0013-7227/91/1294-1933$03.00/0 Endocrinology Copyright © 1991 by The Endocrine Society
Vol. 129, No. 4 Printed in U.S.A.
Expression of the Low Km GLUT-1 Glucose Transporter Is Turned on in Perivenous Hepatocytes of InsulinDeficient Diabetic Rats* MICHAEL TALt, BARBARA B. KAHN, AND HARVEY F. LODISH Whitehead Institute for Biomedical Research (M.T., H.F.L.), Cambridge, Massachusetts 02142; the Department of Biology, Massachusetts Institute of Technology (H.F.L.), Cambridge, Massachusetts 02139; and the Charles A. Dana Research Institute and Harvard-Thorndike Laboratory of Beth Israel Hospital, Department of Medicine, Beth Israel Hospital and Harvard Medical School (B.B.K.), Boston, Massachusetts 02215
ABSTRACT. In normal fed rats the low Km glucose transporter GLUT-1 is expressed only in one row of hepatocytes immediately surrounding a terminal hepatic venule, while the high Km GLUT-2 is expressed in every hepatocyte. Previously, we showed that additional perivenous hepatocytes express GLUT-1 in fasting animals. In diabetes, as in starvation, the liver functions to release glucose into the circulation, but unlike starvation, circulating extracellular glucose is high in diabetes. By immunofluorescence and Western blotting we studied whether glucose or insulin is the primary extracellular signal for inducing GLUT-1 expression in hepatocytes. We observed that streptozocin-induced diabetes causes induction of GLUT-1
expression in the plasma membrane of hepatocytes within four cell rows of a terminal hepatic venule; GLUT-2 expression is unaltered. Chronic insulin treatment of diabetic rats reduces the number of rows of hepatocytes expressing GLUT-1 from ~four to ~two. In contrast, chronic insulin infusion into nondiabetic rats does not affect the number of hepatocytes expressing GLUT-1. Thus, both fasting and diabetes induce GLUT-1 expression in specific hepatocytes that normally do not express this gene. This induction correlates with low insulin levels in the blood, and not with circulating glucose levels. {Endocrinology 129: 1933-1941,1991)
T
HE LIVER functions as a glucostat by removing excess glucose after a meal and releasing glucose between meals; after a meal, 50% of blood glucose is absorbed during one passage through a liver acinus (1). An acinus is a mass of liver parenchymal cells that is perfused in the same direction. Blood enters a liver acinus from the intestine through a portal venule and mixes in a sinus with blood from a hepatic arteriole. Blood diffuses along several rows of hepatocytes and then exits from the acinus into a terminal hepatic venule (2). GLUT-2 is the predominant glucose transporter in the liver and is expressed in the basolateral blood-facing plasma membrane of every hepatocyte (3-5). Unlike that in muscle and adipocytes, hepatic glucose transport is not affected by insulin (6). The high Km for glucose of Received April 16,1991. Address all correspondence and requests for reprints to: Dr. Michael Tal, Whitehead Institute for Biomedical Research, 9 Cambridge Center, Cambridge, Massachusetts 02142. * This work was supported by NIH Grants GM-40916 and HL-41484 (to H.F.L.) and NIH Physician Scientist Award AG-00294 and Juvenile Diabetes Foundation Grant 189833 (to B.B.K.). t Supported by a fellowship from the European Molecular Biology Organization.
GLUT-2 (15-20 mM) allows glucose influx or efflux to respond passively to changes in extracellular glucose from the 5-mM resting state (7). Control of glucose metabolism in hepatocytes involves both short and long term effects of hormones, primarily insulin and glucagon. In poorly controlled type I diabetes, insulin levels are low, and the control of liver, muscle, and fat glucose metabolism is severely disturbed. Paradoxically, in many forms of diabetes, while blood glucose is high, the liver is induced to synthesize glucose via gluconeogenesis and to export it. This worsens the diabetic state by further raising the blood glucose level. In livers from both diabetic and starved rats there is an over 200% increase in the activity of key gluconeogenic enzymes, which are expressed predominantly by hepatocytes in the periportal zone (glucose 6-phosphatase, fructose 1,6-bisphosphatase, and phosphoenolpyruvate carboxykinase). In contrast, in the perivenous zone the activity of key glycolytic enzymes (glucokinase and pyruvate kinaseiJ is reduced only 25% (8-17). In contrast to GLUT-2, the GLUT-1 glucose transporter isoform is localized in rat liver to a subset of hepatocytes that form the first row of cells around each terminal hepatic venule (18). GLUT-1 is a glucose trans-
1933
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
1934
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
porter with a low Km for glucose (1-2 HIM) (19). GLUT1 is expressed mainly by cells oxidizing glucose, where the direction of glucose transport is out to in. On the other hand, expression of GLUT-2 is found in all hepatocytes, including those active in gluconeogenesis, where glucose transport is predominantly from the cytoplasm to the extracellular space. In the kidney there is a similar correlation between the glycolytic activities of the different segments of the nephron and the relative expression of GLUT-1, and a correlation between expression of GLUT-2 and the gluconeogenic activity of cells (4, 20). After 3 days of starvation, there is a 3- to 4-fold elevation of expression of GLUT-1 mRNA and protein in the liver as a whole (18). This correlates well with expression of GLUT-1 in more hepatocytes than in the fed state, namely those forming the first three or four rows around the terminal hepatic venules. Importantly, starvation-induced expression of GLUT-1 on the plasma membrane of these additional hepatocytes is lost within 3 h of glucose injection. In streptozocin-induced diabetes in the liver as a whole, the level of GLUT-1 mRNA is increased, and insulin treatment restores GLUT-1 mRNA levels to normal (5). However, consistent changes in GLUT-1 protein levels were not seen in the few animals studied. Here we studied GLUT-1 expression by immunofluorescence microscopy, defining cells expressing this glucose transporter in the plasma membrane, the site at which glucose transport occurs. We induced diabetes by a single streptozocin injection and studied rats over a course of diabetes and insulin treatment. In addition, we studied insulin-infused normal rats to address the question of whether abnormal levels of glucose or insulin affect GLUT-1 expression. We show that in diabetic rats GLUT-1 is expressed in three or four rows of hepatocytes around each terminal hepatic venule. The level of GLUT-2, as visualized by immunofluorescence, is unaffected, and GLUT-2 remains expressed in every hepatocyte. Chronic insulin infusion of diabetic rats reduces GLUT-1 expression to ~1.8 rows of cells around the terminal hepatic venules. In contrast, insulin infusion of normal rats does not change GLUT-1 expression. Taking these results together with those on GLUT-1 expression in starved rats, we conclude that catabolic states associated with low levels of insulin are accompanied by induction of GLUT1 expression in perivenous hepatocytes. This may be due to the lack of insulin directly or be a metabolic consequence of low insulin levels. Importantly, GLUT-1 expression in liver does not correlate with changes in glucose levels per se. In contrast, hepatic GLUT-2 level and localization appear to be relatively unaffected by starvation or diabetes.
Endo • 1991 Vol 129 • No 4
Materials and Methods Experimental design Male Sprague-Dawley rats (CD strain, Charles River Breeding Laboratories, Wilmington, MA), weighing 150-165 g, were maintained with ad libitum feeding of standard chow for several days before initiation of the experimental period. Rats were divided into the following experimental groups: standard chow, streptozocin injection, streptozocin injection followed by insulin infusion, sucrose-enriched chow, and sucrose-enriched chow followed by insulin infusion. For studies of diabetes and insulin treatment, rats were injected ip with 80-85 mg/kg anhydrous streptozocin and citric acid (Zanosar, Upjohn, Kalamazoo, MI) reconstituted with 0.9% saline. Rats were maintained diabetic for 7-14 days (postprandial blood glucose, >20 HIM). After 7 days of hyperglycemia, some rats were treated for 7 days with insulin. After 13-14 days of hyperglycemia, rats were treated for 3, 6, or 24 h with insulin. Rats treated for 3-6 h received one sc injection of 100 U human crystalline insulin (Eli Lilly Co., Indianapolis, IN). Rats treated for 1 day received a combination of crystalline and NPH insulins by sc injection every 12 h. For 7 days of treatment of diabetic and normal rats, 500 U pork crystalline insulin (Lilly), prepared as described by Bringer et al. (21) to prevent aggregation, was infused via sc osmotic minipumps (Alzet 2001, Alza, Palo Alto, CA) at 4 U/ day. Normal rats infused with insulin were allowed free access to sugar cubes to avoid severe hypoglycemia. A corresponding control group was fed sugar cube supplements. Blood glucose was measured in whole blood samples from the tail vein with Chemstrips bG (Bio-Dynamics, Indianapolis, IN), read in an AccucheckbG reflectance meter (Bio-Dynamics). Animals were killed by CO2 inhalation followed by decapitation, and portions of the liver were rapidly excised and immediately fixed and processed for microscopy, For immunoblots, the livers were frozen in liquid N2 and stored at 70 C before lysates were prepared. Serum was frozen for determination of insulin levels by standard RIA. Detection of GLUT-1 by immunoblotting Liver lysates were prepared by homogenizing frozen liver with a polytron (Brinkmann Instruments, Westbury, NY) in a solution containing 5% sodium dodecyl sulfate, 80 mM TrisHC1 (pH 6.8), 5 mM EDTA, 1 mM phenylmethylsulfonylfluoride, and 0.2 mM iV-ethylmaleimide. The protein content of the lysates was measured by the BCA protein assay (Pierce, Rockford, IL), using BSA as the standard. Proteins were separated by electrophoresis on 10% sodium dodecyl sulfate-polyacrylamide gels and electrophoretically transferred to nitrocellulose filters overnight at 4 C. Uniform loading and transfer of protein from all lanes of the gel were assessed by staining the filters with Ponceau S (Sigma, St. Louis, MO) and the gel with Coomassie blue. Immunoblotting was carried out, as previously described (22), with a polyclonal antiserum (10 /ug/ml) prepared against a synthetic peptide consisting of the 16 carboxy-terminal amino acids (477-492) of GLUT-1 (a generous gift from Dr. B. Thorens, Whitehead Institute, Cambridge, MA), followed by [125I]protein-A and in some experiments also by [125I] goat antirabbit immunoglobulin G (IgG; New England Nuclear,
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT Boston, MA). Filters were autoradiographed using XAR-5 film (Eastman Kodak Co., Rochester, NY) and intensifying screens at -70 C. GLUT-1 protein levels were quantitated by scanning densitometry, using a GS 300 densitometer (Hoefer Scientific Instruments, San Francisco, CA) and the GS 350 computer program (Hoefer Scientific Instruments). Statistical analyses of densitometry scans were performed on the Beth Israel Hospital analyzer system. Differences between experimental groups were determined using analysis of variance in conjunction with the Newman-Keuls test. When comparing experimental groups to controls, a one-sample comparison test was used, since these results are evaluated as a percentage of the control value due to the arbitrary nature of optical density units. Immunofluorescence Liver fragments were fixed by immersion in paraformaldehyde-lysine-periodate fixative (23) overnight at 4 C, then kept at 4 C in PBS until use. For preparing S-nm frozen sections, tissue fragments were equilibrated for several hours at 4 C in PBS containing 0.6 M sucrose and 0.02% sodium azide, subsequently embedded in O.C.T. compound (Miles, Elkhart, IN), frozen in 2-methylbutane in liquid nitrogen for 10 sec, and kept at -70 C until sectioning in a Reichert Frigocut cryostat. Sections were placed on polylysine-coated glass slides, air dried, and kept at -70 C until staining. Tissue sections were incubated for 10 min in 1% BSA-PBS. Sections were then incubated for 40 min with one of the following antisera (diluted 1:100 in 1% BSA/PBS): affinity-purified sera raised against the COOH-terminus of GLUT-1 (original concentration, 0.32 mg/ml) (24) or against a peptide corresponding to the COOHterminous of GLUT-2 (3) (both generous gifts from Dr. B. Thorens, Whitehead Institute, Cambridge, MA). Thereafter, sections were washed three times for 5 min each with PBS. The sections were incubated with fluoresceinated goat antirabbit IgG (Accurate Chemical and Scientific Corp., Westbury, NY) diluted 1:100 in 1% BSA-PBS. The sections were washed, as described above, and mounted in 60% glycerol, 2% n-propyl gallate, and 0.2 M Tris-HCl (pH 8.1). Sections were observed and photographed with Kodak TMAX film on a Zeiss Photomicroscope III (New York, NY). For double staining, the coverslips were removed after photography, and the sections were washed for 5 min in PBS. The sections were then incubated for 1 h with an affinity-purified goat antirabbit IgG (Cappel, Malvem.PA; 0.1 mg/ml) in 1% BSA/PBS. Thereafter, the sections were washed four times for 15 min each with PBS. The second antibody (anti-GLUT-2) was applied as described for the initial staining, and detection was performed with rhodamine-conjugated goat antirabbit IgG (Cappel) diluted 1:100 in 1% BSA-PBS. The statistical analysis of the number of rows of hepatocytes expressing GLUT-1 around the terminal hepatic venules was performed by pairwise Student's t test on a Microsoft Excel spreadsheet using a Macintosh computer.
Results Immunostaining of GLUT-1 in liver sections from diabetic rats and diabetic rats infused with insulin Liver sections were stained for GLUT-1 using a specific antibody raised against a peptide corresponding to
1935
its carboxy-terminus. This antibody does not cross-react with GLUT-2, the major glucose transporter expressed in the liver (5). Figure 1 depicts the expression of GLUT1 in livers of insulin-responsive diabetic rats. In normal rats, GLUT-1 is localized to the first row of hepatocytes around the terminal hepatic venule (Fig. 1A) and nowhere else (18). In diabetic rats 7 and 14 days after streptozocin injection, GLUT-1 is expressed in several rows of hepatocytes around the terminal hepatic venule (Fig. 1, B and D, respectively). GLUT-1 is not expressed in other hepatocytes, such as those surrounding the portal triad, in livers from both diabetic (Fig. IF) and normal (18) rats. By analysis of more animals we quantified the number of rows of hepatocytes around the terminal hepatic venules that express GLUT-1 (3.1 ± 0.4 rows of hepatocytes 7 days after streptozocin injection and 4.3 ±1.3 rows after 14 days, compared to 1.3 rows in normal rats; Table 1). The effect of streptozocin was largely reversed in diabetic rats treated with insulin for 7 days (Fig. 1C and Table 1); only 1.8 ± 0.9 rows expressed GLUT-1, which was different from the number in both control and streptozocin-treated rats (P < 0.05). Insulin treatment of diabetic rats for short times [3 and 6 h (three rats per time point; data not shown) and 24 h (Fig. IE and Table 1)] did not significantly affect the expression of GLUT-1. At these early time periods of insulin injection, serum insulin levels were substantially greater than control values, and expression of GLUT-1 was unaltered from that in untreated 14-day diabetic rats. Figure 2 shows that insulin infusion of a normal rat for 7 days has no effect on expression of GLUT-1, although it lowers blood glucose from 115-130 mg/dl (control) to 40 mg/dl (Table 1). Insulin levels were at least 10 times greater than normal in insulin-infused rats, as we previously reported (25). In the insulininfused normal rats (Fig. 2A) and in the normal control rats (Fig. 2B), only one row of hepatocytes around the terminal hepatic venule expressed GLUT-1 (see Table 1). By double staining liver sections from diabetic rats for GLUT-1 and GLUT-2, we showed that cells around the terminal hepatic venules express both glucose transporters on the plasma membrane (Fig. 3, A and B). As was shown previously, both GLUT-2 (4) and GLUT-1 (18) are expressed only on the basolateral or blood-facing surface. Although not shown here, we immunostained GLUT-2 liver sections from rats in all of the metabolic states summarized in Table 1 and found no difference in the staining intensity or the staining pattern. In summary, in diabetes, GLUT-2 is expressed normally on the blood-facing surface of every hepatocyte, while expression of GLUT-1 protein is induced in hepatocytes that are in two to four rows around each terminal hepatic
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
1936
Endo • 1991 Vol 129 • No 4
FIG. 1. Immunofluorescence localization of GLUT-1 in livers from diabetic rats and diabetic rats treated with insulin. Liver sections are from a normal rat (A), a diabetic rat 7 days after streptozocin injection (B), a diabetic rat 7 days after streptozocin injection and 7 days of insulin treatment (C), a diabetic rat 14 days after streptozocin injection (D), a diabetic rat 13 days after streptozocin injection and 1 day of insulin treatment (E), and a diabetic rat 14 days after streptozocin injection (F). The GLUT-1 protein is localized to the plasma membrane of hepatocytes that are in one row (A) or several rows (B-E) around the terminal hepatic venule (T). No GLUT1 expression is observed (F) in hepatocytes around the portal triad (P). Bar represents 30 fim.
TABLE 1. GLUT-1 expression in hepatocytes surrounding terminal hepatic venules
Mean no. of rows Mean venule diameter (/*m) No. of venules Blood glucose (mg/dl) BW(g) Serum insulin levels (/uU/ml) No. of animals
Normal rat
Diabetes 7 d
Diabetes 7 days, insulin 7 d
Diabetes 14 days
Diabetes 13 days, insulin 1 d
Normal sucrose
Normal sucrose, insulin 7 d
1.3 ± 0.4 56 ±23 20 132 ± 7 297 ± 5 72±2 C 15(6)
3.1 ± 1.7" 63 ±35 11 467 ± 17 255 ± 7 48 ± 7°'c 15(5)
1.8 ± 0.9* 48 ±36 52 71 ± 7 265 ± 8 178 ± 28°-c 15(9)
4.3 ± 1.3° 52 ±36 39 580 ± 37 182 ± 10 ND 12(8)
3.9 ± 1.7° 65 ± 3.6 56 236 ± 61 212 ± 15 >200a'c 5(5)
1.4 ± 0.4° 50 ±34 25 115 ± 14 275 ± 5 ND 2(2)
1.4 ± 0.3° 61 ±37 17 40 ± 6 280 ± 9 ND 3(3)
Livers were sectioned from Sprague-Dawley rats that were in the following metabolic states: normal, diabetic 7 and 14 days after streptozocin injection (diabetes 7 d, diabetic 14 d), diabetic rats after insulin infusion for 1 or 7 days (diabetes 7 d, insulin 7 d; diabetes 13 d, insulin 1 d), normal sucrose-supplemented rats (normal sucrose), and normal sucrose-supplemented after 7 days of insulin infusion (normal sucrose, insulin 7 d ). Sections were immunostained for GLUT-1 in the region around the hepatic venules. The first row indicates the mean number (±SD) of rows of hepatocytes surrounding a venule that express GLUT-1; the mean diameter of venules is in the second row, and the number of venules tested is in the third row. Blood glucose levels (±SEM) and body weight (±SEM) for all rats studied are in the fourth and the fifth rows. Serum insulin levels (±SEM) are in the sixth row, and were performed using a human insulin standard curve because human insulin was administered to treated rats. The last row shows the number of rats used in all experiments (i.e. immunofluorescence and immunoblotting), while in parentheses are the number of rats used for immunofluorescence. ND, not done. " P < 0.05 vs. normal rats (by Student's t test). 6 P < 0.05 vs. diabetic 14 d (by Student's t test). c Use of the human standard curve accounts for values in control and diabetic rats which are slightly higher than previously reported (25).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
FIG. 2. Immunofluorescence localization of GLUT-1 in livers from insulin-infused rats. The liver sections are from a normal rat infused for 7 days with insulin (A) and from a normal rat (B). In both cases GLUT-1 protein is localized to the plasma membrane of hepatocytes that are in one row around a terminal hepatic venule (T). Bar represents 30/xm.
venule. Insulin is a negative regulator for GLUT-1, but not GLUT-2, expression. Immunoblotting analysis of GLUT-1 expression in livers from diabetic rats Immunoblotting of gel-separated proteins by antisera to GLUT-1 quantifies the total amount of this transporter, both on the plasma membrane and in internal membranes. Autoradiograms, as represented by the one in Fig. 4, A and B, were quantified in Fig. 5 and showed that after 7 or 14 days of diabetes the overall level of hepatic GLUT-1 was increased 2-fold (P < 0.05). This agrees with the increased number of cells expressing GLUT-1 (Fig. 1). Strikingly, when 7-day diabetic rats were infused with insulin for 7 days, the total level of GLUT-1 tended to decrease, but was not statistically different from 7-day diabetic levels (Figs. 4A and 5), while the immunofluorescence studies showed a significant (P < 0.05) over 2-fold-decrease in the number of cells expressing GLUT-1 on their plasma membrane. We do not know the reason for this difference, but we hypothesize that in insulin-infused diabetic rats, GLUT-1
1937
FIG. 3. Double immunofluorescence detection of GLUT-1 and GLUT2 in liver from a diabetic rat. The liver section is from a diabetic rat 7 days after the injection of streptozocin and is stained with a specific antibody for GLUT-1 (A) or GLUT-2 (B). A, GLUT-1 protein is localized to the plasma membrane of hepatocytes that are in several rows around a terminal hepatic venule (T), while GLUT-2 is expressed in every hepatocyte, including those that express GLUT-1 (B). Bar represents 30 ^m.
in some hepatocytes may be in internal membranes, where the carboxy-terminal epitope is inaccesible to our antibody and, therefore, not visible by immunofluorescence. We also studied by immunobloting the short (324 h) term effect of insulin on diabetic rats; by immunofluorescence these rats have ~four rows of hepatocytes expressing GLUT-1 around each terminal hepatic venule. Figure 4C shows that the total hepatic levels of GLUT-1 protein remain as high as those in nontreated diabetic rats, and were significantly higher than those in normal control rats. Immunoblots of liver lysates from normal control rats and normal rats after 7 days of insulin infusion showed the same amount of GLUT-1 (Fig. 4D) and were consistent with the immunofluorescence data, indicating no change in the number of cells expressing GLUT-1 (Table 1). In the same experiment, diabetic rats that were infused with insulin had 2- to 3-fold higher levels of GLUT1 protein than control rats, with or without insulin
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
1938
Endo•1991 Vol 129 • No 4
6LUT-1 45 kDa
B
D
FIG. 4. Immunoblotting analysis of GLUT-1 protein in livers from diabetic or normal rats before and after insulin treatment. Cell lysates (150 ng protein) from liver tissues of rats under the following metabolic conditions were used: normal rats (normal), diabetic rats 7 (7 d diabetes) or 14 days after streptozocin injection (14 d diabetes), diabetic rats 7 days after streptozocin injection and after 7 days of insulin infusion (7 d diabetes 7 d insulin), 13- to 14-day diabetic rats 3, 6, and 24 h after insulin injection (13-14 d diabetes + insulin 3h, 6h, and 24h), and sucrose-supplemented control rats (normal/sucrose) or after 7 days of insulin infusion in normal rats (normal/sucrose/7 d insulin). Each lane represents a different rat. Lysates were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (10% acrylamide) and transferred to a nitrocellulose filter. The filter was incubated for 1 h with a polyclonal serum raised against a peptide corresponding to the COOH-terminal amino acids of GLUT-1 (24), and bound antibody was detected with [125I]protein-A and, in some cases, [125I]goat antirabbit IgG. Exposure time with Kodak XAR-5 film at —70 C using an intensifying screen was 1 day for A and D, and 2 days for B and C. The arrowhead on the left represents the expected position of the GLUT-1 polypeptide.
infusion (Fig. 4D). This agrees with the immunofluorescence data of Fig. 1 and Table 1, showing that in livers from diabetic animals infused with insulin for 7 days the number of cells expressing GLUT-1 is significantly higher than normal (1.8 rows of cells vs. 1.4 rows).
300
Discussion
normal
7d
14d
7d
diabetes
diabetes
diabetes 7 d insulin
FIG. 5. Relative expression of GLUT-1 in livers from diabetic or normal rats before and after insulin treatment. Immunoblots were from nine normal rats (normal), six rats 7 days after streptozocin injection (7 d diabetes), seven rats 14 days after streptozocin injection (14 d diabetes), and eight rats 7 days after streptozocin injection and 7 days of insulin infusion (7 d diabetes 7 d insulin). Data are presented as a percentage of control expression and are from densitometric scans of immunoblots similar to the one presented in Fig. 4. All experimental groups show significant differences from the control (P < 0.05), but not from each other.
The expression of the low Km glucose transporter GLUT-1 is increased 2- to 3-fold in livers from both starved and diabetic rats. Strikingly, in both cases GLUT-1 expression is not uniform in all hepatocytes; rather, it is the number of cells expressing GLUT-1 that is increasing. We used a morphometric assay which measures the number of rows of hepatocytes surrounding each hepatic venule that express GLUT-1 (18). Studying many venules in multiple animals in different metabolic states gives us a reliable assay for GLUT-1 expression, but not for expression of GLUT-2, the high Km glucose transporter which is expressed in every hepatocyte. We conclude that the increased GLUT-1 expression in diabetes involves an induction of GLUT-1 mRNA in additional hepatocytes surrounding each hepatic venule. The increase we observed in total GLUT-1 expression in livers from diabetic rats correlates with the number of cells expressing GLUT-1. In control rats GLUT-1 expression is confined to 1.3 rows of hepatocytes around
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
1939
Liver acinus FlG. 6. GLUT-2 and GLUT-1 expression in liver acini from normal and diabetic rats. In a liver acinus the blood (small dots) enters through the portal triad, which consists of a bile duct (B), arteriole (A), and portal venule (PV). The blood is mixed in the liver sinus and diffuses (arrowheads) toward the terminal hepatic venule (THV) and exits the liver. GLUT-2 is expressed in every hepatocyte (open circles, including closed and dotted circles). Normally, GLUT-1 is expressed only in hepatocytes in the first row around the terminal hepatic venule (closed circles). In diabetic or starved rats GLUT-1 expression is switched on in three or four rows of hepatocytes around the terminal hepatic venule (dotted circles).
THV
< Terminal Hepatic Venule
PV Portal Venule
each hepatic venule; at 7 and 14 days of diabetes this is increased to three or four rows of hepatocytes around each hepatic venule. Chronic insulin treatment of diabetic rats reduces the number of hepatocytes that express GLUT-1 to —two rows of hepatocytes around each venule. This change in the number of hepatocytes expressing GLUT-1 is significantly (P < 0.05) different from that in both control and 14-day diabetic rats. The reduction we found in the number of cells expressing GLUT-1 is accompanied by a slight, although not statistically significant, reduction in the total amount of GLUT-1 protein measured by immunoblotting of liver lysates. In the insulin-infused animals, GLUT-1 protein could be in internal hepatocyte membranes, which may not be detected by immunofluorescence. We observed a similar phenomenon in livers from starved and refed rats (18). In that case, total liver GLUT-1 protein was increased, but there was no change in the number of hepatocytes expressing GLUT-1 on their plasma membrane. During the Western blotting procedure GLUT-1 protein is denatured and dissociated from other proteins and lipids, while in sections, GLUT1 is fixed, and its COOH-terminus, the site of binding of our antisera, might be bound to other proteins or otherwise masked. In the case of starved and refed rats, we attemped to dissociate such possible epitope masking by sodium dodecyl sulfate treatment of sections, but there was no visible change in GLUT-1 immunostaining (Tal, M., and H. F. Lodish, unpublished). In a previous report (5) GLUT-1 mRNA in liver from diabetic rats was increased 2.4-fold compared to that in control rats, an observation in accord with the results
—
shown here. However, in that early study (5), the level of GLUT-1 protein was not increased statistically, in contradiction to our present data; the difference may result from the smaller number of rats analyzed for GLUT-1 protein in this early report. Also, immunoblotting was performed on total liver membranes (5), while here we used total liver cell lysates, which represent all glucose transporters in the tissue. However, the quantitation of GLUT-1 should be the same using both preparations, since glucose transporters are integral membrane proteins, unless there is selection of specific membranes using that protocol (5). A short period (24 h) of insulin treatment of diabetic rats did not significantly reduce GLUT-1 expression, nor did treatment between 3 and 24 h (data not shown). Therefore, the effects of insulin are not directly on GLUT-1 protein and probably involve other insulindependent mechanisms, such as the subcellular localization or stability of GLUT-1. The effects of insulin could be indirect, mediated by some other hormone or metabolite which changes in both starvation and insulin-deficient diabetes. A common feature of diabetes and starvation is the low level of circulating insulin. Table 2 summarizes data indicating a correlation between low circulating insulin and induction of GLUT-1 expression in additional perivenous hepatocytes, but no correlation with circulating glucose. Noteworthy are findings that glucose injection of starved animals and chronic insulin infusion of diabetic ones reduce the number of perivenous hepatocytes that express GLUT-1. In addition, GLUT-1 is expressed only in the first row of hepatocytes around each terminal
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
1940
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
TABLE 2. Low insulin levels correlate with induction of GLUT-1 protein, and normal and high insulin levels correlate with normal expression of GLUT-1
Fed Starved Starved/glucose-injected Streptozocin-diabetes Streptozocin-diabetes/insulin Normal/insulin-infusion fa/fa Zucker rats
Glucose levels
Insulin levels
No. of rows expressing GLUT-1
Normal Low High High Normal/low Low Normal
Normal Low High Low High High High
1.3° 3.4° 1.4° 3-46 1.86 1" lc
" Data from Ref. 18. 4 Data presented here. c Data from Tal, M., and Lodish, H. F., unpublished observation.
hepatic venule of liver tissue from fa/fa Zucker rats (Tal, M., and H. F. Lodish, unpublished observation). The Zucker fa/fa rat is a model of genetic obesity, with naturally occurring hyperinsulinemia and normal blood glucose (26). Thus, in all cases studied, the first row of hepatocytes around each venule continues to express GLUT-1. Expression in these cells, in particular, is not affected by hyperinsulinemia. GLUT-2 is the predominant hepatic glucose transporter. By immunofluorescence we could see no difference in GLUT-2 expression in hepatocytes in the metabolic states studied here, nor was there any change in GLUT-2 expression in starved or starved/glucose-injected rats or in hepatocytes from fa/fa Zucker rats (18) (Tal, M., and H. F. Lodish, unpublished observations). We detect by immunofluorescence only plasma membrane expression of GLUT-2; we have seen no morphological changes in the localization of GLUT-2 within the liver (periportal vs. perivenous). We do feel that a 2-fold difference in GLUT-2 expression levels would be detected by our immunofluorescence studies. Our immunofluorescence results are in agreement with those of a previous report, which showed by Western and Northern blotting that GLUT-2 levels were unchanged in diabetes (5). These results contrast with another study which showed that GLUT-2 levels were induced by diabetes and reduced to normal levels after chronic insulin treatment (27). The differences may be due to difference in the strain of rats, treatment protocols, and/or procedures for membrane preparation. Finally, our studies add to the increasing body of literature, summarized in the Introduction, that hepatocytes are heterogeneous in their expression of proteins required for gluconeogenesis and glycolysis. In normal fed rats GLUT-1 expression occurs in those hepatocytes most active in glycolysis. In other tissues there is a correlation of GLUT-1 expression with utilization of
Endo• 1991 Voll29»No4
glucose as an energy substrate (see Introduction). This suggests that in diabetes, the additional perivenous hepatocytes expressing GLUT-1 are also transporting glucose inward and functioning to reduce the high extracellular glucose. Certainly, hepatocytes are heterogeneous in expression of glucose transporter isoforms and in their responses to the presence or lack of insulin. Figure 6 summarizes our current knowledge of the heterogeneity of liver hepatocytes with respect to glucose transporter expression. The parenchymal cells in each acinus can be divided into three groups: 1) hepatocytes nearest the portal venule and hepatic arteriole that express only GLUT-2, regardless of any alteration in metabolic state tested, including those cells most active in gluconeogenesis; 2) hepatocytes within one row of each hepatic venule that express both GLUT-2 and GLUT-1, regardless of any alteration in metabolic state tested; this group of hepatocytes is most active in glucose oxidation; and 3) hepatocytes in rows 2-4 around each terminal hepatic venule that always express GLUT-2. Expression of GLUT-1 in these cells occurs in at least two conditions in which the blood insulin level is low: in streptozocininduced diabetes and starvation. In diabetes, the increased expression of GLUT-1 may facilitate entry of glucose into more hepatocytes, and thus lower blood glucose. Unlike under other conditions, specifically starvation, expression of GLUT-1 in these additional cells might allow them to release glucose at an increased rate into the blood.
Acknowledgments We thank Dr. B. Thorens for GLUT-1 and GLUT-2 antibodies, A. Rosen for excellent technical assistance, and Dr. Amihod Dotan, Tel-Aviv University, for statistical evaluation of our immunofluorescence results.
References 1. Thurman RG, Kauffman FC, Jungermann K 1986 Regulation of hepatic glucose metabolism. Plenum Press, New York and London 2. Rappaport AM, Borowy ZJ, Lougheed WM, Lotto WN 1954 Subdivision of hexagonal liver lobules into a structural and functional unit: role in hepatic physiology and pathology. Anat Rec 119:1134 3. Thorens B, Sarkar HK, Kaback HR, Lodish HF 1988 Cloning and functional expression in bacteria of a novel glucose transporter present in liver, intestine, kidney, and beta-pancreatic islet cells. Cell 55:281-290 4. Thorens B, Cheng ZQ, Brown D, Lodish HF 1990 Liver glucose transporter: a basolateral protein in hepatocytes and intestine and kidney cells. Am J Physiol 2:1 5. Thorens B, Flier FS, Lodsh HF, Kahn BB 1990 Differential regulation of two glucose transporters in rat liver by diabetes and insulin treatment. Diabetes 39:712-719 6. Williams TF, Exton JH, Park CR, Regen DM 1968 Stereospeclfic transport of glucose in the perfused rat liver. Am J Physiol 215:1200-1209 7. Craik JD, Elliott KRF 1979 Kinetics of 3-O-methyl-D-glucose transport in isolated rat hepatocytes. Biochem J 182:503-508
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT 8. Bartels H, Vogt B, Jungermann K 1988 Glycogen synthesis via the indirect gluconeogenic pathway in the periportal and via the direct glucose-utilizating pathway in the perivenous zone. Histochemistry 89:253-260 9. Fischer W, Ick M, Katz N 1982 Reciprocal distribution of hexokinase and glucokinase in periportal and perivenous rat liver tissue. Hoppe Seylers Z Physiol Chem 363:375-380 10. Guder WG, Schmidt U 1976 Liver cell heterogeneity; the distribution of pyruvate kinase and phosphoenolpyruvate carbosykinase (GTP) in the lobule of fed and starved rats. Hoppe Seylers Z Physiol Chem 357:1793-1800 11. Jungermann K, Katz N 1989 Functional specialization of different hepatocyte populations. Physiol Rev 69:708-764 12. Katz N, Jungermann K 1975 Autoregulatory switch from glycolysis to gluconeogenesis in rat hepatocyte suspension. Physiol Chem 356:244 13. Katz NR, Teutsch HF, Jungermann K, Sasse D 1977 Heterogeneous reciprocal localization of fructose-1,6-bis-phosphatase and of glucokinase in microdissected periportal and perivenous rat liver tissue. FEBS Lett 83:272-276 14. Probst I, Schwartz P, Jungermann K 1982 Induction in primary culture of 'gluconeogenic' and 'glycolytic' hepatocytes resembling periportal and perivenous cells. Eur J Biochem 126:271-278 15. Quistorff B 1985 Gluconeogenesis in periportal and perivenous hepatocytes of rat liver, isolated by a new high-yield digitonincollagenase perfusion technique. Biochem J 229:221-226 16. Sasse D, Katz NR, Jungermann K 1975 Functional heterogeneity of rat liver parenchyma and of isolated hepatocytes. FEBS Lett 57:83-88 17. Trus MH, Zawalich H, Gaynor D, Matschinsky F 1980 Hexokinase and glucokinase distribution in the liver lobule. J Histochem Cytochem 28:579-581
1941
18. Tal M, Schneider DL, Thorens B, Lodish HF 1990 Restricted expression of the "erythroid/brain" glucose transporter isoform to perivenous hepatocytes in rats: modulation by glucose. J Clin Invest 86:982-992 19. Wheeler TJ, Hinkle PC 1981 Kinetic properties of the reconstituted glucose transporter from human erythrocytes. J Biol Chem 256:8907-8914 20. Lawrence GM, Jepson MA, Trayer IP, Walker DG 1986 The compartmentation of glycolytic and gluconeogenic enzymes in rat kidney and liver and its significance to renal and hepatic metabolism. Histochem J 18:45-53 21. Bringer J, Hedt A, Grodsky GM 1981 Preventon of insulin aggregation by dicarboxylic amino acids during prolonged infusion. Diabetes 30:83-85 22. Kahn BB, Charron MJ, Lodish HF, Cushman SW, Flier JS 1989 Differential regulation of two glucose transporters in adipose cells from diabetic and insulin-treated diabetic rats. J Clin Invest 84:404-11 23. McLean IW, Nakane FP 1974 Periodate-lysine paraformaldhyde fixative: a new fixative for immunoelectron microscopy. J Hystochem 1077-1083 24. Thorens B, Lodish HF, Brown D 1990 Differential expression of two glucose transporter isoforms in rat kidney. Am J Physiol 259:C286-C294 25. Kahn BB, Horton ES, Cushman SW 1987 Mechanism for enhanced glucose transport response to insulin in adipose cell from chronically hyperinsulinemic rats. J Clin Invest 79:853-858 26. Zucker LM, Antoniades HN 1972 Insulin and obesity in the Zucker genetically obese rat "fatty." Endocrinology 90:1320-1330 27. Oka Y, Asano T, Shibasaki Y, Lin JL, Tsukuda K, Akanuma Y, Takaku F 1990 Increased liver glucose-transporter protein and mRNA in streptozocin-induced diabetic rats. Diabetes 39:441-446
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
Vol. 129, No. 4 Printed in U.S.A.
Expression of the Low Km GLUT-1 Glucose Transporter Is Turned on in Perivenous Hepatocytes of InsulinDeficient Diabetic Rats* MICHAEL TALt, BARBARA B. KAHN, AND HARVEY F. LODISH Whitehead Institute for Biomedical Research (M.T., H.F.L.), Cambridge, Massachusetts 02142; the Department of Biology, Massachusetts Institute of Technology (H.F.L.), Cambridge, Massachusetts 02139; and the Charles A. Dana Research Institute and Harvard-Thorndike Laboratory of Beth Israel Hospital, Department of Medicine, Beth Israel Hospital and Harvard Medical School (B.B.K.), Boston, Massachusetts 02215
ABSTRACT. In normal fed rats the low Km glucose transporter GLUT-1 is expressed only in one row of hepatocytes immediately surrounding a terminal hepatic venule, while the high Km GLUT-2 is expressed in every hepatocyte. Previously, we showed that additional perivenous hepatocytes express GLUT-1 in fasting animals. In diabetes, as in starvation, the liver functions to release glucose into the circulation, but unlike starvation, circulating extracellular glucose is high in diabetes. By immunofluorescence and Western blotting we studied whether glucose or insulin is the primary extracellular signal for inducing GLUT-1 expression in hepatocytes. We observed that streptozocin-induced diabetes causes induction of GLUT-1
expression in the plasma membrane of hepatocytes within four cell rows of a terminal hepatic venule; GLUT-2 expression is unaltered. Chronic insulin treatment of diabetic rats reduces the number of rows of hepatocytes expressing GLUT-1 from ~four to ~two. In contrast, chronic insulin infusion into nondiabetic rats does not affect the number of hepatocytes expressing GLUT-1. Thus, both fasting and diabetes induce GLUT-1 expression in specific hepatocytes that normally do not express this gene. This induction correlates with low insulin levels in the blood, and not with circulating glucose levels. {Endocrinology 129: 1933-1941,1991)
T
HE LIVER functions as a glucostat by removing excess glucose after a meal and releasing glucose between meals; after a meal, 50% of blood glucose is absorbed during one passage through a liver acinus (1). An acinus is a mass of liver parenchymal cells that is perfused in the same direction. Blood enters a liver acinus from the intestine through a portal venule and mixes in a sinus with blood from a hepatic arteriole. Blood diffuses along several rows of hepatocytes and then exits from the acinus into a terminal hepatic venule (2). GLUT-2 is the predominant glucose transporter in the liver and is expressed in the basolateral blood-facing plasma membrane of every hepatocyte (3-5). Unlike that in muscle and adipocytes, hepatic glucose transport is not affected by insulin (6). The high Km for glucose of Received April 16,1991. Address all correspondence and requests for reprints to: Dr. Michael Tal, Whitehead Institute for Biomedical Research, 9 Cambridge Center, Cambridge, Massachusetts 02142. * This work was supported by NIH Grants GM-40916 and HL-41484 (to H.F.L.) and NIH Physician Scientist Award AG-00294 and Juvenile Diabetes Foundation Grant 189833 (to B.B.K.). t Supported by a fellowship from the European Molecular Biology Organization.
GLUT-2 (15-20 mM) allows glucose influx or efflux to respond passively to changes in extracellular glucose from the 5-mM resting state (7). Control of glucose metabolism in hepatocytes involves both short and long term effects of hormones, primarily insulin and glucagon. In poorly controlled type I diabetes, insulin levels are low, and the control of liver, muscle, and fat glucose metabolism is severely disturbed. Paradoxically, in many forms of diabetes, while blood glucose is high, the liver is induced to synthesize glucose via gluconeogenesis and to export it. This worsens the diabetic state by further raising the blood glucose level. In livers from both diabetic and starved rats there is an over 200% increase in the activity of key gluconeogenic enzymes, which are expressed predominantly by hepatocytes in the periportal zone (glucose 6-phosphatase, fructose 1,6-bisphosphatase, and phosphoenolpyruvate carboxykinase). In contrast, in the perivenous zone the activity of key glycolytic enzymes (glucokinase and pyruvate kinaseiJ is reduced only 25% (8-17). In contrast to GLUT-2, the GLUT-1 glucose transporter isoform is localized in rat liver to a subset of hepatocytes that form the first row of cells around each terminal hepatic venule (18). GLUT-1 is a glucose trans-
1933
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
1934
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
porter with a low Km for glucose (1-2 HIM) (19). GLUT1 is expressed mainly by cells oxidizing glucose, where the direction of glucose transport is out to in. On the other hand, expression of GLUT-2 is found in all hepatocytes, including those active in gluconeogenesis, where glucose transport is predominantly from the cytoplasm to the extracellular space. In the kidney there is a similar correlation between the glycolytic activities of the different segments of the nephron and the relative expression of GLUT-1, and a correlation between expression of GLUT-2 and the gluconeogenic activity of cells (4, 20). After 3 days of starvation, there is a 3- to 4-fold elevation of expression of GLUT-1 mRNA and protein in the liver as a whole (18). This correlates well with expression of GLUT-1 in more hepatocytes than in the fed state, namely those forming the first three or four rows around the terminal hepatic venules. Importantly, starvation-induced expression of GLUT-1 on the plasma membrane of these additional hepatocytes is lost within 3 h of glucose injection. In streptozocin-induced diabetes in the liver as a whole, the level of GLUT-1 mRNA is increased, and insulin treatment restores GLUT-1 mRNA levels to normal (5). However, consistent changes in GLUT-1 protein levels were not seen in the few animals studied. Here we studied GLUT-1 expression by immunofluorescence microscopy, defining cells expressing this glucose transporter in the plasma membrane, the site at which glucose transport occurs. We induced diabetes by a single streptozocin injection and studied rats over a course of diabetes and insulin treatment. In addition, we studied insulin-infused normal rats to address the question of whether abnormal levels of glucose or insulin affect GLUT-1 expression. We show that in diabetic rats GLUT-1 is expressed in three or four rows of hepatocytes around each terminal hepatic venule. The level of GLUT-2, as visualized by immunofluorescence, is unaffected, and GLUT-2 remains expressed in every hepatocyte. Chronic insulin infusion of diabetic rats reduces GLUT-1 expression to ~1.8 rows of cells around the terminal hepatic venules. In contrast, insulin infusion of normal rats does not change GLUT-1 expression. Taking these results together with those on GLUT-1 expression in starved rats, we conclude that catabolic states associated with low levels of insulin are accompanied by induction of GLUT1 expression in perivenous hepatocytes. This may be due to the lack of insulin directly or be a metabolic consequence of low insulin levels. Importantly, GLUT-1 expression in liver does not correlate with changes in glucose levels per se. In contrast, hepatic GLUT-2 level and localization appear to be relatively unaffected by starvation or diabetes.
Endo • 1991 Vol 129 • No 4
Materials and Methods Experimental design Male Sprague-Dawley rats (CD strain, Charles River Breeding Laboratories, Wilmington, MA), weighing 150-165 g, were maintained with ad libitum feeding of standard chow for several days before initiation of the experimental period. Rats were divided into the following experimental groups: standard chow, streptozocin injection, streptozocin injection followed by insulin infusion, sucrose-enriched chow, and sucrose-enriched chow followed by insulin infusion. For studies of diabetes and insulin treatment, rats were injected ip with 80-85 mg/kg anhydrous streptozocin and citric acid (Zanosar, Upjohn, Kalamazoo, MI) reconstituted with 0.9% saline. Rats were maintained diabetic for 7-14 days (postprandial blood glucose, >20 HIM). After 7 days of hyperglycemia, some rats were treated for 7 days with insulin. After 13-14 days of hyperglycemia, rats were treated for 3, 6, or 24 h with insulin. Rats treated for 3-6 h received one sc injection of 100 U human crystalline insulin (Eli Lilly Co., Indianapolis, IN). Rats treated for 1 day received a combination of crystalline and NPH insulins by sc injection every 12 h. For 7 days of treatment of diabetic and normal rats, 500 U pork crystalline insulin (Lilly), prepared as described by Bringer et al. (21) to prevent aggregation, was infused via sc osmotic minipumps (Alzet 2001, Alza, Palo Alto, CA) at 4 U/ day. Normal rats infused with insulin were allowed free access to sugar cubes to avoid severe hypoglycemia. A corresponding control group was fed sugar cube supplements. Blood glucose was measured in whole blood samples from the tail vein with Chemstrips bG (Bio-Dynamics, Indianapolis, IN), read in an AccucheckbG reflectance meter (Bio-Dynamics). Animals were killed by CO2 inhalation followed by decapitation, and portions of the liver were rapidly excised and immediately fixed and processed for microscopy, For immunoblots, the livers were frozen in liquid N2 and stored at 70 C before lysates were prepared. Serum was frozen for determination of insulin levels by standard RIA. Detection of GLUT-1 by immunoblotting Liver lysates were prepared by homogenizing frozen liver with a polytron (Brinkmann Instruments, Westbury, NY) in a solution containing 5% sodium dodecyl sulfate, 80 mM TrisHC1 (pH 6.8), 5 mM EDTA, 1 mM phenylmethylsulfonylfluoride, and 0.2 mM iV-ethylmaleimide. The protein content of the lysates was measured by the BCA protein assay (Pierce, Rockford, IL), using BSA as the standard. Proteins were separated by electrophoresis on 10% sodium dodecyl sulfate-polyacrylamide gels and electrophoretically transferred to nitrocellulose filters overnight at 4 C. Uniform loading and transfer of protein from all lanes of the gel were assessed by staining the filters with Ponceau S (Sigma, St. Louis, MO) and the gel with Coomassie blue. Immunoblotting was carried out, as previously described (22), with a polyclonal antiserum (10 /ug/ml) prepared against a synthetic peptide consisting of the 16 carboxy-terminal amino acids (477-492) of GLUT-1 (a generous gift from Dr. B. Thorens, Whitehead Institute, Cambridge, MA), followed by [125I]protein-A and in some experiments also by [125I] goat antirabbit immunoglobulin G (IgG; New England Nuclear,
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT Boston, MA). Filters were autoradiographed using XAR-5 film (Eastman Kodak Co., Rochester, NY) and intensifying screens at -70 C. GLUT-1 protein levels were quantitated by scanning densitometry, using a GS 300 densitometer (Hoefer Scientific Instruments, San Francisco, CA) and the GS 350 computer program (Hoefer Scientific Instruments). Statistical analyses of densitometry scans were performed on the Beth Israel Hospital analyzer system. Differences between experimental groups were determined using analysis of variance in conjunction with the Newman-Keuls test. When comparing experimental groups to controls, a one-sample comparison test was used, since these results are evaluated as a percentage of the control value due to the arbitrary nature of optical density units. Immunofluorescence Liver fragments were fixed by immersion in paraformaldehyde-lysine-periodate fixative (23) overnight at 4 C, then kept at 4 C in PBS until use. For preparing S-nm frozen sections, tissue fragments were equilibrated for several hours at 4 C in PBS containing 0.6 M sucrose and 0.02% sodium azide, subsequently embedded in O.C.T. compound (Miles, Elkhart, IN), frozen in 2-methylbutane in liquid nitrogen for 10 sec, and kept at -70 C until sectioning in a Reichert Frigocut cryostat. Sections were placed on polylysine-coated glass slides, air dried, and kept at -70 C until staining. Tissue sections were incubated for 10 min in 1% BSA-PBS. Sections were then incubated for 40 min with one of the following antisera (diluted 1:100 in 1% BSA/PBS): affinity-purified sera raised against the COOH-terminus of GLUT-1 (original concentration, 0.32 mg/ml) (24) or against a peptide corresponding to the COOHterminous of GLUT-2 (3) (both generous gifts from Dr. B. Thorens, Whitehead Institute, Cambridge, MA). Thereafter, sections were washed three times for 5 min each with PBS. The sections were incubated with fluoresceinated goat antirabbit IgG (Accurate Chemical and Scientific Corp., Westbury, NY) diluted 1:100 in 1% BSA-PBS. The sections were washed, as described above, and mounted in 60% glycerol, 2% n-propyl gallate, and 0.2 M Tris-HCl (pH 8.1). Sections were observed and photographed with Kodak TMAX film on a Zeiss Photomicroscope III (New York, NY). For double staining, the coverslips were removed after photography, and the sections were washed for 5 min in PBS. The sections were then incubated for 1 h with an affinity-purified goat antirabbit IgG (Cappel, Malvem.PA; 0.1 mg/ml) in 1% BSA/PBS. Thereafter, the sections were washed four times for 15 min each with PBS. The second antibody (anti-GLUT-2) was applied as described for the initial staining, and detection was performed with rhodamine-conjugated goat antirabbit IgG (Cappel) diluted 1:100 in 1% BSA-PBS. The statistical analysis of the number of rows of hepatocytes expressing GLUT-1 around the terminal hepatic venules was performed by pairwise Student's t test on a Microsoft Excel spreadsheet using a Macintosh computer.
Results Immunostaining of GLUT-1 in liver sections from diabetic rats and diabetic rats infused with insulin Liver sections were stained for GLUT-1 using a specific antibody raised against a peptide corresponding to
1935
its carboxy-terminus. This antibody does not cross-react with GLUT-2, the major glucose transporter expressed in the liver (5). Figure 1 depicts the expression of GLUT1 in livers of insulin-responsive diabetic rats. In normal rats, GLUT-1 is localized to the first row of hepatocytes around the terminal hepatic venule (Fig. 1A) and nowhere else (18). In diabetic rats 7 and 14 days after streptozocin injection, GLUT-1 is expressed in several rows of hepatocytes around the terminal hepatic venule (Fig. 1, B and D, respectively). GLUT-1 is not expressed in other hepatocytes, such as those surrounding the portal triad, in livers from both diabetic (Fig. IF) and normal (18) rats. By analysis of more animals we quantified the number of rows of hepatocytes around the terminal hepatic venules that express GLUT-1 (3.1 ± 0.4 rows of hepatocytes 7 days after streptozocin injection and 4.3 ±1.3 rows after 14 days, compared to 1.3 rows in normal rats; Table 1). The effect of streptozocin was largely reversed in diabetic rats treated with insulin for 7 days (Fig. 1C and Table 1); only 1.8 ± 0.9 rows expressed GLUT-1, which was different from the number in both control and streptozocin-treated rats (P < 0.05). Insulin treatment of diabetic rats for short times [3 and 6 h (three rats per time point; data not shown) and 24 h (Fig. IE and Table 1)] did not significantly affect the expression of GLUT-1. At these early time periods of insulin injection, serum insulin levels were substantially greater than control values, and expression of GLUT-1 was unaltered from that in untreated 14-day diabetic rats. Figure 2 shows that insulin infusion of a normal rat for 7 days has no effect on expression of GLUT-1, although it lowers blood glucose from 115-130 mg/dl (control) to 40 mg/dl (Table 1). Insulin levels were at least 10 times greater than normal in insulin-infused rats, as we previously reported (25). In the insulininfused normal rats (Fig. 2A) and in the normal control rats (Fig. 2B), only one row of hepatocytes around the terminal hepatic venule expressed GLUT-1 (see Table 1). By double staining liver sections from diabetic rats for GLUT-1 and GLUT-2, we showed that cells around the terminal hepatic venules express both glucose transporters on the plasma membrane (Fig. 3, A and B). As was shown previously, both GLUT-2 (4) and GLUT-1 (18) are expressed only on the basolateral or blood-facing surface. Although not shown here, we immunostained GLUT-2 liver sections from rats in all of the metabolic states summarized in Table 1 and found no difference in the staining intensity or the staining pattern. In summary, in diabetes, GLUT-2 is expressed normally on the blood-facing surface of every hepatocyte, while expression of GLUT-1 protein is induced in hepatocytes that are in two to four rows around each terminal hepatic
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
1936
Endo • 1991 Vol 129 • No 4
FIG. 1. Immunofluorescence localization of GLUT-1 in livers from diabetic rats and diabetic rats treated with insulin. Liver sections are from a normal rat (A), a diabetic rat 7 days after streptozocin injection (B), a diabetic rat 7 days after streptozocin injection and 7 days of insulin treatment (C), a diabetic rat 14 days after streptozocin injection (D), a diabetic rat 13 days after streptozocin injection and 1 day of insulin treatment (E), and a diabetic rat 14 days after streptozocin injection (F). The GLUT-1 protein is localized to the plasma membrane of hepatocytes that are in one row (A) or several rows (B-E) around the terminal hepatic venule (T). No GLUT1 expression is observed (F) in hepatocytes around the portal triad (P). Bar represents 30 fim.
TABLE 1. GLUT-1 expression in hepatocytes surrounding terminal hepatic venules
Mean no. of rows Mean venule diameter (/*m) No. of venules Blood glucose (mg/dl) BW(g) Serum insulin levels (/uU/ml) No. of animals
Normal rat
Diabetes 7 d
Diabetes 7 days, insulin 7 d
Diabetes 14 days
Diabetes 13 days, insulin 1 d
Normal sucrose
Normal sucrose, insulin 7 d
1.3 ± 0.4 56 ±23 20 132 ± 7 297 ± 5 72±2 C 15(6)
3.1 ± 1.7" 63 ±35 11 467 ± 17 255 ± 7 48 ± 7°'c 15(5)
1.8 ± 0.9* 48 ±36 52 71 ± 7 265 ± 8 178 ± 28°-c 15(9)
4.3 ± 1.3° 52 ±36 39 580 ± 37 182 ± 10 ND 12(8)
3.9 ± 1.7° 65 ± 3.6 56 236 ± 61 212 ± 15 >200a'c 5(5)
1.4 ± 0.4° 50 ±34 25 115 ± 14 275 ± 5 ND 2(2)
1.4 ± 0.3° 61 ±37 17 40 ± 6 280 ± 9 ND 3(3)
Livers were sectioned from Sprague-Dawley rats that were in the following metabolic states: normal, diabetic 7 and 14 days after streptozocin injection (diabetes 7 d, diabetic 14 d), diabetic rats after insulin infusion for 1 or 7 days (diabetes 7 d, insulin 7 d; diabetes 13 d, insulin 1 d), normal sucrose-supplemented rats (normal sucrose), and normal sucrose-supplemented after 7 days of insulin infusion (normal sucrose, insulin 7 d ). Sections were immunostained for GLUT-1 in the region around the hepatic venules. The first row indicates the mean number (±SD) of rows of hepatocytes surrounding a venule that express GLUT-1; the mean diameter of venules is in the second row, and the number of venules tested is in the third row. Blood glucose levels (±SEM) and body weight (±SEM) for all rats studied are in the fourth and the fifth rows. Serum insulin levels (±SEM) are in the sixth row, and were performed using a human insulin standard curve because human insulin was administered to treated rats. The last row shows the number of rats used in all experiments (i.e. immunofluorescence and immunoblotting), while in parentheses are the number of rats used for immunofluorescence. ND, not done. " P < 0.05 vs. normal rats (by Student's t test). 6 P < 0.05 vs. diabetic 14 d (by Student's t test). c Use of the human standard curve accounts for values in control and diabetic rats which are slightly higher than previously reported (25).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
FIG. 2. Immunofluorescence localization of GLUT-1 in livers from insulin-infused rats. The liver sections are from a normal rat infused for 7 days with insulin (A) and from a normal rat (B). In both cases GLUT-1 protein is localized to the plasma membrane of hepatocytes that are in one row around a terminal hepatic venule (T). Bar represents 30/xm.
venule. Insulin is a negative regulator for GLUT-1, but not GLUT-2, expression. Immunoblotting analysis of GLUT-1 expression in livers from diabetic rats Immunoblotting of gel-separated proteins by antisera to GLUT-1 quantifies the total amount of this transporter, both on the plasma membrane and in internal membranes. Autoradiograms, as represented by the one in Fig. 4, A and B, were quantified in Fig. 5 and showed that after 7 or 14 days of diabetes the overall level of hepatic GLUT-1 was increased 2-fold (P < 0.05). This agrees with the increased number of cells expressing GLUT-1 (Fig. 1). Strikingly, when 7-day diabetic rats were infused with insulin for 7 days, the total level of GLUT-1 tended to decrease, but was not statistically different from 7-day diabetic levels (Figs. 4A and 5), while the immunofluorescence studies showed a significant (P < 0.05) over 2-fold-decrease in the number of cells expressing GLUT-1 on their plasma membrane. We do not know the reason for this difference, but we hypothesize that in insulin-infused diabetic rats, GLUT-1
1937
FIG. 3. Double immunofluorescence detection of GLUT-1 and GLUT2 in liver from a diabetic rat. The liver section is from a diabetic rat 7 days after the injection of streptozocin and is stained with a specific antibody for GLUT-1 (A) or GLUT-2 (B). A, GLUT-1 protein is localized to the plasma membrane of hepatocytes that are in several rows around a terminal hepatic venule (T), while GLUT-2 is expressed in every hepatocyte, including those that express GLUT-1 (B). Bar represents 30 ^m.
in some hepatocytes may be in internal membranes, where the carboxy-terminal epitope is inaccesible to our antibody and, therefore, not visible by immunofluorescence. We also studied by immunobloting the short (324 h) term effect of insulin on diabetic rats; by immunofluorescence these rats have ~four rows of hepatocytes expressing GLUT-1 around each terminal hepatic venule. Figure 4C shows that the total hepatic levels of GLUT-1 protein remain as high as those in nontreated diabetic rats, and were significantly higher than those in normal control rats. Immunoblots of liver lysates from normal control rats and normal rats after 7 days of insulin infusion showed the same amount of GLUT-1 (Fig. 4D) and were consistent with the immunofluorescence data, indicating no change in the number of cells expressing GLUT-1 (Table 1). In the same experiment, diabetic rats that were infused with insulin had 2- to 3-fold higher levels of GLUT1 protein than control rats, with or without insulin
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
1938
Endo•1991 Vol 129 • No 4
6LUT-1 45 kDa
B
D
FIG. 4. Immunoblotting analysis of GLUT-1 protein in livers from diabetic or normal rats before and after insulin treatment. Cell lysates (150 ng protein) from liver tissues of rats under the following metabolic conditions were used: normal rats (normal), diabetic rats 7 (7 d diabetes) or 14 days after streptozocin injection (14 d diabetes), diabetic rats 7 days after streptozocin injection and after 7 days of insulin infusion (7 d diabetes 7 d insulin), 13- to 14-day diabetic rats 3, 6, and 24 h after insulin injection (13-14 d diabetes + insulin 3h, 6h, and 24h), and sucrose-supplemented control rats (normal/sucrose) or after 7 days of insulin infusion in normal rats (normal/sucrose/7 d insulin). Each lane represents a different rat. Lysates were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (10% acrylamide) and transferred to a nitrocellulose filter. The filter was incubated for 1 h with a polyclonal serum raised against a peptide corresponding to the COOH-terminal amino acids of GLUT-1 (24), and bound antibody was detected with [125I]protein-A and, in some cases, [125I]goat antirabbit IgG. Exposure time with Kodak XAR-5 film at —70 C using an intensifying screen was 1 day for A and D, and 2 days for B and C. The arrowhead on the left represents the expected position of the GLUT-1 polypeptide.
infusion (Fig. 4D). This agrees with the immunofluorescence data of Fig. 1 and Table 1, showing that in livers from diabetic animals infused with insulin for 7 days the number of cells expressing GLUT-1 is significantly higher than normal (1.8 rows of cells vs. 1.4 rows).
300
Discussion
normal
7d
14d
7d
diabetes
diabetes
diabetes 7 d insulin
FIG. 5. Relative expression of GLUT-1 in livers from diabetic or normal rats before and after insulin treatment. Immunoblots were from nine normal rats (normal), six rats 7 days after streptozocin injection (7 d diabetes), seven rats 14 days after streptozocin injection (14 d diabetes), and eight rats 7 days after streptozocin injection and 7 days of insulin infusion (7 d diabetes 7 d insulin). Data are presented as a percentage of control expression and are from densitometric scans of immunoblots similar to the one presented in Fig. 4. All experimental groups show significant differences from the control (P < 0.05), but not from each other.
The expression of the low Km glucose transporter GLUT-1 is increased 2- to 3-fold in livers from both starved and diabetic rats. Strikingly, in both cases GLUT-1 expression is not uniform in all hepatocytes; rather, it is the number of cells expressing GLUT-1 that is increasing. We used a morphometric assay which measures the number of rows of hepatocytes surrounding each hepatic venule that express GLUT-1 (18). Studying many venules in multiple animals in different metabolic states gives us a reliable assay for GLUT-1 expression, but not for expression of GLUT-2, the high Km glucose transporter which is expressed in every hepatocyte. We conclude that the increased GLUT-1 expression in diabetes involves an induction of GLUT-1 mRNA in additional hepatocytes surrounding each hepatic venule. The increase we observed in total GLUT-1 expression in livers from diabetic rats correlates with the number of cells expressing GLUT-1. In control rats GLUT-1 expression is confined to 1.3 rows of hepatocytes around
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
1939
Liver acinus FlG. 6. GLUT-2 and GLUT-1 expression in liver acini from normal and diabetic rats. In a liver acinus the blood (small dots) enters through the portal triad, which consists of a bile duct (B), arteriole (A), and portal venule (PV). The blood is mixed in the liver sinus and diffuses (arrowheads) toward the terminal hepatic venule (THV) and exits the liver. GLUT-2 is expressed in every hepatocyte (open circles, including closed and dotted circles). Normally, GLUT-1 is expressed only in hepatocytes in the first row around the terminal hepatic venule (closed circles). In diabetic or starved rats GLUT-1 expression is switched on in three or four rows of hepatocytes around the terminal hepatic venule (dotted circles).
THV
< Terminal Hepatic Venule
PV Portal Venule
each hepatic venule; at 7 and 14 days of diabetes this is increased to three or four rows of hepatocytes around each hepatic venule. Chronic insulin treatment of diabetic rats reduces the number of hepatocytes that express GLUT-1 to —two rows of hepatocytes around each venule. This change in the number of hepatocytes expressing GLUT-1 is significantly (P < 0.05) different from that in both control and 14-day diabetic rats. The reduction we found in the number of cells expressing GLUT-1 is accompanied by a slight, although not statistically significant, reduction in the total amount of GLUT-1 protein measured by immunoblotting of liver lysates. In the insulin-infused animals, GLUT-1 protein could be in internal hepatocyte membranes, which may not be detected by immunofluorescence. We observed a similar phenomenon in livers from starved and refed rats (18). In that case, total liver GLUT-1 protein was increased, but there was no change in the number of hepatocytes expressing GLUT-1 on their plasma membrane. During the Western blotting procedure GLUT-1 protein is denatured and dissociated from other proteins and lipids, while in sections, GLUT1 is fixed, and its COOH-terminus, the site of binding of our antisera, might be bound to other proteins or otherwise masked. In the case of starved and refed rats, we attemped to dissociate such possible epitope masking by sodium dodecyl sulfate treatment of sections, but there was no visible change in GLUT-1 immunostaining (Tal, M., and H. F. Lodish, unpublished). In a previous report (5) GLUT-1 mRNA in liver from diabetic rats was increased 2.4-fold compared to that in control rats, an observation in accord with the results
—
shown here. However, in that early study (5), the level of GLUT-1 protein was not increased statistically, in contradiction to our present data; the difference may result from the smaller number of rats analyzed for GLUT-1 protein in this early report. Also, immunoblotting was performed on total liver membranes (5), while here we used total liver cell lysates, which represent all glucose transporters in the tissue. However, the quantitation of GLUT-1 should be the same using both preparations, since glucose transporters are integral membrane proteins, unless there is selection of specific membranes using that protocol (5). A short period (24 h) of insulin treatment of diabetic rats did not significantly reduce GLUT-1 expression, nor did treatment between 3 and 24 h (data not shown). Therefore, the effects of insulin are not directly on GLUT-1 protein and probably involve other insulindependent mechanisms, such as the subcellular localization or stability of GLUT-1. The effects of insulin could be indirect, mediated by some other hormone or metabolite which changes in both starvation and insulin-deficient diabetes. A common feature of diabetes and starvation is the low level of circulating insulin. Table 2 summarizes data indicating a correlation between low circulating insulin and induction of GLUT-1 expression in additional perivenous hepatocytes, but no correlation with circulating glucose. Noteworthy are findings that glucose injection of starved animals and chronic insulin infusion of diabetic ones reduce the number of perivenous hepatocytes that express GLUT-1. In addition, GLUT-1 is expressed only in the first row of hepatocytes around each terminal
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
1940
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT
TABLE 2. Low insulin levels correlate with induction of GLUT-1 protein, and normal and high insulin levels correlate with normal expression of GLUT-1
Fed Starved Starved/glucose-injected Streptozocin-diabetes Streptozocin-diabetes/insulin Normal/insulin-infusion fa/fa Zucker rats
Glucose levels
Insulin levels
No. of rows expressing GLUT-1
Normal Low High High Normal/low Low Normal
Normal Low High Low High High High
1.3° 3.4° 1.4° 3-46 1.86 1" lc
" Data from Ref. 18. 4 Data presented here. c Data from Tal, M., and Lodish, H. F., unpublished observation.
hepatic venule of liver tissue from fa/fa Zucker rats (Tal, M., and H. F. Lodish, unpublished observation). The Zucker fa/fa rat is a model of genetic obesity, with naturally occurring hyperinsulinemia and normal blood glucose (26). Thus, in all cases studied, the first row of hepatocytes around each venule continues to express GLUT-1. Expression in these cells, in particular, is not affected by hyperinsulinemia. GLUT-2 is the predominant hepatic glucose transporter. By immunofluorescence we could see no difference in GLUT-2 expression in hepatocytes in the metabolic states studied here, nor was there any change in GLUT-2 expression in starved or starved/glucose-injected rats or in hepatocytes from fa/fa Zucker rats (18) (Tal, M., and H. F. Lodish, unpublished observations). We detect by immunofluorescence only plasma membrane expression of GLUT-2; we have seen no morphological changes in the localization of GLUT-2 within the liver (periportal vs. perivenous). We do feel that a 2-fold difference in GLUT-2 expression levels would be detected by our immunofluorescence studies. Our immunofluorescence results are in agreement with those of a previous report, which showed by Western and Northern blotting that GLUT-2 levels were unchanged in diabetes (5). These results contrast with another study which showed that GLUT-2 levels were induced by diabetes and reduced to normal levels after chronic insulin treatment (27). The differences may be due to difference in the strain of rats, treatment protocols, and/or procedures for membrane preparation. Finally, our studies add to the increasing body of literature, summarized in the Introduction, that hepatocytes are heterogeneous in their expression of proteins required for gluconeogenesis and glycolysis. In normal fed rats GLUT-1 expression occurs in those hepatocytes most active in glycolysis. In other tissues there is a correlation of GLUT-1 expression with utilization of
Endo• 1991 Voll29»No4
glucose as an energy substrate (see Introduction). This suggests that in diabetes, the additional perivenous hepatocytes expressing GLUT-1 are also transporting glucose inward and functioning to reduce the high extracellular glucose. Certainly, hepatocytes are heterogeneous in expression of glucose transporter isoforms and in their responses to the presence or lack of insulin. Figure 6 summarizes our current knowledge of the heterogeneity of liver hepatocytes with respect to glucose transporter expression. The parenchymal cells in each acinus can be divided into three groups: 1) hepatocytes nearest the portal venule and hepatic arteriole that express only GLUT-2, regardless of any alteration in metabolic state tested, including those cells most active in gluconeogenesis; 2) hepatocytes within one row of each hepatic venule that express both GLUT-2 and GLUT-1, regardless of any alteration in metabolic state tested; this group of hepatocytes is most active in glucose oxidation; and 3) hepatocytes in rows 2-4 around each terminal hepatic venule that always express GLUT-2. Expression of GLUT-1 in these cells occurs in at least two conditions in which the blood insulin level is low: in streptozocininduced diabetes and starvation. In diabetes, the increased expression of GLUT-1 may facilitate entry of glucose into more hepatocytes, and thus lower blood glucose. Unlike under other conditions, specifically starvation, expression of GLUT-1 in these additional cells might allow them to release glucose at an increased rate into the blood.
Acknowledgments We thank Dr. B. Thorens for GLUT-1 and GLUT-2 antibodies, A. Rosen for excellent technical assistance, and Dr. Amihod Dotan, Tel-Aviv University, for statistical evaluation of our immunofluorescence results.
References 1. Thurman RG, Kauffman FC, Jungermann K 1986 Regulation of hepatic glucose metabolism. Plenum Press, New York and London 2. Rappaport AM, Borowy ZJ, Lougheed WM, Lotto WN 1954 Subdivision of hexagonal liver lobules into a structural and functional unit: role in hepatic physiology and pathology. Anat Rec 119:1134 3. Thorens B, Sarkar HK, Kaback HR, Lodish HF 1988 Cloning and functional expression in bacteria of a novel glucose transporter present in liver, intestine, kidney, and beta-pancreatic islet cells. Cell 55:281-290 4. Thorens B, Cheng ZQ, Brown D, Lodish HF 1990 Liver glucose transporter: a basolateral protein in hepatocytes and intestine and kidney cells. Am J Physiol 2:1 5. Thorens B, Flier FS, Lodsh HF, Kahn BB 1990 Differential regulation of two glucose transporters in rat liver by diabetes and insulin treatment. Diabetes 39:712-719 6. Williams TF, Exton JH, Park CR, Regen DM 1968 Stereospeclfic transport of glucose in the perfused rat liver. Am J Physiol 215:1200-1209 7. Craik JD, Elliott KRF 1979 Kinetics of 3-O-methyl-D-glucose transport in isolated rat hepatocytes. Biochem J 182:503-508
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
GLUT-1 IN HEPATOCYTES OF DIABETIC RAT 8. Bartels H, Vogt B, Jungermann K 1988 Glycogen synthesis via the indirect gluconeogenic pathway in the periportal and via the direct glucose-utilizating pathway in the perivenous zone. Histochemistry 89:253-260 9. Fischer W, Ick M, Katz N 1982 Reciprocal distribution of hexokinase and glucokinase in periportal and perivenous rat liver tissue. Hoppe Seylers Z Physiol Chem 363:375-380 10. Guder WG, Schmidt U 1976 Liver cell heterogeneity; the distribution of pyruvate kinase and phosphoenolpyruvate carbosykinase (GTP) in the lobule of fed and starved rats. Hoppe Seylers Z Physiol Chem 357:1793-1800 11. Jungermann K, Katz N 1989 Functional specialization of different hepatocyte populations. Physiol Rev 69:708-764 12. Katz N, Jungermann K 1975 Autoregulatory switch from glycolysis to gluconeogenesis in rat hepatocyte suspension. Physiol Chem 356:244 13. Katz NR, Teutsch HF, Jungermann K, Sasse D 1977 Heterogeneous reciprocal localization of fructose-1,6-bis-phosphatase and of glucokinase in microdissected periportal and perivenous rat liver tissue. FEBS Lett 83:272-276 14. Probst I, Schwartz P, Jungermann K 1982 Induction in primary culture of 'gluconeogenic' and 'glycolytic' hepatocytes resembling periportal and perivenous cells. Eur J Biochem 126:271-278 15. Quistorff B 1985 Gluconeogenesis in periportal and perivenous hepatocytes of rat liver, isolated by a new high-yield digitonincollagenase perfusion technique. Biochem J 229:221-226 16. Sasse D, Katz NR, Jungermann K 1975 Functional heterogeneity of rat liver parenchyma and of isolated hepatocytes. FEBS Lett 57:83-88 17. Trus MH, Zawalich H, Gaynor D, Matschinsky F 1980 Hexokinase and glucokinase distribution in the liver lobule. J Histochem Cytochem 28:579-581
1941
18. Tal M, Schneider DL, Thorens B, Lodish HF 1990 Restricted expression of the "erythroid/brain" glucose transporter isoform to perivenous hepatocytes in rats: modulation by glucose. J Clin Invest 86:982-992 19. Wheeler TJ, Hinkle PC 1981 Kinetic properties of the reconstituted glucose transporter from human erythrocytes. J Biol Chem 256:8907-8914 20. Lawrence GM, Jepson MA, Trayer IP, Walker DG 1986 The compartmentation of glycolytic and gluconeogenic enzymes in rat kidney and liver and its significance to renal and hepatic metabolism. Histochem J 18:45-53 21. Bringer J, Hedt A, Grodsky GM 1981 Preventon of insulin aggregation by dicarboxylic amino acids during prolonged infusion. Diabetes 30:83-85 22. Kahn BB, Charron MJ, Lodish HF, Cushman SW, Flier JS 1989 Differential regulation of two glucose transporters in adipose cells from diabetic and insulin-treated diabetic rats. J Clin Invest 84:404-11 23. McLean IW, Nakane FP 1974 Periodate-lysine paraformaldhyde fixative: a new fixative for immunoelectron microscopy. J Hystochem 1077-1083 24. Thorens B, Lodish HF, Brown D 1990 Differential expression of two glucose transporter isoforms in rat kidney. Am J Physiol 259:C286-C294 25. Kahn BB, Horton ES, Cushman SW 1987 Mechanism for enhanced glucose transport response to insulin in adipose cell from chronically hyperinsulinemic rats. J Clin Invest 79:853-858 26. Zucker LM, Antoniades HN 1972 Insulin and obesity in the Zucker genetically obese rat "fatty." Endocrinology 90:1320-1330 27. Oka Y, Asano T, Shibasaki Y, Lin JL, Tsukuda K, Akanuma Y, Takaku F 1990 Increased liver glucose-transporter protein and mRNA in streptozocin-induced diabetic rats. Diabetes 39:441-446
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 November 2015. at 13:54 For personal use only. No other uses without permission. . All rights reserved.
