doses of APAP. Hepatic histopathology was performed, and biochemical parameters for APAP metabolism and necrosis were investigated, including depletion of glutathione/glutathione-disulfide (GSH+GSSG), β-nicotinamide adenine dinucleotide (NADH+NAD+), and adenosine triphosphate (ATP); release of aminotransferases and Dnase1; and occurrence of DNA fragmentation. As expected, an APAP overdose in WT mice led to massive hepatocellular necrosis characterized by the release of aminotransferases and depletion of hepatocellular GSH+GSSG, NADH+NAD+, and ATP. These metabolic events were accompanied by extensive DNA degradation. In contrast, Dnase1 KO mice were considerably less affected. In conclusion, whereas the innermost pericentral hepatocytes of both mouse strains underwent necrosis to the same extent independent of DNA damage, the progression of necrosis to more outwardly located cells was dependent on DNA damage and only occurred in WT mice. Dnase1 aggravates APAP-induced liver necrosis. (HEPATOLOGY 2006;43:297–305.)

Acetaminophen, a widely used analgesic, is one of the most commonly overdosed pharmaceuticals.1 Overdose leads to liver necrosis caused by the electrophilic metabolite N-acetyl-p-benzoquinone-imine (NAPQI), which is generated via oxidation of acetaminophen (APAP) by a microsomal cytochrome P450 containing mixed-function oxidase system.2 Subsequent detoxication of NAPQI by cellular glutathione leads to depletion of glutathione/glutathione-disulfide (GSH+GSSG) followed by NAPQI accumulation and induction of necrosis due to protein arylation and severe oxidative stress.2, 3 Nuclear accumulation of Ca2+ and DNA damage have been described to occur during APAP-induced cytotoxicity of murine hepatocytes in vitro and in vivo.4–6 DNA damage seems to be an important event, because APAP toxicity was prevented by known inhibitors of Ca2+-dependent endonucleases, including Ca2+-chelators such as aurintricarboxylic acid, glycol ether diamine tetra-acetic acid, and the Ca2+-calmodulin antagonist chlorpromazine.5, 7, 8 However, the nucleases involved in APAP-induced DNA damage have not yet been identified.

Deoxyribonuclease 1 (DNASE1), a Ca2+/Mg2+-dependent secreted endonuclease with a pH optimum of approximately 7.5, hydrolyzes double-stranded DNA generating 5′-phospho-tri- and/or tetra-oligonucleotides.9, 10DNASE1 gene expression has been demonstrated in a number of different organs of the urogenital and alimentary tracts as well as in the murine liver.11 In a previous report, we showed that Dnase1 deficiency in the mouse promotes anti-DNA autoimmunity, suggesting a protective role of Dnase1 through its ability to degrade chromatin presumably liberated during cell death.12 In addition, we have demonstrated that extracellular DNASE1 diffuses into necrotic cells, and together with the plasminogen system facilitates necrotic chromatin breakdown in vitro.13 Interestingly, we found that DNASE1 not only facilitates chromatin breakdown of necrotic cells post mortem but is also a critical factor for the induction of necrotic cell death.14 Thus, administration of cisplatin to wild-type (WT) and Dnase1 knockout (KO) mice resulted in a high rate of necrosis of proximal tubular epithelial cells in WT mice but was strongly reduced in Dnase1 KO mice, indicating that DNA damage is a critical event for the induction and promotion of necrosis.14

In the present study, we compared the role of Dnase1 during APAP-induced liver necrosis in WT and Dnase1 KO mice. The data obtained indicate that compared with WT mice, APAP in Dnase1 KO mice led to a strongly reduced progression of pericentral hepatocellular necrosis. We attribute this difference to almost complete energy depletion in WT hepatocytes as a consequence of the stimulation of DNA repair mechanisms, such as poly–adenosine diphosphate–ribose polymerase (PARP) activation due to DNA fragmentation catalyzed by Dnase1. Our data suggest that Dnase1-induced DNA damage aggravates energy depletion in WT hepatocytes, resulting in necrosis of a large number of pericentral hepatocytes.


PAPAP, acetaminophen; DNASE1, deoxyribonuclease 1; KO, knockout; WT, wild-type; GSH+GSSG, glutathione/glutathione-disulfide; NADH+NAD+, β-nicotinamide adenine dinucleotide; ATP, adenosine tri-phosphate; NAPQI, N-acetyl-p-benzoquinone-imine; PARP, poly–adenosine diphosphate–ribose polymerase; HE, hematoxylin/eosin; TUNEL, terminal deoxyribonucleotidyl transferase dUTP nick end-labeling; ALT, alanine aminotransferase; AST, aspartate aminotransferase.

Materials and Methods

Animal Treatments.

All mice were bred in our animal facility. The animals were allowed free access to standard laboratory chow and water and were kept on a light/dark cycle of 12 hours (light from 7:00 A.M. to 7.00 P.M.). All animal procedures performed during the study received prior approval from the local animal protection committee and were in agreement with the criteria outlined in the National Academy of Sciences Guide for the Care and Use of Laboratory Animals. Male CD-1 (ICR) WT and Dnase1 KO mice (3–6 months old) weighing 28–35 g were fasted for 12 hours before application of a single dose of 600 or 800 mg/kg body weight APAP (dissolved in 50% [v/v] propylene glycol/water) via intraperitoneal injection at 9.00 A.M.. After different time points of APAP exposure as indicated in the figures, the mice (4 animals per time point) were anesthetized with ether, and blood was collected via cardiac puncture for serum preparation. Animals were subsequently killed via cervical translocation and the livers were removed, frozen in N2, and stored at −80°C for further analysis. As control animals, we used mice that were fasted for 12 hours and killed at 9.00 A.M. and mice that were fasted for 12 hours, given a vehicle injection at 9.00 A.M., and fasted for further 10 hours. For induction of hepatocellular apoptosis, mice fasted overnight received an intraperitoneal injection of 80 μg of a monoclonal hamster antimurine CD95 antibody (Jo2; BD Pharmingen, Heidelberg, Germany) dissolved in phosphate-buffered saline. The mice were killed 2 to 3 hours later.

Histopathology by Hematoxylin/Eosin Staining and TUNEL.

Parts of the left hepatic lobe were fixed in 10% (w/v) paraformaldehyde dissolved in phosphate-buffered saline for 24 hours; embedded in paraffin; and sectioned and stained with hematoxylin/eosin (HE). For terminal deoxyribonucleotidyl transferase dUTP nick end-labeling (TUNEL), rehydrated sections were permeabilized with 0.5% (w/v) pepsin dissolved in 0.01 mol/L HCl for 10 minutes at 37°C, and endogenous peroxidase was quenched by 3% (v/v) H2O2 diluted in methanol for 5 minutes at room temperature. Labeling of 3′-OH ends with biotinylated nucleotides was performed with a terminal deoxyribonucleotidyl transferase in situ kit from R&D Systems (Wiesbaden, Germany). Detection was achieved using a streptavidin–biotin complex conjugated to horseradish peroxidase employing 3-amino-9-ethylcarbazole as a chromogen substrate (Dako Cytomation, Hamburg, Germany). Bovine DNASE1 was used as a positive control for the procedure using sections pretreated with 10 μg/mL. Stained sections were examined via light microscopy; photographs were taken with an AxioCam HRc using AxioVision 3.0 software (Zeiss, Göttingen, Germany).

DNA Fragmentation.

Hepatic DNA breakdown was measured spectrophotometrically following the procedure described by Ray and Jena15 with slight modifications. Liver samples were homogenized (10% w/v) in ice-cold lysis buffer (10 mmol/L Tris/HCl, 20 mmol/L EDTA, 0.5% [v/v] Triton X-100 [pH 8.0]), the cell debris was sedimentated (22,000g for 15 min at 4°C), and the DNA content of the cytoplasmic (supernatant) and nuclear (pellet) fractions was determined. DNA fragmentation is expressed as a percentage of the total cellular DNA appearing in the supernatant fraction. For qualitative analysis of DNA fragmentation, the DNA of 500 μL of liver cytoplasmic fraction was isolated using the DNeasy Tissue Kit (Qiagen, Hilden, Germany) and analyzed via 1.5% (w/v) tris/borate/EDTA buffer agarose gel electrophoresis.16

Analytical Procedures for Quantification of Hepatotoxicity.

Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels were determined using the DG159-UV kit and DG158-UV kit, respectively, from Sigma Aldrich (Deisenhofen, Germany). Total GSH+GSSG content of 10% (w/v) liver homogenates prepared in 10% (w/v) tri-chloro-acetic acid was measured according to a method originally described by Tietze17 and modified by Griffith.18 Adenosine triphosphate (ATP) levels of neutralized 1% (w/v) liver homogenates prepared in 0.5 N perchloric acid were determined using the ATP Determination Kit (A-22066) from Molecular Probes Europe BV (Leiden, The Netherlands). Total β-nicotinamide adenine dinucleotide (NADH+NAD+) levels of 2% (w/v) liver homogenates prepared in 40 mmol/L NaOH and 0.5 mmol/L cysteine were determined via fluorescence measurements using the enzymatic cycling method described by Klingenberg.19 Serum Dnase1 activity was quantified using the single radial enzyme diffusion assay and the native polyacrylamide gel electrophoresis zymogram technique as previously described.11, 12


Immunoblotting was performed on liver homogenates prepared as described above for quantitative DNA fragmentation assay. Samples were subjected to SDS-PAGE followed by immunoblotting using an alkaline phosphatase detection system.16 The following primary antibodies and dilutions were used: murine anti–β-actin immunoglobulin M ascites (Oncogene, Schwalbach, Germany), 1:500; rabbit anti–BCL-XL/S (Santa Cruz Biotechnology, Heidelberg, Germany), 1:200; rabbit anti–caspase-3 (active) (R&D Systems), 1:500; and rabbit anti-PARP (Santa Cruz Biotechnology), 1:200. As secondary antibodies, we used either a chicken anti-rabbit or chicken anti-mouse antibody conjugated with alkaline phosphatase in a dilution of 1:2,000 (Santa Cruz Biotechnology).

Statistical Analysis.

For comparison of difference of several time points versus single control and genotypes (WT vs. KO), two-way ANOVA with Fisher LSD post hoc analysis was performed. In parallel, comparison of parameters in WT vs. KO, as well as the nuclease activity in WT animals, was performed using two-tailed Student t test with Bonferroni's adjustment and Mann-Whitney U test. All of the above-mentioned comparisons were confirmed by Levene's test of the homogeneity of variances. A P value of less than .05 was considered statistically significant. The datasets that passed all applied tests were considered significant.


Quantification of APAP-Induced Liver Necrosis in Dnase1+/+ and Dnase1−/− Mice.

To elucidate the role of Dnase1 in APAP-induced liver necrosis, we exposed male CD-1 WT and Dnase1 KO mice with a single dose of 600 or 800 mg APAP/kg body weight and quantified the ensuing hepatocellular damage by measuring the release of liver aminotransferases into the serum. After exposure to 600 mg/kg APAP for 6 to 12 hours, we observed no difference between WT and Dnase1 KO animals; however, at 18 hours, the AST/ALT serum levels reached values of 2,697 ± 1,664 U/L/2,940 ± 1,131 U/L in WT mice and 1,318 ± 707 U/L/538 ± 242 U/L in KO mice (Fig. 1A–B). These data demonstrate that between 12 and 18 hours after APAP application, hepatocellular necrosis progressed more dramatically in WT mice compared with Dnase1 KO mice. Statistically significant AST/ALT increases were observed at 18 hours in both WT and KO animals (P < .001). No significant difference between WT and KO was observed with AST at 0 through 18 hours, while ALT was significantly different at 18 hours after 600 mg/kg APAP injection (P < .001).

Figure 1.

Quantification of APAP hepatotoxicity in Dnase1+/+ and Dnase1−/− mice treated with 600 mg/kg APAP. Time-dependent release of (A) AST and (B) ALT illustrating first a comparable (0–12 h) and afterward a reduced (12–18 h) sensitivity of Dnase1−/− (KO) mice toward APAP-induced hepatotoxicity compared with Dnase1+/+ (WT) mice. (C) Time-dependent increase in serum nuclease activity as analyzed via single radial enzyme diffusion assay in WT mice treated with APAP. *P < .05 compared with WT untreated 0-hour control. #P < .05 in comparison with KO 0-hour control. ΔP < .05 in comparison with treated WT mice at the same time point. AST, aspartate aminotransferase; WT, wild-type; KO, knockout; Ctrl, control; ALT, alanine aminotransferase; Dnase1, deoxyribonuclease 1.

Application of the higher dose of 800 mg/kg APAP accelerated hepatocellular damage by approximately 10 hours as deduced from AST/ALT measurements (compare Fig. 1A–B with Fig. 2A–B). Although significant increases of AST and ALT occurred at 2 and 4 hours (P < .05), we observed no significant difference between WT mice and Dnase1 KO mice in the initial increase in AST/ALT levels (first 2 hours for AST and 4 hours for ALT) (Fig. 2A–B). Thereafter, the liver damage progressed more severely in WT mice, reaching maximal values after 8 hours of 2,872 ± 186 U/L/2,162 ± 212 U/L in WT mice but only 1,557 ± 480 U/L/911 ± 412 U/L in the Dnase1 KO cohort. Between 12 and 24 hours after application of 800 mg/kg APAP, most of the WT mice became moribund and died, whereas the Dnase1 KO animals survived the 24-hour period.

Figure 2.

Quantification of APAP hepatotoxicity in Dnase1+/+ and Dnase1−/− mice treated with 800 mg/kg APAP. Time-dependent release of (A) AST and (B) ALT illustrating first a comparable (0–2 h) and afterward a reduced (2–10 h) sensitivity of Dnase1−/− (KO) mice toward APAP-induced hepatotoxicity compared with Dnase1+/+ (WT) mice. Note that signs of hepatotoxicity as indicated by AST and ALT levels were detected approximately 10 hours earlier than for mice treated with 600 mg/kg APAP (compare with Fig. 1A–B). *P < .05 in comparison with WT untreated 0-hour control. #P < .05 in comparison with KO 0-hour control. ΔP < .05 in comparison with treated WT mice at the same time point. AST, aspartate aminotransferase; WT, wild-type; KO, knockout; Ctrl, control; ALT, alanine aminotransferase.

In WT mice, we also analyzed serum Dnase1 activity after 600 mg/kg APAP exposure. Paralleling AST/ALT release, we found a continuous increase in serum Dnase1 activity as measured via single radial enzyme diffusion assay, resulting in a 40-fold higher activity after 18 hours (Fig. 1C). Because the single radial enzyme diffusion assay is a general DNase assay, we confirmed the increase in serum Dnase1 activity via native polyacrylamide gel electrophoresis zymograms11 (data not shown). These data indicate that Dnase1 must have been liberated during hepatocellular necrosis, preferentially from hepatocellular secretory compartments, and/or from further APAP-sensitive, Dnase1-synthesizing cells.

Histopathology of APAP-Induced Liver Necrosis in Dnase1+/+ and Dnase1−/− Mice.

We next tried to confirm the difference in the sensitivity toward APAP-induced hepatotoxicity of WT mice versus Dnase1 KO mice by examining the liver histopathology of both cohorts. As an example, we present the data of an experiment applying 800 mg/kg APAP. APAP-induced necrosis was evaluated via HE staining and correlated to DNA damage as visualized via TUNEL, because it was previously shown to be a prominent parameter of APAP toxicity.4, 6, 20 For this, serial liver sections of each time point were treated with either HE staining or TUNEL.

We observed that the necrotic alterations (i.e., the increased plasma membrane permeability that led to AST/ALT release) are undetectable via light microscopy during the first 2 hours (compare Fig. 2A–B with Fig. 3A). The first morphological hepatocellular alterations visible via light microscopy became detectable at 4 hours in both mouse strains; however, they were more prominent and developed more progressively in WT mice than in Dnase1 KO mice (Fig. 3A). These alterations comprised a loss of cytoplasmic eosinophilia due to glycogen depletion and of basophilia due to chromatolysis (i.e., loss of ribosomes and RNA).21 Furthermore, cytoplasmic and perinuclear vacuoles appeared (arrows 1 and 2, Fig. 3B), which on the ultrastructural level seemed to originate from swollen cisternae of the rough endoplasmic reticulum (data not shown). With increasing time, we observed a massive outward enlargement (i.e., progression of the pericentral necrotic area) in WT mice, whereas in Dnase1 KO mice only 1 or, at most, 2 pericentral cell layers were affected (Fig. 3A). At higher magnification, the morphological changes after 8 hours (Fig. 3B) demonstrated disintegration of cell nuclei frequently leading to a punctated hematoxylin staining (karyorrhexis) (Fig. 3B, arrow 3) and finally to their complete disappearance (karyolysis) (Fig. 3B, arrow 4). In WT mice, damaged cells finally condensed to a pale amorphous mass accompanied by dilation of the sinusoids and by aggregated erythrocytes (hemostasis) (Fig. 3B, arrow 5). After 10 to 24 hours, most of the WT animals became moribund and died. In contrast, Dnase1 KO mice did not develop hemostasis, and all of them survived. Their livers displayed pericentral necrotic areas encompassing only 1 or, at most, 2 cell layers. Phagocytes were observed infiltrating the necrotic region of some Dnase1 KO mice, probably facilitating removal of cell debris (Fig. 3B, arrow 6). Signs of tissue regeneration were observed 4 to 6 hours after APAP exposure in Dnase1 KO mice but not in WT mice. These signs included enlargement of hepatocellular nuclei, which might indicate polyploidy and enhanced gene expression (Fig. 3B, arrow 7). Furthermore, immunohistochemistry demonstrated upregulation of the proliferation marker pKi-6722 (data not shown).

Figure 3.

Histopathology of APAP hepatotoxicity in Dnase1+/+ and Dnase1−/− mice. (A) Time-dependent progress of APAP-induced necrosis in Dnase1+/+ in comparison to Dnase1−/− mice given 800 mg/kg APAP. Histopathology was investigated via HE staining (left rows) and TUNEL for fragmented genomic DNA (right rows) using serial liver sections. For vehicle-treated control mice, HE staining and positive controls for the TUNEL procedure are given (see Materials and Methods). Four hours after APAP application, an increasing number of pericentral layers of hepatocytes became visibly necrotic in Dnase1+/+ mice as revealed by light microscopy. The necrotic areas (HE stain) and zones of DNA degradation (TUNEL) displayed a strong concordance 4 hours after APAP application. The liver sections of Dnase1−/− animals showed a later onset and dramatically reduced necrotic and TUNEL-positive areas. (B) Typical signs of APAP-induced necrosis after 8 hours presented at higher magnification compared with a vehicle-treated Dnase1+/+ control mouse. Necrotic cells surrounding the central veins displayed loss of eosinophilia, cytoplasmic (arrow 1) and perinuclear (arrow 2) vacuoles, karyorrhexis (arrow 3), and karyolysis (arrow 4). In Dnase1+/+ mice, erythrocytes accumulated in the midzonal regions (arrow 5). In Dnase1−/− mice, progression of necrosis was dramatically reduced, phagocytes infiltrated the pericentral region (arrow 6), and hepatocytes harboring giant cell nuclei (arrows 7) appeared. Additionally, a few apoptotic cells were detectable (arrows 8). Dnase1, deoxyribonuclease 1; Ctrl, control.

DNA damage detected via TUNEL was first observed after 4 hours of APAP exposure and revealed an almost perfect concordance of its location and size with the visible necrotic area (HE staining) in WT mice (Fig. 3A). This concordance also persisted during the time-dependent enlargement of the necrotic area. In contrast, DNA damage was first detected after 6 hours in Dnase1 KO mice, also coinciding with the considerably smaller necrotic area that did not enlarge with time. As indicated by AST/ALT release, the necrotic insult was similar in WT and Dnase1 KO mice during the first 2 hours. However, DNA damage detected by TUNEL was delayed in both mouse strains compared with the initial AST/ALT release. Because DNA damage was considerably reduced in Dnase1 KO mice, our data demonstrate that the endonuclease responsible for DNA degradation during APAP intoxication is Dnase1. Other nucleases may play only a minor role due to their reduced presence and activity.

In addition to necrotic hepatocytes, we also observed apoptotic hepatocytes in APAP-treated Dnase1 KO mice (Fig. 3B, arrow 8). Because the occurrence of hepatocellular apoptosis during an APAP overdose has previously been discussed,15 we evaluated the extent of apoptosis by immunoblotting liver tissue extracts for apoptotic markers using vehicle-treated controls, APAP-exposed mice, and mice displaying massive liver apoptosis after treatment with agonistic anti-CD95 antibodies.23, 24 In contrast to anti-CD95 treated mice, we detected no activation of caspase-3, no cleavage of PARP-1, and no prominent loss of the antiapoptotic Bcl-XL—typical indicators of apoptosis—after APAP treatment of both WT and Dnase1 KO mice (Fig. 4), suggesting that the extent of apoptosis in Dnase1 KO mice must have been below the detection limit of the immunoblotting technique and that the APAP overdose predominantly induced necrosis.

Figure 4.

APAP overdoses predominantly induce necrosis rather than apoptosis. Immunoblotting for markers of apoptosis in liver homogenates of Dnase1+/+ and Dnase1−/− mice challenged with 800 mg/kg APAP are shown. As an apoptosis control, homogenates of anti-CD95–treated mice (α-CD95) were employed. In contrast to anti-CD95–treated mice, no caspase-3 activation, PARP-1 cleavage, or loss of Bcl-XL was detected either in Dnase1+/+ or Dnase1−/− mice treated with APAP. The weak time-dependent loss of PARP-1 and Bcl-XL in Dnase1+/+ mice could be explained by the general loss of nuclear and cytoplasmic cell material as observed via both HE staining (Fig. 3A) and immunoblotting against β-actin. Dnase1, deoxyribonuclease 1; APAP, acetaminophen; Ctrl, control; Parp-1, poly–adenosine diphosphate–ribose polymerase 1.

Internucleosomal DNA Degradation Occurs During APAP Intoxication.

We next analyzed whether the amount of TUNEL-positive cells correlated with the extent of DNA fragmentation as estimated by a biochemical method. Therefore, we determined the percentage of DNA released from the hepatocyte nuclei into cytoplasms during necrosis as described by Ray et al.6 In accordance with TUNEL, we found a clear and significant (P < .001) difference in the extent of DNA fragmentation 10 hours after APAP application between WT mice (10.6 ± 4.1%) and Dnase1 KO mice (1.4 ± 0.9%) (Fig. 5A). However, this method did not detect the large difference in DNA fragmentation observed between WT mice and Dnase1 KO mice via TUNEL at earlier time points (Fig. 3A). This difference may be due to the fact that Dnase1 induces only TUNEL-positive single-strand breaks during the initial phase of DNA degradation. Apparently, at earlier time points single-strand DNA breaks dominated, and only after 10 hours sufficient amounts of small DNA fragments had been generated that were able to diffuse into the cytoplasm. This hypothesis was supported by agarose gel electrophoresis that demonstrated the appearance of considerably higher amounts of nucleosomal DNA fragments in WT mice compared with Dnase1 KO mice 10 hours after APAP exposure (Fig. 5B). Indeed, internucleosomal DNA degradation was repeatedly shown to also occur during necrosis.5, 13

Figure 5.

DNA fragmentation during APAP intoxication. (A) Quantitative analysis of hepatic DNA fragmentation in WT mice compared with Dnase1 KO mice given 800 mg/kg APAP. The amount of fragmented DNA appearing in the cytoplasm of necrotic cells was estimated and is expressed as the percentage of total cellular DNA. In contrast to TUNEL (Fig. 3A), a significant difference in DNA fragmentation was not detectable until 10 hours after APAP administration. (B) Qualitative analysis of DNA prepared from liver cytoplasm of APAP-treated and untreated control mice via conventional agarose gel electrophoresis revealed a higher degree of internucleosomal DNA fragmentation after 10 hours in WT mice. ΔP < .05 compared with WT mice treated at the same time point. WT, wild-type; KO, knockout; APAP, acetaminophen; Ctrl, control; Dnase1, deoxyribonuclease 1.

Hepatic Pathophysiology During APAP Intoxication.

Finally, we compared a number of hepatic metabolic parameters known to be representative for necrosis and APAP intoxication in liver samples collected from WT mice and Dnase1 KO mice treated with 800 mg/kg APAP. First, we indirectly estimated the extent of APAP metabolism by measuring the change of total hepatic glutathione level (GSH+GSSG [mg]/liver wet weight [g]), which is the main detoxification system for the reactive metabolite NAPQI (Fig. 6).2 We found significant (P < .001) glutathione depletion during the first 4 hours after APAP application (Fig. 6), that occurred with an almost identical time course for both mouse strains. After 2 hours, minimal glutathione values were attained in both groups (0.75 ± 0.07 mg/g in WT mice, 0.66 ± 0.06 mg/g in Dnase1 KO mice). In comparison, the GSH+GSSG levels of untreated control animals were 5.63 ± 0.52 mg/g in WT mice and 5.76 ± 0.77 mg/g in Dnase1 KO mice. In both mouse cohorts treated with APAP, the GSH+GSSG levels did not further decrease up to 10 hours but remained at these significantly lower values (P < .001) (data not shown). In contrast, vehicle-treated controls did not show glutathione depletion within 10 hours (not shown). The difference of the GSH+GSSG level in WT mice and Dnase1 KO mice did not pass 2-way ANOVA, the Student t test with Bonferroni's adjustment, or the Mann-Whitney U test for significance at any time point. These data indicate a comparable APAP metabolism in both mouse strains.

Figure 6.

Hepatic glutathione depletion during APAP intoxication. Quantification of the total hepatic GSH+GSSG content (GSH+GSSG [mg]/liver wet weight [g]) indirectly indicates that mice of both Dnase1 genotypes treated with 800 mg/kg APAP for 15 minutes to 4 hours metabolized APAP with a comparable efficiency. Glutathione levels in both mouse cohorts reached a constant minimal value 2 hours after APAP administration. *P < .05 compared with WT untreated 0-hour control. #P < .05 compared with KO 0-hour control. GSH + GSSG, glutathione/glutathione-disulfide; WT, wild-type; KO, knockout; Ctrl, control.

APAP toxicity is known to induce massive DNA damage. In general, DNA damage is rapidly sensed by nuclear proteins such as PARP-1, which binds to single- and double-strand DNA breaks and catalyzes poly–adenosine diphosphate–ribosylation of nuclear acceptor proteins, thereby consuming NAD+.25, 26 Consistent with this observation, we detected a prominent decrease of the hepatic NADH+NAD+ content (NADH+NAD+ [μmol]/liver wet weight [g]) in APAP-treated WT mice with ongoing exposure time starting from 6 hours after treatment (P < .001), which did not occur in Dnase1 KO mice (Fig. 7A). The NADH+NAD+ content of WT livers decreased from 0.90 ± 0.16 μmol/g (0-hour control) to 0.29 ± 0.04 μmol/g during 10 hours after APAP treatment. These data correlate well with the difference in the occurrence of DNA degradation found between WT mice and Dnase1 KO mice and support the assumption that DNA damage leads to NAD+ and therefore energy depletion by activating DNA repair mechanisms. Consistent with this assumption, no PARP-1 degradation (i.e., inactivation) was detected via immunoblotting of liver homogenates of APAP-treated mice (Fig. 4).

Figure 7.

Hepatic ATP and NADH+NAD+ depletion during APAP intoxication. Measurement of total liver (A) NAD++NADH and (B) ATP levels in WT and Dnase1 KO mice treated with 800 mg/kg APAP revealed a time-dependent energy depletion in WT mice that was not detectable to the same extent in Dnase1 KO mice. As controls, data of untreated mice (0-hour controls) and mice treated for 10 hours with vehicle only were presented. *P < .05 in comparison with WT 0-hour control. #P < .05 in comparison with KO 0-hour control. ΔP < .05 in comparison with WT mice treated at the same time point. NADH+NAD+, β-nicotinamide adenine dinucleotide; WT, wild-type; KO, knockout; Ctrl, control; ATP, adenosine triphosphate.

Finally, we estimated the liver ATP level (ATP [μmol]/liver wet weight [g]), because its depletion was previously shown to induce necrotic cell death.27 Loss of ATP can result from mitochondrial dysfunction,28, 29 but might also be a consequence of NADH+NAD+ consumption by PARP-1 induced by DNA damage. We found that application of 800 mg/kg APAP to WT mice led to a significant (P < .05) decrease of ATP from 0.42 ± 0.13 μmol/g (0-hour control) to 0.18 ± 0.05 μmol/g after 8 hours (Fig. 7B). In contrast, Dnase1 KO mice and vehicle-treated control animals did not display a significant reduction in the hepatic ATP level within a period of 10 hours (Fig. 7B).


Overdose of APAP is known to cause extensive centrilobular hepatic cell damage, which is caused by APAP oxidation to the highly reactive metabolite NAPQI by the microsomal cytochrome P450 oxidase system. During this metabolic phase, NAPQI is detoxified by conjugation to glutathione.2 However, if glutathione is depleted, NAPQI accumulates and causes cell damage by covalently binding to cellular proteins (protein arylation).2 Subsequent to the metabolic phase, a toxic phase occurs. It is characterized by mitochondrial permeability transition and oxidative stress leading to a decrease in mitochondrial ATP production followed by cell damage.3, 28 Furthermore, the cellular Ca2+ hemostasis is known to be disturbed, leading to a rise in cytoplasmic and nuclear Ca2+.2 This Ca2+ increase has been shown to be responsible for increased DNA damage4, 6 by nucleases probably released from cellular compartments whose integrity is impaired by oxidative events. Indeed, nuclease inhibition by chelators of divalent cations as well as Zn2+ ions protects against APAP toxicity,5, 7, 8 as does application of PARP-1 inhibitors.8 PARP-1 is known to sense DNA damage and to induce DNA repair mechanisms by ADP ribosylation of DNA-binding proteins, thereby consuming NAD+.25, 26 Therefore, nicotinamide supplementation also protects against APAP toxicity.8 These and our data are in agreement with the assumption that DNA damage induced by a Ca2+- and/or Mg2+-dependent nuclease represents a critical step during APAP-induced cell death.

Necrosis is known to occur when cellular ATP is depleted.27, 30 During APAP-induced hepatocellular necrosis, 3 different mechanisms might converge to result in a loss of ATP: (1) mitochondrial dysfunction, (2) DNA damage, and (3) a gradient in oxygen supply from the periportal to the inner pericentral region. The contribution of the first 2 mechanisms during APAP-induced reduction of the intracellular ATP below a level necessary for survival therefore might depend on hepatocyte localization. Indeed, it is known that the periportal hepatocytes are not affected during APAP intoxication.2

Our data directly indicate a crucial role of Dnase1 during APAP-induced necrosis as revealed by the strongly reduced sensitivity of Dnase1 KO mice. DNASE1 is a Ca2+/Mg2+-dependent endonuclease that is translocated into the endoplasmic reticulum during translation and subsequently secreted by the liver into the blood plasma.10, 11 Because the cytochrome P450 oxidase complex is located in the endoplasmic reticulum membrane, it is conceivable that NAPQI generation impaired the integrity of this compartment and induced the release of Dnase1 from the endoplasmic reticulum. Indeed, in addition to mitochondrial damage, electron microscopy indicated a rapid vacuolization of the endoplasmic reticulum after APAP application (Jacob et al., unpublished data). The observed increase of serum activity of Dnase1 and ALT/AST strongly suggests that Dnase1 had gained access not only to the extracellular space but also to the cytoplasm and the nucleus, where it subsequently catalyzed DNA degradation.

The earliest signs of DNA damage in WT animals via TUNEL occurred after 4 hours (i.e., after the first wave of aminotransferase release) and affected the innermost pericentral hepatocytes. This observation could be interpreted to mean that DNA damage always occurs after plasma membrane rupture; however, at later time points the increase in the amount of TUNEL-positive necrotic cells paralleled the further increase in serum aminotransferases. Therefore, a difference in APAP sensitivity between the innermost and the more outwardly located pericentral hepatocytes seems to exist. We hypothesize that the combination of lower oxygen supply and APAP-induced mitochondrial dysfunction sufficed to induce necrosis in the innermost cells. Additional loss of ATP due to DNA damage does not seem to be necessary for these cells to undergo necrosis. This assumption is supported by the observed necrotic death of the pericentral hepatocytes in Dnase1 KO mice. In these hepatocytes, the DNA damage measured via TUNEL became detectable at later time points (6 hours) and was probably due to nucleases other than Dnase1 that might be less active or abundant.

In contrast to the necrotic death of the innermost hepatocytes, the progression of necrosis to more periportal layers was dependent on the presence of Dnase1, because it only occurred in WT animals. Therefore, necrosis of these cells seems to depend on an increased ATP depletion due to PARP-1 stimulation by Dnase1-induced DNA damage. This assumption is supported by the observed decrease in NADH+NAD+ and ATP levels at later time points (4–10 h) in WT mice that was absent in Dnase1 KO mice. We assume that additional ATP depletion is necessary to induce necrosis in these outer hepatocytes, because they have a higher basal energy level that is guaranteed by a better oxygen supply.

In conclusion, our data indicate that the metabolic phase of APAP intoxication is nearly identical in both mouse strains and almost completed between 30 minutes and 2 hours as revealed by the comparable rate of glutathione consumption. However, the outcome of the toxic phase appeared to depend on the oxygen gradient and therefore the hepatocyte location. The presence of Dnase1 is crucial for the development of necrosis at this stage. For the innermost hepatocytes, APAP-induced mitochondrial damage seems to be sufficient to undergo necrosis, whereas necrosis of the more periportal hepatocytes seems to be dependent on further ATP depletion, apparently caused by DNA damage induced by Dnase1. Our results are in agreement with previous data on APAP metabolism and intoxication; in addition, they offer a new explanation of the mechanism of the ensuing aggravation of hepatocellular damage. This knowledge may prove useful in the development of therapeutic approaches to effectively diminish the hepatotoxic phase of APAP overdose.


The authors acknowledge the expert technical assistance of Swantje Wulf and Kristina Klar. We thank Dr. Monika Jacob for performing ultrastructural analysis, Dr. Dirk Eulitz for organizing our animal facility, and Eric R. Siegel and Project Director Sudhir V. Shah for advice on the statistical analysis of data and critical reading of the manuscript, respectively.