443 High concentration of glucose decreases glucose transporter-1 expression in mouse placenta in vitro and in vivo

K Ogura, M Sakata, M Yamaguchi, H Kurachi and Y Murata Department of Obstetrics and Gynecology, Osaka University Medical School, 2–2 Yamadaoka Suita, Osaka, 565–0871 Japan (Requests for offprints should be addressed to M Sakata)

Abstract Facilitative glucose transporter-1 (GLUT1) is expressed protein levels on day 4 of culture were significantly abundantly and has an important role in glucose transfer in reduced in cells incubated with 22 mM glucose compared placentas. However, little is known about the regulation of with control. Next, we rendered mice diabetic by admin- GLUT1 expression in placental cells. We studied the istering 200 µg/g body weight streptozotocin (STZ) on changes in placental GLUT1 levels in relation to changes day 8 of pregnancy. Animals were killed on day 12 of in glucose concentration in vitro and in vivo.Inin vitro pregnancy and placental tissues were obtained. [3H]Cyto- experiments, dispersed mouse placental cells were incu- chalasin B binding study was carried out to assess total bated under control (5·5 mM) and moderately high GLUTs, and GLUT1 mRNA and protein were measured (22 mM) glucose concentrations, and 2-deoxyglucose as above. [3H]Cytochalasin B binding sites in placentas uptake into cells was studied on days 1–5 of culture. from STZ-treated mice were significantly less than those After 4 days of incubation under both conditions, in control mice. Northern and immunoblot analyses GLUT1 mRNA and proten levels were examined by revealed a significant decrease in GLUT1 mRNA and Northern and immunoblot analyses. Treatment of cells protein levels in diabetic mice compared with the controls. with 22 mM glucose resulted in a significant decrease in These findings suggest that the glucose concentration may 2-deoxyglucose uptake compared with control, from day 2 regulate the expression of placental GLUT1. to day 5 of culture. Moreover, GLUT1 mRNA and Journal of Endocrinology (1999) 160, 443–452

Introduction level expression of GLUT3 in human placenta (Shepherd et al. 1992). These observations suggest that GLUT1, not Transfer of glucose from the maternal circulation to the GLUT3, is the major GLUT isoform in human placenta. fetus is fundamental in the utero–placental–fetal transfer Two reports in humans ( Jansson et al. 1993, Sakata et al. system (Danics & Schneider 1975). Fetal plasma glucose 1995), a report in rats (Zhou & Bondy 1993), and a report concentrations have been shown to correlate with mater- in mice (Yamaguchi et al. 1996) describe the changes in nal glucose concentrations (Economides & Nicolaides placental GLUT gene expression during pregnancy. In the 1989). The glucose transfer across the placenta is facili- human placenta, GLUT1 is a dominantly expressed tated, and is independent of the Na+ gradient ( Johnson & glucose transporter and its levels increase during preg- Smith 1980). The expression of glucose transporters nancy ( Jansson et al. 1993, Sakata et al. 1995). The data (GLUTs) shows a tissue-specific distribution. For example, in rats (Zhou & Bondy 1993) showed that placental GLUT1 has been found in erythrocytes, kidney and GLUT1 mRNA decreased throughout pregnancy, blood–tissue barriers, including those in the brain and whereas placental GLUT3 mRNA showed little change placenta (Mueckler et al. 1985, Asano et al. 1988). after midpregnancy. However, in mice, we have shown GLUT3 has been shown to be abundant in organs with that the level of GLUT1 mRNA increased after midpreg- high glucose requirements, such as the brain (Haber et al. nancy, not only in the placenta, but also in the decidua 1993), and to be expressed in the placenta in rodents (Yamaguchi et al. 1996). An increase in the level of (Shepherd et al. 1992, Haber et al. 1993, Zhou & Bondy GLUT1 in the utero–placental unit during pregnancy may 1993, Boileau et al. 1995). Although GLUT3 mRNA has be helpful in the increased requirement of glucose in the been reported to be expressed in many tissues, including fetus, especially after midpregnancy. human and rat placentas (Kayano et al. 1990), it was not It is known also that the fetal plasma glucose concen- easy to detect GLUT3 protein by immunoblot analysis tration increases when the mother is diabetic and hyper- (Haber et al. 1993, Jansson et al. 1993) because of the low glycemic (Pedersen 1977). Despite the importance of

Journal of Endocrinology (1999) 160, 443–452  1999 Society for Endocrinology Printed in Great Britain 0022–0795/99/0160–443 Online version via http://www.endocrinology.org

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access 444 K OGURA and others · Inhibition of placental GLUT1 expression by glucose

GLUTs, the changes in the level of placental GLUT in the Cells were plated in multiwell plates at a density of presence of a high glucose concentration have not been 1#106/cm2 and were allowed to attach for 2 h under an clearly elucidated. In this study we analyzed the effect of atmosphere of 95% air/5% CO2 at 37 )C. Then the high glucose concentration on placental glucose transport medium was exchanged for a fresh one containing 2% and the levels of GLUT1 protein and mRNA, using a FCS. Glucose concentrations were adjusted and the cells primary mouse placental cell culture and the placental were incubated for 5 days. The medium was changed tissues from pregnant mice with streptozotocin-induced daily. The day on which the cells were plated was diabetes. considered as day 0.

Animal models for diabetes melitus and preparation of Materials and Methods placental membrane Chemicals Time-pregnant female ICR mice were housed under standard conditions and allowed free access to standard Cytochalasin B, - and -glucose, EDTA, aprotinin, food and water. Diabetes was induced by intraperitoneal leupeptin, pepstatin A, antipan, soybean trypsin inhibitor (i.p.) injection of 200 µg/g body weight of streptozotocin and phenylmethylsulfonyl fluoride (PMSF) were obtained (STZ) dissolved in 50 mM sodium citrated buffer, pH 4·5 from Sigma (St Louis, MO, USA). 2-[1,2-3H]Deoxy-- 3 on day 8 of gestation. Mice were killed by spinal disloca- glucose (30 Ci/mmol) and 5-[1,2- H]hydroxytryptamine tion on day 12, and placentas and fetuses were collected, (serotonin) (28 Ci/mmol) were purchased from Du frozen in liquid nitrogen, and kept at "80 )C until Pont New England Nuclear (Boston, MA, USA). required for use. Diabetes was ascertained by a blood [3H]Cytochalasin B, -[4,5-3H]leucine (140 Ci/mmol), 32 glucose concentration greater than 250 mg/dl. All subse- [á- P]CTP (3000 Ci/mmol) and multiprime DNA quent procedures for placental membrane preparation labeling kit were purchased from Amersham (Amersham, were perfomed at 4 )C. Pieces of placentas were hom- Bucks, UK). ogenized using a Dounce homogenizer in a buffer con- taining 250 mM sucrose, 10 mM Tris–HCl (pH 7·4), 5 mM EDTA, 10 mg/ml of a mixture of leupeptin, Animals and placental cell culture pepstatin A, aprotinin and antipan, 50 µg/ml soybean The Institute of Cancer Research (ICR) pregnant mice trypsin inhibitor and benzamide and 1 mM PMSF as (10 weeks old), which originated from Charles River described previously (Sakata et al. 1992). The homogenate Laboratories Inc., (Wilmington, MA, USA), were pur- was centrifuged at 2600 g for 10 min and the resultant chased from Japan SLC Inc. (Hamamatsu, Shizuoka, supernatant was centrifuged at 45 000 g for 1 h (Sakata Japan). Tissue for cell culture was obtained on day 12 of et al. 1992). The final pellet was resuspended in a pregnancy (vaginal plug=day 0 of pregnancy) as described buffer containing 20 mM Tris–HCl, 1 mM EDTA, previously (Yamaguchi et al. 1995) and suspended in a cell 100 mM NaCl and 4 mM MgCl2, to determine protein culture medium (NCTC-135, 20 mM HEPES, pH 7·2, concentration (Sakata et al. 1992). 50 µg/ml streptomycin, 50 U/ml penicillin G) containing 5% fetal calf serum (FCS). The tissue was then minced and Measurement of 2-deoxyglucose uptake incubated with a dissociation medium (Medium 199, 3 20 mM HEPES, 10 mM NaHCO3, 5 µg/ml streptomycin Uptake of 2-[ H]deoxyglucose, a non-metabolizable ana- and 50 U/ml penicillin G, pH 7·2) containing 0·1% log of glucose, was measured in 12 wells per group in each (w/v) collagenase (Clostridium histolyticum, type 1, CLS; experiment on mouse placental cells plated in 24-well Worthington Biochemical Co., Malvern, PA, USA) and plates as described previously (Kitagawa et al. 1991), with 0·002% (w/v) bovine pancreatic DNase (type 1, EC modifications. Each experiment was repeated twice with 3·1.21·1) at 37 )C for 1 h. After centrifugation at 600 g for similar results; representative results are shown. After the 5 min, the tissue was dispersed by pipetting, in a calcium- medium was removed, the cells were washed three times and magnesium-free Hanks’ solution containing 0·1% with 1·0 ml prewarmed PBS and incubated in 1·0 ml (w/v) BSA, and the cells were filtered through 150 µm prewarmed transport buffer containing 25 mM HEPES, Nitex (Tetko, Elmsford, NY, USA). The cell suspension pH 7·4, 0·8 mM MgSO4, 140 mM NaCl, 5·4 mM 3 was centrifuged as above, resuspended in a 2-ml dissocia- KCl, 1·8 mM CaCl2, and 1 µCi/ml 2-[ H]deoxy- tion medium containing 0·015% DNase, and then frac- glucose (1 µM) at 37 )C for 10 min. To terminate tionated on a 40% Percoll (Pharmacia Fine Chemicals, 2-[3H]deoxyglucose uptake, the radioactive buffer con- Uppsala, Sweden) gradient. Finally, the cells were col- taining 2-[3H]deoxyglucose was removed and rapidly lected and washed in a dissociation medium. These cells washed three times with ice-cold transport buffer. Incor- were trophoblast cells, as described previously (Yamaguchi poration of the radioactivity into the cells was linear during et al. 1995). this period (data not shown). The cells were solubilized in

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access Inhibition of placental GLUT1 expression by glucose · K OGURA and others 445

0·6 ml 0·03% sodium dodecyl sulfate (SDS). A 0·4-ml amounts of GLUT1 mRNA were divided by those of aliquot was counted in a liquid scintillation counter. â-actin mRNA. A 0·1-ml aliquot was used to determine protein concentration using a DC protein assay kit (Bio-Rad, Immunoblot analysis Richmond, CA, USA). Non-carrier-mediated uptake of 2-deoxyglucose was determined in parallel wells Cells incubated with 5·5 mM or 22 mM glucose for 5 days containing either 10 µM cytochalasin B or 100 mM were washed with NCTC-135 medium and treated as -glucose. described previously (Sakata et al. 1992). After treatment with detachment buffer (250 mM sucrose, 10 mM Tris, pH 7·4, 1 mM EDTA, 0·3 mM PMSF, 1 µg/ml each Measurement of serotonin and leucine uptake pepstain and leupeptin), cells were scraped off the plates, collected by centrifugation at 12 000 g at 4 C for 10 min, Serotonin uptake was studied as described previously ) and solubilized with 1% Triton X-100. After preparation (Cool et al. 1991). Cells were incubated for 45 min at of placental membranes, aliquots were denatured with an room temperature before the measurement of serotonin equal amount of buffer containing 200 mM dithiothreitol, uptake, in a buffer containing 25 mM HEPES, pH 7·5, 20% glycerol, 0·04% bromophenol blue, 10% SDS and 0·8 mM MgSO , 140 mM NaCl, 5·4 mM KCl, 1·8 mM 4 120 mM Tris–HCl (pH 6·8) at 60 C for 15 min to avoid CaCl ,5mM-glucose and 0·1 mM iproniazid (transport ) 2 aggregation of GLUT proteins (Morita et al. 1992). In buffer). After the incubation, serotonin uptake was studied in vivo experiments, placentas were homogenized in a by adding 1 ml transport buffer containing 50 nM radio- ice-cold sucrose buffer (250 mM sucrose, 10 mM Tris, pH labeled serotonin. After 20 min of incubation, the radio- 7·4, 5 mM EDTA, 0·3 mM PMSF, 1 µg/ml each apro- active buffer was removed and each well was rapidly tinin and leupeptin) (Sakata et al. 1992), then centrifuged washed three times with transport buffer containing at 2600 g for 10 min at 4 C. The supernant was centri- 0·1 mM imipramine. Cells were solubilized and the ) fuged at 45 000 g for 60 min at 4 C. The crude mem- radioactivity incorporated into the cells was counted as ) brane pellet was resuspended in a cold resuspension buffer described above. Leucine uptake was studied using a (20 mM Tris, pH 7·4, 1 mM EDTA, 100 mM NaCl, and procedure similar to that described above, except that the 4 mM MgCl ) (Sakata et al. 1992). The protein concen- preincubation with iproniazid was omitted and the wash- 2 tration was determined using the Bradford method. ing buffer did not contain imipramine. Uptake of sero- Twenty-five micrograms denatured samples were sub- tonin and leucine was measured in 12 wells per group in jected to 10% SDS–PAGE (Laemmli 1970), and trans- each experiment on mouse placental cells plated in 24- ferred to a nitrocellulose filter (Wilson et al. 1984) well plates. Each experiment was repeated twice. (Bio-Rad). The filter was incubated with a rabbit anti- GLUT1 antiserum (Oka et al. 1988) at a dilution of 1:8000, followed by a second incubation with goat anti- Northern blot analysis rabbit immunoglobulin G conjugated to horseradish peri- Total RNA was prepared from cells cultured with 5·5 mM oxidase (Bio-Rad) at a dilution of 1:16 000. Then the filter or 22 mM glucose for 4 days. Total RNA was also was stained with an enhanced chemiluminescence (ECL) prepared from placentas of pregnant mice with or without system (Amersham, Arlington Heights, IL, USA) and the STZ treatment. Twenty micrograms total RNA was exposed to a Kodak X-Omat film (Kodak, Rochester, denatured and electrophoresed on a 1% agarose–0·66 M NY, USA) for 15 s. The band intensities were analyzed as formaldehyde gel and blotted onto a nylon membrane described above. filter (Zeta-Probe; Bio-Rad). The filter was hybridized for 6 18hat43)C in a hybridizing buffer containing 2#10 3 c.p.m./ml 32P-labeled rabbit GLUT1 cDNA probe [ H]Cytochalasin B binding assay (Asano et al. 1988). The coding region of this rabbit cDNA -Glucose-protectable binding of cytochalasin B is estab- (Asano et al. 1988) shared 93·2% and 89·4% nucleotide lished as a reliable assessment of the total number of homology with that of human (Mueckler et al. 1985) and glucose transporters (Gorga & Lienhard 1981). [3H]Cyto- rat (Birnbaum et al. 1986) glucose transporter respectively. chalasin B binding assay was performed as described The EcoRI-digested insert of this cDNA can be used as a previously (Gorga & Lienhard 1981). To determine probe for Northern blot analysis in mouse (Asano et al. [3H]cytochalasin B-specific binding simultaneously in all 1988). The filter was washed and autoradiographed for 3 the membrane preparations from placentas of control and days at "80 )C. The filter was stripped, rehybridized STZ treated mice, 400 nM [3H]cytochalasin B was used with 32P-labeled â-actin (Cleveland et al. 1980) probes in the binding assay as described previously (Gorga & and autoradiographed for 1 day at "80 )C. The band Lienhard 1981). Membrane-bound and free [3H]cytocha- intensities were analyzed with a densitometer (Imaging lasin B were separated by a rapid filtration system using Reserch Inc., St Catharines, Ontario, Canada) and the GF/B glass fiber filters (Whatman International Ltd,

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access 446 K OGURA and others · Inhibition of placental GLUT1 expression by glucose

Figure 1 (a) Time-dependent effect of 22 mM glucose on 2-deoxyglucose uptake by mouse placental cells. Placental cells (1#106 cells/well) were incubated with 5·5 mM (control) or 22 mM glucose for the periods indicated. 2-Deoxyglucose uptake was assayed in each well. Results are expressed as the mean&S.E. of data from 12 wells per group in each experiment. *P<0·05, **P<0·01: statistically significant difference from control on that day (unpaired t-test). (b) Concentration-dependent effect of glucose on 2-deoxyglucose uptake. Placental cells (1#106 cells/well) were incubated with various concentrations of glucose for 4 days and assayed for 2-deoxyglucose uptake in each well. Results are expressed as the mean&S.E. of data from 12 wells per a group in each experiment. **P<0·01: statistically significant difference from control (Scheffe’s test).

Maidstone, UK). The filters were dried and counted for glucose on 2-deoxyglucose uptake may not be the result of radioactivity in 10 ml liquid scintillation fluid. glucose toxicity. To study whether the GLUT1 synthesis is decreased by 22 mM glucose, we analyzed the changes in steady-state Statistical methods GLUT1 mRNA levels by Northern blot analysis. After Data were expressed as the mean&.., and were statisti- the mouse placental cells had been incubated with 5·5 mM cally analyzed using analysis of variance followed by (control) or 22 mM glucose for 4 days, total RNA samples unpaired t-test or Scheffe’s multiple range test. Statistical were obtained. Hybridization with a 32P-labeled GLUT1 significance was determined at P<0·05. cDNA probe (Asano et al. 1988) yielded a single band of approximately 2·8 kb (data not shown). The amounts of GLUT1 mRNA were normalized with reference to those Results of â-actin mRNA. The GLUT1 band intensity in cells treated with 22 mM glucose was significantly (P<0·01) Incubation of mouse placental cells with 22 mM glucose decreased compared with control (normalized band inten- significantly (P<0·01) reduced the 2-deoxyglucose uptake sity (arbitrary units): control, 2·3&0·3; 22 mM, 1·5&0·1) from day 2 to day 5 of culture (Fig. 1a). This inhibitory (Fig. 3). The filter was rehybridized with a 32P-labeled effect of 22 mM glucose on 2-deoxyglucose uptake was GLUT3 cDNA. GLUT3 mRNA could not be detected, not observed on day 1. Figure 1b shows that the inhibitory even in an autoradiograph exposed for more than 10 days effect of glucose was dose-dependent. at "80 )C (data not shown). To rule out the possibility that the inhibitory effect of To examine the effect of 22 mM glucose on the 22 mM glucose on 2-deoxyglucose uptake was a toxic GLUT1 protein level, immunoblot analysis was per- effect, the effect of 22 mM glucose on leucine and formed. Mouse placental cells were incubated with serotonin uptake was examined. Leucine and serotonin 5·5 mM (control) or with 22 mM glucose for 4 days and uptakes are, respectively, independent of Na+ and depen- solubilized with 1% Triton X-100. Amounts of 49 000-Da dent on Na+. Incubation of mouse placental cells with protein detected by a rabbit anti-GLUT1 antiserum were 22 mM glucose for 4 days had no effect on leucine and also decreased by 22 mM glucose (Fig. 4a), consistent with serotonin uptake compared with the control (Fig. 2). the results from 2-deoxyglucose uptake and Northern blot These results suggest that the inhibitory effect of 22 mM analysis. GLUT1 protein band intensity in cells treated

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access Inhibition of placental GLUT1 expression by glucose · K OGURA and others 447

Figure 2 Effects of 22 mM glucose on uptakes of 2-deoxyglucose, Figure 3 leucine and serotonin. Cells were incubated with 5 mM (control) Northern blot analysis of GLUT1 mRNA in placental cells from 5·5 mM (control) and 22 mM glucose-treated cells. Total or 22 mM glucose for 4 days and assayed for 2-deoxyglucose, 32 leucine and serotonin uptake. Results are expressed as a RNA (20 µg) was electrophoresed and hybridized with P-labeled percentage of the control. Each group consisted of data from 12 GLUT1 and with â-actin cDNA probes. Band intensities were wells. **P<0·01: statistically significant difference from control quantitated by densitometry; that of GLUT1 mRNA was (unpaired t-test). normalized with respect to that of â-actin. Densitometry data obtained from three individual Northern blots are shown in arbitrary units. Results are expressed as mean&S.E. of three separate experiments. **P<0·01: statistically significant difference with 22 mM glucose was significantly (P<0·01) decreased from control (unpaired t-test, n=3). compared with control (normalized band intensity (arbi- trary units): control, 0·42&0·05; 22 mM, 0·23&0·04) (Fig. 4b). A significant (P<0·01) time-dependent decrease [3H]cytochalasin B-specific binding was determined in in GLUT1 protein in response to 22 mM glucose treat- placentas from control and STZ-treated mice on days 10, ment compared with that in controls (5·5 mM glucose) 12, 16, and 18 of pregnancy. On day 10 of pregnancy, we was observed on days 3 and 4 of culture (Fig. 4c). failed to find a significant difference in [3H]cytochalasin B Placental glucose transport has been suggested to be binding levels between control and STZ-treated mice independent of insulin stimulation (Challier et al. 1986), in (data not shown). On day 12 of pregnancy, the spite of the presence of large amounts of placental insulin levels of [3H]cytochalasin B binding in STZ-treated receptors ( Jones et al. 1993). Therefore, we investigated mice (1·83&0·45 pmol/mg protein) were significantly the effect of 1 and 5 µg/ml insulin on glucose uptake by (P<0·05) lower than those in the controls (3·52& placental cells. Treatment of mouse placental cells with 0·48 pmol/mg protein) (Fig. 6). On day 16 and 18 of insulin for 4 days had no significant effect on the level of pregnancy, the levels of [3H]cytochalasin B binding in 2-[3H]deoxy--glucose uptake (Fig. 5), consistent with STZ-treated mice were substantially lower than those in the results in previous reports (Challier et al. 1986). the controls, but not statistically significant, because of a STZ-treated pregnant mice were sacrificed on day 12 large variation in the [3H]cytochalasin B binding level. of pregnancy. Maternal and fetal weight, crown–rump This may have resulted from the difficulty in separating length (CRL) of the fetuses, maternal and fetal blood placental tissues from decidua, which was easily performed glucose concentrations, and maternal insulin concen- on day 12 of pregnancy. Therefore, we performed the trations are shown in Table 1. STZ-treated pregnant mice [3H]cytochalasin B binding assay on day 12 of pregnancy. were significantly (P<0·01) hyperglycemic and hypoin- The mothers’ brain levels of [3H]cytochalasin B binding in sulinemic compared with controls. Fetal blood glucose STZ-treated mice (1·32&0·48 pmol/mg protein) were concentrations of STZ-treated pregnant mice were signifi- also significantly (P<0·05) lower than those in the controls cantly (P<0·01) greater than those of controls. Fetuses (2·55&0·13 pmol/mg protein) (Fig.6). showed significant (P<0·01) growth retardation under [3H]Cytochalasin B recognizes not only GLUT1, but these conditions (Table 1). Figure 6 shows the levels of also other GLUTs (Mueckler et al. 1985, Kayano et al. [3H]cytochalasin B binding to placental and maternal brain 1990). Therefore, the results observed in Fig. 6 are the membranes, both of which are rich in GLUT1 (Mueckler changes in levels of total GLUTs, and may not specifically et al. 1985, Asano et al. 1988). In preliminary experiments, reflect the changes in GLUT1. To determine the effect of

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access 448 K OGURA and others · Inhibition of placental GLUT1 expression by glucose

Figure 4 Immunoblot analysis of GLUT1 in the mouse placental cells. (a) Cells were incubated with 5·5 mM (control) (lanes 1, 3) or 22 mM glucose (lanes 2, 4) for 4 days. Protein (25 µg) extracted from cells was applied to an SDS–10% polyacrylamide gel and transferred to a nitrocellulose filter. The filter was incubated with a rabbit anti-GLUT1 antiserum (lanes 1, 2) or with a non-immune rabbit serum (lanes 3, 4). Both antiserum and non-immune serum were used at a dilution of 1:8000. The positions of molecular markers (Da) are indicated on the left. (b) GLUT1 protein band intensities were quantitated by densitometry. Densitometry data obtained from three individual immunoblots are shown in arbitrary units. Results are expressed as mean&S.E. of three separate experiments. **P<0·01: statistically significant difference from control (unpaired t-test, n=3). (c) Time-dependent effect of glucose on GLUT1 protein expression. Cells were incubated with 5·5 mM (control) or 22 mM glucose for the periods indicated. Densitometry data obtained from three individual Western blots are shown in arbitrary units. Results are expressed as mean&S.E. of three separate experiments. **P<0·01: statistically significant difference from control on that day (unpaired t-test, n=3).

maternal diabetes on placental GLUT1 mRNA expres- RNA samples examined (data not shown). The GLUT1 sion, we analyzed the amounts of the steady-state band intensity of placentas from STZ-treated mice was GLUT1 mRNA in placentas from STZ-treated and con- significantly (P<0·01) decreased compared with control trol mice. GLUT1 mRNA was observed at 2·8 kb in all (normalized band intensity (arbitrary units): control, 1·26&0·12; STZ-treated, 0·76&0·23) (Fig. 7). The filter was rehybridized with a 32P-labeled GLUT3 cDNA. GLUT3 mRNA could not be detected, even in an auto- radiograph exposed for more than 10 days at "80 )C (data not shown). To determine the effect of maternal diabetes on placen- tal GLUT1 protein level, amounts of GLUT1 protein were determined in placentas from STZ-treated and control mice by immunoblot analysis using a specific antiserum against GLUT1 (Oka et al. 1988). When the filter was incubated with a non-immune serum, no specific bands were observed (data not shown). GLUT1 protein level in placentas from diabetic mice was signifi- cantly (P<0·01) decreased compared with that in controls (normalized band intensity (arbitrary units): control, 5·66&0·35; STZ-treated, 3·23&0·20) (Fig. 8), consistent with the results in the [3H]cytochalasin B binding assay and Northern blot analysis.

Discussion

Figure 5 Effect of insulin on 2-deoxyglucose uptake. Placental cells In mouse placentas, immunohistochemical studies reveal (1#106 cells/well) were incubated without and with various that GLUT1 protein is observed mainly in the membranes concentrations of insulin for 4 days and assayed for 2-deoxy- found on both sides of the trilaminar epithelial cells that glucose uptake in each well. Results are expressed as the mean&S.E. of data from 12 wells per group. Values do not show predominantly line the labyrinthine maternal vascular statistically significant differences (Scheffe’s test). spaces, and less abundantly in those of trophoblastic giant

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access Inhibition of placental GLUT1 expression by glucose · K OGURA and others 449

Table 1 Physiological parameters in control and STZ-treated pregnant mice on day 12 of gestation. (STZ was injected i.p. on day 8 of gestation)

Placental Fetal Maternal blood Maternal serum Fetal blood weight weight CRL glucose insulin glucose (mg) (mg) (mm) (mg/dl) (ng/ml) (mg/dl) Group Control (n=17) 81·6&8·5 214·4&28·1 10·9&0·64 156·4&28·9 1303·9&347·5 40·9&16·4 STZ-treated (n=11) 59·9&12·2** 153·4&44·5** 9·6&1·0** 436·9&169·3** 425·2&172·6** 232·2&56·3**

CRL, crown–rump length. **P<0·01 compared with control. cells (Devaskar et al. 1994). These findings suggest that To date, the regulation of placental GLUT1 has not GLUT1 expression in mouse placenta is specifically local- been extensively investigated. In the mouse placental cell ized to cells responsible for transplacental transport of culture, we have demonstrated that cAMP inhibited glucose. GLUT3 immunoreactivity is observed in the glucose transport activity (Sakata et al. 1996), suggesting trilaminar epithelial and basal trophoblastic cells (Devaskar that cAMP was a potent modulator of mouse placental et al. 1994). The GLUT1 mRNA level in the mouse GLUT1. In this report, incubation with a moderately high placenta increases as gestation proceeds, whereas the concentration (22 mM=400 mg/dl) of glucose decreased amount of GLUT3 mRNA is unchanged throughout mouse placental 2-[3H]deoxy--glucose uptake (Fig. 1a, gestation (Yamaguchi et al. 1996). These observations b) and expression of GLUT1 protein (Fig. 4) and mRNA suggest that GLUT1 is a main glucose transporter in (Fig. 3). GLUT1 protein level in mouse placental cells on mouse placenta, from midpregnancy through to term. day 4 of culture in 22 mM glucose medium was dimin- In this study, we failed to detect GLUT3 mRNA in ished by 45% (Fig. 4b), but 2-deoxyglucose uptake was mouse placenta on day 12 of pregnancy. Although a diminished by only 30% compared with control (Fig. 1a). previous study using 2 µg poly(A)+ RNA demonstrated Although it remains unclear why the changes observed in the presence of GLUT3 mRNA in mouse placenta, immunoblot analysis were larger than those in glucose GLUT3 mRNA expression on day 13 of pregnancy was uptake experiments, this phenomenon has been observed faint (Yamaguchi et al. 1996). These results suggest that previously (Sakata et al. 1996). the amount of GLUT3 mRNA expression in mouse placenta was insufficient to be detected by Northern blot analysis using 20 µg total RNA.

Figure 7 Northern blot analysis of GLUT1 mRNA in mouse placentas from the control and the STZ-treated mice on day 12 of gestation. Total RNA (20 µg) was electrophoresed and hybridized with 32P-labeled GLUT1 or with â-actin cDNA probes. Band intensities were quantitated by densitometry; that of GLUT1 Figure 6 Levels of [3H]cytochalasin B binding to placental (n=13) mRNA was normalized with respect to that of â-actin. and brain (n=3) membranes in the control and the STZ-treated Densitometry data obtained from three individual Northern blots mice on day 12 of gestation. Results are expressed as mean&S.E. are shown in arbitrary units. Results are expressed as mean&S.E. of three separate experiments. **P<0·01: statistically significant of three separate experiments. **P<0·01: statistically significant difference from control (unpaired t-test). difference from control (unpaired t-test, n=3).

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access 450 K OGURA and others · Inhibition of placental GLUT1 expression by glucose

(Sakata et al. 1996) demonstrated a time-dependent increase in glucose uptake in mouse placental cells that was consistent with the data in the present Fig. 1a and with those from human first-trimester chorionic villi (Gordon et al. 1995). Insulin is an important modulator of GLUT4 in muscle and adipose tissues. Although placentas have abundant insulin binding sites ( Jones et al. 1993), it has been accepted that the placenta is insulin-insensitive for glucose transport (Challier et al. 1986), which is consistent with our data (Fig. 5). STZ, a nitrosourea derivative of Streptomyces acromogenes that destroys â-cells in pancreas (Schein et al. 1967), has been used for induction of diabetes in rats (Younes et al. 1980) and mice (Schein et al. 1967). A previous study demonstrated that brain expression of GLUT1 mRNA in Figure 8 Immunoblot analysis of GLUT1 protein in mouse STZ-treated adult rats was significantly decreased com- placentas from control and STZ-treated mice on day 12 of pared with that in controls, which is similar with our data gestation. Protein (25 µg) extracted from placentas of control and (Fig. 6), whereas liver and kidney GLUT1 levels were not STZ-treated mice was applied to an SDS–10% polyacrylamide gel changed by STZ treatment (Koranyi et al. 1991). These and transferred to a nitrocellulose filter. The filter was incubated with a rabbit anti-GLUT1 antiserum (1:8000 dilution) or with a findings suggest that the concentration of GLUT1 in the non-immune rabbit serum. GLUT1 protein band intensities were brain was decreased to protect the brain – an organ quantitated on a densitometry. Densitometry data obtained from indispensable for life – from the toxicity of hyperglycemia. three individual immunoblots are shown in arbitrary units. Results It is known that placental glucose transport was closely are expressed as mean&S.E. of three separate experiments. **P<0·01: statistically significant difference from control (unpaired related to maternal plasma glucose concentration (Battaglia t-test, n=3). & Meschia 1978) and that glucose transport from the mother to the fetus is enhanced in diabetic rats (Thomas et al. 1990). Our in vitro finding that the placental GLUT1 A previous study demonstrated that incubation with level was decreased by a moderately high glucose concen- extremely low concentrations of -glucose (0–18·2 mg/dl) tration may enable us to speculate that placental GLUT1 enhanced 2-[3H]deoxy--glucose uptake in human first- expression was down-regulated by a hyperglycemia, to trimester trophoblast-like cells, and that the effects of protect fetuses from the glucose toxicity during pregnancy. -glucose on GLUT1 mRNA level paralleled the uptake To confirm this speculation, we investigated the effect of of 2-[3H]deoxy--glucose (Gordon et al. 1995). Other hyperglycemia on placental GLUT1 expression, fetal studies, and our data, suggest that the glucose concen- growth and fetal plasma glucose concentration, using an tration may be a potent regulator of GLUT1 expression in STZ-treated pregnant mouse model. We found that the placental cells. placental GLUT1 levels decreased in STZ-treated preg- In this study, we observed an increase in glucose uptake nant mice compared with those in controls, which was and in GLUT1 levels from day 1 to day 4 of culture in the consistent with our speculation. However, this protective presence of a 5·5-mM glucose concentration (Figs 1a,4c). mechanism may not be established within the fetal tissues The mechanisms responsible for this time-dependent because the levels of GLUT1 in the fetal muscle of increase are not fully understood. A previous study using STZ-treated pregnant mice (36·1 mM blood glucose) cultured human trophoblast cells showed that the level of were greater than those of the controls (Atkins et al. 1994). GLUT1 mRNA was significantly increased during 3 h of These results suggest that the placenta, but not the fetus, incubation with 20% FCS, and declined to basal values might have a mechanism of protection from maternal after 20 h of culture (Mouzon et al. 1994). We examined hyperglycemia. the acute effect of a high concentration of glucose on A previous study in rats demonstrated that glucose uptake in mouse placenta. Within 24 h, treatment GLUT1 mRNA and protein levels in placentas from with 22 mM glucose had no significant effects on glucose STZ-treated animals remained unchanged compared with uptake compared with controls (data not shown). Another controls, whereas GLUT3 mRNA and protein expression investigation using an HRP.1 rat placental cell line in placentas from diabetic rats increased compared with revealed that treatment of 0 and 5 mM glucose did not controls (Boileau et al. 1995). The findings in another affect GLUT1 protein levels during 48 h of incubation report (Das et al. 1998), that there was no significant (Das et al. 1998). The differences between these data and change in rat placental GLUT1 protein levels after STZ ours may derive from the fact that different species and cell treatment, were different from our data. This may be types were used in these studies. Our previous report attributable to the facts that the STZ-treated rats did not

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access Inhibition of placental GLUT1 expression by glucose · K OGURA and others 451 exhibit intrauterine fetal growth retardation, and that rat glucose uptake and transport in the non obese diabetic mouse placentas from day 20 of pregnancy were used in the pregnancy. American Journal of Obstetrics and Gynecology 171 1316–1323. previous report, whereas the fetuses were growth-retarded Economides DL & Nicolaides KH 1989 Blood glucose and oxygen and mouse placentas from day 12 of pregnancy were used tension levels in small-for-gestational-age fetuses. American Journal of in our study. Obstetrics and Gynecology 160 385–389. Finally, these results may enable us to accept the Gordon MC, Zimmerman PD, Landon MB, Gabbe SG & Kniss DA concept that the placental expression of GLUT1 in the 1995 Insulin and glucose modulate glucose transporter messenger ribonucleic acid expression and glucose uptake in trophoblasts diabetic mouse may be down-regulated to inhibit further isolated from first-trimester chorionic villi. American Journal of excessive glucose supply from the mother to the fetus. Obstetrics and Gynecology 173 1089–1097. Further investigation will be necessary to clarify the Gorga FR & Lienhard GE 1981 Equilibria and kinetics of ligand regulatory mechanisms of placental glucose transport. binding to the human erythrocyte glucose transporter. Evidence for an alternating conformation model for transport. Biochemistry 20 5108–5113. Haber RS, Weinstein SP, O’BoyleE&Morgello S 1993 Tissue Acknowledgements distribution of the human GLUT3 glucose transporter. Endocrinology 132 2538–2543. This work was partially supported by a grant for Scientific Jansson T, Wennergren M & Illsley NP 1993 Glucose transporter protein expression in human placenta throughout gestation and Research from the Ministry of Education, Science and intrauterine growth retardation. Journal of Clinical Endocrinology and Culture of Japan. Metabolism 77 1554–1562. Johnson LW & Smith CH 1980 Monosaccharide transport across microvillous membrane of human placenta. American Journal of Physiology 238 C160–C168. References Jones CJP, Hartmann M, BlaschitzA&DesoyeG1993 Ultrastructural localization of insulin receptors in human placenta. Asano T, Shibasaki Y, Kasuga M, Kanazawa Y, Takaku F, Akanuma American Journal of Reproductive Immunology 30 136–145. Y & Oka Y 1988 Cloning of a rabbit brain glucose transporter Kayano T, Burant CF, Fukumoto H, Gould GW, Fan Y-S, Eddy RL, cDNA and alteration of glucose transporter mRNA during tissue Byers MG, Shows TB, SeinoS&BellGI1990Human facilitative development. Biochemical and Biophysical Research Communications glucose transporters. Journal of Biological Chemistry 265 154 1204–1211. 13276–13282. Atkins V, Flozak AS & Ogata ES 1994 The effects of severe maternal Kitagawa T, MasumiA&Akamatsu Y 1991 Transforming growth

diabetes on glucose transport in the fetal rat. Endocrinology 135 factor-â1 stimulates glucose uptake and the expression of glucose 409–415. transporter mRNA in quiescent Swiss mouse 3T3 cells. Journal of Battaglia FC & Meschia G 1978 Principal substrates of fetal Biological Chemistry 266 18066–18071. metabolism. Physiological Reviews 58 499–527. Koranyi L, Bourey RE, James D, Mueckler M, Fiedorek FT & Birnbaum MJ, Haspel HC & Rosen OM 1986 Cloning and Permutt MA 1991 Glucose transporter gene expression in rat brain: characterization of a cDNA encoding the rat brain glucose pretranslational changes associated with chronic insulin-induced transporter protein. Proceedings of the National Academy of Sciences of hypoglycemia, fasting, and diabetes. Molecular and Cellular the USA 83 5784–5788. Neurosciences 2 244–252. Boileau P, Mrejen C, GirardJ&MouzonSH1995 Overexpression of Laemmli UK 1970 Cleavage of structural proteins during the assembly GLUT3 placental glucose transporter in diabetic rats. Journal of of the head of bacteriophage T4. Nature 277 680–685. Clinical Investigation 96 309–317. Morita Y, Tsutsumi O, Hosoya I, Taketani Y, OkaY&KatoT1992 Bradford MM 1976 A rapid and sensitive method for the quantitation Expression and possible function of glucose transporter protein of microgram quantities of protein utilizing the principle of GLUT1 during preimplantation mouse development from oocytes protein-dye binding. Analytical Biochemistry 72 248–254. to blastocysts. Biochemical and Biophysical Research Communications Challier JC, Hauguel-de MouzonS&Desmaizieres V 1986 Effect of 188 8–15. insulin on glucose transport and metabolism in the human placenta. Mouzon SH, Leturque A, Alsat E, Loizeau M, Brion DE & Girard J Journal of Clinical Endocrinology and Metabolism 62 803–807. 1994 Developmental expression of Glut1 glucose transporter and Cleveland DW, Lopata MA, MacDonald RJ, Cowan NJ, Rutter WJ c-fos genes in human placental cells. Placenta 15 35–46. & Kirschner MW 1980 Number and evolutionary conservation of Mueckler M, Caruso C, Baldwin SA, Panico M, Blench I, Morris á-andâ-tubulin and cytoplasmic â-andã-actin genes using specific cloned cDNA probes. Cell 20 95–105. HR, Allard WJ, Lienhard GE & Lodish HF 1985 Sequence and 229 Cool DR, Leibach FH, Bhalla VK, Mahesh VB & Ganapathy V 1991 structure of a human glucose transporter. Science 941–945. Expression and cyclic AMP-dependent regulation of a high affinity Oka Y, Asano T, Shibasaki Y, Kasuga M, KanazawaY&Takaku F serotonin transporter in the human placental choriocarcinoma cell 1988 Studies with antipeptide antibody suggest the presence of at line ( JAR). Journal of Biological Chemistry 266 15750–15757. least two types of glucose transporter in rat brain and adipocyte. Danics J & Schneider H 1975 Physiology: transfer and barrier Journal of Biological Chemistry 263 13432–13439. function. In The Placenta and its Maternal Supply Line, pp 98–124. Pedersen J 1977 The Pregnant Diabetic and Her Newborn, edn 2, Ed P Gruenwald. Lancaster: Medical and Technical Publishing pp 1–280. Baltimore: Williams and Wilkins. Co. Sakata M, Farooqui SM & Prasad C 1992 Post-transcriptional Das UG, Sadiq HF, Soares MJ, Hay WW Jr & Devaskar SU 1998 regulation of loss of rat striatal D2 dopamine receptor during aging. Time-dependent physiological regulation of rodent and ovine Brain Research 575 309–314. placental glucose transporter (GLUT-1) protein. American Journal of Sakata M, Kurachi H, Imai T, Tadokoro C, Yamaguchi M, Physiology 274 R339–R347. YoshimotoY&Miyake A 1995 Increase in human placental Devaskar SU, Devaskar UP, Schroeder RE, deMello D, Fiedorck FT glucose transporter-1 during pregnancy. European Journal of Jr & Mueckler M 1994 Expression of genes involved in placental Endocrinology 132 206–212.

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access 452 K OGURA and others · Inhibition of placental GLUT1 expression by glucose

Sakata M, Yamaguchi M, Imai T, Tadokoro C, Yoshimoto Y, Oka Yamaguchi M, Imai T, Maeda T, Sakata M, Akira M & Linzer DIH Y, KurachiH&Miyake A 1996 8-Bromo-cAMP inhibits glucose 1995 cAMP stimulation of placental proliferin and proliferin-related transport activity in mouse placental cells in culture. Journal of protein secretion. Endocrinology 136 2040–2046. Endocrinology 150 319–327. Yamaguchi M, Sakata M, OguraK&Miyake A 1996 Gestational Schein PS, Cooney DA & Vernon ML 1967 The use of nicotinamide changes of glucose transporter gene expression in the mouse to modify the toxicity of streptozotocin diabetes without loss of placenta and decidua. Journal of Endocrinological Investigation 19 antitumour activity. Cancer Research 27 2324–2332. 567–569. Shepherd PR, Gould GW, Colville CA, McCoid SC, Gibbs EM & Younes M, Schlichting R & Segers CP 1980 Glutathione S-transferase Kahn BB 1992 Distribution of GLUT3 glucose transporter protein activities in rat liver: effect of some factors influencing the in human tissues. Biochemical and Biophysical Research Communications metabolism of xenobiotics. Pharmacological Research Communications 188 149–154. 12 115–129. ZhouJ&BondyCA1993Placental glucose transporter gene ThomasC,ErikssonG&ErikssonU1990Effects of maternal diabetes expression and metabolism in the rat. Journal of Clinical Investigation on placental transfer of glucose in rats. Diabetes 39 276–282. 91 845–852. Wilson PT, Gershoni JM, HawrotE&Lentz TL 1984 Binding of á-bungarotoxin to proteolytic fragments of the alpha subunit of Torpedo acetylcholine receptor analyzed by protein transfer on Received 20 April 1998 positively charged membrane filters. Proceedings of the National Revised manuscript received 28 September 1998 Academy of Sciences of the USA 81 2553–2557. Accepted 27 October 1998

Journal of Endocrinology (1999) 160, 443–452

Downloaded from Bioscientifica.com at 09/30/2021 02:46:15AM via free access