In spite of the histological homogeneity of the liver parenchyma, hepatocytes display profound functional differences depending on their position within the liver lobule, a phenomenon known as the “metabolic zonation” of the liver. It was established more than 30 years ago that important metabolic activities in the adult liver were separated into distinct regions in a gradient along the portocentral axis. According to their position with respect to the vascular structures, parenchymal cells near the portal triad are known as periportal (PP) hepatocytes, while those located close to the efferent centrilobular vein are known as the pericentral (PC) hepatocytes. Most interestingly, metabolic processes are distributed in finely tuned gradients between PP and PC regions in such a way that pathways performing opposing functions follow an inverse representation through the portocentral axis. For instance, gluconeogenesis, cholesterol and urea synthesis predominate in the PP region, while glycolysis, bile acid, and glutamine synthesis and drug metabolism are most prominent in the PC zone. For some processes, such as carbohydrate metabolism, this pattern of zonation can contract or expand depending on nutritional needs, while others like ammonia metabolism display a more strict location.[4, 5] This complex topology of metabolic activities is achieved through the precise regulation of the messenger RNA (mRNA) levels of key enzymes involved in the different pathways. The mechanisms underlying the regulatory network responsible for liver zonation have been the subject of active research over the past decades. It was initially thought that gradients in the concentrations of oxygen, hormones, cytokines, and nutrients in the blood along the portocentral axis could determine zonation.[4, 6] Indeed, changes in blood oxygen, hormones, or xenobiotics levels may influence the patterns of some processes like glycolysis and gluconeogenesis, or drug metabolism. However, there are other less dynamic systems such as ammonia detoxification that display a more stable zonation. Nevertheless, evidence produced during the last few years has identified the Wnt/β-catenin pathway as a critical system for the preservation of the hepatic function and as a master regulator of hepatic zonation. Early observations found that the expression of glutamine synthetase (GS), a marker of distal PC hepatocytes, was strongly up-regulated in liver tumors with activating β-catenin mutations.[2, 9] Subsequent investigations using genetically modified mice with targeted activation or inactivation of β-catenin signaling further demonstrated that this system plays a central role in the control of the expression of a wide range of zonated genes.[10-13] Among these studies the work of Benhamouche et al. provided critical insights. First, they showed the complementary localization of active β-catenin in PC hepatocytes and that of its negative regulator adenomatous polyposis coli (Apc) in PP parenchymal cells. Moreover, using mice with liver-targeted inactivation of Apc, the same authors demonstrated that panlobular β-catenin signaling led to the activation of a PC genetic program throughout the liver, while the ectopic expression of the Wnt inhibitor Dickkopf-1 reversed this effect.
While together these findings cogently demonstrated a predominant role for β-catenin in liver metabolic zonation, the precise molecular mechanisms mediating β-catenin hepatic gene regulation have remained more elusive. In the canonical Wnt pathway β-catenin translocates to the nucleus, where it interacts with DNA-bound Tcf/Lef (T-cell factor / lymphoid enhancer factor) transcription factors on consensus Wnt response elements (WRE), leading to the activation of target gene expression. However, recent developments indicate that the mechanisms underlying β-catenin control of hepatic gene expression could be far more complex. Besides Tcf/Lef, β-catenin may certainly interact with other transcription factors and hormone receptors potentially influencing zonated gene expression.[2, 15] Additionally, an important crosstalk was found between Lef1 and the transcription factor hepatocyte nuclear factor 4α (HNF4α), another major keeper of liver zonation, with Lef1 also binding HNF4α consensus sites (HREs) and driving gene expression upon β-catenin activation.
To unravel the intricate molecular determinants of β-catenin-dependent zonal transcription, Gougelet et al. decided to undertake an unbiased and more holistic approach in an elegant study published in this issue of Hepatology. Using hepatocytes with constitutively activated or inactivated β-catenin (Apc- and β-catenin-null mice, respectively), and taking advantage of state-of-the art mRNA-seq, ChiP-seq, and metabolomic analyses, they shed new light on this critical issue. One unexpected and important finding in this report was the realization that β-catenin expression has a strong influence on Tcf4 recruitment to WRE sites. Indeed, ChIP-seq studies revealed that chromatin was much more occupied by Tcf4 in hepatocytes from Apc-null mice than from β-catenin-deficient animals, which is in contrast with the current view that Tcf4 constitutively binds WREs even in the absence of β-catenin. This finding suggests that β-catenin would facilitate the opening of chromating around WRE sites, an aspect that merits additional mechanistic studies in order to identify the chromatin remodeling factors recruited by β-catenin. ChIP-seq profiles also revealed that most of the β-catenin binding sites were included in Tcf4-bound chromatin, although admittedly lower affinity interactions of β-catenin with other transcription factors could have escaped this ChIP assay, and therefore important crosstalks may have been missed. Of interest, independently of β-catenin, it was found that Tcf4 could also bind HRE sites in genes positively regulated by HNF4α, particularly in the PP region, allowing hepatic glucose production and lipogenesis. Furthermore, in vitro biochemical evidence is provided demonstrating a reciprocal antagonistic effect of β-catenin and HNF4α on the transactivation of their respective target promoter elements. These findings could be at variance with the previously described positive cooperation between HNF4α and Wnt signaling in the expression of PP genes such as Cyp1a1. Further studies are certainly needed to clarify this complex crosstalk. For the moment, Gougelet et al. provide mechanistic evidence showing that the negative effect of β-catenin on HNF4α DNA binding and gene transcription could be explained in part by the physical interaction between these two transcriptional regulators. Collectively, these novel observations, together with the previously known but also validated here interaction between Tcf4 and HNF4α proteins, provide an attractive integrative model explaining how β-catenin contributes to shape zonated gene expression in the liver, as summarized in Fig. 1.
By combining mRNA-seq and ChIP-seq data both in β-catenin- and Apc-deficient mice, this study also provides important information on the transcriptional network governed by β-catenin/Tcf4. A clear picture emerged emphasizing the role of β-catenin on the regulation of genes involved in bile acid, cholesterol, and drug/xenobiotic metabolism, including the key xenobiotic receptor genes AhR and CAR. This role was corroborated by accompanying metabolomic studies in Apc-null mice showing increased bile acid levels and reduced lipogenesis, metabolic alterations reminiscent of findings in CTNNB1-mutated liver tumors. In this respect, the functional antagonism between β-catenin and HNF4α reported here may be of particular relevance to the carcinogenic process. The fact that HNF4α is essential for the preservation of hepatocellular differentiation and quiescence makes this novel crosstalk certainly worthy of further investigations.
The current findings of Gougelet et al. represent indeed a substantial advancement in our knowledge of the molecular effectors of the metabolic mosaic of the liver, placing β-catenin together with Tcf4 and HNF4α together under the spotlight. In view of the transcendent role of β-catenin, it is even more important now to address the old question of “who regulates the regulator?” Characterization of the source and identity of the Wnt agonists driving zonal gene expression, understanding the mechanisms underlying the predominantly periportal expression of Apc, and exploring the crosstalk of Wnt signaling with that of growth factors and other morphogens will surely deliver important fundamental knowledge with high translational potential.
Carmen Berasain, Ph.D.1,2 and Matías A. Avila, B.Pharm., Ph.D.1,2
1Division of Hepatology and Gene Therapy, Centro de Investigación Médica Aplicada (CIMA), Universidad de Navarra, Pamplona, Spain
2Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd), Clínica Universidad de Navarra, 31008 Pamplona, Spain