• Open Access

Tumor progression locus 2/Cot is required for activation of extracellular regulated kinase in liver injury and toll-like receptor–induced TIMP-1 gene transcription in hepatic stellate cells in mice


  • Potential conflict of interest: Nothing to report.

  • The present study was supported by the Wellcome Trust (grant no.: WT086755MA) and a European Commission FP7 program grant “INFLA-CARE” (EC contract no.: 223151; http://inflacare.imbb.forth.gr/) (both to D.A.M.). The authors thank Dr. Steven C Ley (MRC National Institute for Medical Research, London, UK) for the generous gift of Tpl2−/− mice.


Toll-like receptors (TLRs) function as key regulators of liver fibrosis and are able to modulate the fibrogenic actions of nonparenchymal liver cells. The fibrogenic signaling events downstream of TLRs on Kupffer cells (KCs) and hepatic stellate cells (HSCs) are poorly defined. Here, we describe the MAP3K tumor progression locus 2 (Tpl2) as being important for the activation of extracellular regulated kinase (ERK) signaling in KCs and HSCs responding to stimulation of TLR4 and TLR9. KCs lacking Tpl2 display defects with TLR induction of cytokines interleukin (IL)-1β, IL-10, and IL-23. tpl2−/− HSCs were unable to increase expression of fibrogenic genes IL-1β and tissue inhibitor of metalloproteinase 1 (TIMP-1), with the latter being the result of defective stimulation of TIMP-1 promoter activity by TLRs. To determine the in vivo relevance of Tpl2 signaling in liver fibrosis, we compared the fibrogenic responses of wild-type (WT) and tpl2−/− mice in three distinct models of chronic liver injury. In the carbon tetrachloride and methionine-choline–deficient diet models, we observed a significant reduction in fibrosis in mice lacking Tpl2, compared to WT controls. However, in the bile duct ligation model, there was no effect of tpl2 deletion, which may reflect a lesser role for HSCs in wounding response to biliary injury. Conclusion: We conclude that Tpl2 is an important signal transducer for TLR activation of gene expression in KCs and HSCs by the ERK pathway and that suppression of its catalytic activity may be a route toward suppressing fibrosis caused by hepatocellular injuries. (HEPATOLOGY 2013)

Liver fibrosis is a pathological process of progressive deposition of collagen-rich extracellular matrix (ECM) scar tissue that is associated with chronic liver inflammation arising from injuries caused by alcohol, hepatic accumulation of lipids, viral infections, and autoimmune and hereditary diseases.1, 2 Currently, fibrogenesis is considered to be a dynamic process, with equal potential for regression as well as progression.3, 4 This relatively recent discovery is encouraging because it suggests that established fibrosis is a tractable therapeutic target if antifibrotics can be developed that stimulate ECM degradation. The major cell type responsible for liver fibrogenesis is the hepatic stellate cell (HSC)-derived myofibroblast, which promotes ECM deposition by synthesis of fibril-forming collagens.5 In addition, myofibroblasts are proposed to actively suppress ECM degradation by secreting tissue inhibitor of metalloproteinase 1 (TIMP-1), which functions as a broad specificity inhibitor of collagenases.6 An understanding of the mechanisms that control the fibrogenic actions of the myofibroblast and, in particular, the signaling pathways that dictate the balance between ECM deposition and turnover should therefore facilitate the design of antifibrotics for chronic liver disease (CLD).

Recent work suggests the importance of signaling cross-talk between the innate immune and fibrogenic systems of the liver. The liver is particularly vulnerable to exposure to bacterial components as a result of it being the first organ that encounters venous blood flowing from the gut through the portal vein.7, 8 In a healthy individual, the degree to which the liver is exposed to bacterial products is limited by the intestinal epithelial barrier and tolerance to the minor quantities that translocate across this barrier.9, 10 Increased permeability of the intestinal epithelium is a feature of many liver diseases, and in some cases, such as nonalcoholic fatty liver disease, is coupled with alterations in the composition of the gut microbiota. As a consequence, the diseased liver is exposed to abnormal quantities of bacterial products and with compositions for which tolerance may not have developed.7, 11 The major parenchymal cells driving fibrogenesis, Kupffer cells (KCs) and HSCs, are responsive to bacterial products through surface-expressed Toll-like receptors (TLRs). In addition, TLRs on these cells provide mechanisms for innate immune responses to alarmins released by dying and damaged liver cells. Clinical studies have reported elevated bacterial-derived lipopolysaccharide (LPS) in the circulation of patients with CLD.11 Moreover, recent experimental studies by Seki et al. indicate an essential role for the LPS receptor, TLR4, and its downstream intracellular signaling molecule, myeloid differentiation primary response gene 88 (Myd88), in liver fibrosis.12 Furthermore, single-nucleotide polymorphisms (SNPs) in the TLR4 gene are predictive for the risk of cirrhosis in hepatitis C virus (HCV)-infected patients13, 14 and appear to modulate the fibrogenic activities of HSC-derived myofibroblasts.


Ab, antibody; ATP, adenosine triphosphate; BDL, bile duct ligation; BMDMs, bone-marrow–derived macrophages; CLD, chronic liver disease; ECM, extracellular matrix; ELISA, enzyme-linked immunosorbent assay; ERK, extracellular regulated kinase; HCV, hepatitis C virus; H&E, hematoxylin and eosin; HSCs, hepatic stellate cells; IHC, immunohistochemistry; IKK, IκB kinase; IκB-α, inhibitor of NF-κB alpha; IL, interleukin; JNK, Jun N-terminal kinase; KCs, kupffer cells; LPS, lipopolysaccharide; MKK; mitogen-activated protein kinase kinase; MCD, methionine and choline deficient; mRNA, messenger RNA; Myd88, myeloid differentiation primary response gene 88; NASH, nonalcoholic steatohepatitis; NF-κB, nuclear factor kappaB; PCR, polymerase chain reaction; PDGF, platelet-derived growth factor; p-ERK, phosphorylated ERK; p-JNK, phosphorylated JNK; SEM, standard error of the mean; α-SMA, alpha smooth muscle actin; SNPs, single-nucleotide polymorphisms; TACE, TNF-α-converting enzyme; TGF-β; transforming growth factor beta; TIMP-1, tissue inhibitor of metalloproteinase 1; TLRs, Toll-like receptors; TNF-α, tumor necrosis factor alpha; Tpl2, tumor progression locus 2; WT, wild type.

Despite this knowledge of the interplay between innate immune signaling and fibrogenesis, we are lacking details of how TLR signaling affects the fibrogenic process, and we also currently have no molecular targets in the TLR-signaling pathways that could be targeted in a safe way for therapeutic purpose. Tumor progression locus 2 (Tpl2; but also known as Cot or MAP3K8) is a serine-threonine kinase with an important role in TLR as well as tumor necrosis factor (TNF), interleukin (IL)-1, and G-protein-coupled receptor-mediated signaling.15 Activation of Tpl2 requires IkappaB kinase (IKK)-β-catalyzed phosphorylation of the p105 nuclear factor kappaB (NF-κB) protein, which is complexed with Tpl2 in its inactive state. Subsequent ubiquitination and proteasome-mediated processing of p105 to its shorter p50 form releases Tpl2, which then becomes catalytically active before its degradation by the proteasome, which terminates Tpl2 signaling.16 The major biochemical function for Tpl2 is activation of extracellular regulated kinase (ERK) by direct phosphorylation of mitogen-activated protein kinase kinase (MKK)1 and MKK2, the ERK kinases.17 tpl2−/− mice provided genetic evidence for a physiological role in innate immunity, these animals being resistant to endotoxin shock and tpl2−/− macrophages displaying defective LPS responsiveness resulting from loss of ERK activation.18 Under LPS stimulation, production of the proinflammatory cytokine, TNF, is attenuated in tpl2−/− macrophages. Tpl2 regulates TNF synthesis by post-transcriptional mechanisms involving control of nucleocytoplasmic transport of the TNF messenger RNA (mRNA) and processing of pre-TNF to the secreted form of the cytokine through the ERK-mediated phosphorylation of the TNF-α-converting enzyme.18, 19 In addition, Tpl2 regulates IL-1β and IL-23 production through transcriptional mechanisms.20, 21 Given the importance of Tpl2 as a downstream mediator of LPS/TLR4 signaling and evidence in the literature for functions of ERK in fibrosis,22 we were interested to determine whether Tpl2 regulates the expression of genes regulating fibrogenic processes in the liver.

Materials and Methods

Experimental Models of Liver Fibrosis.

Wild-type (WT) and tpl2−/− male mice (25-30 g) were intraperitoneally injected once- (acute) or twice-weekly for 8 weeks (chronic) with CCl4 at 2 μL/g body weight (CCl4/olive oil at 1:1 [vol:vol] [acute] and 1:3 [vol:vol] [chronic]). Bile duct ligation (BDL) was performed as previously described.23 Mice were allowed to develop cholestatic disease and fibrosis over 15 days. Mice were fed methionine-choline–deficient (MCD) or control diet (Research Diets, New Brunswick, NJ) for 6 weeks. Animal care and procedures were approved by the Newcastle Ethical Review Committee, and experiments were performed under a UK Home Office project license.

Isolation, culture, and treatment of Mouse Bone-Marrow–Derived Macrophages, Hepatocytes, KCs, and HSCs.

Details are described in the Supporting Materials and Methods.

RNA Isolation and Quantitative RT-PCR Analysis.

Total RNA was extracted from liver tissue or primary cells using the TRI reagent (Sigma-Aldrich, St. Louis, MO). Reverse transcription was performed as previously described.23 Real-time polymerase chain reaction (PCR) was performed with SYBR-Green JumpStart Taq ReadyMix (Sigma-Aldrich), following the manufacturer's instructions. Primer sequences are included in Supporting Table 1. PCR reactions were normalised to the internal control and relative level of transcriptional difference calculated using the equation [1/(2A)]×100.

Sodium Dodecyl Sulfate/Polyacrylamide Gel Electrophoresis and Immunoblotting.

Details and antibodies (Abs) tested are described in the Supporting Materials.

IL-1β Detection by Enzyme-Linked Immunosorbent Assay.

For the quantitative determination of mouse IL-1β concentrations in culture supernatant, the Quantikine Mouse IL-1β enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN) was used, following the manufacturers instructions.

Immunohistochemistry and Image Analysis.

Staining was performed on formalin-fixed mouse liver sections. We performed hematoxylin and eosin (H&E), Sirius Red, alpha smooth muscle actin (α-SMA), and neutrophil marker NIMP staining as previously described.23 Details for Tpl2 staining and dual immunofluorescence staining are described in the Supporting Materials and methods.

Plasmid DNAs and Reporter Assays.

Details are described in the Supporting Materials and methods.

Statistical Analysis.

Statistical analysis was performed with GraphPad Prism version 5.00 software (GraphPad Software Inc., San Diego, CA). Experiments were performed at least twice in triplicates. Data are expressed as means ± standard error of the mean (SEM). Data were compared among groups using a two-tailed unpaired Student t test or one-way analysis of variance with Tukey's post-hoc or Bonferroni's test for individual subgroup comparison, if required. Statistically significant data are represented in figures with p values of <0.05, <0.01, and <0.001, respectively.


Hepatic Tpl2 Expression Is Elevated in Fibrotic Disease and Restricted to KCs and HSCs.

Before beginning experimental studies, we employed immunohistochemistry (IHC) to determine hepatic Tpl2 expression. Analysis of mouse liver sections showed that in controls, healthy liver Tpl2 expression is restricted to monocytic cells; however, expression was substantially increased in diseased liver, with staining localized to fibrotic tissue (Fig. 1A). Dual immunostaining with anti-Tpl2 and anti-α-SMA revealed colocalization, indicating expression in activated HSCs (Supporting Fig. 1). To further define the cellular expression of Tpl2, we determined Tpl2 protein expression by immunoblotting using protein extracts from isolated primary mouse hepatocytes, KCs, and culture-activated HSCs. Both KC and myofibroblastic HSCs expressed 58- (Tpl2L) and 52-kDa (Tpl2S) isoforms of Tpl2 (Fig. 1B). Hepatocytes lacked expression of either isoform, which confirms the absence of detection in these cells by IHC (Fig. 1B). However, hepatocytes did express an unknown higher molecular-weight protein that has immunoreactivity with the Tpl2 Ab. From these investigations, we propose that Tpl2 isoforms are expressed in KCs and HSCs, but most likely not in hepatocytes.

Figure 1.

Tpl2 kinase expression in liver. (A) Representative photomicrographs of liver sections (at ×100 magnification) showing Tpl2 IHC staining in 8-week CCl4-injured mouse liver BDL injured mouse liver, and healthy control liver (olive oil control). Arrows denote positive staining in macrophages and activated HSCs. (B) Whole-cell protein extract (30-μg) from primary mouse hepatocytes, KCs, or activated HSCs (day 7) were separated by SDS-PAGE and immunoblotted for Tpl2 and GAPDH. SDS-PAGE, sodium dodecyl sulfate/polyacrylamide gel electrophoresis; GADPH, glyceraldehyde-3-phosphate dehydrogenase.

Tpl2 Is Required for TLR-Induced ERK and Cytokine Responses in KCs.

Next, we were interested to determine the requirement for Tpl2 in KC and HSC for responsiveness to LPS. First, we confirmed the role of Tpl2 in bone bone-marrow–derived macrophages (BMDMs) by comparing LPS induction of ERK phosphorylation (p-ERK) and cytokines between WT and tpl2−/− cells. As expected, tpl2−/− BMDMs lacked expression of Tpl2 proteins and were unable to activate ERK in response to LPS, whereas processing of NF-κB1 p105 to p50, phosphorylation of p65, degradation of inhibitor of NF-κB alpha (IκB-α), and phosphorylation of Jun N-terminal kinase (p-JNK) were similar between WT and knockout cells (Fig. 2A). We were unable to confirm an effect of Tpl2 depletion on LPS induction of TNF-α and also observed no effect on induction of IL-6; however, tpl2−/− BMDMs displayed blunted LPS responses for IL-1β and IL-10 (Fig. 2B). Because Tpl2 has not yet been functionally investigated in KCs, we next isolated WT and tpl2−/− KCs and compared their responses to LPS treatment. Under unstimulated conditions, WT and tpl2−/− KCs expressed low levels of active p-ERK, p-JNK, and p-p38 (Fig. 3A). All three kinases in WT KCs underwent rapid phosphorylation in response to LPS treatment, with maximal responses observed after 30 minutes, and we also observed increased p50 expression, phosphorylation of p65, and degradation of IκB-α, all of which are indicative of expected NF-κB activation. LPS-induced p-ERK was defective in tpl2−/− KCs, with no detectable increase in p-ERK at any treatment time point, suggesting that the kinase is essential for LPS-induced ERK in KCs. Additionally, we observed a modest reduction in LPS-induced p-p38 in tpl2−/− KC, but found no effect on activation of JNK or activation of NF-κB. Because Tpl2 is reported to be responsive to several TLRs, we were interested to determine whether KCs also require Tpl2 for activating ERK after stimulation of the TLR9 receptor, given previous reports for functions of TLR9 in hepatic injury and fibrosis.24-26 Engagement of TLR9 with CpG resulted in activation of ERK, with a maximal response at 30 minutes; by contrast, no activation of ERK was found with a control ligand (−CpG) or in tpl2−/− KCs (Fig. 3B). We conclude that Tpl2 functions as an important regulator of TLR-induced ERK activation in KCs. We observed no differences in CpG-induced activation of NF-κB, p38, or JNK (Supporting Fig. 2A). LPS induction of TNF-α and IL-6 were modestly blunted in tpl2−/− KCs, compared to WT, whereas activation of IL-1β, IL-10, and IL-23 expression was not observed in Tpl2-deficient cells (Fig. 3C). By contrast, IL-12 was significantly induced in tpl2−/− KC, compared to WT, in agreement with previous observations made in other types of macrophages under TLR stimulation.27 Therefore, Tpl2 is essential for KC expression of key inflammatory cytokines; however, fibrogenic growth factors platelet-derived growth factor (PDGF)B and transforming growth factor beta (TGF-β)1 were unaffected by Tpl2 deletion (Supporting Fig. 2B).

Figure 2.

MAPK signaling and cytokine responses in WT and tpl2−/− BMDMs stimulated with LPS. (A) Mouse WT and tpl2−/− BMDMs were serum-starved for 12 hours and LPS treated (100 ng/mL) for the indicated time points, and the expression of Tpl2, NF-κB1 p105, and p50 subunits, IκB-α and phosphorylation of p65 (Ser536), ERK1/2, p38, and JNK was determined by western blotting. (B) Cytokine expression profile in BMDMs after LPS stimulation (100 ng/mL). Data represent mean ± SEM. **P < 0.01; ***P < 0.001. mRNA levels are expressed in arbitrary units (AU).

Figure 3.

Tpl2 is required for TLR-induced ERK phosphorylation and cytokine responses in KCs. (A) WT and tpl2−/− KCs were serum starved for 12 hours and treated with 100 ng/mL of LPS and (B) CpG-ODN or non-CpG-ODN (−CpG) for the indicated time points, and Tpl2 and p50 expression, IκB-α and phosphorylation of p65 (Ser536), ERK1/2, p38, and JNK was determined by western blotting. (C) TNF-α, IL-6, IL-1β, IL-10, IL-12, and IL-23 mRNA expression in KCs stimulated with 100 ng/mL of LPS for 30 minutes and 1, 3, and 6 hours. Data represent mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.

Tpl2 Is Essential for TLR4 and TLR9 Activation of ERK in HSCs.

To investigate a function for Tpl2 in activated HSCs, we examined LPS-induced signaling events in cultured primary WT and tpl2−/− mouse HSCs. LPS treatment was not associated with NF-κB p105/p50 subunit processing in WT or tpl2−/− HSC (Fig. 4A); however, in their basal state, HSCs already expressed high levels of p50, supporting our previous reports of constitutive active NF-κB in HSC-derived myofibroblasts.28 Phosphorylation of ERK, p38, and JNK was strongly stimulated by LPS in WT HSCs, although with later kinetics than observed for KCs, with maximal effects being observed at 60 minutes post-treatment for all three kinases. tpl2−/− HSC failed to induce p-ERK in response to LPS; we also observed a slight delay in activation of p38, but no effect on p-JNK. As observed with KCs, there was a complete failure of LPS induction of IL-1β in Tpl2-deficient HSCs and a modest blunting of the IL-6 response (Fig. 4B). To determine whether activated HSCs secrete functional IL-1β, cells from WT mice were treated with LPS for 8 hours, followed by adenosine triphosphate (ATP) treatment to activate caspase-1. Supernatants from WT HSCs were positive for functional IL-1β, as determined by specific ELISA; however, by contrast, almost undetectable levels were found in supernatants of tpl2−/− HSCs (Fig. 4C). It has been reported that the TGF-β1 pseudoreceptor, Bambi, is suppressed by LPS in HSCs, enabling their responsiveness to TGF-β1.12 Bambi expression was diminished in response to LPS in a similar manner between WT and tpl2−/− HSCs (Fig. 4D). tpl2−/− HSCs were also unable to phosphorylate ERK in response to CpG engagement of TLR9 (Fig. 4E); however, this did not reflect an inherent defect in ERK activation, because the kinase was phosphorylated in response to treatment with PDGF (Fig. 4F). These findings support a role for Tpl2 in HSCs as a regulator of ERK activation and cytokine expression downstream of TLRs and, together with the data from KCs, were suggestive of a profibrogenic function for Tpl2.

Figure 4.

Tpl2 is essential for LPS, but not PDGF, activation of ERK in activated HSCs. (A) WT and tpl2−/− HSCs were serum starved for 12 hours and treated with LPS (100 ng/mL) for the indicated time points, and protein expression levels of Tpl2, p105, p50, IκB-α, and phosphorylation of p65 (Ser536), ERK1/2, p38, and JNK was detected by western blotting. (B) IL-1β and IL-6 mRNA levels were measured by quantitative PCR in WT and tpl2−/− HSCs stimulated with LPS (100 ng/mL) for 1, 3, and 6 hours. Data represent mean ± SEM. **P < 0.01; ***P < 0.001. (C) To convert active IL-1β from proIL-1β, HSCs were treated with 2 mmol/L of ATP for 30 minutes after 8 hours of incubation with LPS (100 ng/mL), and secreted IL-1β levels in supernatant were measured by ELISA. (D) Bambi mRNA expression was measured in WT and tpl2−/− HSCs stimulated with LPS (100 ng/mL) for 6 hours. (E) WT and tpl2−/− HSCs were treated with CpG-ODN (5 μg/mL) or (F) PDGF-BB (10 ng/mL), and phosphorylation of ERK1/2 and Akt was measured by western blotting.

Deletion of Tpl2 Modulates Fibrogenesis In Vivo.

It has previously been reported that TLR4 signaling is important for a fibrogenic response to liver injury by a variety of distinct mechanisms, including hepatotoxic (e.g., CCl4) and cholestatic (e.g., BDL) injuries.12 Because Tpl2 is required for the activation of ERK downstream of TLR4 in both KCs and HSCs, we therefore determined whether disruption of TLR4/Tpl2/ERK signaling in tpl2−/− mouse liver had an effect on the development of fibrosis. First, we performed chronic CCl4 injury and determined differences for fibrosis at peak injury (day 1 after the final CCl4 administration) and in postinjury spontaneous recovery (days 3 and 7). Alanine aminotransferase and aspartate aminotransferase measurements revealed no differences in injury between WT and tpl2−/− mice (Supporting Fig. 3). However, staining for neutrophils (NIMP and neutrophil elastase) showed substantially reduced recruitment of neutrophils in tpl2−/− livers, suggestive of a depressed innate immune response as might be predicted from disruption of TLR signaling (Supporting Fig. 4). Fibrosis was determined by Sirius Red (Fig. 5A) and α-SMA (Fig. 5B) staining. As expected, staining for both was high at peak injury (day 1) and underwent stepwise reductions as the liver recovered from days 3 to 7, this being with similar kinetics between WT and tpl2−/− mice. At peak injury, both Sirius Red and α-SMA were modestly reduced in tpl2−/− livers, compared to WT (Fig. 5C,D, respectively). This influence of Tpl2 deletion on fibrosis was confirmed by quantification of mRNAs for Collagen I, α-SMA, and TIMP-1, all of which were expressed at reduced levels in tpl2−/− livers (Fig. 6A). Analysis of fibrogenic cytokines/growth factors revealed defective induction of IL-1β in tpl2−/− livers (Fig. 6A), but no differences for expression of TGF-β1 and connective tissue growth factor (Supporting Fig. 5). Analysis of whole-liver p-ERK revealed substantially raised levels in CCl4-injured versus olive oil control WT mice; by contrast, CCl4 injury only modestly increased levels of p-ERK in tpl2−/− livers (Fig. 6B). Therefore, Tpl2 is critical for in vivo activation of hepatic ERK signaling and for an optimal fibrogenic response in the CCl4 liver injury model.

Figure 5.

Attenuated hepatic fibrogenesis in chronic CCl4-injured tpl2−/− mice. (A) Sirius Red– and (B) α-SMA-stained liver sections from WT and tpl2−/− mice treated with CCl4 for 8 weeks and killed 1, 3, or 7 days after the last CCl4 injection. Between 10 and 15 high-power field (×100 magnification) images were taken of each Sirius Red (C) or α-SMA (D) section, and images were analyzed. Quantification of fibrosis was demonstrated as a function of mean percentage of field staining. Results displayed are mean ± SEM for each group of mice. Statistical analysis was performed using the two-tailed unpaired t test. *P < 0.05 versus values in WT mice.

Figure 6.

Quantitative real-time PCR analysis of the expression of fibrosis-related genes in livers of WT and tpl2−/− mice after chronic CCl4 treatment. (A) Expression of α-SMA, α1(I)procollagen, TIMP-1, and IL-1β mRNA. mRNA levels are expressed in arbitrary units (AU). *P < 0.05 versus values in WT mice. (B) Western blotting analysis of Tpl2 and p-ERK expression in total liver extracts from WT and tpl2−/− mice sacrificed 1 day after the last CCl4 injection.

To determine whether there is a general requirement for Tpl2 in liver fibrogenesis, we employed two additional models: BDL as a model of cholestasis and the MCD diet as a model of steatosis. In the BDL model, we observed a modest reduction in numbers of α-SMA+ cells in Tpl2-deficient livers, but no difference in severity of fibrosis, as determined by Sirius Red staining (Fig. 7A,B). Gene-expression analysis showed that absence of Tpl2 reduced levels of transcripts for α-SMA and TIMP-1, but had no effect on expression of Collagen I (Fig. 7C). MCD diet caused a 3-fold increase in Sirius Red–stained fibrotic matrix, compared to animals on the control diet, indicating the expected fibrogenic response (Fig. 7D,E). Comparison of WT and tpl2−/− mice revealed a significant attenuation of fibrosis in the latter. Western blotting revealed defective MCD induction of TIMP-1 protein expression in tpl2−/− livers (Fig. 7F). However, no differences were observed for Collagen I or α-SMA expression between WT and tpl2−/− (Supporting Fig. 6). We conclude that Tpl2 contributes toward fibrogenic response to chronic injury in the CCl4 and MCD models, but has a less-prominent role in cholestatic injury.

Figure 7.

tpl2−/− mice showed no differences in hepatic fibrogenesis in a BDL fibrosis model, but showed attenuated hepatic fibrogenesis in an MCD model. (A) Representative Sirius Red– and α-SMA-stained liver sections of WT and tpl2−/− BDL mice. (B) Fibrosis densitometry and α-SMA quantification are shown in graphs. (C) Quantitative real-time PCR analysis of Collagen I, α-SMA, and TIMP-1 expression in WT and tpl2−/− BDL mice. (D) Representative H&E- and Sirius Red–stained liver sections of WT and tpl2−/− mice fed with a control diet or MCD diet for 6 weeks. (E) Fibrosis densitometry of Sirius Red–stained sections (**P<0.01). (F) Western blotting analysis of TIMP-1 expression in total liver extracts from WT and tpl2−/− mice fed with a control diet or MCD diet.

Tpl2 Is Required for LPS Induction of p-ERK and TIMP-1 Gene Transcription.

Hepatic injury with a single administration of CCl4 is a useful model for studying inflammation-driven HSC activation in vivo. Using this model, we were able to confirm that induction of TIMP-1 is defective in the absence of Tpl2 signaling, with loss of expected increases in transcript expression observed at 48 and 72 hours postinjury (Supporting Fig. 7). We also noted a modest, but significant, blunting of the induction of IL-1β; however, no differences between WT and tpl2−/− were observed for induction of Collagen I, α-SMA, IL-6, TNF-α, IL-10, or IL-12. These data confirm that Tpl2 is not required for HSC activation per se, but that it is necessary for induction of TIMP-1, an important regulator of ECM turnover and fibrosis progression.6 Next, we investigated whether Tpl2 is required for LPS-induced regulation of TIMP-1 in culture-activated HSCs. LPS treatment of WT HSCs resulted in 5- and 6-fold induction of TIMP-1 transcript at 3 and 6 hours, respectively; by contrast, a modest 2-fold induction was observed at these time points in LPS-treated tpl2−/− HSC (Fig. 8A). Western blotting demonstrated reduced basal and LPS-induced TIMP-1 protein expression in tpl2−/− HSCs (Fig. 8B). LPS stimulation of WT HSCs caused a time-dependent loss of Tpl2 expression (Fig. 8C). This resembles LPS-induced degradation of active Tpl2 in macrophages, which ensures transient activation of ERK signaling, suggesting that Tpl2 is regulated by similar mechanisms in HSCs. Of note, p-ERK was induced by LPS maximally at 30 minutes, but by 180 minutes, was close to basal levels, which coincided with loss of Tpl2. The direct targets for catalytically active Tpl2 are the ERK kinases, MKK1 and MKK2.15 Phosphorylation of ERK in LPS-treated HSCs was inhibited by the MKK1/2 inhibitor, U0126; however, the drug had no effect on Tpl2 turnover, confirming that Tpl2 operates upstream of MKK1/2 and ERK (Fig. 8C). U0126 also suppressed LPS induction of TIMP-1 mRNA, confirming that MKK1/2/ERK signaling is essential for TLR regulation of TIMP-1 in HSCs (Fig. 8D). Finally, we were able to show that a transfected TIMP-1 promoter/Luc reporter was responsive to LPS in WT, but not in tpl2−/− HSCs (Fig. 8E). Miura et al. described how IL-1β enhances HSC expression of TIMP-126; to determine whether this paracrine and autocrine fibrogenic cytokine pathway requires Tpl2, we investigated effects of IL-1β on TIMP-1 transcript expression in cultures of activated WT and tpl2−/− HSCs (Fig. 8F). IL-1β-induction of TIMP-1 was confirmed in WT HSCs; however, basal and IL-1β-stimulated TIMP-1 transcript levels were reduced in tpl2−/− HSCs. We conclude that LPS and IL-1β are potent stimulators of TIMP-1 gene transcription in activated HSCs and require signaling through the Tpl2/MKK1/2/ERK signal-transduction pathway.

Figure 8.

Tpl2 is required for LPS induction of p-ERK and TIMP-1 gene transcription. (A) Quantitative real-time PCR analysis of the expression of TIMP-1 in WT and tpl2−/− activated HSCs treated with LPS (100 ng/mL) for 1, 3, and 6 hours. (B) TIMP-1 protein expression in HSCs treated with LPS (100 ng/mL) for 24 hours. (C) Western blotting analysis of Tpl2 and p-ERK in activated HSCs treated with LPS (100 ng/mL) for 30 minutes and 1 or 3 hours and pretreated with MEK inhibitor U0126 for 30 minutes. (D) TIMP-1 mRNA expression in WT activated HSCs pretreated with the MEK inhibitor, U0126 (15 μmol/L), for 30 minutes. (E) Seven-day–cultured activated WT and tpl2−/− HSCs transfected with 1 μg of TIMP-1 (pTIMP-1) and 0.1 μg of Renilla luciferase vector (pRLTK). Twenty-four hours after transfection, serum was removed and HSCs were treated for 6 hours with LPS (100 ng/mL). Luciferase activity was determined, and results are expressed as mean percentage (normalized to Renilla values) ± SEM of three independent transfection experiments. (F) WT and tpl2−/− HSCs were treated with IL-1β for 6 hours, and TIMP-1 mRNA expression was measured by quantitative real-time PCR. ***P≤0.001.


TLRs are expressed on KCs and HSCs and are strongly implicated in the initiation and progression of liver fibrogenesis, both in humans and most of the commonly used rodent models. In particular, TLR3, TLR4, and TLR9 have been described as stimulators of fibrogenesis in response to microbial components that are elevated in the diseased versus healthy liver, mainly as a consequence of disruption of the intestinal barrier and leakage of bacteria into the bloodstream.29 TLR4 has received the most attention, with Seki et al. providing genetic evidence in mouse models that TLR4 and the TLR adaptor molecule, Myd88, are required for the development of liver fibrosis in multiple injury models.12 TLR4 is expressed on parenchymal and nonparenchymal liver cells; however, Seki et al. demonstrated that TLR4 on HSCs is crucial for liver fibrosis. In support of this animal model work, there is evidence that functional SNPs in the TLR4 gene reduce the risk of cirrhosis in patients with chronic HCV infection.13 Furthermore, Guo et al. showed that TLR4 SNPs D299G and T399I are associated with blunted TLR4 signaling and cell death in HSCs.14 These data raise the potential for inhibitors of TLR4 (and other TLRs) to be developed as antifibrotics; however, such inhibitors may affect appropriate control of infections, which raises concerns about such an approach.

An alternative approach is to identify signaling molecules and events downstream of the TLRs that might be targeted to more selectively suppress fibrogenic TLR signaling. Here, we focused on the MAP3K molecule, Tpl2, which is required for TLR-induced activation of MEK1/2/ERK signaling in macrophages.15 Preliminary data revealed increased expression of Tpl2 in diseased rodent liver and restricted expression of Tpl2 to nonparenchymal liver cells. We went on to provide in vitro evidence that Tpl2 is critical for LPS and CpG induction of ERK in KCs and HSCs. This led us to ask whether Tpl2 is required for liver fibrosis, which we tested using three independent chronic liver injury models comparing WT to tpl2−/− mice. With the CCl4 and MCD models, in which HSC-derived myofibroblasts are thought to be primary drivers of fibrosis, we observed reduced fibrosis. By contrast, there was no effect of Tpl2 deletion on fibrosis in the BDL model, this principally being driven by myofibroblasts derived from portal fibroblasts.30 Therefore, Tpl2 may play a more-restricted role in fibrogenic processes focused around the HSCs and where there is an influence of TLR signaling in these cells.

Although we have yet to fully characterize differences in basal and TLR-induced gene expression in tpl2−/− HSCs, we showed that LPS and CpG stimulation of IL-1β, IL-6, and TIMP-1 is defective in these cells. Miura et al. recently described how, in the context of a mouse model of nonalcoholic steatohepatitis (NASH), stimulation of TLR9 on KCs increases the secretion of IL-1β, which then acts on bystander hepatocytes to promote lipid accumulation and on HSCs to induce collagen and TIMP-1.26 Here, we confirmed that LPS can induce IL-1β expression in KCs and HSCs and show that this is dependent on Tpl2, which suggests a role for Tpl2/ERK as a critical component of the signal-transduction pathways that link the engagement of TLRs with IL-1β expression. Selective small-molecule inhibitors of Tpl2 are available and might be used to block TLR/IL-1β pathways in NASH.31, 32 We also observed suppression of TIMP-1 expression in tpl2−/− diseased livers, which may be the result of loss of paracrine or autocrine IL-1β. Alternatively reduced TIMP-1 induction may result from the perturbation of ERK activation downstream of TLR-Tpl2. An ERK/JunD pathway controls the transcription of TIMP-1 and IL-6 in HSCs,40 and, as shown here, LPS stimulation of TIMP-1 promoter activity requires MKK1/2/ERK.

Although Tpl2 deletion has a strong inhibitory effect on TLR-induced KC and HSC gene expression and on levels of TIMP-1 in diseased liver, its effects on fibrosis were modest and model specific. With all three injury models, there was a substantial accumulation of fibrotic matrix in tpl2−/− livers, albeit modestly reduced in the CCl4 and MCD injuries. This was surprising, given that TIMP-1 plays multiple roles in promoting fibrosis, including inhibition of collagenases, prevention of HSC apoptosis, and suppression of liver regeneration.33-36 A possibility is that genetic ablation of Tpl2 may trigger developmental mechanisms to compensate for reduced expression of TIMP-1. Alternatively, TIMP-1 may play a less-critical role in liver fibrosis than previously suggested, and, in this respect, it is noteworthy that Wang et al. recently reported on increased hepatocyte death and enhanced fibrosis in CCl4-injured timp1−/− mice.37 This contrasted with earlier work showing that transgenic overexpression of hepatic TIMP-1 enhances CCl4-induced fibrosis.34 Therefore, there is complexity regarding TIMP-1 and liver homeostasis and, as yet, no clear message has from mouse models.

Tpl2 helps couple IKK/NF-κB and ERK inflammatory signaling pathways downstream of TLRs because its catalytic activity relies upon IKK-regulated proteolytic degradation of p105.16, 38 The nfkb1−/− mouse that lacks p105 expresses low levels of Tpl2 because of the fact that the p105/Tpl2 complex prevents proteasomal degradation of Tpl2; in addition, in comparison to WT animals, these mice develop exaggerated neutrophillic hepatitis and enhanced fibrosis in response to chronic CCl4 injury.39 By contrast, similar injury in the tpl2−/− background is associated with reduced numbers of hepatic neutrophils and less fibrosis. Therefore, we can rule out perturbation of Tpl2 signaling as a contributor to the extreme pathology observed in the nfkb1−/− mouse.

In summary, Tpl2 is important for TLR4 activation of ERK-dependent expression of fibrogenic IL-1β and TIMP-1 and appears to play a contributory role in liver fibrosis. Therefore, selective pharmacological inhibitors of Tpl2 may be of value to explore in animal models of liver disease.