OTUB1 prevents lethal hepatocyte necroptosis through stabilization of c-IAP1 during murine liver inflammation

In bacterial and sterile inflammation of the liver, hepatocyte apoptosis is, in contrast to necroptosis, a common feature. The molecular mechanisms preventing hepatocyte necroptosis and the potential consequences of hepatocyte necroptosis are largely unknown. Apoptosis and necroptosis are critically regulated by the ubiquitination of signaling molecules but especially the regulatory function of deubiquitinating enzymes (DUBs) is imperfectly defined. Here, we addressed the role of the DUB OTU domain aldehyde binding-1 (OTUB1) in hepatocyte cell death upon both infection with the hepatocyte-infecting bacterium Listeria monocytogenes (Lm) and D-Galactosamine (DGal)/Tumor necrosis factor (TNF)-induced sterile inflammation. Combined in vivo and in vitro experiments comprising mice lacking OTUB1 specifically in liver parenchymal cells (OTUB1LPC-KO) and human OTUB1-deficient HepG2 cells revealed that OTUB1 prevented hepatocyte necroptosis but not apoptosis upon infection with Lm and DGal/TNF challenge. Lm-induced necroptosis in OTUB1LPC-KO mice resulted in increased alanine aminotransferase (ALT) and lactate dehydrogenase (LDH) release and rapid lethality. Treatment with the receptor-interacting serine/threonine-protein kinase (RIPK) 1 inhibitor necrostatin-1s and deletion of the pseudokinase mixed lineage kinase domain-like protein (MLKL) prevented liver damage and death of infected OTUB1LPC-KO mice. Mechanistically, OTUB1 reduced K48-linked polyubiquitination of the cellular inhibitor of apoptosis 1 (c-IAP1), thereby diminishing its degradation. In the absence of OTUB1, c-IAP1 degradation resulted in reduced K63-linked polyubiquitination and increased phosphorylation of RIPK1, RIPK1/RIPK3 necrosome formation, MLKL-phosphorylation and hepatocyte death. Additionally, OTUB1-deficiency induced RIPK1-dependent extracellular-signal-regulated kinase (ERK) activation and TNF production in Lm-infected hepatocytes. Collectively, these findings identify OTUB1 as a novel regulator of hepatocyte-intrinsic necroptosis and a critical factor for survival of bacterial hepatitis and TNF challenge.

The Gram-positive bacterium Listeria monocytogenes (Lm) can induce cell type-dependent apoptosis and necroptosis of infected target cells in the liver. Within a few hours after systemic infection, Lm homes to the liver, infects Kupffer cells, and induces necroptosis [11]. Subsequently, released bacteria infect hepatocytes and may induce hepatocyte apoptosis [12]. In nonlethal listeriosis, bacteria are cleared from the liver by the immune response without any residual liver damage. Control of Lm in the liver is critically dependent on TNF produced by Kupffer cells, macrophages, and hepatocytes as well as TNFR1 expression of myeloid cells but not of hepatocytes [13,14].
In contrast to listeriosis, treatment with TNF or lipopolysaccharide (LPS) in combination with D-Galactosamine (D-Gal) mediates hepatocyte apoptosis and acute liver failure [15,16]. In this model, D-Gal inhibits hepatocyte transcription and TNF signals through TNFR1 resulting activation of caspases and apoptosis of hepatocytes [15][16][17][18]. In contrast, hepatocytes are relatively resistant to the induction of necroptosis. However, inhibition of caspase-8 and c-IAP1, respectively, can force hepatocyte necroptosis [2,19] indicating that necroptosis is actively inhibited by yet incompletely defined mechanisms.
Post-translational modification of apoptosis and necroptosis-associated proteins by ubiquitination critically regulates these cell death pathways. Ubiquitination is a multi-step process catalyzed by ubiquitin-activating (E1), ubiquitin-conjugating (E2), and ubiquitin-ligating (E3) enzymes leading to covalent attachment of single or multiple ubiquitin molecules to target substrates. Of note, substrates linked with K48-linked polyubiquitin chains are targeted for proteasomal degradation while K63-linked or linear polyubiquitin chains confer non-degradative functions such as protein trafficking and signal transduction [20]. With respect to cell death signaling, the E3 ligase c-IAP1 prevents RIPK1-mediated apoptosis by autoubiquitination and K63-linked polyubiquitination of RIPK1 supporting activation of NF-κB instead of cell death pathways.
Based on the regulation of both cell death and proinflammatory signaling pathways by OTUB1 we explored the in vivo function of hepatocyte-specific OTUB1 during bacterial hepatitis and acute liver injury induced by DGal/ LPS and DGal/TNF. Here, we demonstrate that OTUB1 prevents in vivo and in vitro Lm and TNF-driven necroptosis of hepatocytes and rescued mice from Lm and D-Gal/ TNF-induced mortality. Mechanistically, OTUB1 prevented necroptosis by reducing K48-dependent proteasomal degradation of c-IAP1, thereby, enabling c-IAP1-mediated K63-linked polyubiquitination of RIPK1. Thus, OTUB1 is an endogenous inhibitor of hepatocyte necroptosis and pro-survival factor during bacterial-and TNF-induced inflammation.

Data reporting
No statistical methods were used to predetermine sample size.
In vitro, cells were infected with Lm (multiplicity of infection (MOI) of 10) at 37°C for 1 h. Thereafter, extracellular Lm were killed by adding 30 µg/ml gentamicin. For analysis of the K48-linked ubiquitination, cells were infected in the presence of MG132 (10 µM, Cat# M7449, Sigma-Aldrich, Steinheim, Germany). Bacterial loads in liver and spleen were determined by plating serial dilutions of liver and spleen homogenates on brain heart infusion agar. To enumerate intracellular bacteria in cell culture, cell lysates were plated on brain heart infusion agar. Inoculated agar plates were incubated at 37°C for 24 h and bacterial colonies were counted microscopically in a randomized blinded manner. Mice were monitored for signs of disease (weight loss, food and water intake, lethargy, activity, decline of body temperature). Mice with severe clinical signs of disease were sacrificed.

Serology
Blood was obtained by cardiac puncture of anesthetized mice. Serum was separated by centrifugation and analyzed for alanine aminotransferase (ALT) activity and lactate dehydrogenase (LDH) release according to the manufacturer's instructions using an ALT activity assay (Cat# MAK052-1KT, Sigma-Aldrich, Steinheim, Germany) and CyQUANT LDH Cytotoxicity Assay (Cat#C20300, Thermo Scientific, Waltham, USA), respectively.

Histopathology
Liver tissue from randomly selected mice was fixed in 4% paraformaldehyde at 4°C. Paraffin sections were counterstained with hematoxylin and eosin (H&E).

HepG2 cell culture
HepG2 cells were obtained from ATCC (ATCC HB 8065). The cells were authenticated by the vendor using STR profiling prior to purchase. The cells were tested free of mycoplasma and cultured in DMEM supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin, 1% L-Glutamine, and 1% nonessential amino acids. OTUB1 was deleted from the HepG2 cells by transfection with lentiviral vectors.

Plasmids and lentivirus preparation
The shRNA plasmids were generated by inserting the hairpin oligonucleotides into lentiviral construct as previously described ( [29]). Oligos contain a 21-mer sequence followed by a loop sequence and the reverse complement: OTUB1-1 (5′-GAT CCC CGC AAG TTC TTC GAG CAC TTT TCA AGA GAA AGT GCT CGA AGA ACT TCG TTT TTT GGA AG-3′) and OTUB1-2 (5′-GAT CCC CCC GAC TAC CTT GTG GTC TAT TCA AGA GAT AGA CCA CAA GGT AGT CGG TTT TTT GGA AG). Virus production was performed by calcium phosphate transfection of HEK293T cells using four plasmids encoding the helper functions (gagpol, rev, VSV-G) and the lentiviral vector. Titers were evaluated on NIH3T3 by flow cytometry.

Cell death analysis
Cell death was determined by incubation of cells with 0.4% trypan blue solution (Cat# 15250061, Thermo Scientific, Waltham, USA). Samples were counted in a randomized blinded manner under a light microscope, and the percentage of nonviable blue cells was determined by the ratio of stained cells versus all cells.

LDH assay
The LDH release was measured in serum of mice and cell culture supernatant using CyQUANT LDH Cytotoxicity Assay (Cat#C20300, Thermo Scientific, Waltham, USA) according to the manufacturer's protocol. The measurements were carried out in a microtiter plate reader at 490 and 680 nm. Values were shown as LDH release relative to untreated mice and cells, respectively.

Deletion of MLKL by siRNA
HepG2 cells were transfected with MLKL siRNA or negative control siRNA (Cat# L-005326-00-0010, Dharmacon, Colorado, USA). Transfection was performed with Lipofectamine RNAiMAX transfection reagent (Cat# 13778075, Thermo Scientific, Waltham, USA) according to the manufacturer's protocol for 48 h in a six-well plate. Knockdown efficiency was controlled by western blotting (WB).

Quantitative real-time PCR (qRT-PCR)
Total RNA was isolated from liver tissue using RNeasy Kit (Cat# 74136, Qiagen, Hilden, Germany). First strand cDNA was synthetized from 1 µg of total RNA with SuperScript Reverse Transcriptase Kit (Cat# 18064022, Thermo Scientific, Waltham, USA). qRT-PCR was performed with TaqMan probes according to the manufacturer's protocol (Applied Biosystems, Foster City, USA) using a Light-Cycler 480 (Roche Diagnostics, Mannheim, Germany). All primers indicated in Supplementary Table 1 were purchased from Thermo Fisher Scientific (Waltham, USA). Gene expression was normalized to the hypoxanthine phosphoribosyltransferase as reference gene and calculated as ratio between the target gene and the reference gene employing the ΔΔCt method.

Immune cell characterization
Liver immune cell populations were characterized by flow cytometry. Therefore, the anesthetized mice were intracardially perfused with 0.9% NaCl and the liver was subsequently minced through a 100 µm cell strainer. After erythrocyte lysis with ammonium chloride, leukocytes were isolated by Percoll density gradient centrifugation (Cat# GE17-0891-01, GE Healthcare, Uppsala, Sweden). Hepatic leukocytes were stained with mouse-specific fluorochrome-labeled antibodies as indicated in Supplementary Table 2. For detecting of intracellular cytokines, cells were re-stimulated with heat-killed Lm (80°C, 1 h) for 5 h in the presence of Brefeldin A solution (Cat# 420601, Biolegend, San Diego, USA) followed by intracellular cytokine staining using the Intracellular Fixation & Permeabilization Kit (eBioscience by affymetrix, San Diego, USA). Samples were acquired on BD FACS Canto II (BD Biosciences, San Jose, USA) followed by data analysis using FlowJo X software (Tree Star, Ashland, USA).

Protein lysate preparation and western blotting
Liver tissue was minced in radioimmunoprecipitation assay buffer (Cat# 9806, Cell Signaling Technology, Danvers, USA) containing protease inhibitor cocktail (Cat# P8340, Sigma-Aldrich, Steinheim, Germany), phenylmethylsulfonyl fluoride (Cat# 8553, Cell Signaling Technology, Danvers, USA) and PhosSTOP (Cat# 04 906 837 001, Roche Diagnostics Mannheim, Germany) incubated for 30 min on ice and subsequently centrifuged at 13,000 × g at 4°C for 15 min. Supernatant was collected and protein concentrations were measured with Bradford assay. Equal amounts of protein were fractionated on SDS-polyacrylamide gel, blotted on a polyvinylidene fluoride membrane, and blocked at room temperature for 1 h with 5% bovine serum albumin (BSA) or 5% milk, respectively. Proteins were probed with primary antibodies as listed in Supplementary Table 3. Blots were developed using Pierce ECL 2 Western Blotting Substrate (Cat# 32132, Thermo Fisher Scientific, Waltham, USA) and images were captured with the Intas Chemo Cam Luminescent Image Analysis system (INTAS Science Imaging Instruments, Göttingen, Germany). For quantification, the LabImage Platform software was used.

Immunoprecipitation
Protein lysates prepared as described before were precleared with GammaBind G Sepharose beads (Cat# 17-0885-01, GE Healthcare, Uppsala, Sweden) at 4°C for 1 h with agitation. Protein samples were centrifuged at 10,000 × g at 4°C for 10 min to remove the beads. Equal protein amounts were incubated with the respective primary antibody overnight on a rocking platform at 4°C, followed by 2 h incubation with the beads at 4°C. The beads were collected by centrifugation at 10,000 × g at 4°C for 30 s and subsequently washed for three times with cold PBS. Immune complexes were eluted with 2 × lane marker reducing sample buffer (Cat# 39000, Thermo Scientific, Waltham, USA) at 99°C for 5 min. To remove the beads, samples were centrifuged and western blot was performed as described above.

Immunofluorescence
HepG2 cells plated on a coverslip were infected with Lm or stimulated with TNF/CHX as described before. Following stimulation, cells were fixed in 4% paraformaldehyde for 10 min at room temperature and thereafter permeabilized with 0.1% Triton-X-100 for 7 min. Nonspecific bindings were blocked by incubating the cells for 45 min in 0.5% BSA, 0.3% Triton-X-100, 5% sucrose, 10% fetal calf serum) before staining the cells with anti-p-MLKL (Cat# ab187091, Abcam, Cambridge, UK) overnight on a rocking platform. The slides were washed three times with PBS and then incubated with the fluorochrome-conjugated secondary antibody (Goat Anti-Rabbit IgG, Alexa Fluor Plus 594, Thermo Fisher Scientific, Waltham, USA) for 2 h protected from light. After washing with PBS, the slides were mounted on a glass slide using ProLong Diamond Antifade with DAPI (Cat# P36962, Thermo Fisher Scientific, Waltham, USA). Images were captured on Microscope Axio Imager Z1 (Carl Zeiss AG, Oberkochen, Germany).

Statistical analysis
All graphs represent the mean; error bars indicate the standard error of the mean (SEM). The statistical significance was determined with the software Prism 8 (GraphPad Software, Inc., San Diego, USA) using the respective test as indicated in the figure legends. All experiments were performed at least twice.

Results
Liver parenchymal cell-specific OTUB1 protects mice from bacterial and DGal/LPS-induced mortality To investigate the function of OTUB1 in the liver, we generated OTUB1 LPC-KO mice lacking OTUB1 expression specifically in liver parenchymal cells (Fig. S1A). OTUB1 LPC-KO mice developed normally and liver architecture was regular without inflammation up to 1 year (Fig. S1B-D). To study the function of hepatocyte-specific OTUB1 during infection, OTUB1 LPC-KO and OTUB1 FL mice were i.v. infected with Lm. While all OTUB1 LPC-KO mice succumbed up to day 6 p.i., 90% of OTUB1 FL mice survived ( Fig. 1A). At day 3 p.i., bacterial loads in liver and spleen did not differ significantly between both strains of mice (Figs. 1B and S1E) indicating that the protective effect of OTUB1 on survival was independent of pathogen control.
OTUB1 LPC-KO mice developed severe listeriosis characterized by large inflammatory foci with severely damaged hepatocytes and central necrosis surrounded by leukocytes (Fig. 1C). In contrast, control mice exhibited numerous small inflammatory foci containing viable hepatocytes but no necrosis throughout the hepatic parenchyma (Fig. 1C). OTUB1 limited liver dysfunction and cell death in Lminfected mice as indicated by lower serum levels of ALT and LDH (Fig. 1D), respectively, in OTUB1 FL mice. Hepatocyte OTUB1 expression did not impact on the total number of inflammatory leukocytes recruited to the liver and hepatic transcription of interleukin (IL)−6, IL-1β, and interferon (IFN)γ (Fig. S1F, G). However, hepatic TNF mRNA expression was significantly increased in the livers of OTUB1 LPC-KO mice at day 3 p.i. (Fig. 1E). Thus, OTUB1 prevented lethal listeriosis, liver necrosis, and production of TNF.
Next, we identified the cell population producing increased amounts of TNF in the absence of OTUB1 during listeriosis. As shown in Fig. S1H, TNF production of hepatic CD45 + F4/80 + CD11b low Ly-6C low Kupffer cells, CD45 + F4/80 + CD11b int Ly-6C high inflammatory monocytes and CD45 + F4/80 + CD11b high Ly-6C int monocyte-derived macrophage was comparable in both Lm-infected mouse strains. However, upon in vitro Lm infection of hepatocytes, TNF mRNA expression and LDH release (Fig. 1F) were significantly increased in OTUB1-deficient hepatocytes. In agreement with the in vivo data (Fig. 1B), bacterial loads were comparable between cultivated Lm-infected hepatocytes of both mouse strains (Fig. S1I) suggesting that OTUB1 may suppress TNF production by hepatocytes without affecting pathogen control.
To further study the role of OTUB1 in TNF responses of hepatocytes, we employed the DGal/LPS model, which induces TNF-dependent hepatocyte death and acute liver failure. Upon i.p. injection of low dose LPS in combination with DGal, all OTUB1 LPC-KO mice succumbed up to 24 h p. i, whereas 70% of OTUB1 FL mice survived ( Fig. 2A). As early as 6 h after DGal/LPS injection, the liver of OTUB1 LPC-KO mice harbored large necrosis evident already macroscopically, which were absent in the liver in OTUB1 FL mice (Figs. 2B and S2A). Histologically, sharply demarcated hepatic necrosis was present in OTUB1 LPC-KO but not in OTUB1 FL mice (Figs. 2B and S2A). Serum ALT was significantly increased in OTUB1 LPC-KO mice. Additionally, OTUB1 LPC-KO mice expressed significantly more TNF mRNA in the liver illustrating an augmented pathogen-independent TNF production in the absence of OTUB1 (Fig. 2C). To directly assess the role of TNF in the enhanced hepatocyte death in OTUB1 LPC-KO mice, we substituted LPS by TNF and challenged mice with equally low amounts of TNF in combination with DGal. In accordance to the DGal/LPS experiments, survival of OTUB1 LPC-KO mice was significantly reduced as compared to OTUB1 FL mice (Fig. 2D). Macroscopically, livers of OTUB1 LPC-KO mice showed liver damage and hemorrhages, which were absent in OTUB1 FL mice (Figs. 2E and S2B). OTUB1 LPC-KO hepatocytes were swollen, the cytoplasm was eosinophilic and nuclei were no longer identified (Figs. 2E and S2B). Small inflammatory infiltrates were associated with damaged hepatocytes in OTUB1 LPC-KO mice (Figs. 2E and S2B) In contrast, livers of OTUB1 FL mice were histologically normal (Fig. 2E) or showed only moderate liver pathology (Fig. S2B). Furthermore, serum ALT and LDH were elevated and hepatic TNF mRNA was increased in OTUB1 LPC-KO mice (Fig. 2F). Taken together, these data indicate that OTUB1 prevents TNF-induced hepatocyte death in vivo.
OTUB1 reduced ERK activation and shifted cell death signaling in Listeria-infected and DGal/TNFstimulated mice Since OTUB1 has been shown to inhibit TWEAK-induced apoptosis and pro-inflammatory signaling [21,23,30], we investigated the impact of hepatocyte-specific OTUB1 expression on NF-κB and cell death pathways. WB analysis of Listeria-infected liver tissue of OTUB1 LPC-KO and OTUB1 FL mice revealed an equal phosphorylation of p65 indicating a similar activation of the pathway (Fig. 3A). Additionally, the NF-κB regulated pro-apoptotic gene Bcl-2-associated x protein (Bax) (Figs. 3A and S3A) as well as the antiapoptotic genes Bcl-2 (Figs. 3A and S3B) and c-FLIP (Fig. S3C) showed equal mRNA expression levels in the livers of both genotypes.
Next, we studied the expression of c-IAP1, apoptosis, and necroptosis regulating signaling molecules in livers of Listeria-infected and D-Gal/TNF challenged mice, respectively. In Lm-infected mice, c-IAP1 proteins were decreased in OTUB1 LPC-KO mice (Fig. 3A, B). Of note, transcription of c-IAP1 mRNA was not affected following infection with Listeria (Fig. S3D) indicating that OTUB1 regulated c-IAP1 protein stability. In contrast, TRAF2, an inhibitor of cell death [31] was equally expressed in both genotypes (Fig. 3A).
Apoptotic cell death is tightly regulated by caspase-8, which either directly initiates apoptosis by cleaving the effector caspases-3, 6, and 7 or indirectly via cleaving BH3interacting-domain death antagonist (BID) to tBID (truncated BID) resulting in Bax and effector caspases activation [2]. In OTUB1 FL mice, hepatic cleavage of procaspase-3, 6, and 7 was increased as compared to OTUB1 LPC-KO mice, while levels of BID and tBID were comparable between both groups (Fig. 3A). Thus, the mild liver pathology of Listeria-infected OTUB1-competent mice was associated with effector caspase cleavage, whereas the lethal liver pathology and hepatocyte death of OTUB1 LPC-KO mice was independent of the cleavage of the executioner caspases. Since decreased c-IAP1 levels in conjunction with caspase-8 inhibition promote necroptosis [19], we analyzed RIPK1/RIPK3 necrosome formation. RIPK1 and RIPK3 were slightly increased upon infection in the livers of uninfected and Lm-infected OTUB1 FL and OTUB1 LPC-KO mice (Fig. 3A). RIPK1 immunoprecipitates from livers of both mouse strains contained RIPK3 with an increased amount in OTUB1 LPC-KO mice (Fig. 3C) indicating that OTUB1-deficiency increases necroptosis during listeriosis. Furthermore, we analyzed the activation of extracellular signal-regulated kinase (ERK), p-38, c-Jun N-terminal kinase (JNK), and the noncanonical NF-κB pathway. Interestingly, phosphorylation of ERK was more pronounced in the livers of OTUB1 LPC-KO mice as compared to OTUB1 FL mice, while activation of p-38, JNK, and the noncanonical NF-κB pathway did not differ between the two groups of mice (Figs. 3A and S3E).

Inhibition of RIPK1 and MLKL but not of caspases rescued liver function and prevented death of Listeria-infected OTUB1 LPC-KO mice
Since RIPK1 kinase activity can mediate both caspase-8dependent apoptosis and RIPK1/RIPK3/MLKL-dependent necroptosis [32], we treated Listeria-infected OTUB1 LPC-KO and OTUB1 FL mice with the RIPK1 kinase inhibitor Nec-1s and the pan-caspase inhibitor zVAD. Nec-1s but not zVAD treatment resulted in a significant decline in serum ALT, LDH levels and hepatic TNF mRNA expression in OTUB1 LPC-KO mice and abolished differences between both genotypes (Fig. 4A). However, Nec-1s treatment had no influence on the bacterial burden (Fig. 4B). In accordance with the reduced liver damage and TNF production, Nec-1s treatment conferred protection against Listeria-induced lethality in the OTUB1 LPC-KO mice (Fig. 4C).
Previous reports identified MLKL as the terminal executor of necroptosis by inducing cell membrane disruption [33,34]. Thus, we crossed OTUB1 LPC-KO mice with MLKL-deficient mice to obtain hepatocytes deficient in both OTUB1 and MLKL (MLKL del/del OTUB1 LPC-KO mice, Fig. S3A). Upon infection, deletion of MLKL in OTUB1 LPC-KO mice significantly reduced ALT and LDH levels as compared to OTUB1 LPC-KO (Fig. 4D) illustrating that necroptosis aggravated liver damage in the absence of OTUB1. Of note, TNF mRNA production was equally increased in both MLKL del/del OTUB1 LPC-KO and OTUB1 LPC-KO mice as compared to OTUB1 FL and MLKL del/del mice, respectively, (Fig. 4D) indicating that OTUB1-mediated suppression of TNF mRNA is upstream of MLKL. These in vivo data suggest that hepatocyte-intrinsic OTUB1 inhibits necroptosis but not apoptosis in Lm-infected mice.

OTUB1 prevents necroptosis by stabilization of c-IAP1
Considering the increased in vivo degradation of c-IAP1 in the absence of OTUB1 (Fig. 3A, B), we studied whether OTUB1 protects hepatic c-IAP1 against proteasomal degradation in listeriosis and upon D-Gal/TNF treatment.  activating signal and (ii) blocks its kinase activity required for RIPK1/RIPK3 necrosome formation [35]. Analysis of K63-linked polyubiquitination of RIPK1 in Listeria-infected livers revealed a strong K63-linked polyubiquitination of RIPK1 in OTUB1 FL mice but not in OTUB1 LPC-KO and MLKL del/del OTUB1 LPC-KO mice (Figs. 5F and S5F). Collectively, these data illustrate that OTUB1 removes K48linked polyubiquitin chains from c-IAP1 and reduces its proteasomal degradation. Hence, in OTUB1 FL mice, increased amounts of c-IAP1 resulted in enhanced K63-linked polyubiquitination of RIPK1, which inhibits RIPK1/ RIPK3 necrosome formation.
To substantiate that the enhanced RIPK1/RIPK3 interaction in Lm-infected and TNF-stimulated OTUB1-deficient HepG2 cells resulted in increased necroptosis, we stained for phospho-MLKL (p-MLKL). Both infection with Lm and stimulation with TNF significantly increased numbers of p-MLKL + cells and the amount of p-MLKL per cell in OTUB1-deficient HepG2 cells as compared to OTUB1-sufficient HepG2 cells (Fig. 7D-F).
OTUB1 inhibits RIPK1/ERK-dependent TNF production RIPK1 and RIPK3 can induce ERK-dependent TNF mRNA production in macrophages [38]. To determine whether OTUB1 regulates RIPK1/ERK-dependent TNF production, we stimulated HepG2 cells with Lm and TNF, respectively, and determined RIPK1/ERK interaction. Both Lm infection (Figs. 8A and S4B) and TNF stimulation (Fig. 8B) induced an interaction of RIPK1 with ERK independent of OTUB1 expression. Confirmatively, inhibition of RIPK1 by Nec-1s and ERK by SCH772984 significantly downregulated TNF mRNA expression in OTUB1-deficient cells and abolished differences between the two genotypes ( Fig. 8C-E). Thus, OTUB1 inhibits hepatocyte necroptosis by two synergistic mechanisms, i.e. inhibition of the RIPK1/RIPK3/MLKL pathway and by the reduction of RIPK1/ERK-dependent TNF production, which further decreases TNF-induced necroptosis.

Discussion
Inflammatory stimuli including TNF and pathogens can induce different types of cell death including apoptosis and necroptosis depending on the underlying cell type and the respective pathogen. With respect to hepatocytes, apoptosis but not necroptosis is induced in human and experimental diseases (reviewed in [3,6,39,40]). Here, we identify OTUB1 as an inhibitor of hepatocyte necroptosis in listeriosis and upon TNF stimulation. The critical in vivo importance of OTUB1 as a hepatocyte-intrinsic inhibitor of necroptosis is illustrated by the lethal exacerbation of bacterial hepatitis and DGal/LPS-induced liver inflammation. Before, it has been demonstrated that inhibition or deletion of c-IAP1 and caspase-8, respectively, results in hepatocyte necroptosis and lethal hepatocyte damage further indicating that hepatocyte necroptosis is intrinsically inhibited to protect the liver [2].
So far it has been exemplified in listeriosis, that infected Kupffer cells undergo RIPK1/MLKL-dependent necroptosis, whereas hepatocytes may become apoptotic [11, 12 and this study]. This illustrates that cell-type-specific but not pathogen-specific factors account for the different forms of cell death in liver resident cells. Apoptosis of hepatocytes has been documented in various human diseases and experimental models including stimulation of mice with DGal/LPS and DGal/TNF, respectively [39,40].
Under homeostatic conditions, interaction of OTUB1 and c-IAP1 was weak. However, infection with Lm and stimulation with TNF strongly increased association of OTUB1 with c-IAP1 and deubiquitination of K48-linked polyubiquitin chains from c-IAP1, thereby reducing its degradation. The reduction of c-IAP1 in OTUB1-deficient hepatocytes resulted in diminished apoptosis, induction of necroptosis and ERK-dependent TNF production upon infection with Lm and stimulation with TNF, respectively. This is in contrast to the rapid and complete degradation of c-IAP1 by c-IAP1 antagonists and activation of TWEAK-FN14 signaling, respectively, which both induce apoptosis but not necroptosis of tumor cell lines by NF-κB-dependent autocrine TNF-production [41][42][43]. These differences in c-IAP1 regulated cell death and pro-inflammatory signaling pathways may be explained by the distinct c-IAP1 levels and cell-type-specific effects, but potentially also by c-IAP1-independent effects of OTUB1 on other signaling molecules.
In TNFR1 signaling, stability of c-IAP1 in the TNFR1signaling complex 1 is important for c-IAP1-dependent K11-, K48-and K63-linked polyubiquitination of RIPK1, which regulate the pro-survival function of RIPK1 in complex I and block the formation of the apoptosis and necroptosis inducing complexes II [44][45][46]. In particular, K63-linked polyubiquitination of RIPK1 is important to inhibit the kinase activity of RIPK1 and to prevent caspase-8 dependent apoptosis and RIPK3/MLKL-dependent necroptosis [32,36]. Thus, in the absence of c-IAP1 and impaired RIPK1 ubiquitination, RIPK1 spontaneously triggers the formation of complex IIb, which can initiate both apoptosis and necroptosis [47]. In accordance, TNF stimulation of RIPK1 Lys376 (K376) mutant mouse embryonic fibroblasts, which exhibit impaired K63-linked polyubiquitination, undergo both RIPK1 kinase-dependent apoptosis and necroptosis, which was further enhanced in the presence of c-IAP inhibitors [32,36]. In good agreement, the increased c-IAP1 degradation in Lm-infected OTUB1 LPC-KO mice and OTUB1-deficient HepG2 cells resulted in reduced K63-linked polyubiquitination of RIPK1 and cell death. In vivo, the critical function of the RIPK1 kinase activity for the exacerbation of listeriosis in OTUB1 LPC-KO mice is illustrated by the reduction of hepatocyte damage, cell death and prevention of lethal listeriosis upon inhibition of RIPK1 kinase activity by Nec-1s.
In addition to K63-linked polyubiquitination of RIPK1 by c-IAP1, linear M1 ubiquitination of RIPK1 by the linearubiquitin assembly complex (LUBAC) inhibits RIPK1induced cell death upon TNF stimulation. LUBAC initiates recruitment of NEMO, IKKα/β, Tank-binding kinase-1 and IKKε to the TNFR1 signaling complex 1, and these kinases phosphorylate RIPK1 and, thereby, prevent cell death [48][49][50][51][52]. Of note, OTUB1 has only specificity for K48-linked but not for K63-linked and M1 ubiquitin chains [53] [27] and, thus, cannot directly reduce K63-linked and M1 ubiquitin-dependent phosphorylation of RIPK1. In contrast to the inhibition of RIPK1 kinase activity by LUBACdependent recruitment of RIPK1-phosporylating kinases, autophosphorylation of RIPK1 at serine 166 licenses RIPK1 Fig. 6 OTUB1 deletion in HEPG2 cells results in c-IAP1 destabilization and promotes necroptosis upon Lm-infection. A-E OTUB1 was stably knocked down in HepG2 cells using lentiviral transduction and MLKL was transiently deleted by siRNA treatment in the indicated OTUB1-deficient HepG2 cells. OTUB1-sufficient, OTUB1deficient, and OTUB1/MLKL-deficient HepG2 cells were infected with Lm with a MOI of 10 for the indicated time points. A LDH release (left) was measured in the supernatant at 8 h.p.i. and calculated as x-fold increase over the respective uninfected controls (n = 10). The percentage of dead cells (right) were enumerated using trypan blue dye. B Cells were harvested at 0, 1, 6, and 24 h post Lm-infection, respectively, and immunoblotted with the indicated antibodies. Total proteins of uninfected and Lm-infected (6 h.p.i.) OTUB1-sufficient, OTUB1-deficient, and OTUB1/MLKL-deficient HepG2 cells were immunoprecipitated with anti-c-IAP1 (C), anti-ubiquitin (Lys48-specific) (D), and anti-RIPK1 (E), respectively, and analyzed by WB with the indicated antibodies. All blots are representative for one of three experiments. Lm Listeria monocytogenes, MOI multiplicity of infection, p.i. post infection, LDH lactate dehydrogenase, c-IAP1 cellular inhibitor of apoptosis 1, p phospho, ERK extracellular-signal-regulated kinase, IκB inhibitor of kappa B, RIPK receptor-interacting serine/ threonine kinase, ub ubiquitin, p-Ser phospho-serine, WB western blot. kinase activity and cell death [37]. In this regard, we detected an increased total serine phosphorylation of RIPK1 in OTUB1-deficient Lm-infected HepG2 cells and it will be interesting to further explore how OTUB1 regulates the complex phosphorylation events of RIPK1 and its kinase activity.
In both Lm-infected and TNF-stimulated OTUB1-deficient mice and HepG2 cells, necroptosis induction was indicated by enhanced RIPK1/RIPK3-necrosome formation and phosphorylation of MLKL. Functionally important, in vivo inhibition of RIPK1 by Nec-1s completely prevented lethal listeriosis in OTUB1 LPC-KO mice and both Nec-1s treatment and combined OTUB1 and MLKL deletion significantly reduced liver damage in Lm-infected OTUB1 LPC-KO mice and cell death of Lm-infected and TNF-stimulated HepG2 cells, respectively. We also detected increased cleavage of caspase-8 and caspase-3 in vivo upon infection of OTUB1 FL mice with Lm and treatment with DGal/TNF, respectively. However, zVAD treatment of OTUB1 FL mice did not alter LDH and ALT levels in Lm infection indicating that hepatocyte apoptosis plays a minor role in the Fig. 7 TNF stimulation triggers the degradation of c-IAP1 and necroptosis in OTUB1-deficient cells. A-D OTUB1-sufficient, OTUB1-deficient, and OTUB1/MLKL-deficient HepG2 cells were left untreated or stimulated with TNF (20 ng/ml)/CHX (10 µg/ml). A LDH release (left) was measured in the supernatant at 6 h and calculated as x-fold increase over the respective uninfected controls (n = 10). The percentage of dead cells (right) were enumerated using trypan blue staining. B OTUB1-sufficient, OTUB1-deficient, and OTUB1/MLKLdeficient HepG2 cells were left untreated or stimulated with TNF (20 ng/ml)/CHX (10 µg/ml) for 1, 3, and 6 h and, thereafter, analyzed by WB with the indicated antibodies. C RIPK1-associated complexes were analyzed in untreated and TNF (20 ng/ml)/CHX (10 µg/ml)-stimulated (  pathogenesis of listeriosis. Since caspase-8 inhibits RIPK1mediated necroptosis [5], the partial cleavage of caspase-8 in combination with the simultaneously impaired K63-linked polyubiquitination of RIPK1, may be sufficient to induce necroptosis in OTUB1-deficient hepatocytes. This is in contrast to Goncharov et al., who demonstrated that destabilization of c-IAP1 in OTUB1-deficient endothelial and epithelial cells induced apoptosis upon in vitro activation of Fn14 by TWEAK. The reason for this discrepancy may relate not only to cell-type specificity but also to the different stimuli leading to divergent activation of pro-inflammatory signaling pathways. In TWEAK/FN14-activated OTUB1deficient cells, c-IAP1 degradation impaired the activation of mitogen-activated protein kinase and canonical NF-κB pathways which sensitized these cells towards endogenously released TNF-dependent and caspases-mediated apoptosis [21]. In contrast, our study shows that OTUB1deficiency in hepatocytes and the subsequent accelerated degradation of c-IAP1 (i) augmented Lm-and TNF-induced ERK activation without affecting canonical and noncanonical NF-κB activation and expression of NF-κBdependent antiapoptotic genes (ii) induced RIPK1 and MLKL-dependent necroptosis. Functionally relevant, the increased RIPK1/ERK-dependent TNF mRNA production of OTUB1-deficient hepatocytes may support an autocrine feedback loop by binding of TNF to TNFR1 and fostering RIPK1-dependent necroptosis and TNF production. This is in accordance with data demonstrating that the RIPK1/ RIPK3 necroptotic complex can associate with ERK to induce pro-inflammatory cytokines including TNF and that binding of TNF to the TNFR1 further amplifies necroptotic death. Collectively, OTUB1 is a hepatocyte-intrinsic factor inhibiting necroptosis upon bacterial infection and TNFinduced liver damage, respectively, thereby, protecting the host from lethal exacerbation of these inflammatory diseases.

Compliance with ethical standards
Conflict of interest The authors declare no competing interests.
Ethics statement This study does not involve human subjects. All animal experiments were performed in accordance to the German Animal Welfare Act and approved by local authorities (Landesverwaltungsamt Sachsen-Anhalt, Germany; file number: 42502-2-1440).
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.