Liver-specific deficiency of unc-51 like kinase 1 and 2 protects mice from acetaminophen-induced liver injury

ANIMAL EXPERIMENTS

Ulk1^+/F mice29 and Ulk2^+/– mice30 were obtained from Jackson Laboratory, ME, USA (B6.129-Ulk1^tm1Thsn/J, stock number: 017976, backcrossed to C57BL/6J mice for at least 10 generations) and MMRRC (stock number: 011678-UNC, C57BL/6J), respectively. Ulk1^+/F mice were crossed with Albumin-Cre mice (Albumin-Cre, C57BL/6J; Jackson Laboratory) and Ulk2^+/– whole-body knockout mice to generate Ulk1^F/FUlk2^–/–Albumin-Cre⁺ (Ulk1/2 LDKO) mice. Mice were fed a normal chow diet and housed with corncob bedding in specific pathogen-free conditions. All animal experiments were approved by the Institutional Animal Care and Use Committee at Xiamen University (Fujian, China). Male mice were used for all the in vivo experiments. For APAP treatment, Ulk1/2 Ctrl (Ulk1^F/FUlk2^−/−Albumin-Cre⁻ and Ulk1^F/+Ulk2^−/−) and Ulk1/2 LDKO mice at 6-8 weeks old were either given saline or APAP (500 mg/kg, Cat. A7085; Sigma-Aldrich, Shanghai, China) for 0-24 hours and were sacrificed. For tumor necrosis factor alpha (TNF-α)/galactosamine (GalN) treatment, mice were injected with phosphate-buffered saline (PBS) or GalN (800 mg/kg, Cat. G0500; Sigma-Aldrich) 30 minutes before PBS or TNF-α (12 μg/kg, Cat. CF09-B; Novoprotein, NJ, USA) intraperitoneally and were sacrificed 6 hours later. Whole blood was collected by cardiac puncture and was centrifuged at 4°C to obtain serum. Mouse liver was fixed in 4% paraformaldehyde and paraffin embedded. Sections of 5 μm were used for hematoxylin and eosin (H&E) staining. Serum was used for determination of alanine aminotransferase (ALT) activities using a commercial kit (Cat. A7526; Pointe Scientific, MI, USA). Levels of triacylglycerol (Cat. 290-63701; Wako, Osaka, Japan), nonesterified fatty acid (NEFA; Cat. 294-63601; Wako), ketone body (Cat. 700190; Caymen, MI, USA ), and cholesterol (Cat. 294-65801; Wako) were measured according to the manufacturer's instructions. Blood glucose values were determined using OneTouch UltraVue automatic glucometers (Johnson & Johnson, NJ, USA). GSH was measured by using the DTNB-glutathione reductase recycling assay, as described.32 APAP-cysteine adducts were measured by high-pressure liquid chromatography with electrochemical detection, as described.33 Immunohistochemical detection of Ki-67 (Cat. 27309-1-AP; Proteintech, IL, USA) was performed on paraffin-embedded liver tissue sections. To assess the function of JNK inhibitor in APAP-induced liver injury, mice were injected with JNK inhibitor SP600125 (10 mg/kg, Cat. S1460; Selleck, TX, USA) 3 hours before APAP administration. Five percent dimethyl sulfoxide (DMSO) in PBS was used as the solvent for SP00125 (SP00125 was first dissolved in DMSO and then diluted with PBS), as described.13 A vehicle control containing the same percentage of DMSO in PBS, but without the inhibitor was used for all experiments using SP00125. For Ulk1/2 rescue experiments, Ulk1/2 LDKO mice were injected with adenovirus-expressing vector WT-Ulk1 or WT-Ulk2 by caudal vein. Five days later, mice were treated with APAP and sacrificed. Total liver lysates were prepared using a lysis buffer as described.34

ANTIBODIES AND DRUGS

Antibodies to phospho-serine in S*/P-motif (Cat. #9477; used for detection of phospho-S403-MKK7) in the KinomeView Profiling Kit (Cat. #9812), Ulk1 (Cat. #8054), phospho-Ser757-Ulk1 (Cat. #6888), FIP200 (Cat. #12436), ATG13 (Cat. #13273), ATG101 (Cat. #13492), phospho-Thr183/Tyr185-JNK (Cat. #4668), P-S6K (S6 kinase; Cat. #9234), S6K (Cat. #9202), MKK4 (Cat. #9152), MKK7 (Cat. #4172; used for western blotting), phospho-Ser271/Thr275-MKK7 (Cat. #4171), and tubule-associated protein 1 light chain 3 beta (LC3B; Cat. #2775) were purchased from CST, MA, USA. Antibodies to hemagglutinin (HA; F-7), JNK (C-17), and MKK7 (E-7; used for immunoprecipitating endogenous MKK7) were from Santa Cruz Biotechnology, TX, USA. Antibody to p62 (Cat. ab56416) and Ulk2 (Cat. ab56736) were from Abcam, London, UK. Antibodies to Flag (Cat. F2555), Actin (Cat. A8481), and anti-Flag beads (Cat. M8823) were from Sigma-Aldrich. Antibodies to proliferating cell nuclear antigen (PCNA; Cat. 10205-2-A), cytochrome P450 2E1 (Cyp2E1; Cat. 19937-1-AP), modulatory subunit of glutamate-cysteine ligase (GCLM; Cat. 14241-1-AP), catalytic subunit of glutamate-cysteine ligase (GCLC; Cat. 12601-1-AP), and SH3 domain-binding protein 5 (SH3BP5 [Sab]; Cat. 11127-2-AP) were from Proteintech.

ELECTRON MICROSCOPY

Liver tissues were fixed with 2.5% glutaraldehyde in 0.05 M of sodium cacodylate buffer (pH 7.2), followed by 1% OsO₄. The average number of autophagosomes from each cell was determined from a randomly selected pool of 20 fields under each condition.

CELL CULTURE, TRANSIENT TRANSFECTION, IMMUNOPRECIPITATION, AND WESTERN BLOTTING

HEK293T were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS), 2 mM of L-glutamine, 100 IU of penicillin, and 100 μg/mL of streptomycin in a humidified incubator with 5% CO₂ at 37 °C. Adenoviruses for infection were packaged and amplified in β5 cells using TurboFect transfection reagent (Cat. #R0531; Thermo Scientific, MA, USA). Polyethylenimine at a final concentration of 10 μM was used to transfect HEK293T cells. Transfected cells were harvested at 16 hours after transfection. Cell lysis, immunoprecipitation (IP), and western blotting were carried out as described.34

qRT-PCR

Total RNA was extracted from livers of Ulk1/2 LDKO and their control mice and was reverse transcribed. Complementary DNA (cDNA) was then used as the template for the detection of the different transcripts using GoTaq qPCR Master Mix (Promega, WI, USA). Expression levels were analyzed by Applied Biosystems 7900HT Sequence Detection System v2.3 software. All mRNA transcript levels are expressed as the ratio of expression of the target gene to Actin. The sequences of the primers were as described.33

BACULOVIRUS EXPRESSION SYSTEM

For recombinant Ulk1 expression, Ulk1 was subcloned into pFastBac containing N-terminal His-tag and expressed in Sf21 cells using the Bac-to-Bac expression system from Invitrogen as described.35

PLASMID CONSTRUCTIONS

Full-length cDNAs encoding mouse Ulk1, Ulk2, and human MKK4, MKK7, JNK1, and JNK2 were obtained by PCR using mouse testis cDNA or human cDNA generated from HeLa cells.

PRIMARY HEPATOCYTES

Primary hepatocytes were isolated from mice with a nonrecirculating perfusion of livers with 0.05% Collagenase Type IV (Sigma-Aldrich). Cells were then plated in six-well plates in Williams' E medium (Cat. 12551-032; Gibco, MA, USA; supplemented with 10% FBS and 1 mM of GlutaMAX, [Cat. 35050061; Gibco], 100 IU penicillin and 100 mg/ml streptomycin) for 4 hours for attachment. After that, cells were cultured in the William's E medium without serum overnight before APAP treatment. TNF-α (20 ng/mL)/actinomycin (ActD; 0.5 μg/mL) were used for primary hepatocyte treatment. For generation of Mkk4/7 double knockdown (DKD) cells, attached hepatocytes were infected with adenovirus-expressing control shRNA or shRNAs targeting mouse Mkk4 (targeted sequence 5′-GCAACTGTGAAAGCACTAA-3′) and mouse Mkk7 (targeted sequence 5′-GCCATGACTGCCCTTACAT-3′) for at least 24 hours before APAP treatment. For generation of reconstituted hepatocytes, hepatocytes were infected with adenovirus expressing vector, WT-Ulk1, WT-Ulk2, WT-MKK7 (human), or MKK7 mutants. Experiments using primary hepatocytes were performed 24 to 48 hours after isolation and harvested within 72 hours. All cells were maintained in a 37°C incubator with 5% CO₂.

LONG-LIVED PROTEIN DEGRADATION ASSAY

Degradation of long-lived proteins from primary hepatocytes was measured as described.36 Briefly, cells were incubated with a medium containing 0.2 μCi/mL of L-[¹⁴C]Valine (Cat. NEC291EU050UC; PerkinElmer, MA, USA) for 20 hours. After washing and incubation with a medium containing 10 mM of unlabelled Valine for 1 hour, the medium was replaced with fresh medium with or without 10 mM of APAP, and the incubation was continued for the indicated durations. The medium was then precipitated by 10% trichloroacetic acid, and the release of [¹⁴C]Valine from protein degradation was calculated as the percentage of the total cell radioactivity.

CELL DEATH DETERMINATION

To determine the cell death rate of hepatocytes, primary hepatocytes seeded in six-well plates were treated with 10 mM of APAP, TNF-α (50 ng/mL), or TNF-α (20 ng/mL)/ActD (0.5 μg/mL, Cat. A9415; Sigma-Aldrich) for 0-6 hours. Cell death and morphological changes were analyzed at different time points with propidium iodide (PI) staining (1 μg/mL, 30 minutes).

IN VITRO KINASE ASSAYS

In vitro kinase assays using Ulk1 as the kinase were performed as described.34 For kinase assays using MKK as the kinase, recombinant JNKα1 (Cat. J2455, Sigma-Aldrich) was incubated with Flag-tagged MKK4 or MKK7 in a kinase buffer (20 mM of HEPES [pH 7.4], 5 mM of MgCl₂, 0.5 mM of ethylene glycol tetraacetic acid, 1 mM of dithiothreitol) supplied with 100 μM of ATP. The mixtures were maintained at 30 °C for 30 minutes and stopped by addition of sodium dodecyl sulfate sample buffer.

STATISTICAL ANALYSIS

For experiments with only two groups, the Student t test was used for statistical comparisons. Analysis of variance (ANOVA) with Tukey's post-test (one-way ANOVA for comparisons between groups; two-way ANOVA for comparisons of magnitude of changes between different groups from different cell lines or treatments) was used to compare values among different experimental groups using the SPSS Statistics (version 17.0; SPSS, Inc., Chicago, IL) program or GraphPad Prism (version 6; GraphPad Software Inc., La Jolla, CA) program. P < 0.05 was considered as statistically significant (^*) and P < 0.01 as highly significant (^**).

HEPATIC DELETION OF ULK1/2 DEMONSTRATES HEPATOMEGALY BUT NORMAL AUTOPHAGIC ACTIVITY IN THE LIVER

To study the physiological functions of Ulk1 and Ulk2 in the liver, we conditionally deleted Ulk1 in the Ulk2^–/– background by crossing Ulk1^+/FUlk2^−/− mice with albumin-Cre transgenic mice, which specifically express the Cre recombinase in hepatocytes. Ulk1^F/FUlk2^−/−Albumin-Cre⁺ mice (Ulk1/2 LDKO mice) were born at the expected Mendelian frequency and did not demonstrate any apparent phenotypes compared to control littermates (Ulk1^F/FUlk2^−/−Albumin-Cre⁻ and Ulk1^+/FUlk2^−/− Albumin-Cre⁺ mice, designated as Ulk1/2 Ctrl mice). Ulk1/2 LDKO mice also showed similar body weights compared to Ulk1/2 Ctrl mice up to 18 weeks of age (Fig. 1A), but displayed significantly larger and heavier liver (hepatomegaly) at as early as 3 weeks of age (Fig. 1B,C). However, histological analysis revealed no evident pathological changes in liver sections of Ulk1/2 liver-specific double knockout (LDKO) mice in comparison to those of WT mice (Supporting Fig. S1A). The protein levels of other Ulk complex components, including FIP200, ATG13, and ATG101, were increased in liver cells from Ulk1/2 LDKO mice (Fig. 1D), perhaps attributed to a compensatory response to the absence of Ulk1/2, similar to the phenomenon in hepatocyte-specific FIP200 deficiency mice.37 Surprisingly, we detected similar autophagic activities in either fed or overnight fasted Ulk1/2 LDKO mice liver compared to those in Ulk1/2 Ctrl mice liver as revealed by accumulation of the levels of p62, lipidation of LC3B and autophagosome formation (Fig. 1D,E). We also analyzed various metabolic parameters in plasma and liver of Ulk1/2 LDKO mice, including glucose, triacylglycerol (TAG), NEFA, ketone body, liver TAG, and cholesterol. It was found that all of the parameters tested were similar between control and Ulk1/2 LDKO mice (Supporting Fig. S1B-G).

ULK1/2 LDKO MICE ARE RESISTANT TO APAP-INDUCED LIVER INJURY INDEPENDENTLY OF THE AUTOPHAGIC PROCESSES

Previous studies showed that autophagy alleviates APAP-induced liver injury, and depletion of autophagy-related genes such as Atg7 in the liver makes mice become more vulnerable to APAP-induced liver injury (AILI).38, 39 However, another study demonstrated that mice with hepatic deletion of Atg5, another autophagy-essential gene, became more resistant to APAP-induced hepatic damage.33 We therefore tested the role of Ulk1/2 in APAP-stimulated liver injury. It was found that, in Ulk1/2 control (Ctrl) mice, typical centrilobular necrosis, as demonstrated by cellular vacuolization, cell swelling, and nuclear disintegration, was induced after APAP treatment (Fig. 2A), as described.4 However, necrotic cell death was markedly attenuated in APAP-treated Ulk1/2 LDKO mice (Fig. 2A). Of note, single knockout of either Ulk1 or Ulk2 did not protect mice from AILI (Supporting Fig. S2). Meanwhile, we found that APAP-induced increase of serum ALT levels was remarkably attenuated in Ulk1/2 LDKO mice (Fig. 2B), indicative of reduced liver injury in these mice. Unexpectedly, we detected similar induction of autophagy processes in the liver of control and Ulk1/2 LDKO mice (Fig. 2C,D), indicating that autophagy was induced through a mechanism independent of Ulk1/2 after APAP treatment. Together, these results show that AILI was strongly blocked in mouse livers with loss of Ulk1/2 independently of autophagy indication.

HEPATIC ULK1/2 ARE ESSENTIAL FOR JNK-MEDIATED LIVER INJURY IN RESPONSE TO APAP TREATMENT

APAP is metabolized mainly by cytochrome P450 2E1 (CYP2E1) in the early hours in hepatocytes, which generates the reactive metabolite, NAPQI, that depletes intracellular GSH and covalently binds to proteins.4-6 It was found that loss of Ulk1/2 did not affect the protein level of CYP2E1, GSH depletion, or covalent binding of APAP to proteins after APAP treatment (Supporting Fig. S3A-C). Activation of JNK is another hallmark of AILI.6, 13, 14 Indeed, we found that treatment of Ulk1/2 Ctrl mice with APAP for 6 hours strongly induced activation of JNK in liver (Fig. 3A). Interestingly, consistent with markedly reduced liver injury in APAP-treated Ulk1/2 LDKO mice (Fig. 2A,B), Ulk1/2 LDKO mice were strongly resistant to APAP-stimulated JNK activation in the liver (Fig. 3A). Furthermore, cotreatment of Ulk1/2 Ctrl mice with JNK inhibitor SP600125 almost completely blocked APAP-induced-JNK activation in liver as well as APAP-stimulated increase of plasma ALT and liver injury (Fig. 3B-D), confirming that JNK activation is an essential event for AILI.13, 14

It has been reported that robust oxidative stress response may play protective roles in AILI. 33 We therefore measured the expression levels of nuclear factor (erythroid-derived 2)-like 2 (Nrf2)-targeting genes, including GCLM, GCLC, and NAD(P)H quinone dehydrogenase 1, by either western blotting or RT-PCR. It was found that loss of Ulk1/2 did not affect Nrf2 signaling under both basal and APAP treatment conditions (Supporting Fig. S3C,D). Liver regeneration is also considered to be important for survival after AILI. 2-4, 33 We thus investigated the proliferation rate of Ulk1/2 Ctrl and LDKO liver cells by using Ki-67 immunostaining. It was found that the number of Ki-67-positive cells was significantly elevated in Ulk1/2 LDKO livers under both basal and APAP-treated conditions (Supporting Fig. S4A). Moreover, increased protein levels of cell proliferation marker PCNA were detected in Ulk1/2 LDKO mouse livers compared to those in Ulk1/2 Ctrl mice (Supporting Fig. S4B). These results suggest that the proliferation rate of liver cells of the Ulk1/2 LDKO mouse is higher than that from the Ulk1/2 Ctrl mouse, which may contribute to the hepatomegaly found in Ulk1/2 LDKO mice and accelerate tissue recovery from AILI.

ULK1/2 DEFICIENCY PROTECTS HEPATOCYTES FROM APAP-INDUCED JNK ACTIVATION AND NECROTIC CELL DEATH

To confirm a pivotal role of Ulk1/2 in APAP-induced necrosis and liver injury, primary hepatocytes were isolated from both Ulk1/2 Ctrl and LDKO mice and treated with APAP. Consistent with the in vivo results (Fig. 3A,B), we found that depletion of Ulk1/2 strongly impaired activation of JNK in response to APAP in primary hepatocytes (Fig. 4A). In contrast, TNF-α, another classical activator of JNK signaling,10 robustly induces JNK phosphorylation in both Ulk1/2 LDKO and Ctrl hepatocytes (Fig. 4B). However, depletion of Ulk1/2 did not affect expression of Sab (SH3BP5; Supporting Fig. S5), which is critical in promoting mitochondrial ROS release that helps sustain JNK activation through activation of MAP kinase kinase kinase (MAP3K), such as apoptosis signal-regulating kinase 1 (ASK1).40 Importantly, in agreement with previous findings on inhibition of mTORC1 by APAP,39 we found that mTORC1-mediated phosphorylation of Ulk1 at Ser⁷⁵⁷, which suppresses activity of Ulk1,24 was markedly decreased during APAP treatment, but not during TNF-α stimulation (Fig. 4A,B), indicating that de-inhibition of Ulk1 after Ser⁷⁵⁷ dephosphorylation may be a critical step for APAP-induced JNK activation. Of note, in line with the normal induction of autophagy in liver of Ulk1/2 LDKO mice (Fig. 2C,D), autophagy was effectively induced in Ulk1/2 LDKO hepatocytes after APAP treatment, through assaying p62 degradation, long-lived protein degradation or LC3B levels with or without lysosome inhibitor treatment (Fig. 4A,C,D). By quantifying PI-positive cells with intact nuclei as necrotic cells, we found that APAP-induced necrotic cell death was remarkably decreased in Ulk1/2 LDKO hepatocytes compared to that in Ulk1/2 Ctrl hepatocytes (Fig. 4E). Collectively, these results reveal that Ulk1/2 affect APAP-induced JNK activation and necrosis in a cell-autonomous manner.

To determine whether the role of Ulk1/2 in liver injury is specific to APAP, TNF-α/GalN (in vivo) or TNF-α/ActD (in vitro) was also used to induce liver injury in Ulk1/2 Ctrl and LDKO mice/hepatocytes.41 Similar extent of liver injury (in vivo) or cell death (in vitro) in Ulk1/2 Ctrl and LDKO mice/hepatocytes treated with TNF-α/GalN or TNF-α/ActD was detected (Fig. 5A-C). Consistently, in the hepatocytes of both mouse strains, JNK signaling was comparably activated (Fig. 5D,E). In addition, contrary to the markedly decreased Ser⁷⁵⁷-Ulk1 phosphorylation after APAP treatment (Fig. 4A), combined treatment of TNF-α and ActD did not affect Ser⁷⁵⁷ phosphorylation of Ulk1 (Fig. 5E), similar to the data from TNF-α single-treatment condition (Fig. 4B). These results suggest that hepatic depletion of Ulk1/2 did not affect TNF-α/GalN- or TNF-α/ActD- induced liver injury/cell death in mice/hepatocytes, and that activation of Ulk1 after Ser⁷⁵⁷ dephosphorylation is a unique step for AILI.

MKK4/7 ARE DIRECT TARGETS OF ULK1/2 AND MEDIATE AILI

Next, we asked the molecular mechanism by which Ulk1/2 participate in JNK signaling. A co-immunoprecipitation assay was performed to test whether Ulk1 would specifically interact with JNK1 and JNK2 or their direct upstream kinases, MKK4 and MKK7.42 We found that MKK4 as well as MKK7, but not JNK1 or JNK2, strongly interacted with Ulk1 (Fig. 6A). We next tested whether MKK4 and MKK7 could be the direct substrates for the serine/threonine kinase, Ulk1. It was found that coexpression of WT, but not the kinase-inactive (KI) form of Ulk1, with MKK4 or MKK7 resulted in different electrophoretic mobility shifts in gels containing Phos-tag that retards the mobility of phosphorylated proteins (Fig. 6B,C).43 Similar results were obtained using Ulk2 as the kinase (Supporting Fig. S6). Treatment of immunoprecipitated MKK4 or MKK7 with calf-intestinal alkaline phosphatase (CIP) effectively diminished the shifts (Fig. 6B,C). In vitro kinase assays using His-tagged Ulk1 expressed and purified from insect cells further confirmed that Ulk1 could directly phosphorylate MKK4 and MKK7 (Fig. 6D). To explore the functional consequences of Ulk1-mediated phosphorylation of MKK4 and MKK7, we tested whether their kinase activities toward JNK were affected by Ulk1-mediated phosphorylation. Intriguingly, we found that coexpression of Ulk1 with MKK7 significantly increased its kinase activity toward JNK by more than 2-fold, whereas the activity of MKK4 was increased by around 30% after Ulk1 coexpression (Fig. 6E), suggesting that MKK7 may play a major role in Ulk1/2-dependent JNK activation. We therefore focused on the phosphorylation event of MKK7 in our following experiments.

ULK1/2-DEPENDENT MKK7 PHOSPHORYLATION IS CRITICAL FOR APAP-INDUCED JNK ACTIVATION AND CELL DEATH

In vitro phosphorylated MKK7 was then subjected to mass spectrometry; four phosphorylated serine residues (Ser³⁵, Ser³⁷⁸, Ser³⁹², and Ser⁴⁰³) were identified as candidate sites targeted by Ulk1 (Fig. 7A), distinct from the conventional activation sites by MAP3Ks.42 Mutants carrying alterations of these serine residues to alanine singly or in combination were constructed. It was found that combined mutation of the four serine residues to alanine (4SA) almost completely blocked the phosphorylation of MKK7 by Ulk1, as indicated by the disappearance of mobility shifts in Phos-tag-containing gels (Fig. 7B). In addition, by using a motif-dependent antibody kit, we found that Ulk1-mediated phosphorylation of Ser⁴⁰³ on MKK7 happened to be recognized by a context-dependent antibody (Phospho-CDK substrate motif [(K/H)pSP] antibody, referred here as P-S403-MKK7 antibody; Fig. 7B). Further work revealed that the activities of Ulk1-nonphosphorylatable 4SA mutant of MKK7 were no longer affected by Ulk1 coexpression (Fig. 7C). Endogenous phosphorylation of MKK7 after APAP treatment was also detected in Ulk1/2 Ctrl, but not in LDKO mice hepatocytes (Fig. 7D). In addition, interaction between endogenous Ulk1 and MKK4/7 was strongly increased in liver of APAP-treated mice (Fig. 7E), supporting a dynamic response of Ulk1/2 in regulating MKK4/7 after APAP treatment.

By infecting primary hepatocytes with adenovirus expressing control shRNA or shRNAs targeting Mkk4 and/or Mkk7, we found that DKD of Mkk4 and Mkk7 (Mkk4/7 DKD) strongly impaired APAP-induced JNK phosphorylation and toxicity (Fig. 8A and Supporting Fig. S7). Importantly, reconstitution of WT-MKK7, but not 4SA-MKK7, effectively rescued the activation of JNK after APAP treatment in Mkk4/7 DKD hepatocytes (Fig. 8B). Finally, Mkk4/7 DKD hepatocytes reconstituted with WT-MKK7 showed severely increased necrotic cell death than that of cells reconstituted with the 4SA-MKK7 mutant (Fig. 8C,D). Several MAP3Ks, including ASK1 and mixed-lineage protein kinase 3 (MLK3), have been shown to be critical in AILI by acting as direct upstream kinases of MKK4/7 and thereby activating JNK signaling.46-48 We found that MAP3Ks-mediated phosphorylation on MKK7-Ser²⁷¹/Thr²⁷⁵ does not alter Ser⁴⁰³ phosphorylation by Ulk1/2, but Ulk1/2-mediated MKK7 phosphorylation is required for a robust phosphorylation of MKK7 on Ser²⁷¹/Thr²⁷⁵ by MAP3Ks after APAP treatment (Fig. 7D and Supporting Fig. S8). Using adenovirus-mediated protein expression methods, we have also reconstituted the expression of either Ulk1 or Ulk2 both in vivo in Ulk1/2 LDKO mice and in isolated Ulk1/2 DKO hepatocytes. It was found that reconstitution of either Ulk1 or Ulk2 could significantly restore APAP-induced toxicity in Ulk1/2 LDKO mice/hepatocytes (Fig. 8E and Supporting Fig. S9), confirming that Ulk1/2 play key roles in AILI. Together, these results provide strong evidence showing that Ulk1/2-dependent MKK7 phosphorylation is a critical step for the activation of JNK and the induction of necrotic cell death during APAP treatment.