Sh(i)pping signals protect against Stat3‐driven liver cancer

The generation of functional hepatocytes independent of donor liver organs is of great therapeutic interest with regard to regenerative medicine and possible cures for liver disease. Induced hepatic differentiation has been achieved previously using embryonic stem cells or induced pluripotent stem cells. Particularly, hepatocytes generated from a patient’s own induced pluripotent stem cells could theoretically avoid immunological rejection. However, the induction of hepatocytes from induced pluripotent stem cells is a complicated process that would probably be replaced with the arrival of improved technology. Overexpression of lineage-specific transcription factors directly converts terminally differentiated cells into some other lineages, including neurons, cardiomyocytes and blood progenitors; however, it remains unclear whether these lineage-converted cells could repair damaged tissues in vivo. Here we demonstrate the direct induction of functional hepatocyte-like (iHep) cells from mouse tail-tip fibroblasts by transduction of Gata4, Hnf1a and Foxa3, and inactivation of p19Arf. iHep cells show typical epithelial morphology, express hepatic genes and acquire hepatocyte functions. Notably, transplanted iHep cells repopulate the livers of fumarylacetoacetate-hydrolase-deficient (Fah -/-) mice and rescue almost half of recipients from death by restoring liver functions. Our study provides a novel strategy to generate functional hepatocyte-like cells for the purpose of liver engineering and regenerative medicine. Sekiya S, Suzuki A. Direct conversion of mouse fibroblasts to hepatocyte-like cells by defined factors. Nature 2011;475:390-393. (Reprinted with permission.)

fusion with hepatocytes. These unexpected findings suggest that factors critical to hepatocyte differentiation exist and become activated to induce hepatocyte-specific properties in different cell types. Here, by screening the effects of twelve candidate factors, we identify three specific combinations of two transcription factors, comprising Hnf4a plus Foxa1, Foxa2 or Foxa3, that can convert mouse embryonic and adult fibroblasts into cells that closely resemble hepatocytes in vitro. The induced hepatocyte-like (iHep) cells have multiple hepatocytespecific features and reconstitute damaged hepatic tissues after transplantation. The generation of iHep cells may provide insights into the molecular nature of hepatocyte differentiation and potential therapies for liver diseases.

Comment
Two groundbreaking papers published recently in Nature provide evidence that somatic cells can be reprogrammed directly into functional hepatocyte-like cells, without a requirement for embryonic stem (ES) cells or induced pluripotent stem cells. This innovative strategy will allow more efficient production of hepatocytes of diverse genetic backgrounds for medical research and transplantation.
The lack of effective cell-culture systems to expand and maintain mature, fully differentiated hepatocytes in vitro presents a major obstacle to basic and translational investigation of the pathogenesis and treatment of liver disease. Mouse and human ES cells can be directed toward a hepatic phenotype using stepwise application of growth factors in vitro to mimic embryonic endodermal development. However, complete hepatocytic differentiation in vitro has proven very difficult to achieve, and available human ES cell lines are limited in their genetic background.
The ability to convert terminally differentiated mouse cells to induced pluripotent stem cells (iPS) by overexpressing just four key transcription factors-Klf4, c-myc, Oct3/4, and Sox2 1 -was later adapted with subtle variations to human cells. 2,3 Cellular reprogramming by these factors occurs by an incompletely understood process involving epigenetic alterations, which ''reset'' regulatory chromatin configurations toward a more permissive embryonic or pluripotent state, resembling that of ES cells. 4 Like ES cells, iPS derived from somatic cells share many characteristics with ES cells and can be directed to differentiate toward a desired cell type in culture (Fig. 1).
Though iPS do not completely mimic ES cells in all respects, they can provide stem cells with any desired genetic background, reflecting a specific patient or disease condition. Reprogramming somatic cells to iPS, then directing them to become mature hepatocytes in vitro, is inefficient. Furthermore, undifferentiated ES and iPS give rise to teratomas after transplantation, so complete elimination of undifferentiated cells will be critical to any human therapeutic application.
Innovative new technologies may overcome these concerns. Two recent reports demonstrate the conversion of mouse fibroblast cells to hepatocyte-like cells (induced-hepatocytes or ''iHep''), using forced expression of endodermal transcription factors. In both cases, phenotypic conversion appears to occur directly, rather than through a pluripotent intermediate. Direct reprogramming of terminally differentiated cells from one phenotype to another has been demonstrated previously in nonhepatic model systems. [5][6][7][8][9][10][11] To generate iHep from fibroblasts, both groups screened a set of 12-14 transcription factors predicted to support hepatocytic differentiation. These factors were narrowed by systematic subtraction. Hepatocytic differentiation in vitro was assessed using reverse-transcriptase polymerase chain reaction, microarray, glycogen production (PAS staining), low-density lipoprotein uptake, albumin secretion, urea production, cytochrome P450 metabolism, and immunolabeling, in comparison to adult hepatocytes. Both groups demonstrated hepato-cytic differentiation and function in vivo by transplanting fibroblast-derived hepatocyte-like cells into fumaryl acetoacetate hydrolase knockout (FAH À/À ) mice, which model the human disease hereditary tyrosinemia. 12 The requirement for at least one member of the Forkhead box A (Foxa) transcription factor family was common to both strategies. This is not surprising, in view of the critical role of Foxa factors in the patterning of embryonic endoderm, which ultimately gives rise to the liver. 13 Huang et al. derived hepatocyte-like cells from mouse embryonic fibroblasts, 3T3 cells, or tail-tip fibroblasts. Fourteen candidate transcription factors were screened by lentiviral infection in fibroblasts and narrowed by systematic subtraction. The combination of Foxa3, GATA-4, and HNF1alpha was best able to support hepatocyte differentiation, but the cells did not proliferate in culture unless p19Arf was inactivated in the tail-tip fibroblasts by gene knockout or RNA interference.
In the accompanying report, Sekiya and Suzuki also derived hepatocyte-like cells from mouse embryonic fibroblasts; but to rigorously exclude any contribution of nascent liver cells from embryonic cells, they also derived hepatocyte-like cells from embryonic limbs, adult dermal fibroblasts, or unfractionated bone marrow cells. Twelve candidate transcription factors were screened by retroviral infection. The minimal combination for successful hepatic reprogramming included HNF4alpha plus one of the Foxa transcription factors can be reprogrammed to an iPS phenotype by transfection with the ''Yamanaka reprogramming factors. '' Both ES cells and iPS can be directed toward an endodermal/hepatocytic phenotype in vitro using stepwise application of growth factors, which recapitulate embryonic development of the liver. The new approach relies on transfection of endodermal transcription factors to force fibroblasts to acquire the hepatocyte phenotype directly, bypassing the requirement for ES or iPS. This innovative strategy will allow more efficient production of hepatocytes of diverse genetic backgrounds for medical research and transplantation.
(Foxa1, Foxa2, or Foxa3). In contrast to the work of Huang et al., these iHeps proliferated spontaneously without a requirement for inactivation of p19Arf.
Both groups achieved therapeutic liver repopulation in the FAH null mouse. Huang et al. reported that transplanted iHep proliferated in vivo for several weeks, then became quiescent by 8 weeks after transplant. No tumors were found in the transplant recipients. Sekiya and Suzuki similarly found that iHep stopped proliferating within 2 months after transplantation into FAH null recipients. The ability of iHeps to reenter the cell cycle was demonstrated by performing a partial hepatectomy in FAH null mice after stable repopulation with iHep. Impressively, iHep could even be made from FAH null embryonic fibroblasts. Retroviral infection of FAH null iHep with a vector expressing wild-type FAH restored FAH expression after transplantation, providing proof of concept that iHep might provide a platform for gene and cell therapy for liver diseases. The FAH null mouse model is unique in that therapeutic repopulation after bone marrow transplantation occurs through fusion between recipient hepatocytes and donor-derived macrophages. 14,15 In contrast, repopulation after hepatocyte transplantation occurs through the proliferation of donor cells, without donor-recipient fusion. 16,17 Sekiya and Suzuki demonstrated that therapeutic liver repopulation using iHep did not involve the fusion mechanism, supporting the notion that iHeps are sufficiently hepatocyte-like to proliferate, repopulate the liver, and correct the defect in tyrosine metabolism.
There can be no doubt that the general approach of reprogramming fibroblasts to hepatocytes will be extended to human cells, in the hope that patient-and disease-specific hepatocytes will become available. Initially, iPS were prepared by delivering the ''reprogramming factors'' using retroviral vectors, which integrate into the target cell genome, potentially disrupting key genes and promoting unpredictable, aberrant behavior. The same concerns apply to these studies, in which the reprogramming factors were delivered by lentiviral or retroviral infection. In future work, alternative strategies to express reprogramming factors transiently may allow reprogramming without genomic integration.
The general approach of direct conversion of fibroblastic cells to iHeps may obviate the need to prepare iPS and differentiate them to hepatocytes in culture. Application of this technology to human cells for the aim of drug discovery and toxicity screening could allow earlier identification of problematic candidates during pharmaceutical development. Unlike mature hepatocytes, iHep proliferate in culture, potentially allowing the production of a genetically diverse array of hepatocytes for research and transplantation.

Abstract
The molecular identification of adult hepatic stem/progenitor cells has been hampered by the lack of truly specific markers. To isolate putative adult liver progenitor cells, we used cell surface-marking antibodies, including MIC1-1C3, to isolate subpopulations of liver cells from normal adult mice or those undergoing an oval cell response and tested their capacity to form bilineage colonies in vitro. Robust clonogenic activity was found to be restricted to a subset of biliary duct cells antigenically defined as CD45(-)/CD11b(-)/CD31(-)/MIC1-1C3(þ)/ CD133(þ)/CD26(-), at a frequency of one of 34 or one of 25 in normal or oval cell injury livers, respectively. Gene expression analyses revealed that Sox9 was expressed exclusively in this subpopulation of normal liver cells and was highly enriched relative to other cell fractions in injured livers. In vivo lineage tracing using Sox9creER(T2)-R26R(YFP) mice revealed that the cells that proliferate during progenitor-driven liver regeneration are progeny of Sox9-expressing precursors. A comprehensive array-based comparison of gene expression in progenitor-enriched and progenitor-depleted cells from both normal and DDC (3,5-diethoxycarbonyl-1,4-dihydrocollidine or diethyl1,4-dihydro-2,4,6-trimethyl-3,5-pyridinedicarboxylate)treated livers revealed new potential regulators of liver progenitors.

Abstract
Isolation of hepatic progenitor cells is a promising approach for cell replacement therapy of chronic liver disease. The winged helix transcription factor Foxl1 is a marker for progenitor cells and their descendants in the mouse liver in vivo. Here, we purify progenitor cells from Foxl1-Cre; RosaYFP mice and evaluate their proliferative and differentiation potential in vitro. Treatment of Foxl1-Cre; RosaYFP mice with a 3,5-diethoxycarbonyl-1,4-dihydrocollidine diet led to an increase of the percentage of YFP-labeled Foxl1(þ) cells. Clonogenic assays demonstrated that up to 3.6% of Foxl1(þ) cells had proliferative potential. Foxl1(þ) cells differentiated into cholangiocytes and hepatocytes in vitro, depending on the culture condition employed. Microarray analyses indicated that Foxl1(þ) cells express stem cell markers such as Prom1 as well as differentiation markers such as Ck19 and Hnf4a Thus, the Foxl1-Cre; RosaYFP model allows for easy isolation of adult hepatic progenitor cells that can be expanded and differentiated in culture.

Comment
The shortage of human donor livers, low engraftment rates, and poor survival of transplanted hepatocytes hamper the use of clinical and experimental hepatocyte transplantation. In healthy organs, liver progenitor cells (LPCs) are generally dormant (or slowly cycling) and are only present in low numbers in different niches of the liver. 1,2 When a liver gets injured and the regenerative capacity of mature hepatocytes and/or cholangiocytes is impaired, these LPCs become activated in humans as well as in animal models of liver disease 3 and can replace dysfunctional or damaged parenchymal cells. Because of their high proliferative ability and differentiation potential toward hepatocytes and cholangiocytes, LPCs are considered as an attractive alternative source for cell therapy. However, their isolation remains challenging. 4 So far, antibodies recognizing cell-surface proteins, such as cluster of differentiation (CD)133, epithelial cell adhesion molecule (EpCAM), CD49f, and CD24, were used to isolate putative adult LPCs mainly from injured mouse livers. [5][6][7][8][9] The use of liver injury models increases the overall yield of LPCs, but gives a mix of dormant and activated LPCs, which complicates the characterization of these cells. An expected ''gold standard'' for the isolation of adult LPCs has, therefore, not yet been established.
Recently, two elegant reports have shed new light on the identity/biology of LPCs, giving new hope for their prospective use in liver cell replacement therapy. 10,11 First, the investigators describe two novel approaches for the successful isolation of bipotential LPCs from normal and DDC (3,5-diethoxycarbonyl-1,4-dihydrocollidine)-induced mouse livers. Second, they both demonstrate the progenitor features of these populations by clonally expanding and differentiating them to functional mature cells. Third, based on hierarchical clustering of gene-array data, they attempt to describe how LPCs react upon liver injury. From this, they conclude that the LPC response appears to be biphasic: primarily, a set of genes awakens the LPCs from a dormant state, whereas in a second phase, the expression of genes involved in metabolism and motility gets dramatically changed, which further favors the reconstitution of the liver mass.
The Dorrell article is unique in the sense that the investigators isolated LPCs from healthy livers and at several time points during liver injury using the monoclonal antibody, MIC1-1C3 (macrophage inhibitory cytokine-1-1C3), which is specific for duct cells. 12 It allowed them to compare the gene-expression profile of dormant LPCs with activated LPCs. 11 In another approach, Shin et al. circumvented the need for LPCspecific antibodies by using a transgenic mouse engineered to express yellow fluorescent protein (YFP) whenever the transcription factor, Foxl1 (Forkhead Box l1), was expressed (Foxl1Cre;Rosa YFP). Because Foxl1 is only expressed in activated LPCs, 13 they could compare gene-array data from LPCs isolated at different time points during DDC treatment. LPCs were separated based on their MIC1-1C3 and YFP positivity from other nonparenchymal cells by flow cytometry for further analysis (Fig. 1A). Both reports are noteworthy because LPCs were isolated at different time points of liver injury and both demonstrate that isolated LPCs can be clonally expanded, even up to 15 passages using conditioned media from E14,5 liver cells. 10,11 It would now be a great advantage to identify those factors that allow the expansion of the progenitor cells. Furthermore, both studies unequivocally show that the isolated LPCs are bipotent by in vitro differentiation of a Gene-expression profiling of progenitor-enriched and progenitordepleted cells from both healthy and DDC-treated livers underscores the pathways that are important for LPC stemness and activation. In the activated state, pathways presented in the blue, pink, and gray arrows are respectively overrepresented in early-and late-activated LPCs and throughout LPC activation.

320
HEPATOLOGY ELSEWHERE HEPATOLOGY, January 2012 clonally expanded LPC (MIC1-1C3 þ or Foxl1 þ ) toward a cholangiocytic and hepatocytic cell type, refuting the existence of two unipotent LPCs. Recently, Okabe et al. demonstrated that EpCAM (TROP1) is expressed both in cholangiocytes of healthy mouse livers and in oval cells (i.e., activated LPCs) when mice were fed with DDC containing chow. Its related protein, TROP2, is expressed exclusively in oval cells and not in the healthy liver. 8 Foxl1, like TROP2, also appears to be an oval cell and not a dormant LPC marker. However, for Foxl1, it is tempting to address a function during LPC activation, because earlier studies using Foxl1 À/À mice showed that Foxl1 promoted liver repair after bile-duct ligation-induced liver injury. 14 Giving DDC chow to Foxl1 À/À mice and determine whether they still exhibit a normal oval cell response is an obvious way to test this hypothesis. Though the MIC1-1C3 antibody can be used to identify dormant LPCs, it is not exclusive for the liver. Its expression in the pancreas 11,12 suggests that MIC1-1C3 might be a common marker for stem cells within endoderm-derived digestive organs, joining Sox9 15 and Sox17. 16 Transcriptome profiling of dormant and activated LPCs undertaken by Shin et al. revealed drastic changes in gene expression of the isolated LPCs at different times of the injury. The wealth of data generated by these experiments need further analysis, but bioinformatic pathway analysis already allowed the investigators to identify processes regulated during LPC activation (illustrated in Fig. 1B). Importantly, although the isolation strategy of the LPCs was different, both studies identified similar relevant pathways. It is not surprising to find that LPCs overexpress drug metabolism and defense response genes because they have to survive several insults during the organism's life. It also makes sense to observe that, in their dormant state, LPCs have low expression of cell-cyclerelated genes, compared to their counterpart during injury. Similarly, the overrepresentation of pathways involved in the remodeling of the LPC niche is conceivable because of the necessity of the progeny to be liberated from the niche. More advanced analysis of the data sets would certainly reveal new potential regulators of LPC activation or stemness.
We now look forward to reports that will use a combination of LPC cell-surface markers, such as EpCAM, Trop2, CD133, Dlk1, CD49f, and MIC1-1C3, to isolate LPCs from healthy and injured livers. Cell-surface markers combined with functional characteristics of LPCs, such as overexpression of pumps 17 or aldehyde dehydrogenase activity, 18 could further specify this pop-ulation. Finally, the report by Shin et al. clearly has set the stage for many investigators to use transgenic mice for the isolation of LPCs. Good candidates for such studies are the already published reports on Sox9-Cre 15 and CK19-Cre mice. 19 Though the use of these mice provides the field with elegant tools to further characterize LPCs, we are still in need of strategies to isolate LPCs from human tissues to verify the findings obtained in mice.

Abstract
The human gene PTPN11, which encodes the tyrosine phosphatase Shp2, may act as a proto-oncogene because dominantly activating mutations have been detected in several types of leukemia. Herein we report a tumor-suppressor function of Shp2. Hepatocyte-specific deletion of Shp2 promotes inflammatory signaling through the Stat3 pathway and hepatic inflammation/ necrosis, resulting in regenerative hyperplasia and development of tumors in aged mice. Furthermore, Shp2 ablation dramatically enhanced diethylnitrosamine (DEN)-induced hepatocellular carcinoma (HCC) development, which was abolished by concurrent deletion of Shp2 and Stat3 in hepatocytes. Decreased Shp2 expression was detected in a subfraction of human HCC specimens. Thus, in contrast to the leukemogenic effect of dominant-active mutants, Ptpn11/Shp2 has a tumorsuppressor function in liver.

Comment
Protein phosphorylation at tyrosine residues is a key mechanism of signal transduction during many cellular processes including inflammation, cell death, and proliferation. Therefore, deregulation or hyperactivation of tyrosine kinases is frequently associated with cancer formation. 1 Tyrosine phosphorylation can be antagonized by tyrosine phosphatases. However, the precise interplay between tyrosine kinases and their respective phosphatases is still poorly understood.
Recent research has suggested that the ubiquitously expressed tyrosine phosphatase Shp2 (SH2 domaincontaining protein tyrosine phosphatase-2, also known as protein tyrosine phosphatase, nonreceptor type 11/ PTPN11) may act as a proto-oncogene. Dominant active mutations of Shp2 have been identified in patients with Noonan syndrome, an autosomal dominant congenital disorder resulting in, e.g., congenital heart defects, dwarfism, and hematologic abnormalities that have a high predisposition toward development of childhood leukemias as well as other types of leukemias. 2,3 Mechanistically, Shp2 seems to play a bivalent role, as it not only antagonizes tyrosine kinases, but also activates the Ras/extracellular signal-regulated kinase (ERK) pathway following stimulation by growth factors and cytokines. However, the underlying mechanisms are still largely unknown and remain controversial. 4 Several lines of evidence have suggested that Shp2 is also an important antagonist of the interleukin-6 (IL-6)/glycoprotein (gp130)/Stat3 signaling pathway and acts by dephosphorylating Janus kinases (JAK) and activated Stat3 and attenuates IL-6 signaling by recruitment to tyrosine 759 on the IL-6 receptor subunit gp130. 5,6 Further insight into the function of Shp2 in the liver came from experiments performed in hepatocyte-specific Shp2 knockout mice subjected to partial hepatectomy (PH). Genetic ablation of Shp2 resulted in impaired liver regeneration following PH, most likely due to reduced ERK activation and down-regulation of immediate early genes such as c-Fos, c-Jun, and c-Myc. 7 Additionally, IL-6/Stat3-dependent hepatoprotective signals were enhanced in Shp-deficient mice, confirming the hypothesis that hepatic Stat3 activity is regulated by Shp2. However, these protective signals are apparently not sufficient to support normal compensatory hepatocyte proliferation in the absence of Shp2. From these data, it could be speculated that inactivation of Shp2 would rather protect from excessive hepatocyte proliferation and thus from hepatocarcinogenesis.
By analyzing the same hepatocyte-specific Shp2 knockout mice in more detail, Bard-Chapeau et al. 8 now aimed to clarify the role of Shp2 for the development of liver inflammation and liver cancer. Very surprisingly, ablation of Shp2 results in onset of spontaneous hepatitis and parenchymal necrosis in mice by the age of 2-3 months along with increased intrahepatic expression of 322 HEPATOLOGY ELSEWHERE HEPATOLOGY, January 2012 IL-6 and tumor necrosis factor alpha (TNF). The proinflammatory properties of Shp2-deficiency are even more apparent in a model of sepsis induced by lipopolysaccharide treatment, where Shp2 knockout mice were found to be hypersensitive toward inflammatory liver injury. In agreement with earlier studies from the same group, loss of Shp2 triggers enhanced and prolonged phosphorylation of Stat3 and c-Jun N-terminal kinases (JNK), but impaired activation of the ERK signaling pathways. Hence, the Shp2 protein appears to mediate the antiinflammatory effect by keeping Stat3 and JNK in check.
The spontaneous hepatitis found in Shp2-deficient mice culminates in the development of hepatocellular adenoma with constitutive Stat3 activation after 1-1.5 years. These findings also have clinical relevance, as Bard-Chapeau et al. demonstrated that >10% of human hepatocellular carcinomas are characterized by decreased Shp2 expression. To further elucidate the postulated pro-oncogenic role of Stat3, the authors compared conditional Shp2 knockout mice with Shp2/ Stat3 double knockout mice and wildtype (WT) controls in a model of chemically induced hepatocarcinogenesis by using the carcinogen diethylnitrosamine (DEN). Apparently, Shp2 acts as a tumor suppressor in the liver, as Shp2 knockout mice show strongly enhanced HCC development following DEN exposure. Bard-Chapeau et al. provided convincing evidence that the tumor-promoting properties of the Shp2 knockout are Stat3-dependent, as not only increased tumor susceptibility but also basal inflammation was largely rescued by concomitant hepatic ablation of Shp2 and Stat3. Thus, Stat3 is the actual proinflammatory mediator and main tumor promoter in Shp2-deficient mice. Unexpectedly, however, hepatic Stat3 inactivation alone was not protective in the DEN model, and even resulted in a slightly enhanced tumor formation as compared to DEN-treated WT controls. This phenomenon is contradictory to a recent study by He et al., 9 who showed a strong protection from DEN-mediated hepatocarcinogenesis in Stat3-deficient mice. However, both groups used different Stat3 knockout alleles, which might explain this discrepancy, although future work will be necessary to clarify these differences.
As usual with excellent research, the work of Bard-Chapeau et al. raises several new questions. First, it would be interesting to examine if Stat3/Shp2 double knockout mice are also protected from spontaneous hepatocarcinogenesis, which is anticipated, as these Fig. 1. The many oncogenic facets of IL-6/gp130/Stat3 signaling in the liver. IL-6 activates the transcription factor Stat3 by way of binding to the gp130/IL-6 receptor complex. Induction of Stat3 target genes may result in a proinflammatory response. However, constitutive activation of Stat3 results in chronic hepatitis and potentially liver cancer. Constitutive active somatic mutations for both Stat3 and gp130 have been identified recently in benign liver tumors. It has been demonstrated that IL-6 triggers hepatocarcinogenesis in mice. Now, it is evident that the oncogenic activity of Stat3 is under the control of the phosphatase Shp2. However, if Shp2 inhibits Stat3 directly, e.g., by dephosphorylation or indirectly by blocking upstream signal transducer such as gp130.
mice have markedly reduced basal liver inflammation. Moreover, the question concerning the relevant target proteins of Shp2 in hepatocytes is still open. The data presented here imply that Shp2 may inhibit Stat3 directly, e.g., by dephosphorylation. However, it cannot be excluded that Shp2 also blocks STAT-signaling further upstream by inhibiting the signal transducer gp130, as previously suggested. 6 In this regard, further in vitro studies with Shp2-deficient hepatocytes may be helpful to identify all dominant-active components of the IL-6 pathway in this scenario.
The exciting data from Bard-Chapeau et al. are in line with increasing evidence that demonstrates that aberrant activation of the IL-6/gp130/Stat3 signaling pathway is a hallmark of liver cancer development in mice and man (Fig. 1). In this context, ablation of IL-6 itself was shown to be highly protective against induction of HCC by DEN in mice. 10 More recently, somatic dominant active mutations of Stat3 and gp130 have been identified in human patients with hepatocellular adenomas, both resulting in constitutive activation of Stat3. 11,12 Here, we learn that hepatic tumor cells have evolved at least one more mechanism to get more activated Stat3 through the down-regulation of Shp2, although the nature of this suppressive mechanism is completely unknown. The diversity of mechanisms and factors leading to Stat3 overactivation in liver tumors is remarkable and emphasizes the need to investigate this important and critical pathway even more intensively in the context of inflammatory liver disease.