Novel control of the position-dependent expression of genes in hepatocytes: the Glut-1 transporter.

Abstract

The purpose of this study was to elucidate the mechanisms responsible for the restricted position‐dependent expression of the Glut‐1 glucose transporter in the plasma membrane to a small population of parenchymal cells in rat liver. On the basis of earlier studies in which the erythroid/brain (Glut‐1) glucose transporter isoform was detected by immunofluorescence only in the last two hepatocytes surrounding the terminal hepatic venule,1 investigators in Jorge Gumucio's laboratory at the University of Michigan sought to determine the molecular basis for the observation. To study the expression of this transporter in individual hepatocyte populations (periportal vs perivenous), the authors used the digitonin/collagenase cell isolation method. Their conclusions rest on two assumptions: (1) the modified digitonin/collagenase method of hepatocyte isolation used in the studies adequately separates periportal and perivenous cells, and (2) the Glut‐1 protein distribution in isolated hepatocyte populations reflects the normal localization in liver tissue. The authors go to great lengths to demonstrate that these two assumptions hold true.Hepatocyte populations enriched in perivenous or periportal cells were obtained by the digitonin/collagenase method2,3 in which digitonin is perfused either antegrade or retrograde to selectively destroy periportal or perivenous cells, respectively, followed by the subsequent isolation of the remaining population of cells by collagenase digestion. In this report, the authors describe a refinement of the original technique in which an additional Percoll gradient step is used to separate damaged from healthy hepatocytes. Once the damaged hepatocytes were isolated, hepatocyte population purity was assessed by Northern blot analysis using albumin, P450 isozymes IIB1 and IIB2, and glutamine synthetase probes. The acinar (cell plate) localization of each of these gene products is well established, and results demonstrated population purity, with glutamine synthetase messenger RNA (mRNA) present in the perivenous isolates and absent in periportal isolates. Thus, the first assumption was correct. Concurrently, Glut‐1 mRNA was probed and, surprisingly, found to be present in approximately equal levels in both cell populations despite the restricted plasma membrane localization to only the terminal two perivenous hepatocytes.To ensure that the second assumption held true, the authors used laser scanning confocal microscopy to verify that the Glut‐1 transporter was expressed in the plasma membrane only in the one or two hepatocytes surrounding the sinusoidal outflow. Diffuse intracellular fluorescence that was more accentuated near the plasma membrane was detected in the periportal cells in the liver slices, an observation that would corroborate both the presence of Glut‐1 mRNA in periportal cells and the subsequent subcellular localization of the Glut‐1 protein to the Golgi fraction. Additionally, because Glut‐1 has been shown to be a “stress” protein, the authors demonstrated that the differential localization of Glut‐1 was not a cell isolation‐induced artifact by measuring the levels of glucose‐regulated protein 78 in parallel with Glut‐1. They also ruled out a contribution to Glut‐1 detection by bile duct cell contamination by the absence of cytokeratin 19 in isolated parenchymal cell populations. Collectively, these observations supported the second assumption that the protein pattern for Glut‐1 observed in isolated hepatocyte populations reflects that seen in native liver tissue.Finally, by probing Western blots of subcellular fractions from each of the populations with anti‐Glut‐1 antibodies, the authors confirmed the data from the laser‐scanning confocal microscopy. The Glut‐1 protein was detected in the plasma membrane fraction only in perivenous cells, whereas it was restricted to the low‐density microsomal (Golgi‐enriched) fraction (200,000 x g pellet) in periportal cells. The authors conclude that the differential insertion (ie, a posttranslational event) of the Glut‐1 transporter into the plasma membrane of only the terminal two perivenous hepatocytes is responsible for the immunohistologic observations in previous studies.