Making the liver what it is: The many targets of the transcriptional regulator HNF4α

Authors

  • Klaus H. Kaestner Ph.D.

    Corresponding author
    1. Department of Genetics, University of Pennsylvania School of Medicine, Philadelphia, PA
    • Department of Genetics, University of Pennsylvania School of Medicine, 752B Clinical Research Building, 415 Curie Boulevard, Philadelphia, PA 19104–6145===

    Search for more papers by this author
    • fax: 215-573-5892


  • See Article on Page 642

  • Potential conflict of interest: Nothing to report.

The late 1980s witnessed a methodological breakthrough in the study of gene expression in the liver, with the purification and subsequent molecular cloning of a number of liver-enriched transcription factors. These proteins were shown to recognize specific DNA sequences and to regulate a subset of target genes needed for the development and function of the liver.1 Among these factors, the hepatocyte nuclear factor 4 alpha (HNF4α, encoded by the HNF4A gene, also known as nuclear receptor NR2A1), which was cloned by Frances Sladek while a fellow with James Darnell at Rockefeller University, is arguably one of the most important players for adult liver function.2 The study by Bolotin and colleagues in the current issue of HEPATOLOGY makes this point emphatically by identifying more than a thousand HNF4α binding sites, and clarifying the classes of genes regulated by this important factor.3

Abbreviation:

HNF4α, hepatocyte nuclear factor 4 alpha.

Why is HNF4α so important? First of all, HNF4α is the most abundant liver-enriched DNA-binding protein, giving it the potential to regulate hundreds or even thousands of targets. The importance of HNF4α to liver differentiation and function was elegantly demonstrated by cell-type–specific gene ablation of the gene in mutant mice. When HNF4α was deleted in fetal development, hepatocyte precursors completely “forgot” that they were epithelial cells.4 In contrast, when the gene was deleted in the adult liver, defects to lipid metabolism were most prominent.5 Second, HNF4α is a member of the nuclear receptor family of transcription factors, making it a “druggable” molecule just like other famous members of this gene's family such as the glucocorticoid or estrogen receptors. This is true despite the fact that the nature of the native ligand for HNF4α is still controversial, because multiple molecules, ranging from palmitoyl-coenzyme A to linoleic acid, have been proposed to fulfill this function.6, 7 Nevertheless, several companies and academic centers are in the process of developing agonists and antagonists to manipulate HNF4α function as a therapeutic approach. Third, dominant mutations in the HNF4A gene cause both maturity onset diabetes of the young8 and neonatal macrosomia and hyperinsulinemic hypoglycemia,9 due to important functions of the gene in pancreatic β-cells, as supported by cell-type–specific gene ablation in mice.10, 11

So how does HNF4α accomplish all these important tasks in the liver? In the past, several attempts have been made to identify all the targets regulated by this factor. Expression profiling, performed by comparing wild-type and HNF4α-deficient mouse livers, identified many differentially expressed genes.5 However, these studies could not differentiate which of the affected genes were direct targets of HNF4, and which were changed in activity only in response to the severe phenotype caused by HNF4α-deficiency. A few years ago, the identification of HNF4α targets was attempted by a technique termed “ChIP-on-chip”, or chromatin immunoprecipitation (ChIP) followed by microarray hybridization.12 In this technique, DNA-protein complexes in a cell are first stabilized by covalent cross-linking. Then, the DNA is sheared to small fragments, and an antibody specific to the factor of interest, e.g., HNF4α, is employed to precipitate the DNA fragments that are bound by the factor of interest. The DNA fragments are then purified, labeled, and hybridized to an array platform representing thousands of mouse or human promoters. This study suggested that more than 40% of the actively transcribed genes in the liver are bound close to the transcriptional start site by HNF4α. However, these types of studies are dependent on the quality and specificity of the antibody employed, and thus it remained uncertain if all of the sites represented functional targets.

In the study published in this issue of HEPATOLOGY, Bolotin and colleagues utilize a different approach to capture all potential binding sequences for HNF4α. This technology, termed “protein binding microarray”, makes use of synthetic, short oligonucleotides representing thousands of potential HNF4α binding sites. These sequences were preselected based on sequences mined from the literature, and on variations of the previously known HNF4 consensus motif. HNF4α protein synthesized in mammalian cells was then allowed to bind to these DNA sequences, and the bound probes were detected with a new and specific antibody to HNF4α. By using extracts from mock-transfected cells, the authors could ensure that all binding events detected were indeed HNF4α-dependent. Bolotin and colleagues identified more than 1400 strong HNF4α binding motifs in their essay, and determined that these motifs can occur anywhere in the human genome, not just near transcriptional start sites. They also showed by suppression of HNF4α function in hepatoma cells that the expression of genes containing an HNF4α binding motif is indeed affected when HNF4α is inhibited. Most importantly, this study sought to integrate all the disparate data sources into a unified picture of HNF4α gene function in the liver. Remarkably, the authors increased the number of direct and significantly regulated targets of HNF4α by 240. These targets fall into the classes we have come to expect from HNF4α, such as enzymes involved in hepatic intermediary metabolism, cellular homeostasis, and solute transport. Remarkably, an even larger group of targets is annotated to function in signal transduction, stress response, and apoptosis, thus expanding the range of HNF4α-controlled processes.

In summary, the study by Bolotin and colleagues has expanded our knowledge of HNF4α targets, and thus HNF4α function, significantly. Given the facts that HNF4α has been proven by human geneticists to contribute to disease in people, and that HNF4α is a target for drug development, this is an important finding. Knowing the targets of HNF4α is a prerequisite for a full understanding of the metabolic defects that occur in people with HNF4α mutations and will aid in the interpretation of effects—and side effects—that might occur with HNF4α agonists once they enter into clinical testing. Future research will have to address the question if and how HNF4α regulates different targets in the multiple metabolic tissues where it is expressed, such as liver, intestine, and endocrine pancreas, because of course the genes and pathways regulated by HNF4α must differ from cell type to cell type. Understanding the combinatorial pattern of gene regulation that must underlie tissue-specific regulations by these common factors remains a major challenge for molecular biologists today.

Ancillary