TAK1 Binding Protein 2 Is Essential for Liver Protection from Stressors

The liver is the first line of defense from environmental stressors in that hepatocytes respond to and metabolize them. Hence, hepatocytes can be damaged by stressors. Protection against hepatic cell damage and cell death is important for liver function and homeostasis. TAK1 (MAP3K7) is an intermediate of stressors such as bacterial moieties–induced signal transduction pathways in several cell types. Tak1 deficiency has been reported to induce spontaneous hepatocellular carcinoma. However, the regulatory mechanism of TAK1 activity in liver stress response has not yet been defined. Here we report that activation of TAK1 through TAK1 binding protein 2 (TAB2) is required for liver protection from stressors. We found that a bacterial moiety, lipopolysaccharides (LPS), activated TAK1 in primary hepatocytes, which was diminished by deletion of TAB2. Mice having hepatocyte-specific deletion of the Tab2 gene exhibited only late-onset moderate liver lesions but were hypersensitive to LPS-induced liver injury. Furthermore, we show that a chemical stressor induced greatly exaggerated liver injury in hepatocyte-specific Tab2-deficient mice. These results demonstrate that TAB2 is a sensor of stress conditions in the liver and functions to protect the liver by activating the TAK1 pathway.


Introduction
The liver is responsible for the first pass metabolism of absorbed exogenous chemicals and the major organ of clearance of toxicants and pathogens, and liver cells are constantly exposed to those foreign substances including bacterial moieties and food-and clinical drug-derived chemicals. These stressors directly impact on liver parenchymal cells, and indirectly through cytokines from activated circulating and residential immune cells such as Kupffer cells [1,2]. Although the stressors and stressor-induced cytokines could activate cell death signaling in hepatocytes, they are normally resistant to those basal levels of stressors and functional under the homeostatic conditions. Excessive stressors derived from alcohol consumption and pathogens lead to liver injury. Furthermore, dysregulation of the protecting responses is likely to be associated with liver tumors in which cell death causes compensatory cell proliferation [3]. However, the mechanism by which hepatocytes respond to stressors and protect against cell death is still largely elusive.
Transcription factor NF-kB and mitogen-activated protein kinase, JNK, have been implicated in the regulatory mechanisms of hepatocyte death [4,5]. NF-kB protects hepatocytes from concanavalin A-induced cell death, and ablation of NF-kB induces spontaneous hepatocyte death and compensatory proliferation [4,6]. Activation of JNK is also closely associated with liver injury [5,7,8]. NF-kB transcriptionally activates several cell survival genes such as caspase inhibitors, cellular FLICE-like inhibitory protein (cFLIPL) and cellular inhibitor of apoptosis proteins (cIAPs), and anti-oxidant genes such as glutathione peroxidase and superoxide dismutase, which block proinflammatory cytokine TNF-induced cell death and oxidative damage [4,[9][10][11]. TAK1 (MAP3K7) is a member of the mitogen-activated protein kinase kinase kinase (MAP3K) family and an intermediate of stress-induced intracellular signaling pathways [12][13][14]. TAK1 activating stressors include proinflammatory cytokines TNF and IL-1 and Toll-like receptor ligands as well as chemical and physical stress, such as osmotic stress. TAK1 is the major upstream signaling mediator leading to activation of NF-kB and upregulation of antioxidant enzymes, thereby participating in cell survival in several tissues in vivo [15,16]. Ablation of TAK1 causes cell death associated with oxidative damage and tissue injury in the epidermis, the intestinal epithelium, and blood vessels [16][17][18]. The major cause of cell death in Tak1-deficient tissues in vivo is TNF-induced cell death, and deletion of TNF receptor 1 largely rescues cell death and tissue injury in these Tak1-deficient tissues [16,18]. Similar to these tissues, hepatocyte-specific deletion of the Tak1 gene causes cell death and liver injury, which further causes compensatory hepatocyte proliferation resulting in development of hepatocellular carcinoma [19,20]. Thus, TAK1-mediated cell signaling presumably of NF-kB and antioxidant gene regulation is indispensable for hepatocyte protection from stressor-induced cell death. However, the pathway through which TAK1 is activated in hepatocytes in vivo has not yet been determined.
TAK1 binding partner proteins, TAK1 binding proteins 1, 2 and 3 (TAB1, TAB2 and TAB3), mediate activation of TAK1 through two different mechanisms. TAB1 mediates TAK1 oligomerization to induce autophosphorylation of TAK1 within the activation loop in the protein kinase domain, which catalytically activates TAK1 [13,21,22]. TAB2 and TAB3 are related proteins, both of which bind to polyubiquitin chains and recruit TAK1 to the polyubiquitin chain protein complexes [23][24][25][26][27]. Polyubiquitin-mediated oligomerization of TAK1 proteins facilitates TAK1 protein kinase activation. The TAB1-dependent mechanism is required for osmotic stress-induced TAK1 activation [13] and the basal TAK1 activity in the epidermis in vivo [28]. The polyubiquitin-mediated mechanism activates TAK1 in response to TNF, IL-1 and TLR ligands. Although both TAB2 and TAB3 can mediate the polyubiquitin mechanism, TAB2 plays a predominant role in TAK1 activation in several cell types including endothelial and immune cells [17,23]. Double deficiency of Tab1 and Tab2 abolished TAK1 activity in the epidermis in vivo, which caused TNF-induced cell death and tissue damage resembling the phenotypes of TAK1 deficiency [28]. However, single deletion of either Tab1 or Tab2 does not cause any abnormalities in the epidermis, demonstrating that TAB1 and TAB2 are functionally redundant in the epidermis. Overall, TAB1 and TAB2 are major activators of TAK1 in several cell types, and contributions of TAB1-and TAB2-dependent mechanisms to TAK1 activity vary depending on cell types. In the current study, we investigated the role of TAB2 in hepatocytes and determined that TAB2 activates TAK1 in response to a TLR ligand lipopolysaccharide (LPS), which is essential for hepatocyte survival.

Immunoblotting
Cell extracts were prepared using a lysis buffer (20 mM HEPES (pH 7.4), 150 mM NaCl, 12.5 mM b-glycerophosphate, 100 nM calyculin A, 1.5 mM MgCl 2 , 2 mM EGTA, 10 mM NaF, 2 mM DTT, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 20 mM aprotinin, and 0.5% Triton X-100). Liver extracts were prepared in the abovementioned lysis buffer containing a protease inhibitor cocktail (G-Biosciences, St Louis, MO). Proteins were resolved on SDS-PAGE and transferred to Hybond-ECL or Hybond-P membranes (GE Healthcare). The membranes were immunoblotted with various antibodies, and the bound antibodies were visualized with horseradish peroxidase-conjugated antibodies against rabbit or mouse IgG using the ECL or ECL advance Western blotting detection kit (GE Healthcare).

Cell Death Assays
Terminal dUTP nick-end labeling (TUNEL) assay was performed on formalin-fixed paraffin sections using an apoptotic cell death detection kit (Promega) according to the manufacturer's instructions. Seven to ten immunofluorescent images per mouse were randomly photographed, and at least 2000 DAPI-stained cells per mouse were counted. Quantitative results were generated from the counted numbers in 3-4 mice from independent experiments. 50 mg proteins from liver or hepatocyte extracts were used for Caspase 3 assay (Promega).

Statistical Analysis
Results are expressed as either mean 6 standard error of the mean (SEM) or mean 6 standard deviation (SD). Differences between two groups were determined by two-tailed unpaired Student t test. Probability values of P are shown in figures. P#0.05 was considered statistically significant.

LPS Activates TAK1 through TAB2 in Primary Hepatocytes
Ablation of TAK1 in hepatocytes is known to spontaneously induce apoptosis in mouse liver (19,20). We hypothesized that TAK1 is activated through TAB1-and/or TAB2-dependent mechanisms by stressors and prevents stressor-induced apoptosis in the liver. To begin to test this hypothesis, we investigated whether stressors can activate TAK1 in hepatocytes. We found that LPS activates TAK1 in primary hepatocytes. TAK1 was activated and the activity peaked at 20-60 min following LPS challenge in primary hepatocytes (Fig. 1A). To determine whether TAB1 and/or TAB2 participate in LPS-induced TAK1 activation, we utilized primary hepatocytes having double deficiency of Tab1 and Tab2, and examined LPS-induced TAK1 activation. Tab1 and Tab2 double-deficient primary hepatocytes failed to activate TAK1 following LPS stimulation (Fig. 1B), demonstrating that TAB1-and/or TAB2-dependent mechanisms participate in the LPS signaling pathway. Furthermore, we found that Tab2 single deficient primary hepatocytes failed to activate TAK1 following LPS stimulation (Fig. 1C). The time course and the level of TAK1  Tab2-(C and D) deficient and control primary hepatocytes were prepared as described in Material and Methods, and stimulated with 1 mg/ml LPS. Cell lysates were analyzed by immunoblotting. In these Tak1-floxed mice, Cre-mediated recombination resulted in deletion of 38 amino acids of TAK1, and the truncated TAK1 (TAK1D) was expressed at a low level presumably due to reduced protein stability as indicated in A. Asterisks indicate non-specific bands. (D) Tab2deficient and control primary hepatocytes were stimulated with 1 mg/ml LPS for 24 h. Cell lysates were subjected to caspase 3 assay. Data are shown as means 6 SD, n = 3. doi:10.1371/journal.pone.0088037.g001 activation upon LPS challenge were slightly varied on each primary preparation, but activation of TAK1 was always detected in wild type hepatocytes. In contrast, Tak1-, Tab1/Tab2 double-or Tab2 single-deficient hepatocytes did not exhibit any increase of TAK1 activity in all preparations. These results suggest that a bacterial moiety LPS activates TAK1 through the TAB2dependent mechanism in hepatocytes. We also examined whether NF-kB activation, which is a well-known downstream target of TAK1, is altered by Tak1 or Tab2 single deletion in hepatocytes. We found that Tak1 deletion moderately reduced LPS-induced NF-kB activation in hepatocytes, suggesting that LPS activates NF-kB through multiple pathways including TAK1 pathway in hepatocytes (Fig. S1A). Similarly, Tab2 deletion partially reduced LPS-induced NF-kB activation (Fig. S1B). These results demonstrate that TAB2 is likely to function as an activator of TAK1 in response to LPS in hepatocytes, which may be important for prevention of liver injury. The LPS-TLR4 pathway has been reported to engage apoptosis and necrosis [36][37][38]. To begin to determine whether TAB2 is important for liver integrity, we asked whether deletion of Tab2 enhances LPS-induced cell death signaling. We examined caspase activity in LPS-treated Tab2deficient primary hepatocytes. Tab2-deficient hepatocytes possessed slightly higher activity of caspase 3 compared to littermate control hepatocytes under unchallenged conditions, and caspase 3 activity was moderately but significantly upregulated by LPS treatment in Tab2-deficient hepatocytes (Fig. 1D). However, we did not observe any detectable increase in cell death in both wild type and Tab2-deficient hepatocytes following LPS stimulation. The weak activation of caspase 3 may be insufficient to induce cell death in Tab2-deficient hepatocytes. Thus, TAB2-dependent TAK1 activation is at least preventive to LPS-induced caspase 3 activation, and LPS is not a strong inducer of cell death in cultured hepatocytes. However, LPS is a strong inducer of liver injury in vivo, which is mediated not only through its direct effect on hepatocytes but through indirect stress through inflammatory responses. We speculate that that TAB2-dependent activation of TAK1 may be important for prevention of LPS-induced liver injury in vivo.

TAB2 Knockout Mice Are Hypersensitive to LPS-Induced Liver Injury
To examine whether TAB2 plays a protective role in the liver in vivo, we generated TAB2 -/adult mice by using a ubiquitously expressing inducible Cre deleter strain, Rosa26.CreERT [32]. Tab2 deletion was induced by intraperitoneal injection of the CreERT activator, tamoxifen, and Tab2 mRNA and protein were greatly diminished at 3 weeks after tamoxifen injection ( Fig. 2A). Although germline deletion of Tab2 is known to cause embryonic lethality around embryonic days 13-14 due to liver degeneration [31], Tab2 deletion in adult mice did not cause overt liver abnormalities. This suggests that TAB2 plays an indispensable role during embryogenesis which is not required for adult liver homeostasis at least under unchallenged conditions. We asked whether Tab2 deletion sensitizes the liver to LPS-induced injury in adult mice. Tab2 deletion alone did not induce cell death in the liver, but LPS challenge greatly upregulated caspase activity in Tab2-deficient liver but not in the control within 6 h (Figs. 2B and 2C). TUNELpositive cells with LPS challenge were highly increased by Tab2deficiency (Figs. 2D and 2E). JNK activity was upregulated in Tab2-deficient liver (Fig. 2F), which is known to be associated with liver damage [8] and has been also observed in Tak1-deficient liver in earlier studies [19,20]. We note that, because Tab2-deficient liver was greatly damaged within 6 h post-LPS injection, we were not able to analyze the consequences beyond 6 h. These results demonstrate that TAB2 is indispensable in the protection of the liver from LPS-induced cell death.

Hepatocyte-Specific Deletion of Tab2 Induces Late Onset Liver Damage and Sensitizes Mice to LPS-Induced Liver Injury
To examine whether hepatocyte-derived TAB2 is responsible for the protection against stress-induced liver injury, we generated hepatocyte-specific Tab2-deficient (Tab2 HepKO ) mice by using hepatocyte-specific deleter, alb.Cre, transgenic mice [33]. The Tab2 expression in the liver but not lung, brain or heart extract was greatly diminished in the Tab2 HepKO mice (Fig. 3A). We also generated Tak1 HepKO mice, which are known to develop hepatocellular carcinoma [19,20]. Tak1 HepKO mice exhibited signs of highly increased liver damage including increased levels of serum aminotransferase (ALT), expression of chemokine (C-X-C motif) ligand 2 (Cxcl2), chemokine (C-C motif) ligand 2 (Ccl2), fibrosis markers, collagen, type I, a1 (Col1a1) and tissue inhibitor of metalloproteinase 1 (Timp1) as well as hepatocellular carcinoma marker, H19 gene within one month even in the absence of any exogenous stressors (Fig. S2), consistent with earlier studies. In contrast, Tab2 HepKO mice developed normally and did not exhibit increased ALT and only marginal or moderate increases in inflammatory chemokine expression in the absence of exogenous stressors as shown in later in Fig. 4. Tab2 HepKO liver was histologically indistinguishable from control littermate liver at least at several months of age (Fig. 3B). The lobular architecture was clear in both control and Tab2 HepKO liver, although small foci of immune cell infiltration were occasionally observed in all genotypes (Fig. 3B). Aged Tab2 HepKO liver exhibited moderately disorganized lobular architectures with relatively large infiltration foci as well as cytoplasmic eosinophilic materials (dashed line) around 10-12 months old of age, while Tak1 HepKO showed severe structural damage including unclear lobular architecture in addition to immune cell infiltrations and profound increase of eosinophilic materials (Fig. 3C). Aged Tab2 HepKO liver exhibited signs of fibrosis including increased Sirius Red staining, although it is much milder than that in Tak1 HepKO (Fig. 3D). Aged Tab2 HepKO liver expressed increased fibrosis marker genes, Col1a1 and Timp1 (Fig. 3E). These suggest that TAK1 basal activity but not TAB2-dependent TAK1 activation is required for liver integrity at least for several months, and that TAB2 may be important for long-term liver integrity by preventing accumulation of stress-induced liver lesions. To test acute stress-induced liver injury, Tab2 HepKO mice were treated with LPS for 24 h. Caspase activity was upregulated in Tab2 HepKO liver compared to control mice at 24 h post LPS injection (Fig. 3F); however, the levels of caspase activation were lower compared with those in the ubiquitous Tab2-deficient liver described above (see Fig. 2B). This suggests that Tab2 deficiency highly elevates liver injury also through cell types other than hepatocytes, but that hepatocytederived TAB2 is still at least in part responsible for liver protection under the LPS-induced stress condition.

Hepatocyte-Specific Deletion of Tab2 Sensitizes Mice to Diethylnitrosamine (DEN)-Induced Liver Injury
To further determine the hepatocyte-specific role of TAB2 in stress protection, we utilized a strong liver carcinogen diethylnitrosamine (DEN), which is activated by cytochrome P450 within hepatocytes and induces oxidative stress [39,40]. DEN-induced stress causes hepatocellular carcinoma in the long term, and acute liver injury in the short term [41]. Tab2 HepKO and control mice were treated with DEN for 24 h. We observed highly increased TUNEL-positive cells in Tab2 HepKO liver (Fig. 4A) and elevation of serum ALT (Fig. 4B), indicating profound liver injury. Furthermore, the expression of inflammatory genes, Cxcl2 and Ccl2, was elevated in Tab2 HepKO liver (Fig. 4C). These results show that hepatocyte-derived TAB2 is required for liver protection from DEN-derived stress. As described earlier, Tab2 deletion alone in primary hepatocytes diminished TAK1 activation upon LPS challenge (Fig. 1C). However, other stressors may also utilize the TAB1-dependent pathway to activate the TAK1-cell protection pathway. We then examined whether Tab1-deficiency sensitizes the liver to DEN. To this end, we generated hepatocyte-specific Tab1-deficient (Tab1 HepKO ) mice using the same system as Tab2 HepKO . Tab1 HepKO mice did not exhibit any overt abnormality. Tab1 HepKO mice were treated with DEN for 24 h, but did not exhibit statistically significant increases in ALT or in the expression levels of Cxcl2 and Ccl2 (Fig. 4E). Thus, TAB1 minimally, if any, contributes to liver protection, whereas TAB2 is indispensable for the protection against stress-induced liver injury.

Discussion
Tak1 deficiency in hepatocytes causes spontaneous liver injury which further engages hepatocellular carcinoma development [19,20]. A previous study in Tak1 deficient liver used a mouse model that expresses a truncated version of TAK1, which has 38 amino acids deleted around the ATP binding site of the protein kinase domain but the other domains intact [19,30]. Thus, TAK1 kinase activity but not TAK1 protein itself is important for hepatocyte survival. Accordingly, it is reasonable to assume that TAK1 is active to some extent in the liver under homeostatic conditions, and constitutively upregulates pro-survival factors such  . Hepatocyte-specific deletion of Tab2 induces late-onset fibrosis and sensitizes the liver to LPS-induced injury. (A) Tab2 expression in the liver, lung, brain, and heart was determined by real-time PCR (liver) and immunoblotting (liver, lung, brain and heart) in hepatocytespecific Tab2 as cIAP and c-Jun [15,16], and anti-oxidants such as Nrf2 and NAD(P)H:quinone oxidoreductase 1 [18,42]. Earlier studies have established that this TAK1 pro-survival signaling pathway is indispensable for liver protection [19,20]. In the current study, we found that a TAK1 activator protein, TAB2, is indispensable for stressor-induced TAK1 activation in hepatocytes, and that TAB2 is required for prevention of hepatocyte death in response to stressor challenge in vivo. We propose that TAK1 activity is required for prevention of hepatocyte cell death under basal conditions, and that, under stressed conditions, TAK1 is activated beyond the basal level through TAB2, which is required for liver protection against stressors as summarized in Fig. 5.
As shown in Fig. 3, hepatocyte-specific Tab2 deletion did not cause any abnormality at least for several months, whereas Tak1 deficiency induced profound liver injury and even tumors within 4 weeks of age (Fig. S2) in agreement with the earlier studies [19,20].
Thus, TAB2 is not required for the basal TAK1 activity to protect hepatocytes under the unstressed condition. TAB2 is known to play a major role in cytokine and Toll-like receptor (TLR)-induced TAK1 activation through a polyubiquitin-dependent mechanism [24]. Hence, the cytokine-or TLR ligand-induced pathway is not essential for basal hepatic TAK1 activity. How is TAK1 activated in the absence of TAB2? Another TAK1 binding partner, TAB1, can activate TAK1 through promoting oligomerization of TAK1 [13,21]. TAB1 has been identified to be essential for maintenance of TAK1 activity in the epidermis [28]. We show that Tab1 deficiency greatly reduces TAK1 activity in the epidermis. TAB1 may be responsible for the basal TAK1 activity in the liver, too. However, Tab1 deletion alone does not cause any abnormalities in the liver. Thus, Tab1 is not essential as is Tab2 for hepatocyte survival in vivo. TAB3 can mediate polyubiquitin-dependent activation of TAK1, but Tab3 deletion alone does not cause any abnormalities during embryogenesis or in adult mice [23]. Collectively, it is likely that TAB1 and TAB2 and possibly TAB3 redundantly function to activate TAK1 under the basal conditions. Thus, single deletion of Tab1, Tab2 or Tab3 does not cause spontaneous liver injury.
Blockade of TAK1 signaling is a powerful method to kill cells in vivo. However, since the TAK1-mediated cell survival pathway is indispensable for liver integrity, TAK1 is not an ideal target to remove undesired hepatocytes such as tumor cells. Tak1 deletion even causes hepatocellular carcinoma [19,20]. Tab2 deficiency alone does not cause any abnormality in unstressed liver at least for several months in the mouse model. Therefore, inhibition of TAB2 would not cause acute liver injury in normal liver. Our results suggest that TAB2-dependent activation of TAK1 beyond the basal level is required for hepatocyte survival only when hepatocytes are under stressed conditions. Thus, we anticipate that inhibition of TAB2 might selectively kill stressed hepatocytes such as hepatocellular carcinoma cells. TAB2 could be a new target to kill stressed or damaged hepatocytes without affecting unstressed cells. Figure S1 LPS-induced NF-kB activation is reduced by Tak1 or Tab2 deficiency in primary hepatocytes.