Loss of BOK Has a Minor Impact on Acetaminophen Overdose-Induced Liver Damage in Mice

Acetaminophen (APAP) is one of the most commonly used analgesic and anti-pyretic drugs, and APAP intoxication is one of the main reasons for liver transplantation following liver failure in the Western world. While APAP poisoning ultimately leads to liver necrosis, various programmed cell death modalities have been implicated, including ER stress-triggered apoptosis. The BCL-2 family member BOK (BCL-2-related ovarian killer) has been described to modulate the unfolded protein response and to promote chemical-induced liver injury. We therefore investigated the impact of the loss of BOK following APAP overdosing in mice. Surprisingly, we observed sex-dependent differences in the activation of the unfolded protein response (UPR) in both wildtype (WT) and Bok-/- mice, with increased activation of JNK in females compared with males. Loss of BOK led to a decrease in JNK activation and a reduced percentage of centrilobular necrosis in both sexes after APAP treatment; however, this protection was more pronounced in Bok-/- females. Nevertheless, serum ALT and AST levels of Bok-/- and WT mice were comparable, indicating that there was no major difference in the overall outcome of liver injury. We conclude that after APAP overdosing, loss of BOK affects initiating signaling steps linked to ER stress, but has a more minor impact on the outcome of liver necrosis. Furthermore, we observed sex-dependent differences that might be worthwhile to investigate.


Introduction
Para-(Acetylamino)phenol (acetaminophen, paracetamol, herein referred to as APAP) is one of the most commonly used antipyretic and analgesic drugs in the world [1,2]. It is safe to use within its rather narrow therapeutic window; however, an overdose can lead to severe hepatic injury and liver failure [1, 3]. As an over-the-counter obtainable drug, excessive self-medication-intentional (due to suicide attempts) or unintentional (by the use of combination products containing APAP)-is the main cause of APAP-intoxication and accounts for~50% of all acute liver failure (ALF) cases in the Western world [1][2][3][4]. The metabolism of APAP takes place mainly in the liver, with a lesser contribution of the kidney and intestine [5]. While the main portion of APAP is readily conjugated with glucuronic acid or sulphate and excreted with the urine, only a small percentage (~10%) is metabolized by cytochrome P450 (predominantly CYP2E1 and CYP1A2), leading to the formation of the highly reactive intermediate N-acetyl-p-benzoquinone imine (NAPQI) [2,4,6,7]. Under physiological conditions, NAPQI is rapidly conjugated to glutathione (GSH) and secreted in the urine as cysteine and mercapturic acid conjugates [6]. However, if GSH stores are depleted, a situation encountered upon APAP overdosing, unconjugated NAPQI binds and reacts with cellular, preferentially mitochondrial proteins, leading to mitochondrial and in modulating UPR responses, we addressed in this study how the loss of BOK affects liver damage in male and female mice upon APAP overdosing.
Surprisingly, we found sex-dependent differences in the activation of the UPR following ER stress, and a decrease in JNK activation in Bok -/compared to WT mice. However, whereas the loss of BOK decreased lobular necrosis in females (and to a smaller extent also in males) based on pathohistological assessment, no significant difference in overall liver damage, as assessed by ALT and AST serum levels, was detected between the genotypes.

Expression of CYP2E1 Is Increased at the mRNA but Not the Protein Level in Bok -/-Livers
The metabolism of APAP into the toxic metabolite NAPQI is mediated by CYPs, especially by CYP2E1 and, to a lesser extent, also by other CYPs, such as CYP1A2 [41]. To test whether expression levels of CYPs differ between Bok -/and WT mice, we analyzed the expression of CYP1A2 and CYP2E1 mRNA in primary hepatocytes or whole livers isolated from WT and Bok -/mice of both sexes, at baseline conditions ( Figure 1A-F) and after APAP treatment (Supplementary Figure S1A,B). Of note, the expression of CYP2E1 at the baseline was increased in both primary hepatocytes and whole liver lysates derived from Bok -/male mice compared with WT mice ( Figure 1A,C). The expression of CYP1A2 was also increased; however, the difference did not reach statistical significance ( Figure 1B,D). On the other hand, no difference in respective CYP expression levels was seen in liver lysates from Bok -/and WT female mice ( Figure 1E,F). Despite the difference seen at the mRNA level, CYP2E1 protein levels in untreated livers were comparable between the genotypes ( Figure 1G,H). After APAP treatment, the expression levels of CYP1A2 and CYP2E1 transcripts did not differ significantly between the two genotypes of either sex. However, CYP2E1, but not CYP1A2, expression decreased in males, while no changes in females were detected (Supplementary Figure S1A,B). CYP2E1 protein levels after APAP treatment were comparable between the genotypes (Supplementary Figure S1C-E).

p53 Expression
Correlates with PUMA Expression but Not with the Severity of Liver Damage P53 has been described to play a protective and attenuative role in APAP-induced hepatic injury, yet the underlying mechanism is still obscure [42][43][44]. Since we previously found that BOK-deficient cells (HCT116 and HepG2) and liver lysates derived from Bok -/mice show a higher baseline expression of p53 at both the transcriptional and protein levels [40,45], we tested p53 expression in the liver after APAP treatment by RT-qPCR. As shown in Figure 2A, similar expression levels of p53 were detected in both sexes of WT and Bok /mice. The BH3 only protein and p53 downstream target PUMA has been reported to contribute to APAP-induced hepatotoxicity [21]. Indeed, our data showed that APAP induces the expression of PUMA in WT and in Bok -/mice ( Figure 2B). Furthermore, p53 expression levels correlated linearly with those of PUMA in both males and females ( Figure 2C,D), but not with the severity of liver injury as assessed by the levels of ALT in the serum ( Figure 2E,F). Figure 1. CYP2E1 mRNA but not protein expression is increased in Bok -/-males at baseline. qPCR analysis revealed increased expression of CYP2E1 and CYP1A2 in primary hepatocytes derived from Bok -/-male mice compared with wildtype (WT) (A,B). Expression of CYP2E1, but not CYP2A1 was significantly increased in whole liver lysates derived from Bok -/males compared with WT (C,D). In whole liver lysates of female WT and Bok -/-mice CYP2E1 (E) and CYP1A2 (F) were expressed at similar levels. CYP2E1 protein levels were comparable in whole liver lysates of untreated WT and Bok -/-mice (G,H). Results are depicted as a fold change compared to sex-matched WT (A-F) or as GAPDH normalized values (H). Data are represented as mean ± S.D. and are derived from 3 mice per group. *: p < 0.05; n.s.: not significant.

p53 Expression Correlates with PUMA Expression but Not with the Severity of Liver Damage
P53 has been described to play a protective and attenuative role in APAP-induced hepatic injury, yet the underlying mechanism is still obscure [42][43][44]. Since we previously found that BOK-deficient cells (HCT116 and HepG2) and liver lysates derived from Bok -/mice show a higher baseline expression of p53 at both the transcriptional and protein  levels [40,45], we tested p53 expression in the liver after APAP treatment by RT-qPCR. As shown in Figure 2A, similar expression levels of p53 were detected in both sexes of WT and Bok /-mice. The BH3 only protein and p53 downstream target PUMA has been reported to contribute to APAP-induced hepatotoxicity [21]. Indeed, our data showed that APAP induces the expression of PUMA in WT and in Bok -/-mice ( Figure 2B). Furthermore, p53 expression levels correlated linearly with those of PUMA in both males and females (Figure 2C,D), but not with the severity of liver injury as assessed by the levels of ALT in the serum ( Figure 2E,F).

APAP Induces ER Stress and Activates the Unfolded Protein Response
Several studies reported APAP overdose to induce ER stress [23,24]. Furthermore, there is increasing evidence pointing towards a role for BOK in modulating the unfolded protein response (UPR) [36,39]. Thus, we hypothesized that the activation of the UPR via APAP-triggered ER stress could be impaired in Bok -/mice. Therefore, we investigated the expression levels of several proteins implicated in ER stress-induced UPR pathways, among them ATF4, CHOP, BIM and PUMA, in livers of WT and Bok -/mice. qRT-PCR analysis revealed that APAP treatment results in an elevation of ATF4 levels in livers of WT and Bok -/mice in a sex-specific manner ( Figure 3A). In fact, the increase of ATF4 in WT males was comparable to that in Bok -/males, but substantially stronger than in females.
Interestingly, in the latter, ATF4 induction was more pronounced in Bok -/mice than in the WT controls ( Figure 3A). This increase in ATF4 in treated Bok -/females was associated with increased levels of CHOP and BIM, whereas respective expression levels in males were comparable between the two genotypes ( Figure 3B-F). PUMA protein levels were increased after 5 h of APAP treatment compared with PBS controls, and the increase was stronger in females than in males; however, no differences were observed between the genotypes of either sex ( Figure 3G) In WT mice, APAP treatment resulted in a 40% downregulation of BOK protein in female but not in male mice at the 5 h timepoint ( Figure 3H).    The JNK signalling pathway was activated in WT and Bok -/mice of both sexes. JNK activation was significantly increased in WT females compared to WT males and a similar, although less pronounced, sex-specific effect was seen in mice lacking BOK (Figure 4). In both sexes, a trend of decreased pJNK activation in Bok -/mice compared with their WT counterparts was observed ( Figure 4). Prolonged activation of the UPR eventually induces apoptosis [46]. However, although our results showed an induction or activation of ER stress-activated proteins, such as JNK, CHOP, BIM and PUMA, all of which have the potential to induce apoptosis, we failed to detect effector caspase-3 or -7 activation at the endpoint of 5 h of APAP treatment, whereas some degree of PARP cleavage was seen ( Figure 5). Taken together, we conclude that despite signs of the initiation of classical ER stress-driven apoptotic pathways, effector caspases may have been activated below the detection level, or only during an early phase of APAP overdosing. Alternatively, PARP might have been cleaved by other proteases activated during the course of liver necrosis.     [40]. Data are derived from 8 to 9 mice per group.

Lack of BOK Slightly Reduces Liver Necrosis after APAP Overdose in a Sex-Dependent Manner
The effect of APAP overdosing on inducing liver necrosis was assessed histologically and by quantifying the serum levels of the hepatocellular enzymes ALT and AST. The data showed that when grouping mice of the same sex together, ALT and AST levels were comparable in APAP-treated Bok -/-and WT mice ( Figure 6A,B). However, when comparing male and female mice within a same genotype, slightly increased ALT levels were seen in WT females compared to males ( Figure 6C,D). This difference was not seen in Bok -/-mice ( Figure 6C), nor were there differences in AST levels between sex or genotypes ( Figure 6D). Histopathological analysis revealed that centrilobular necrosis was found in 100% of WT and Bok -/-males, whereas in females, necrosis was present in the majority but not all of the WT (75%) and Bok -/-(62.5%) mice (Table 1). The necrosis was confluent, i.e., bridging, in all animals except one female Bok -/-animal. The mean extent of lobular necrosis indicated in the percentage of affected centrilobular-portal vein distance was generally higher in WT mice compared with Bok -/-mice (Table 1). Furthermore, lobular necrosis in Bok -/-was less extensive in females than in males (21.9% vs. 30.5%) ( Figure 6E and Table  1). The hepatic necrosis was haemorrhagic, i.e., admixed with extravasated erythrocytes, in a significant number (50%) of female WT mice (Table 1). This feature was rare or absent in males and female Bok -/-mice. Altogether, the data indicate that the protective impact upon the loss of BOK on the severity of liver damage is minor, based on histology and A B Figure 5. 5 h APAP treated livers do not display active effector caspases while PARP cleavage is detectable. No caspase-3 or -7 cleavage was detected in the total liver lysates of males (A) and females (B) while low degree PARP cleavage was seen after 5 h of APAP treatment. Liver lysate from DEN-treated mice was used as positive control for apoptosis induction [40]. Data are derived from 8 to 9 mice per group.

Lack of BOK Slightly Reduces Liver Necrosis after APAP Overdose in a Sex-Dependent Manner
The effect of APAP overdosing on inducing liver necrosis was assessed histologically and by quantifying the serum levels of the hepatocellular enzymes ALT and AST. The data showed that when grouping mice of the same sex together, ALT and AST levels were comparable in APAP-treated Bok -/and WT mice ( Figure 6A,B). However, when comparing male and female mice within a same genotype, slightly increased ALT levels were seen in WT females compared to males ( Figure 6C,D). This difference was not seen in Bok -/mice ( Figure 6C), nor were there differences in AST levels between sex or genotypes ( Figure 6D). Histopathological analysis revealed that centrilobular necrosis was found in 100% of WT and Bok -/males, whereas in females, necrosis was present in the majority but not all of the WT (75%) and Bok -/-(62.5%) mice (Table 1). The necrosis was confluent, i.e., bridging, in all animals except one female Bok -/animal. The mean extent of lobular necrosis indicated in the percentage of affected centrilobular-portal vein distance was generally higher in WT mice compared with Bok -/mice (Table 1). Furthermore, lobular necrosis in Bok -/was less extensive in females than in males (21.9% vs. 30.5%) ( Figure 6E and Table 1). The hepatic necrosis was haemorrhagic, i.e., admixed with extravasated erythrocytes, in a significant number (50%) of female WT mice (Table 1). This feature was rare or absent in males and female Bok -/mice. Altogether, the data indicate that the protective impact upon the loss of BOK on the severity of liver damage is minor, based on histology and ALT/AST levels; however, it becomes more evident when comparing the extent of microscopic lobular necrosis. Of note, these effects in Bok -/mice are more strongly pronounced in females, supporting the notion of sex-specific effects in the APAP intoxication model. ALT/AST levels; however, it becomes more evident when comparing the extent of microscopic lobular necrosis. Of note, these effects in Bok -/-mice are more strongly pronounced in females, supporting the notion of sex-specific effects in the APAP intoxication model.

Discussion
APAP intoxication is one of the most common causes for severe liver injury and failure and for subsequent liver transplantation [3]. Although APAP is an over-the-counter drug and one of the most commonly used antipyretic and analgesic drugs in the world, its exact mode of action is unknown. Even less is known about its toxic effect on hepatocytes and the main mechanisms leading to massive cellular damage after ingesting a single overdose (>150 mg/kg or~7.5 g in adults) or after repetitive intake of APAP or APAP-containing drugs within a short period of time (>200 mg/kg or 10 g within 24 h) [47]. Yet, it is well established that the formation of the highly reactive by-product NAPQI and its binding to cellular, preferentially mitochondrial proteins initiates cellular stress pathways, eventually resulting in hepatocellular necrosis [2,4,6,7].
One of these cellular stress responses is the UPR following ER stress [23,24]. BOK is predominantly localized at the ER and ER-associated membranes and has already been assigned UPR-modulating functions [37,39]. Along these lines, BOK has been shown to promote cell death in a chemical-induced liver injury model [40]. BOK-deficient cells were, on the other hand, reported to be more susceptible to Brefeldin A-induced cell death, pointing towards a UPR modulating function of BOK that can enhance either prosurvival or proapoptotic pathways in a context-dependent manner [37]. We thus hypothesized that after an APAP challenge, Bok -/mice would present with a defective ER stress translating either into increased or decreased liver damage. While we did indeed observe slight alterations in the upstream events of ER stress-triggered UPR responses in Bok -/mice, this eventually did not result in major differences in APAP-induced liver necrosis compared with WT controls. Nevertheless, our study revealed sex-relevant differences in the activation of ER stress pathways. While the activation of ATF4, a downstream target of PERK, was increased in males, females showed higher expression of pJNK, which is involved in the IRE-1α pathway. Interestingly, it has been shown that estradiol can induce JNK-dependent apoptosis [48,49], suggesting that the activation of the JNK pathway may be sex-dependent with increased activation in females compared to males. Although BOK is known to be important for the regulation of the PERK > ATF4 arm, along with CHOP and BIM [39,50], we did not detect significant differences in the expression levels of any of these members in Bok -/males compared to WT males. Whereas we found a correlation between p53 and PUMA transcripts ( Figure 2C,D), a direct role of p53 in inducing PUMA in the APAP model is not clear. Actually, Chen et al. proposed a RIPK1 and JNK-dependent, but p53 independent, transcriptional activation of PUMA upon APAP overdosing [21]. On the other hand, the JNK activation in Bok -/females was reduced as compared with WT mice, while ATF4 induction, BIM and CHOP expression levels were increased in Bok -/females. Since JNK activation was also reduced in Bok -/males, we propose that BOK is important for the regulation of the IRE-1α pathway, along with JNK, rather than the PERK > ATF4 arm along with CHOP. Furthermore, the increase of ATF4, CHOP and BIM detected in Bok -/compared to WT females can be proposed as an alternative ER stress pathway activation in Bok -/females, since the lack of BOK is associated with a reduction in the activation of the IRE-1α arm. The observation that BOK is substantially downregulated in females but not in males supports a sex-dependent role of BOK in the activation of ER stress pathways and for the development of hepatic injury.
It is still unclear and controversially discussed how important apoptosis is for the development of hepatic injury after APAP overdose (reviewed in [14]). Despite the induc-tion of the two pro-apoptotic BH3-only proteins BIM and PUMA in response to APAP, the development of liver damage implicates ER stress likely as a late event [51] in APAP intoxication. In fact, our findings support this view since the expression levels of pJNK and CHOP are still rather low after 3 h intoxication (Supplementary Figure S2).
Another unexpected and sex-related finding was the increased expression levels of CYP2E1 mRNA in livers from Bok -/males compared to WT males, which was not seen in females. Increased expression levels of CYPs, induced by e.g., alcohol intake, typically correlate with an increased conversion of APAP into NAPQI and, hence, increased liver damage would be expected for Bok -/males [41]. However, CYP2E1 increase was modest and no difference was seen between the genotypes at the protein level. We conclude that the slightly less extensive centrilobular necrosis found in Bok -/mice compared with their WT counterparts cannot be explained by changes in CYP2E1 levels. Importantly, the difference in centrilobular necrosis between WT and Bok -/mice was higher in females than in males. Thus, BOK seems to be more important for the development of liver damage in females than in males. It has previously been shown that APAP-induced liver damage is sex-specific in mice and that females have a lower susceptibility to APAP hepatotoxicity compared to males [8,9,52]. Du et al. reported a strong protection of C57BL/6 females compared with males in the APAP model. We did not see such a clear difference while using a higher dose of APAP. Of note, in the Du et al. study, the cytosolic pJNK/JNK ratio in females dropped only at a timepoint later than our endpoint of 5 h [8]. However, protection in females is supported by our findings, in which liver necrosis was present in 100% of the males but only 69% (11/16 mice) of the females. This phenomenon may be explained by the finding that females are able to restore GSH faster and thus scavenge more of the toxic NAPQI than males despite similar kinetics of GSH depletion after APAP overdose [8]. In line with these observations, we found total and reduced GSH to be increased in females compared to males; however, no significant difference between WT and Bok -/mice was observed (Supplementary Figure S3). This finding suggests that there is no major difference in the metabolism of APAP between WT and Bok -/mice.
In conclusion, our data support a role of BOK in the regulation of the UPR after APAP intoxication, especially by modulating the IRE1α > JNK arm that follows a sex-relevant activation pattern reflected by a substantial increase in its activation in female as compared to male mice. However, since apoptosis may not play a major role in the development of necrotic liver damage in this acute APAP overdosing model, the contribution of BOK in regulating the severity of liver damage is more minor after 5 h of intoxication. This view is further supported by our findings that loss of BOK does not alter the sensitivity of mouse and human hepatocellular cell lines towards various doses of APAP in vitro (Supplementary Figure S4). However, all cell models tested except for Tib-75 were rather insensitive to APAP-induced killing, requiring high concentrations to see a decrease in viability (Supplementary Figure S4).
In summary, we show that BOK modulates upstream events linked to ER stress in response to APAP intoxication, and that the impact of BOK on hepatic injury following APAP overdose is only moderate. However, and surprisingly, this difference is sex-dependent, probably due to the differential activation of the UPR in males and females. Since the UPR following ER stress is a rather late event in APAP toxicity, later timepoints could be analysed to identify the involvement of BOK in a subacute intoxication model.

Mice Experiments
Age and sex-matched C57BL/6 WT and Bok -/-(kindly provided by A. Strasser, WEHI, Parkville, Australia) mice were fasted overnight with free access to water. Following the fasting period, mice were injected with an overdose of APAP (400 mg/kg) dissolved in warm PBS. Control animals were injected with the respective volume of warm PBS. After 5 h, mice were sacrificed, blood was taken by retro-orbital bleeding and livers were collected. The left lobe of the liver of each mouse was fixed in 4% PFA for histology while the others were snap frozen and used later for protein and RNA expression analyses. All animal experiments were approved by the Animal Experimentation Review Board of the Canton of Bern, Switzerland (BE43/19).

Cell Viability Measurements by MTT and Flow Cytometry
HepaRG, IHH and Tib-75 cells were seeded in 96-well flat-bottom plates. Primary hepatocytes were seeded in collagen-coated 96-well flat-bottom plates. After overnight adherence, cells were treated with the indicated concentrations of APAP (Sigma-Aldrich A5000, Buchs, Switzerland) for 24 h. Cell viability of HepaRG, Tib-75 and primary hepatocytes was assessed by 3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazoliumbromid (MTT) (Sigma-Aldrich M2128, Buchs, Switzerland). In brief, medium was removed and cells were washed once with PBS. 100 µL MTT solution (5 mg/mL) was added to the cells and incubated for 30 min at 37 • C. After incubation, MTT solution was aspirated and 100 µL DMSO was added into each well. Absorbance was read at 562 nm and cell viability was normalised to untreated control.
For flow cytometric analysis, cells were washed with FACS buffer (150 mM NaCl, 4 mM KCl, 2.5 mM CaCl2, 1 mM MgSO4, 15 mM HEPES pH 7.2, 2% FCS and 10 mM NaN3) and incubated with FITC-Annexin V diluted in FACS buffer. After 20 min of incubation on ice in the dark, cells were washed with FACS buffer and resuspended in 100 µL FACS buffer containing 2 µg/mL propidium iodide. Cells were examined by a BD FACSLyric™ Research System (BD Biosciences, San Jose, CA, USA). Double negative cells were considered as viable cells.