INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

WOMEN WITH POORLY CONTROLLED insulin-dependent diabetes mellitus (IDDM) have a much higher incidence of early pregnancy complications (17, 30, 42). These include spontaneous miscarriages, early growth delay, and congenital malformations. The incidence of these complications decreases somewhat with improved maternal glycemic control during the period of fetal organogenesis. This reduced incidence, however, still remains three to four times higher than control. Because of this persistent elevation, it is possible that the metabolic insult leading to these complications may be occurring earlier in development, before organogenesis, specifically during the preimplantation period.

Maternal hyperglycemia adversely affects preimplantation progression from a one-cell to a blastocyst stage in a streptozotocin-induced or a nonobese diabetic (NOD) mouse model (11, 32, 33). In the NOD model at 96 h after superovulation and mating, only 20% of the recovered embryos reach a blastocyst stage compared with 90% among the nondiabetic mice. This developmental delay is reversible by treating the mothers with insulin before superovulation and mating and during the first 96 h of gestation. These findings suggested that some metabolite, elevated during periods of poor glycemic control, is responsible for the developmental retardation. This early preimplantation delay may be manifested later in gestation as a fetal loss or early growth delay. Alternatively, this early perturbation in development may predispose the fetus to a congenital malformation.

The purpose of this work, therefore, was to compare glucose utilization as defined by the combination of glucose uptake and subsequent metabolism in preimplantation embryos from control and streptozotocin-induced hyperglycemic mice. We hypothesized that maternal hyperglycemia or other metabolic factors may alter the preimplantation embryo's ability to utilize glucose and that this may lead to impaired development, akin to fuel-mediated teratogenesis (14).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Induction of hyperglycemia and recovery of embryos. The study was conducted using B6 × SJL F1 female mice, 4-6 wk of age (Jackson Laboratories, Bar Harbor, ME). Mice were given free access to food and water and were maintained on a 12:12-h light-dark cycle. Hyperglycemia was induced by a single intraperitoneal injection of streptozotocin (Sigma Chemical, St. Louis, MO) dissolved in sodium acetate, pH 4.4, at a dose of 190 mg/kg. Blood glucose levels were checked with tail blood by use of a Hemocue B glucose analyzer (Angelholm, Sweden) 4 days after the injection. Blood glucose levels of >250 mg/dl were considered hyperglycemic. In all mice, superovulation was achieved with an intraperitoneal injection of 10 IU/animal of pregnant mare serum gonadotropin (Sigma Chemical), followed 48 h later by 5 IU/animal of human chorionic gonadotropin (hCG; Sigma Chemical). Female mice were mated with males of proven fertility overnight after hCG injection. Mating was confirmed by identification of a vaginal plug.

General analytic procedures. Embryos were recovered from control vs. diabetic mice at 24, 48, 72, or 96 h after hCG and immediately freeze-dried by a procedure described elsewhere (31). Each embryo was washed twice in a simple salt solution containing no glucose and then was transferred with 0.5-1 µl of the same salt solution onto a glass slide with a braking pipette. The embryo was then quick-frozen immediately by dipping the slide into Freon-12 (CCl₂F₂) brought to its freezing point (170°C) with liquid N₂. The specimens were freeze-dried on the slide at 35°C in a glass vacuum tube at a vapor pressure of 0.01 mmHg. The slides were then stored at 20°C under reduced pressure. The general microanalytic procedures have been described for the analysis of single mouse ova and embryos (31, 38). All enzymes and reagents were obtained from Sigma Chemical unless otherwise noted. First, the freeze-dried embryo is transferred with a specially shaped hair point into a microliter droplet of the extraction reagent. For glucose, 2-deoxyglucose, and glucose 6-phosphate assays, the embryos were extracted for 20 min at room temperature in 0.5 µl of a weak acid (0.02 N HCl). For hexokinase activity assays, the embryos were extracted for 120 min at room temperature in 1 µl of a phosphate buffer consisting of 20 mM sodium phosphate, pH 7.4, 0.02% BSA, 0.5 mM EDTA, 5 mM mercaptoethanol, 25% glycerol, and 0.5% Triton X-100. These droplets are placed in wells 5 mm deep by 2 mm wide drilled in a piece of Teflon and covered with a layer of oil. For the metabolite assays, the extract is then heated to 80°C for 20 min. Enzymes and preformed reduced pyridine nucleotides, which might interfere with later measurements, are destroyed by heating in acid. The extract is returned to pH 8.1, 28 mM Tris HCl, by adding 0.2 µl of 0.1 M Tris base before proceeding to the assay. The final volume of this extraction aliquot is ~0.7 µl. At this point the extraction aliquots are either assayed immediately or stored at 70°C in the oil well in a vacuum tube placed under reduced pressure to about one-fourth atmosphere. For the microanalytic assays (see Fig. 1) in the first step, a 0.1- to 0.2-µl aliquot is removed from the acid extraction and used in the specific reaction sequence, ending in oxidation or reduction of a pyridine nucleotide. The excess pyridine nucleotide from the first step is destroyed with acid if it is in the reduced form or with alkali if it is in the oxidized form. The second step is the enzymatic cycling or amplification step. The principle of these reactions is illustrated in Fig. 1. NADPH is alternatively oxidized and reduced. In each oxidation/reduction cycle, 1 mol each of 6-phosphogluconate and glutamate is produced. Cycling rates up to 200,000 are achieved depending on the temperature and concentration of the two enzymes. After the desired multiple of amplification, the enzymes are inactivated with heat, and in the third step, the 6-phosphogluconate is measured by the fluorescence of the NADPH generated in the conversion of 6-phosphogluconate to ribulose 5-phosphate via 6-phosphogluconate dehydrogenase (EC 1.1.1.43). Fluorescence is measured directly in a 1-ml volume in 10 × 75-mm fluorometer tubes by use of a Farrand fluorometer. Calculations are based on internal standards and are therefore independent of variation in enzyme activities, temperature, or incubation time. Exact proportionality between readings and NADP concentration is usually achieved, because the nucleotide concentrations are kept far below the Michaelis constants for the enzymes. The general methodology and each of the specific assays for measuring glucose, glucose 6-phosphate, and hexokinase used in these methods have been described (38).

View larger version (24K): [in this window] [in a new window]   Fig. 1.   General microanalytic procedures. By use of the example of glucose in step I, the metabolite to be measured is linked to a reaction, generating an oxidized or reduced pyridine nucleotide. In step II, the pyridine nucleotide is amplified by adding a reaction mixture containing -ketoglutarate, ammonia chloride, and glucose 6-phosphate (G-6-P), with enzymes glutamate dehydrogenase (GDH) and glucose-6-phosphate dehydrogenase (G-6-PDH). By variation of the concentration of these enzymes and the temperature, reactions were cycled to obtain amplification of as much as 20,000- to 250,000-fold (see MATERIALS AND METHODS for details). Finally, in step III, one of the cycling reaction products, 6-phosphogluconate (6-P-gluconate), is measured by a simple fluorometric assay using 6-phosphogluconate dehydrogenase (6-PGDH). With the use of internal standards, the original concentration of the metabolite can be calculated. HK, hexokinase; ribulose-5-P, ribulose 5-phosphate.

2-Deoxyglucose uptake assays. Embryos at different embryonic stages were incubated at 25°C in 200 µM 2-deoxyglucose (DG) for several different time points ranging from 10 s to 60 min. After the incubation times, the embryos were washed in a DG-free, BSA-free buffer for 1 min and then quick-frozen on a glass slide as described in General analytic procedures. The embryos were then freeze-dried overnight and extracted as described above. To assay the embryonic DG, the following enzymatic steps were performed (Fig. 2). These assays have been described for other systems and embryos elsewhere (9, 10). First, in the oil well apparatus, a 0.1-µl aliquot of the extraction sample was added to a 0.1-µl aliquot of a 2× 6-phosphate removal reagent [60 mM Tris acetate, 0.04% BSA, 6 mM MgCl₂, 100 µM NADP⁺, and 100 µg/ml of glucose6-phosphate dehydrogenase (G-6-PDH)], which converted all the 6-phosphate compounds [2-deoxyglucose-6-phosphate (DG-6-P) and glucose 6-phosphate (G-6-P)] to 6-phosphogluconates via an excess of the enzyme glucose-6-phosphate (Leuconostoc mesenteroides) dehydrogenase. This reaction occurred at room temperature over 20 min. After completion of the first reaction, 0.05 µl of 0.2 N HCl was added and the reaction mix was heated to 80°C for 20 min to destroy the formed NADPH to avoid interference with subsequent steps. NaOH (0.2 N, 0.05 µl) was then added to neutralize the solution. In the third step, 0.1 µl of a 4× removal reagent (50 mM Tris acetate, 0.04% BSA, 200 mM potassium acetate, 1.2 mM ATP, 4 mM phosphoenolpyruvate, 40 µg/ml phosphoglucosisomerase, 65 µg/ml phosphofructokinase, 20 µg/ml pyruvate kinase, and 40 µg/ml hexokinase) was added to the reaction, and the reaction was allowed to occur over 20 min at room temperature. In this two-step reaction, hexokinase is added to convert the remaining free glucose and free DG to the 6-phosphate compounds. Phosphoglucosisomerase then selectively converts the formed G-6-P to fructose 6-phosphate but does not convert DG-6-P. This phosphoglucosisomerase reaction is then driven to completion by adding phosphofructokinase to convert the fructose 6-phosphate to fructobisphosphate. ATP, pyruvate kinase, and phosphoenolpyruvate are added to drive both the phosphofructokinase and hexokinase reactions to completion by replenishing ATP levels. After 20 min at room temperature, the reaction is heated to 80°C for 20 min to destroy the enzymes and prevent the back reactions. In the next step, 0.1 µl of a 5× DG-6-P reagent (50 mM Tris acetate, 0.04% BSA, 10 mM MgCl₂, 20 µm NADP⁺, and 250 µg/ml G-6-PDH) is added to convert the remaining DG-6-P from the previous step to deoxyglucose-6-phosphogluconate, generating an equimolar amount of NADPH. This reaction is performed at room temperature for 40 min. In the final step, 0.1 µl of 0.3 N NaOH is added, and the reaction is heated to 80°C for 20 min to destroy the enzymes. The NADPH generated in the final enzymatic reaction is then cycled overnight for 150,000 cycles as has been described. As previously reported, mouse blastocysts accumulate DG above equilibrium values (9). This phenomenon, which is not seen at any other preimplantation developmental stage, is thought to be due to the presence of a putative active hexose transporter.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   2-Deoxyglucose (DG) assay for measuring glucose uptake in embryos (see MATERIALS AND METHODS for details and definitions). Step 1: 6-phosphate removal step. Step 2: glucose removal step/deoxyglucose conversion to deoxyglucose-phosphate step. Step 3: desired substrate reaction generating NADPH. Step 4: amplification step.

Competitive reverse transcription-polymerase chain reaction techniques. Groups of 100-150 embryos were collected, pooled, and homogenized by centrifugation over a QIAshredder (Qiagen, Chatsworth, CA). Total RNA was extracted from the homogenate by denaturing with a -mercaptoethanol buffer, adjusting the sample to the appropriate binding conditions with 70% ethanol, and then applying the samples to a silica gel column (RNeasy total RNA kit, Qiagen). Successive washes and centrifugation were then used to eliminate contaminants, and the isolated RNA was eluted in 30 µl of diethyl pyrocarbonate-treated water. An aliquot of this embryonic RNA (11.2 µl) was then added to 500 ng of random hexamers (GIBCO, Gaithersburg, MD), heated to 65°C for 10 min, and placed on ice for 5 min. A master mix was then added to produce a 20-µl reaction with the following final concentrations: 50 mM Tris · HCl, pH 8.3, 50 mM KCl, 1 mM dNTP, 10 mM dithiothreitol (DTT), 1 U/ml RNAsin (Promega), and 10 U/ml M-MLV reverse transcriptase (Superscript, GIBCO). The reaction was incubated for 50 min at 42°C and terminated by heating at 70°C for 10 min.

Immunofluorescent labeling to quantitate protein expression. Immunofluorescence staining techniques have been described in embryo preparations previously (19). All labeling was performed in microdroplets. The embryos at different stages were fixed in 3% neutral buffered formaldehyde for 30 min and then permeabilized with 0.1% Tween 20 for 10 min. The embryos were then blocked by incubating for 60 min in 20% donkey serum in PBS containing 2% BSA. Embryos were then washed three times for 10 min each in PBS-BSA and incubated in the affinity-purified primary antibody (polyclonal mouse GLUT-1, GLUT-2, or GLUT-3) at dilutions ranging from 6 to 15 µg/ml for 30 min. The GLUT-2 antibody was a generous gift from Dr. Bernard Thorens, University of Lausanne (44). For negative controls, embryos were incubated in preimmune serum at a dilution of 1:200-1:500, or, as for GLUT-3, the original peptide was used to generate the antibody at a concentration of 20 µg/ml. The embryos were then washed three times for 10 min each in PBS-BSA and incubated with the secondary antibody, FITC-labeled goat anti-rabbit IgG (Chemicon, Temecula, CA) at a concentration of 1:80, and rhodamine-conjugated phalloidin (Molecular Probes, Eugene, OR), to stain cytoskeletal structures, at a concentration of 1:80. Finally, the embryos were washed three times for 10 min each in PBS-BSA and mounted in drops of Vectoshield (Vecto Laboratories, Burlingame, CA) under a supported coverslip. Fluorescence was detected with a Bio-Rad MRC-600 laser-scanning confocal microscope. Confocal images were taken at ×63 magnification. Total fluorescence per embryo was expressed as a number per area with NIH Image (version 1.60). Similar fluorescence ratios were derived for preimmune serum images and subtracted from the GLUT images to generate a total fluorescence value. The average of three of these mean values was then compared between control and diabetic embryos.

Statistical analysis. Unpaired t-tests were used to compare diabetic and control groups. The Statview 4.0 (Abacus Concepts, Berkeley, CA) statistical package was used for all analyses. Significance was defined as P < 0.05. Error bars represent ± SE.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Intracellular free glucose and G-6-P. Because previous studies had found negligible hexokinase activity (2, 40) and no glycolytic activity until the blastocyst stage in mouse preimplantation embryos, intraembryonic free glucose levels were measured at different embryonic stages. These experiments were done to determine whether, under hyperglycemic conditions, free glucose was increasing in these embryos or, alternatively, whether the transported glucose was being metabolized by some other metabolic pathway that could then be investigated. Free glucose and G-6-P were measured using the microanalytic techniques described in MATERIALS AND METHODS, as seen in Fig. 3, and expressed as millimoles per kilogram wet weight. These results represent three different sets of embryos with 5-7 embryos for each time point in each experiment. Free glucose levels among the embryos from diabetic mice were consistently higher at 24 and 72 h, corresponding to the one-cell and morula stages (24 h: diabetic, 2.59 ± 0.25 vs. control, 1.6 ± 0.35 mmol/kg wet wt; P < 0.05; 72 h: diabetic, 3.59 ± 0.1 vs. control, 1.32 ± 0.27 mmol/kg wet wt; P < 0.01). Conversely, free glucose levels consistently dropped to barely detectable levels at 48 and 96 h, corresponding to two-cell and blastocyst stages (48 h: diabetic, 0.23 ± 0.09 vs. control, 2.30 ± 0.43 mmol/kg wet wt; P < 0.001; 96 h: diabetic, 0.31 ± 0.29 vs. control, 5.12 ± 0.17 mmol/kg wet wt; P < 0.001). G-6-P levels were measured at the same intervals and were found to be barely detectable and not significantly different between embryos from control vs. diabetic mice (data not shown). This significant drop in free glucose at 48 and 96 h could represent either a decrease in glucose transport or an increase in hexokinase activity in response to the hyperglycemic conditions.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Intraembryonic free glucose concentration in embryos from diabetic vs. control mice. Free glucose levels are expressed as mmol/kg wet weight in embryos recovered at different time points after human chorionic gonadotropin (hCG) and mating from control mice () vs. streptozotocin-induced hyperglycemic mice (). Values are mean ± SE data from ~15-20 individual embryos from each time point of embryo recovery. Significance at 48 and 96 h, * P < 0.001; at 72 h, + P < 0.01; at 24 h, # P < 0.05.

Hexokinase activity. As seen in Fig. 4, hexokinase activity was not significantly different at any stage in embryos from control vs. diabetic mice. The overall pattern of activity, low levels at the early cleavage stages with an exponential increase at the morula and blastocyst stages, has been reported elsewhere in the literature and is thought to parallel the increase in glycolytic capacity experienced by these embryos at the blastocyst stage (2, 40).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Intraembryonic hexokinase activity in embryos from diabetic vs. control mice. Hexokinase activity expressed as pmol · embryo1 · min1 in embryos recovered at different time points after hCG and mating from control mice () vs. streptozotocin-induced hyperglycemic mice (). Results are not significant at any time point.

Glucose transport. To determine whether the decrease in glucose utilization was attributable to a decrease in uptake, glucose transport into single embryos at different developmental stages from diabetic vs. control mice was measured. 2-DG uptake experiments were conducted in three separate experiments for each of the different embryonic stages. Approximately 7-10 single embryos were assayed at each time point from each stage. Ten minutes was chosen as the representative time point because the uptake curve was still linear at this time. There was a significant difference in the DG uptake between the control and diabetic embryos at 48 and 96 h, corresponding temporally to the decrease in intraembryonic free glucose levels (48 h: diabetic, 0.037 ± 0.003 vs. control, 0.091 ± 0.021 mmol · kg wet wt¹ · 10 min¹; P < 0.05; 96 h: diabetic, 0.249 ± 0.008 vs. control 0.389 ± 0.007 mmol · kg wet wt¹ · 10 min¹; P < 0.02; Fig. 5). Because hexokinase activity was unchanged at 48 and 96 h, whereas transport dropped significantly, the change in free glucose levels and thus glucose utilization at these time points can be attributed to the decrease in transport activity.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   DG uptake in embryos from diabetic vs. control mice. Intraembryonic uptake of DG per 10 min in embryos from control mice (open bars) and diabetic mice (closed bars). Significance at 48 h, * P < 0.05; at 96 h, # P < 0.02.

Competitive reverse transcription-PCR. The levels of mRNAs for GLUT-1, GLUT-2, and GLUT-3 were determined by quantitative PCR using three different groups of embryos for each time point. The mean values ± SE were compared among the three groups of embryos. The quantitative PCR data indicate that the levels of GLUT-1 mRNA in control embryos were 1.8-fold greater than those of diabetic embryos recovered at 48 h, or a 44% reduction in the GLUT-1 mRNA (48 h: diabetic, 0.105 ± 0.006 vs. control, 0.189 ± 0.012 fg/embryo; P < 0.001; Fig. 6). At 96 h, control blastocysts expressed 3.2-fold more GLUT-1 mRNA, or a 68% reduction in the diabetic embryos (96 h: diabetic, 0.125 ± 0.021 vs. control, 0.394 ± 0.033 fg/embryo; P < 0.005). Expression of GLUT-2 and GLUT-3 mRNA was not detected until ~68 h after hCG. The levels of GLUT-2 and GLUT-3 mRNA were not significantly different at 72 h after hCG; however, at 96 h, control blastocysts expressed 2.7- and 4.3-fold more GLUT-2 and GLUT-3 mRNA, respectively, than blastocysts recovered from diabetic mice, corresponding to a 63 and 77% reduction in mRNA, respectively (GLUT-2 diabetic, 0.091 ± 0.009 vs. control, 0.243 ± 0.01 fg/embryo; GLUT-3 diabetic, 0.05 ± 0.009 vs. control, 0.213 ± 0.031 fg/embryo; P < 0.001, GLUT-2 and GLUT-3; Fig. 7). These data support the hypothesis that the facilitative glucose transporters are downregulated at the mRNA level in the preimplantation embryo from diabetic hyperglycemic mice.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   GLUT-1 mRNA levels in embryos from control vs. diabetic mice. Embryonic amount of GLUT-1 mRNA in fg/embryo in embryos recovered at different time points from control (open bars) vs. hyperglycemic (closed bars) mice. Significance at 48 h, * P < 0.001; at 96 h, + P < 0.005.

View larger version (12K): [in this window] [in a new window]   Fig. 7.   GLUT-2 (A) and GLUT-3 (B) mRNA levels in embryos from control vs. diabetic mice. Embryonic amount of GLUT-2 or GLUT-3 mRNA in fg/embryo in embryos recovered at different time points from control (open bars) vs. hyperglycemic (closed bars) mice. Significance: 96 h, * P < 0.001.

Immunofluorescence microscopy. To determine whether this downregulation was also seen at a protein level, quantitative immunofluorescence microscopy was used to measure GLUT protein in single embryos from hyperglycemic vs. control mice. Two-cell embryos recovered from diabetic mice at 48 h post-hCG expressed 49 ± 6% less GLUT-1 protein than control embryos at the equivalent stage. Embryos recovered at 96 h expressed a 66 ± 4% reduction (Fig. 8). Embryos recovered at 72 h post-hCG from diabetic mice expressed equivalent amounts of GLUT-2 and GLUT-3 protein; however, 24 h later at 96 h post-hCG, the diabetic embryos demonstrated a 90 ± 5 and 84 ± 6% reduction in the amount of GLUT-2 and GLUT-3 protein (Fig. 9). These quantitative protein data confirm that the downregulation of the facilitative glucose transporters seen at the mRNA level was also seen at the protein level and thus could account for the decrease in glucose transport and free glucose seen in the embryos from hyperglycemic mice at 48 and 96 h after mating.

View larger version (93K): [in this window] [in a new window]   Fig. 8.   Confocal immunofluorescent labeling to quantitate GLUT-1 protein expression. Embryos recovered 48 h after hCG and mating from control mice (A) vs. diabetic mice (B) labeled with polyclonal GLUT-1 antibody. Embryos recovered 48 h after hCG and mating from control mice (C) vs. diabetic mice (D) labeled with preimmune serum antibody. Embryos recovered 96 h after hCG and mating from control mice (E) vs. diabetic mice (F) labeled with polyclonal GLUT-1 antibody.

View larger version (74K): [in this window] [in a new window]   Fig. 9.   Confocal immunofluorescent labeling to quantitate GLUT-2 and GLUT-3 protein expression. Embryos recovered 72 h after hCG and mating from control mice (A, E) vs. diabetic mice (B, F) labeled with polyclonal GLUT-2 (A, B) or GLUT-3 (E, F) antibody, respectively. Embryos recovered 96 h after hCG and mating from control mice (C, G) vs. diabetic mice (D, H) labeled with polyclonal GLUT-2 (C, D) or GLUT-3 (G, H) antibody, respectively.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study suggests that maternal hyperglycemia induces a decrease in intracellular free glucose in preimplantation embryos, specifically at 48 and 96 h postmating. This decrease was due to a decrease in glucose transport and not an increase in hexokinase activity. DG uptake in single embryos was decreased in those from diabetic mothers compared with those from controls at 48 and 96 h after the hCG time points. This decrease in transport was reflected by the simultaneous decrease in facilitative glucose transporters at both the mRNA and protein levels, suggesting downregulation of these transporters due to the hyperglycemic state.

Two questions need to be investigated. First, how does glucose regulate the transporter expression? Second, how does a decrease in glucose transport lead to abnormal embryonic development, eventually manifested as growth delay, congenital malformations, or spontaneous miscarriage? To address the first question, it is important to note that similar downregulation of the facilitative glucose transporters in response to high ambient glucose is seen in vivo and in vitro in many different cell types (16, 25, 39). In some cell culture systems, glucose deprivation causes a rapid and sustained increase in GLUT-1 mRNA and protein, reflected in the acute increase in glucose transport (18, 45). Conversely, high glucose conditions induce an acute downregulation of transport (26, 41). The mechanisms for the regulation of GLUT-1 by glucose availability are not clear and differ depending on the cell type. The metabolic control of glucose transport can occur at multiple levels, including intrinsic activity of the transporter, localization and content of the transporter protein, and transcriptional and pre- and posttranslational regulation. The control of GLUT-1 expression in preimplantation embryos seen in response to maternal hyperglycemia in this study occurs at the mRNA level and thus may reflect a complex series of events orchestrated by transcription factors acting on upstream DNA sequences of the GLUT-1 gene. This may occur directly or indirectly via posttranscriptional modulation of the transcription factors. Such speculation is founded, given that glucose regulates the expression of the insulin receptor in a similar fashion (12).

Next, further investigation is needed to determine whether a decrease in glucose transport at these critical time points in preimplantation development, namely the two-cell and blastocyst stages, leads to abnormalities in further development. Mouse preimplantation embryos preferentially metabolize pyruvate until the late morula/early blastocyst stage, at which time glucose becomes the predominant energy substrate (5). Glucose, however, is required earlier than the blastocyst stage for optimal development (7, 8, 29). Embryos from F1 hybrid strains need to be exposed to glucose for a period of time as short as 22 h between 24 and 72 h post-hCG to develop to blastocysts in vitro (29). Clearly, preimplantation embryos require some degree of glucose exposure at ~48 h post-hCG, the same time the diabetic embryos in this study are experiencing a decrease in glucose transport, leading to significantly lower intraembryonic levels of free glucose. Glucose deprivation at this critical time of glucose need may be responsible for the later problems these embryos experience.

These findings raise the question, what role does glucose play at this early stage in development, before the activation of anaerobic metabolism? Glycogen synthesis occurs during this period. Glycogen levels rise 10-fold between the one- and two-cell stages in mouse embryos (20, 35). Adequate glycogen stores may be necessary to meet the embryos' energy needs later during compaction or implantation. Alternatively, glucose may be required for the glycosylation of key cellular glycoproteins. Previous studies have shown that, in the presence of tunicamycin, an inhibitor of glycosylation, embryos can undergo compaction but cannot maintain compaction and thus degenerate (46). It has been postulated that the preimplantation embryo relies on a Ca⁺-independent mechanism of cell adhesion to maintain compaction (34). Furthermore, fucosylated cell surface glycoproteins are required for this Ca⁺-independent stabilization of the compacted state (24). Thus deprivation of glucose at this stage could impair later development of the embryo by impairing surface glycoprotein synthesis. Such an early insult in the development of the embryo most likely would lead to embryo loss before implantation.

In those embryos reaching a blastocyst stage, the next question is, how would the decrease in glucose transport affect further development? Several recent studies have linked a decrease in glucose transport to the initiation of apoptosis (4, 23). In models of neuronal development and trophic factor deprivation, a reduction in glucose uptake is one of the earliest changes observed in the cascade of apoptotic events leading to programmed cell death (PCD) (22). Moreover, previous studies have shown that maternal diabetes and hyperglycemia per se adversely affect single cell viability at the blastocyst stage. Pampfer et al. (36) reported a decrease in the number of inner cell mass (ICM) and trophectoderm (TE) cells in blastocysts from diabetic rats. The mechanism behind this decrease and the timing of this cellular loss were not pursued. Similarly, Lea et al. (27) found a 20% reduction in the number of cells in the ICM of blastocysts recovered from the spontaneous IDDM rat model, the BB/E rat. These embryos were morphologically characterized as having cellular blebbing and nuclear condensation, suggestive of apoptosis. It is possible that the hyperglycemia-induced downregulation of the GLUTs and subsequent fall in free glucose as seen at the blastocyst stage in this study trigger the premature onset of an increase in PCD, leading to loss of key progenitor ICM and TE cells. Because a critical threshold in the number of ICM cells is required for normal postimplantation development (43), such an apoptotic event may predispose these diabetic embryos to dysmorphogenesis, fetal loss, or early growth delay, all events occurring at an increased incidence in women with IDDM.