In the Zone: How a Hepatocyte Knows Where It Is
Article Outline
See “Liver zonation occurs through a beta-catenin-dependent, c-Myc-independent mechanism” by Burke ZD, Reed KR, Phesse TJ, et al, on page 2316 in the June 2009 issue of Gastroenterology; and “Convergence of Wnt signaling on the HNF4α-driven transcription in controlling liver zonation” by Colletti M, Cicchini C, Conigliaro A, et al, on page 660.
Although 1 hepatocyte might resemble another in standard histologic sections, functionally, these cells can be quite distinct, depending on their location within the hepatic acinus. Dictated by the unique vasculature of the liver, hepatocytes near the portal triad experience a very different environment than those in close proximity to the central vein. The former, or periportal hepatocytes, receive a mixture of blood from the portal vein, which is high in nutrients, and the hepatic artery, which is high in oxygen and circulating hormones. The latter, or perivenous hepatocytes, are exposed to lower oxygen tension as well as nutrient and hormone levels. This portocentral axis sets up the distinct metabolic zonation within the parenchyma upon which liver function relies.
Not all hepatic processes are clearly zonated. Synthesis and secretion of serum proteins and VLDL particles, for instance, appear to occur in all hepatocytes. However, as was discovered >20 years ago, major liver functions are restricted locally.1 Thus, gluconeogenic enzymes are found at their highest levels in periportal hepatocytes. Likewise, the first enzyme of the urea cycle, carbamoyl phosphate synthetase, is expressed in periportal and midzonal hepatocytes, but absent from the last 2 rows of perivenous hepatocytes. By contrast, glutamine synthesis is present in only 1 or 2 layers of hepatocytes surrounding the central vein.2, 3 Thus, the 2 systems of ammonia removal are expressed in complementary locations in the liver. The high concentration of glutamine synthetase near the central vein is thought to remove any remaining toxic ammonia from the blood stream before it rejoins the general circulation.
For 3 decades, scientists have attempted to answer the question of how this fundamental property of the hepatic architecture is established and maintained. Early hypotheses on different embryonic origins of periportal and perivenous hepatocytes were disproven by lineage tracing experiments and the fact that the newborn liver is not zonated.4 Complex hormonal cues were demonstrated to contribute to activation of genes periportally, as exemplified by the case of carbamoyl phosphate synthetase, which is controlled by both glucocorticoid and cAMP signals.5
A major breakthrough in our understanding of hepatic zonation came with the seminal discovery that the Wnt/β-catenin signaling system is critical for the maintenance of hepatic function and regional gene activation.6 Benhamouche et al6 first determined using immunohistochemistry that the stabilized and active form of β-catenin is localized to perivenous hepatocytes. By contrast, the adenomatous polyposis coli (APC) protein, a negative regulator of β-catenin, was expressed in a complementary fashion to activated β-catenin in periportal hepatocytes.6 They next employed conditional gene ablation of APC to activate the β-catenin pathway in the entire liver. Strikingly, this led to an expansion of the expression of glutamine synthetase and ornithine aminotransferase, 2 perivenous enzymes, to all hepatocytes. Conversely, when these authors blocked β-catenin signaling in the liver, perivenous hepatocytes adopted a periportal phenotype.
This concept is taken further and integrated mechanistically with other known regulators of liver function in complementary studies by Colletti et al7 and Burke et al.8 Colletti et al7 take advantage of fetal rodent hepatocytes, termed “resident liver stem cells,” which can be differentiated in culture toward a hepatocyte phenotype. When the authors analyzed these in vitro differentiated cells for markers of periportal and perivenous gene expression, only genes of the periportal phenotype were active. Next, they were in a position to test directly whether activation of the Wnt/β-catenin pathways could redirect the expression profile of the cultured hepatocytes toward the perivenous phenotype, as would be expected from the in vivo studies by Benhamouche et al.6 Indeed, pharmacologic activation of β-catenin led to a complete, and reversible, induction of the perivenous expression program. The authors then investigated the transcriptional machinery downstream of β-catenin that mediates this dramatic change in gene activation. Nuclear β-catenin is a co-activator of transcription factors of the T-cell factor/lymphoid enhancer factor class (Tcf/Lef), and the authors could show that Lef1 interacts with hepatocyte nuclear factor 4-α (HNF4α), a major transcriptional activator in the liver.9 Using chromatin immunoprecipitation assays to evaluate occupancy of regulatory elements in vivo, the authors found that for glutamine synthetase activity, both HNF4α and Lef1 have to be engaged, while HNF4α binding alone seems to be repressive. Conversely, for periportal promoters normally only HNF4α is bound and activates these genes. If, however, β-catenin is artificially activated in periportal hepatocytes, Lef1 replaces HNF4α and leads to gene silencing. The salient features of this model are summarized in Figure 1.
Figure 1.
The molecular regulation of metabolic zonation in the liver. Hepatocytes near the portal vein are exposed to a very different environment and fulfill different functions than those near the central vein. In perivenous hepatocytes, a local Wnt signal of yet unknown nature binds to frizzled receptors (Frz) on the hepatocyte surface and activates the Wnt/APC/β-catenin pathway. Stabilization of β-catenin leads to its nuclear translocation, activation of the transcription factor Lef1, and activation of glutamine synthetase (GS) and cytochrome p450 enzymes (Cyp1A1). Conversely, in periportal hepatocytes, which are not exposed to a Wnt signal, β-catenin is degraded via the APC complex, and periportal genes such as glutaminase 2 (Gls2) are activated by HNF4α, whereas perivenous genes seem to be repressed by the same factor.7 Action of the Wnt/APC/β-catenin pathway in metabolic zonation of the liver does not require activation of c-Myc.8
Burke et al8 focused on the downstream mediators of Wnt signaling in the establishment of metabolic zonation. In addition to the aforementioned stabilization of β-catenin, engagement of the Wnt pathway, or deletion of its inhibitor APC, also activates the transcriptional regulator c-Myc. In fact, in the intestine, ablation of c-Myc reverts the phenotypic consequences of APC deficiency.10 Therefore, Burke et al8 investigated whether β-catenin or c-Myc is the most relevant player in the control of zonation by APC. Using conditional gene ablation they found that removal of c-Myc is of no consequence to metabolic zonation in the liver, while confirming that deletion of APC leads to expansion of the perivenous phenotype to all hepatocytes, with concomitant loss of periportal markers. Thus, a much clearer picture has emerged of how the Wnt/APC/β-catenin pathway controls the metabolic zonation of the liver: A Wnt-signal that must be high near the central vein activates the pathway in perivenous hepatocytes (Figure 1). Stabilization of β-catenin leads to its nuclear translocation and activation of Lef1 target genes, but does not require c-Myc activation. In the nucleus, Lef1 interacts with the nuclear receptor HNF4α to both activate and repress the appropriate targets, and to establish zonal gene expression.
A pressing question is the identity and source of the Wnt ligand. Although multiple Wnt genes have been shown to be expressed in the liver,11 it is not clear which among them plays a role in zonal gene expression. Likewise, the study by Colletti et al7 raises questions regarding the molecular specificity of the proposed regulatory interactions of Lef1 and HNF4α. How is it that, for some genes, binding by HNF4α is not sufficient to activate transcription (such as glutamine synthetase and carbamoyl phosphate synthetase), whereas for many others it is? How does Lef1 replace HNF4α on the promoters of normally periportally expressed genes such as transthyretin and glutaminase 2 in perivenous hepatocytes? It seems, then, that the fascinating problem of metabolic zonation will continue to occupy investigators for some time to come.
References
- . Functional hepatocellular heterogeneity. Hepatology. 1982;2:385–395
- Reciprocal regulation of glutamine synthetase and carbamoylphosphate synthetase levels in rat liver. Biochim Biophys Acta. 1987;908:231–240
- Heterogeneous distribution of glutamine synthetase during rat liver development. J Histochem Cytochem. 1987;35:49–54
- . Positional and developmental regulation of glutamine synthetase expression in mouse liver. Mol Cell Biol. 1988;8:4966–4971
- Glucocorticoid receptor, C/EBP, HNF3, and protein kinase A coordinately activate the glucocorticoid response unit of the carbamoylphosphate synthetase I gene. Mol Cell Biol. 1998;18:6305–6315
- Apc tumor suppressor gene is the “zonation-keeper” of mouse liver. Dev Cell. 2006;10:759–770
- Convergence of Wnt signaling on the HNF4α-driven transcription in controlling liver zonation. Gastroenterology. 2009;137:661–672
- Liver zonation occurs through a beta-catenin-dependent, c-Myc-independent mechanism. Gastroenterology. 2009;137:2316–2324
- Hepatocyte nuclear factor 4alpha controls the development of a hepatic epithelium and liver morphogenesis. Nat Genet. 2003;34:292–296
- Myc deletion rescues Apc deficiency in the small intestine. Nature. 2007;446:676–679
- Wnt'er in liver: expression of Wnt and frizzled genes in mouse. Hepatology. 2007;45:195–204
Conflicts of interest The author discloses no conflicts.
PII: S0016-5085(09)00989-5
doi:10.1053/j.gastro.2009.06.020
© 2009 AGA Institute. Published by Elsevier Inc. All rights reserved.
Refers to article:
-
Convergence of Wnt Signaling on the HNF4α-Driven Transcription in Controlling Liver Zonation
, 18 May 2009
