Maternal Diabetes Adversely Affects Preovulatory Oocyte Maturation, Development, and Granulosa Cell Apoptosis
http://www.100md.com
内分泌学杂志 2005年第5期
Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology, Washington University in St. Louis, St. Louis, Missouri 63110
Address all correspondence and requests for reprints to: Dr. Kelle Moley, Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology, Washington University in St. Louis, 4911 Barnes Plaza Road, Box 8064, St. Louis, Missouri 63110-1094. E-mail: moleyk@wustl.edu.
Abstract
Maternal diabetes adversely affects preimplantation embryo development and pregnancy outcomes. The objective of this study was to determine whether diabetes has an impact at an earlier stage of development, the preovulatory oocyte. Models of both acute and chronic insulin-dependent diabetes were used. Acute hyperglycemia was induced by a single streptozotocin injection. Akita mice, which harbor an autosomal dominant mutation causing them to be chronically hypoinsulinemic and hyperglycemic, were used. In both models, preovulatory oocytes were markedly smaller when compared with control animals. A significantly greater number of control oocytes had progressed to meiotic maturation before diabetic oocytes. Both models were found to have smaller, less developed ovarian follicles with a greater number of apoptotic foci by histological evaluation as well as by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling staining. Immunohistochemistry displayed a greater amount of TNF-related apoptosis-inducing ligand (TRAIL) and KILLER, a key murine ligand and receptor involved in the extrinsic pathway, expressed in cumulus cells from hyperglycemic mice compared with controls, suggesting that this apoptotic pathway may be up-regulated under diabetic stress. Elevated KILLER expression was also confirmed through Western blotting. Connexin-43 expression was found to be lower by immunohistochemistry and Western blot analysis in the diabetic samples. Both models of maternal hyperglycemia and hypoinsulinemia may have a detrimental effect on oocyte maturation and development as detailed by the smaller sizes of oocytes and developing ovarian follicles, the lowered percentage reaching germinal vesicle breakdown, and the greater amount of apoptosis. In addition, there may be dysfunctional or decreased communication in diabetic oocytes, as demonstrated by lower expression of connexin-43.
Introduction
IN HUMANS, TYPE 1 OR insulin-dependent diabetes has been found to negatively affect pregnancy by causing poor prenatal outcomes such as an increased risk of congenital anomalies and early miscarriage (1, 2). Likewise, maternal diabetes has been found to adversely affect murine preimplantation embryo development in models of type 1 diabetes (3, 4, 5, 6). Previous studies from our laboratory have suggested that an insult occurring before the one-cell zygote stage may have an impact on the outcome of the resulting embryo. Zygotes removed from streptozotocin- or alloxan-induced hypoinsulinemic and hyperglycemic mice demonstrate retarded in vivo development to a two-cell stage with a lower percentage of two-cell zygotes recovered at 48 h after human chorionic gonadotropin (hCG) compared with nondiabetic controls (7). Similarly, two-cell embryos recovered from diabetic mice and cultured in vitro in control media also experience significant delay in their progression to the blastocyst stage compared with embryos from control mice. As a result of these findings, the oocytes from these diabetic mice were also examined. These oocytes from chemically induced diabetic mice experience significant delays in germinal vesicle breakdown (GVBD) and resumption of meiosis I. This previous study was limited, however, and focused primarily on the effects of hyperglycemia on the embryo, not oocyte. Since those initial studies, another group has confirmed these findings (8, 9). To further investigate the developmental impact of hyperglycemia on oocytes and folliculogenesis, including the effect on maturation, growth, and the surrounding granulosa cells, the following studies were undertaken. We hypothesize that maternal diabetes would have a detrimental effect on oocyte maturation and that the follicular environment of the enclosed cumulus complex and oocyte immediately after hCG administration would be adversely affected.
In the developing oocyte, granulosa cells surround the oocyte, support its growth, and provide hormonal supplementation. They form a multilayered structure around the oocyte. Paracrine signaling and gap junctional communication occur between the granulosa cells and the oocyte. The presence of normal granulosa cell communication and development are critical for differentiation and oocyte growth to occur (10). Because of the importance of this relationship, we also chose to examine whether the diabetic state influenced granulosa cell intercellular communication. It is possible that the maturational delay suspected in oocytes from diabetic mice may be a result of poor paracrine communication between these two compartments because of poor intercellular talk among granulosa cells. Connexin-43 is a key gap junction protein expressed in granulosa cells that is necessary for intercellular communication between granulosa cells and normal folliculogenesis (10). A lack of connexin-43 expression, however, results in abnormal oocytes as well as poor folliculogenesis. Furthermore, diabetic human retinal pericytes display lower levels of connexin-43 when compared with control pericytes (11). Finally, recent studies have shown that connexin-43 expression in granulosa cells is inversely related to apoptotic cell death in the avian follicle (12). Because of these links between diabetes, decreased connexin-43 expression, and apoptosis, one of our objectives in this study was to see whether connexin-43 levels were altered in diabetic cumulus-enclosed oocytes (CEOs).
It is well established that the vertebrate ovarian follicles can undergo apoptosis, via a process called atresia, and this apoptosis is initiated within the granulosa cell layer. In previous studies, we have established that hyperglycemia induces apoptosis in the murine blastocyst. One objective of this study is to examine whether an external stress such as hyperglycemia could cause higher levels of apoptosis in the granulosa cells of developing follicles that might influence oocyte development, maturation (13), and perhaps communication between the granulosa cells and the oocyte. Poor outcomes have previously been associated with granulosa cell apoptosis. Granulosa cell apoptosis has been found to be accelerated in human patients with unexplained infertility (14). Increased apoptosis of surrounding cumulus cells has been correlated with oocyte maturational delay and poor pregnancy outcomes (15, 16).
Two major pathways are involved with apoptosis initiation in vertebrates. They include the intrinsic apoptotic pathway whereby mitochondrial disruption occurs as a result of death-promoting members of the Bcl-2 family, leading to the release of factors that promote caspase-9 activation and downstream apoptosis. The extrinsic apoptotic pathway involves activation of the cell-death receptors by ligands that belong to the TNF superfamily and downstream activation of caspase-8. The two pathways appear to be connected by the protein Bid, which contains Bcl-2 homology domains. Caspase-8, which is usually activated in the extrinsic pathway, can trigger mitochondrial cytochrome c release and thus cause downstream activation of the intrinsic pathway (17). Previously, our laboratory had determined that the intrinsic pathway was active during hyperglycemia-induced apoptosis in the preimplantation blastocyst-stage embryo via the overexpression of Bax, a death-promoting member of the Bcl-2 family (3, 18). It was also discovered that apoptosis was increased in the blastocysts obtained in vivo vs. those cultured in vitro in high-glucose conditions, suggesting that this apoptosis may be occurring under the combined influence of both the extrinsic and intrinsic pathways. In this study, the effects of hyperglycemia on activation of the extrinsic pathway, specifically the TNF-related apoptosis-inducing ligand (TRAIL) (TNF superfamily) in the ovarian follicle, are examined. TRAIL mediates apoptosis through stimulation of DR5 or murine KILLER receptor.
In this study, we find that hyperglycemia, both short and long term, has a significant impact on the developing preovulatory oocyte, causing slower growth, delayed maturation, increased apoptosis in the surrounding granulosa cells, up-regulation of the extrinsic apoptotic pathway, and diminished levels of a key gap junction protein.
Materials and Methods
Oocyte retrieval
All mouse studies were approved by the animal Studies Committee at Washington University School of Medicine and conform to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health. Female immature C57BL/6JXSL/J F1 mice (age 20–24 d) were used for most experiments. Diabetic and age-matched controls received 10 IU equine chorionic gonadotropin, which is also known as pregnant mare’s serum gonadotropin, by ip injection, and 48 h later, they either were killed or given an injection of 5 IU hCG. At the appropriate time points after hCG injection (time 0, 2, and 6 h), the ovaries were removed and placed in a dish containing 1.5 ml culture medium. Oocytes were isolated by puncturing the antral follicles with sterile needles, and then washed through several changes of medium. Before puncturing, the ovaries were stored briefly in an incubator with the settings of 5% CO2, 5% O2, 90% N2 atmosphere at 37 C.
Induction of hyperglycemia
To generate a diabetic model, female mice (age 20–24 d) received a single injection of streptozotocin at a dose of 190 mg/kg (dissolved in sodium citrate buffer, pH 4.4; Sigma Chemical Co., St. Louis, MO). Four days after injection, a tail-blood sample was measured for glucose concentrations via a Hemocue B glucose analyzer (Stockholm, Sweden). If glucose levels were more than 240 mg/dl, these mice were selected and received a priming injection of equine chorionic gonadotropin. A few control mice were also randomly selected; their blood sugar was checked to ensure that it was less than 240 mg/dl.
Akita mice
Akita mice have an autosomal dominant mutation, or Mody mutation, resulting in hyperglycemia and notable pancreatic ?-cell dysfunction (19, 20). Diminished levels of both proinsulin and insulin are produced with resultant hyperglycemia. These mice were maintained at our animal facility, received the standard murine chow diet, and kept on a 12-h light, 12-h dark cycle. These mice were not growth retarded at any stage of development compared with control mice. All Akita mice had glucose levels checked at approximately 5–6 wk through a tail-blood sample by a Hemocue B glucose analyzer. If glucose levels were more than 300 mg/dl, these mice were considered to have the mutation. Age-matched controls were C57BL/6 females; these mice also had glucose levels that were checked by the glucometer.
Oocyte maturation and size
Oocytes from the ruptured antral follicles were observed using Hoffman optics on an inverted Nikon microscope (TMS scope). GVBD was assessed, and the oocytes that retained a germinal vesicle and/or the nucleolus failed to demonstrate maturation. Maturation was expressed as a percentage of GVBD. Oocyte diameters were measured excluding the zona pellucida with an eyepiece graticule (x200 magnification). The volume of each oocyte was calculated based on the formula for the volume of a sphere [(4/3)(r3)]. These measurements were done at 2 and 6 h after hCG.
Ovarian sectioning
Mice were killed, and the ovaries were obtained at specific time points (2 and 6 h). The ovaries were quickly submerged in liquid nitrogen and flash frozen. The ovaries were sectioned using a cryostat machine (Leica 1850, CM) in 11-μm sections. Each ovary was sectioned throughout, and one of every eight ovarian sections was used for immunohistochemical staining.
Evaluation of apoptosis by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay
The CEOs and cryopreserved ovarian slices were fixed in 3% paraformaldehyde, permeabilized with 0.1% Tween 20, and then incubated in fluorescein-labeled dUTP and terminal transferase in the dark for 1 h at 37 C to label fragmented 3' DNA (TUNEL, Cell Death in Situ Kit; Roche Molecular Biochemicals, Mannheim, Germany) as previously described. Counterstaining of all nuclear DNA was achieved by incubating the tissue in 4 μM To-Pro-3-iodide (Molecular Probes, Eugene, OR) (blue channel) for 20 min. CEOs were visualized using confocal immunofluorescent microscopy with a Nikon C1 laser scanning confocal microscope at x63 magnification. Apoptosis was expressed as the percentage of TUNEL-positive nuclei per CEO or the number of TUNEL-positive follicles per total number of follicles. Student’s t test was additionally used for statistical analysis.
Immunohistochemistry
CEOs were fixed on glass slides in 3% paraformaldehyde for 20 min and permeabilized with 0.1% Tween 20 for 10 min. The CEOs were first incubated in 20% normal donkey serum (Pierce, Rockford, IL) in PBS containing 2% BSA (PBS/BSA) for 1 h to block any nonspecific binding. They were then washed in PBS/BSA and then incubated in the primary antibody [anti-DR5 (Santa Cruz Biotechnology, Santa Cruz, CA), N-19; mouse anti-actin (Chemicon International, Temecula, CA), 1:1000; anti-Trail (Santa Cruz Biotechnology), H-257, 1:200; and rabbit anti-connexin-43 (Zymed Laboratories, San Francisco, CA), 1:50] for 1 h at room temperature. The CEOs were then washed with PBS/BSA and incubated in the appropriate secondary antibody, goat antirabbit Alexa Fluor 488 antibody or goat antimouse Alexa Fluor 488 antibody, for approximately 30 min. The slides were washed, and nuclear staining was performed in 4 μM To-Pro-3-iodide. After extensive washing in PBS/BSA, confocal immunofluorescent microscopy (Nikon C1 laser scanning microscope) was then used to detect fluorescence as described above.
Expression of DR5 (KILLER) receptor and connexin-43 by Western analysis
CEOs were collected at specific time points (0, 2, and 6 h) for DR5 and (t = 0 h, 6 h) for anti-connexin 43 in equivalent groups (n = 70). The pooled samples were washed in several different washes of human tubal fluid and were frozen in human tubal fluid. (HTF) (Irvine Scientific, Irvine, CA). HeLa cell extracts were used as a positive control for the DR5 receptor. The CEOs were placed in Laemmli buffer with ?-mercaptoethanol. All samples were boiled for at least 5 min at 100 C. The samples were subjected to 10% SDS-PAGE and transferred to nitrocellulose. Mouse DR5 or KILLER protein was then detected with a rabbit polyclonal DR5 antibody (Santa Cruz Biotechnology; 1:500) as the primary antibody and a horseradish peroxidase (HRP)-labeled goat antirabbit antibody (Santa Cruz Biotechnology) as the secondary antibody. Connexin-43 protein was detected by a rabbit polyclonal antibody (Zymed Laboratories; 1:1000), and an HRP-labeled donkey antirabbit antibody (Pierce; 1:20,000) was used as the secondary antibody. HRP-labeled bands were quantified using NIH Image (version 1.6). All experiments were performed in triplicate. Mouse anti-actin (Chemicon International; 1:2000) was used as the loading control for all blots.
Ovarian sections
Ovaries were obtained at specific time points and submerged in Bouin’s fixation solution (Sigma). They were then paraffin embedded, cut, and stained with hematoxylin and eosin at the histology core at Washington University in St. Louis Medical School. The sections were then examined and evaluated by a blinded animal pathologist.
Statistics
Differences between the two groups with protein expression and percent TUNEL staining were analyzed using Student’s t test. If multiple groups were being compared, ANOVA was used with Fisher post hoc test. Results are expressed as means ± SE of at least three separate experiments. For each set of oocyte experiments, at least five mice were used for each group, and approximately 15 oocytes were obtained from each animal. All experiments were performed at least two times, and the majority of the experiments were performed three times. The numbers of ovaries sectioned are described below.
Results
Oocytes from streptozotocin-induced diabetic mice were smaller and had a delay in maturation
To test our hypothesis that diabetic mice had delays in oocyte growth and development, oocytes from antral follicles of diabetic mice were collected and examined at different time points. At 2 h after hCG injection, control oocytes (n = 102) were 31% larger than the diabetic oocytes (n = 132). At 6 h, control oocytes (n = 76) were 23% larger than diabetic oocytes (n = 109). These differences are depicted in bar graph in Fig. 1A.
FIG. 1. Oocytes from streptozotocin-induced diabetic mice are smaller, and GVBD is significantly delayed compared with nondiabetic controls. A, At 2 and 6 h after hCG, control oocytes were significantly larger than diabetic oocytes; *, P < 0.01. B, At 2 and 6 h after hCG injection, a significantly lower percentage of oocytes had reached GVBD; *, P < 0.01.
Oocytes from diabetic mice were also found to have a delay in maturation or completion of meiosis I, often referred to as GVBD. Two hours after hCG injection, 27.2% of diabetic oocytes had reached GVBD, whereas 37.2% of nondiabetic oocytes had completed this maturational stage. Similarly, at 6 h, 45.9% of nondiabetic oocytes had reached GVBD, whereas significantly fewer, 36.7% of diabetic oocytes, had matured to this stage as seen in Fig. 1B.
Increased apoptosis was visualized in the hyperglycemic streptozotocin-induced model
At time 0, in animals that had received ovarian priming with pregnant mare serum gonadotropin, but no hCG, there was no significant difference in the number of TUNEL-positive nuclei in the control vs. the diabetic CEOs. The percentage of TUNEL-positive nuclei or apoptotic nuclei was fairly low in both groups with the nondiabetic group at 6.24% and the diabetic group at 9.36%. This difference was not significant. However, at 6 h after hCG injection, there was a trend of higher levels of apoptosis in the diabetic group as predicted. In the diabetic group, 13.81% of the nuclei were TUNEL positive, whereas in the nondiabetic CEOs, approximately 5.2% of the nuclei were TUNEL positive (P = 0.0517). Each study group had approximately 12 animals.
Ovaries from diabetic (n = 4) and nondiabetic mice (n = 4) were also sectioned to assess in situ apoptosis. After TUNEL-stained ovarian sections were examined with confocal microscopy, each of the follicles was examined, and the percentage of TUNEL-positive follicles was counted. At 6 h after hCG, 68.3% of the diabetic follicles were TUNEL positive, whereas only 15.1% of the nondiabetic follicles had TUNEL staining. Similar to what had been visualized in the CEOs, more of the diabetic follicles were TUNEL positive. Representative examples of nondiabetic and diabetic sections are shown in Fig. 2, A and B, respectively. The diabetic follicles had remarkably more TUNEL-positive follicles when compared with the nondiabetic group. These data are demonstrated in bar graph form in Fig. 2C. This difference was statistically significant with a P value of 0.012.
FIG. 2. Ovaries from diabetic mice display increased apoptosis compared with nondiabetic controls. Control ovaries (A) demonstrate decreased TUNEL staining (pink) compared with a streptozotocin-induced diabetic ovaries (B). C, This increase in apoptosis in diabetic, nonatretic follicles is represented graphically at 6 h after hCG; *, P = 0.012. DM, Diabetic; NDM, nondiabetic.
Histology displayed more apoptotic foci in the streptozotocin-induced diabetic model
Ovarian sections were also paraffin embedded and sectioned 6 h after hCG injection. The slides were stained with hematoxylin and eosin. A blinded pathologist then looked at ovaries (n = 4) from streptozotocin-injected mice compared with control mice. The pathologist first examined 16 sections per ovary, picked the five most representative sections, and evaluated these sections. Under light microscopy, the pathologist found that there were more apoptotic-appearing nuclei in the diabetic follicles compared with the nondiabetic follicles. The diabetic follicles, on average, displayed five apoptotic foci per follicle. The nondiabetic follicles, on the other hand, were larger and demonstrated approximately two apoptotic foci per follicle. This correlated with the higher levels of apoptosis that were found in the diabetic follicles.
Increased apoptosis was visualized in the chronic hyperglycemic Akita model
Akita mice harbor an autosomal dominant mutation in the insulin 2 gene that causes them to have chronic insulinopenia and hyperglycemia because of a conformational problem in insulin secretion (19, 20). As a result, these mice have chronic hyperglycemia and hypoinsulinemia with glucose levels in the range of 300 mg/dl and above at the age of 6–9 wk. Ovarian sections from Akita mice (n = 3) and control C57BL/6 mice (n = 3) were obtained 6 h after hCG injection, and TUNEL staining was performed. Approximately 64.07% of the mutant hyperglycemic mice follicles were found to be TUNEL positive, and 36.7% of the control mice follicles demonstrated TUNEL-positive staining. This difference was significant with a P value of 0.035. Examples of control vs. Akita ovarian sections are displayed in Fig. 3, E and D, respectively. This increase in follicle apoptosis confirmed the increased apoptosis seen in the short-term streptozotocin-induced hyperglycemia model.
FIG. 3. Ovaries from Akita hyperglycemic mice demonstrate smaller follicles and increased apoptosis compared with nondiabetic controls. A, Table showing smaller numbers of large, medium/small, and antral follicles and corpora lutea and greater numbers of apoptotic foci in the Akita mice vs. the controls. B and C, Ovarian section, hematoxylin and eosin stained, from an Akita mouse (B) displays several small and midsize follicles, whereas an ovarian section from a control B6 mouse (C) displays a higher number of larger, more well developed follicles. D and E, Ovarian section from an Akita mouse (D) demonstrates increased apoptosis compared with the ovarian section from a control mouse (E).
Oocytes from Akita mice were smaller and had a delay in maturation
To confirm our hypothesis that hyperglycemia may delay growth and maturation, oocytes from the antral follicles of Akita mice were obtained 6 h after hCG injection. In parallel to the streptozotocin-injected mice, the control B6 oocytes (n = 44) were 26.1% significantly larger than the Akita mice (n = 75) (P < 0.05). Similarly, at this time point, only 22.7% of the Akita oocytes had reached GVBD, whereas 34.1% of the control B6 mice had already completed this stage of development. Each experimental group had five animals.
Histology revealed smaller follicles and increased apoptosis in the Akita mice
Paraffin-embedded ovarian sections from Akita mice were visualized 6 h after hCG by a blinded pathologist. Sixteen sections from each ovary (n = 4) were examined. The five most representative sections were recorded, and an average of these five was reported. Comparing the population of follicles from the Akita with those of the control mice, which were strain-matched B6 nondiabetic mice, fewer large follicles were seen in the ovaries from the Akita ovaries. On average, the control mice had seven large follicles per ovary, whereas the Akita mice had 2.5 large follicles per ovary. The average largest diameter of an Akita follicle was 0.44 μm, which was markedly smaller than the average that was seen in a control mouse, 0.62 μm. In addition, there were fewer antral follicles as well as medium and small follicles. An example of the larger, more developed follicles in the control vs. the smaller, less mature follicles from the Akita mice is displayed in Fig. 3, C and B, respectively. Finally, the pathologist noted that the number of apoptotic foci per field was higher in the Akita follicles compared with the strain-matched control ovaries. Again, this was confirmed as seen in the TUNEL experiments on the ovarian sections, showing increased TUNEL-positive, nonatretic follicles in the Akita mice (Fig. 3D) compared with nondiabetic controls (Fig. 3E).
Increased TRAIL and KILLER expression was visualized in the streptozotocin-induced diabetic model
To test the hypothesis that TRAIL/KILLER interaction was involved in diabetes-induced follicular apoptosis, CEOs from the streptozotocin-induced diabetic mice were examined for TRAIL protein expression. The protein appeared to be localized to the plasma membrane as well as the cytoplasm. Higher levels of expression were seen in the diabetic CEOs (n = 4 mice) at 2 and 6 h after hCG, compared with the nondiabetic control CEOs (n = 4 mice) at both time points. A representative example of the increased expression in the diabetic CEO 6 h after hCG administration is displayed in Fig. 4, A and B.
FIG. 4. Increased TRAIL and its receptor, KILLER, protein expression are seen in CEOs from diabetic vs. nondiabetic mice. Significantly greater TRAIL protein is seen by confocal immunofluorescent microscopy in the CEO and oocyte of the diabetic mouse (B) compared with that seen in the nondiabetic mouse (A). Likewise, KILLER protein expression is significantly higher in the diabetic CEO (D) compared with the nondiabetic control (C) with a high expression seen in the surrounding granulosa cells.
Similarly, KILLER protein was examined by confocal immunohistochemistry. This protein appeared to be localized predominantly in the area of the plasma membrane, as predicted by the distribution of receptor expression. More KILLER protein was present in the diabetic mouse CEOs (n = 4 mice) at 2 h after hCG compared with the nondiabetic CEOs (n = 4 mice) with a representative example seen in Fig. 4, D and C. To confirm this finding, protein expression was also quantified by Western immunoblotting using CEOs from diabetic and nondiabetic mice. Approximately 70 CEOs from each different test group were used as samples. The experiment was performed in triplicate at 0, 2, and 6 h. At time 0, no significant difference in the expression levels of KILLER was discerned with a P value of 0.336; however, significant differences were seen at 2 h as displayed in Fig. 5. The values are expressed as X-fold difference over the control value. The difference by ANOVA with the Fisher post hoc test at 2 h was significant with a P value < 0.01; however, at 6 h there was a trend toward significance with a P value of 0.07 by ANOVA with the Fisher post hoc test. As with other TRAIL/KILLER apoptotic signaling, it is possible that this peak in expression at 2 h and decrease at 6 h reflects the initiation and progression of the apoptotic cascade.
FIG. 5. Increased KILLER protein expression in diabetic CEOs by Western immunoblot. The values are expressed as X-fold difference over the control value. The difference at 2 h was significant with a P value < 0.01, whereas at 6 h, the P value was not significant at 0.07 by ANOVA combined with the Fisher post hoc test.
Lower expression of connexin-43 was found in CEOs from streptozotocin-induced diabetic mice
Connexin-43 expression was examined 6 h after hCG in CEOs from diabetic (n = 12) and nondiabetic (n = 12) mice. There appeared to be a greater amount of punctate immunopositive staining surrounding the cumulus cells and the oocytes in CEOs from nondiabetic control samples. An example of increased staining is seen in Fig. 6A with a CEO from a nondiabetic control mouse compared with the immunopositivity seen in Fig. 6B of a CEO from a diabetic mouse. The quantity of immunopositivity was verified using Western blots in triplicate. At 6 h, significantly more connexin-43 protein was present in equal numbers of pooled CEOs from nondiabetic mice compared with the diabetic mice (P = 0.0085). An example of a Western blot is shown in Fig. 6C.
FIG. 6. Connexin-43 expression is decreased in CEOs from diabetic mice. A, In CEOs from nondiabetic mice, connexin-43 protein is seen as punctate immunostaining in the intercellular spaces between granulosa cells. B, CEOs from diabetic mice shows markedly less staining. C, Connexin-43 protein is higher by immunoblotting in the control CEOs compared with the diabetic CEOs.
Discussion
In this study, we demonstrate several changes that occur in the oocyte-cumulus cell complex as a result of exposure to two different types of hyperglycemic and hypoinsulinemic conditions. Specifically, type 1 diabetes in a murine model appears to be associated with 1) oocyte developmental alterations including a decrease in oocyte size and a delay in meiotic maturation; 2) higher levels of granulosa/cumulus cell apoptosis in the CEOs and in the ovarian sections by both TUNEL assay and morphological examination; 3) up-regulation in expression of cell-death signaling proteins such as TRAIL and KILLER, initiated at the period just before the detection of apoptosis; and 4) a decrease in the expression level of key gap junction proteins that are necessary for communication. These are all novel findings that may explain in part the reproductive problems experienced by mice as well as women with hyperglycemia combined with hypoinsulinemia.
The development of ovarian follicles requires a complex set of cell-cell interactions, and the communication between somatic and germ cells involves endocrine, autocrine, paracrine, and gap junction pathways. In order for both the oocyte and granulosa cell components of the follicle to develop successfully, bidirectional communication is essential (21). Colton et al. (9) demonstrated a loss of metabolic communication between granulosa cells and oocytes in CEOs from diabetic mice. It is possible that a decrease in gap junction communication between granulosa cells leads to changes in paracrine communication between oocyte and granulosa cell that is related to the increase in apoptosis. High-glucose conditions have been found to decrease expression of connexin-43 in human retinal pericytes (11) and simultaneously increase apoptosis (22, 23). In addition, studies in rodent myocytes have demonstrated that interrupted cell-cell communication by down-regulation of connexin-43 is associated with accelerated apoptosis (24). Similarly, we have found that hyperglycemia is associated with a decrease in connexin-43 protein expression, which may impact granulosa cell intercellular communication and accelerate granulosa cell apoptosis. A similar inverse correlation between apoptosis and connexin-43 expression has recently been reported in avian granulosa cells (12). It is difficult to discern whether the deficient communication causes differences in cell signaling that then result in higher levels of apoptosis or whether apoptosis acts as a precursor for downstream problems in communication. In either case, this study links the diabetic condition with apoptosis, a decrease in connexin-43 expression, and an up-regulation of the TRAIL-KILLER cell-death pathway, providing some evidence that a relationship between these phenomena may exist in cell-cell communication in the CEO complex.
Delays in oocyte maturation are seen in the oocytes from both types of diabetic mice. We hypothesize that an insult or a preprogramming event may occur at the oocyte stage secondary to maternal hyperglycemia that permanently alters the course of normal development, and this manifests first as a maturational delay. Several prior studies from our laboratory and others support this hypothesis (5, 7, 8, 9). Embryos recovered in vivo at 48 h after fertilization from chemically induced diabetic mice experience an in vivo developmental delay. Only approximately 70% of the diabetic embryos had reached a two-cell stage, whereas 90% of the control embryos had already reached this maturational stage. Similarly, this delay in development persisted in vitro when these same embryos were cultured in control media for 72 h and assessed at each 24-h interval. A significant impairment in development was seen in the embryos from diabetic mice in their rates of progression to blastocyst compared with nondiabetic mice, despite the fact that both were cultured in identical medium conditions (7, 18). We postulate that this later effect on embryo development is related in part to the oocyte maturation delays described and that these changes are also associated with the increase in apoptosis in the CEOs. Similar conclusions have been drawn in experiments involving exposure to certain toxic agents during the preorganogenic period, from the time of sperm entry and zygote development to late blastocyst stage. Generoso and colleagues (25, 26, 27, 28, 29) published a series of papers demonstrating that brief exposure to ethylene oxide resulted in an increased incidence of fetal death and certain types of fetal malformations, such as craniofacial abnormalities, abdominal wall defects, limb defects, and stillbirths. These effects were not associated with induced chromosomal abnormalities or gene mutations (29). Therefore, there is evidence that oocyte-directed insults may be the result of poorly controlled diabetes during folliculogenesis in animal models as well as humans and that these insults may result in later reproductive failures, such as a higher incidence of malformation and miscarriages in this population of patients
In both the streptozotocin-induced and Akita diabetic mice, antral follicle oocyte size is significantly smaller compared with nondiabetic controls. In addition, in the Akita mice, the follicles at all stages of development are smaller compared with the follicles from the control animals. This smaller follicle and oocyte size may reflect abnormal cell growth and survival, which would correlate with the increased apoptosis seen in the CEOs from the diabetic mice. Similar findings of apoptosis and smaller cell size have been described in Drosophila as well as hematopoietic cell lines as a result of decreased Akt expression and activity (30, 31). Akt determines cell size and survival by modulating mammalian target of rapamycin activity and protein synthesis in the leukocytic cell line FL5.12 (31). TRAIL/KILLER interaction has also been linked to decreased Akt/phosphatidylinositol 3-kinase activity (32, 33). Therefore, it is possible that TRAIL/KILLER up-regulation in the CEOs may be modulating the activity of Akt, leading to decreased cell size and apoptosis. Dysregulation of the Akt pathway would also have downstream effects on protein synthesis and transcriptional regulation of genes perhaps involved in further embryonic development. Using this paradigm, up-regulation of TRAIL/KILLER, increased apoptosis, and decreased cell size may all be contributing to the poor oocyte quality and developmental potential of the resulting embryos.
Separating the physiological effects of high serum glucose levels from a lack of insulin is difficult in both of these models of type 1 diabetes, as would be expected in an animal model of human disease. Destruction of the ?-cells of the pancreas by streptozotocin or prevention of insulin secretion by the Ins2 mutation in the Akita mice causes significantly lower insulin and IGF-I levels as well as hyperglycemia. Previous studies from Kezele et al. (34) have shown that insulin acts as a paracrine factor to facilitate transition from primordial to primary follicle at the level of the oocyte. This group has suggested that abnormally low insulin levels as seen in a type 1 diabetic state may inhibit or retard primordial to primary follicle transition. In other studies, Demeestere et al. (35) have shown that IGF-I has a stimulatory effect on follicular steroidogenesis without any effect on oocyte maturation. They did report, however, improved embryo development as a result of exposure to IGF-I during follicular maturation, suggesting that IGF-I improved the quality of the oocyte. Other studies have suggested that IGF-I and insulin enhance granulosa cell proliferation and increase follicle diameter (36). Thus, it is possible that the effects of maternal diabetes seen in this study, specifically granulosa cell apoptosis and decreased oocyte and follicle size, as well as delayed oocyte maturation might be a combined effect of hyperglycemia and hypoinsulinemia. Colton et al. (8) found that the delay in FSH-stimulated oocyte maturation experienced by the streptozotocin-induced diabetic mice could be only partially recreated in vitro with high-glucose conditions, suggesting other factors may be involved. Although the presence of both simultaneous conditions in the animal models may be considered a limitation of this study, these physiological models may more accurately reflect the true follicular milieu caused by this disease, rather than the in vitro experiments examining the two conditions in isolation.
In both the acute chemically induced diabetic mouse model and in the chronically hyperglycemic Akita mouse, detrimental effects are seen in oocyte development and the degree of apoptosis in the surrounding cumulus cells. For these reasons, women with type I diabetes mellitus may be placing their oocytes and offspring at risk. It is known from human studies on patients undergoing in vitro fertilization that increased cumulus cell apoptosis correlates with poorer-quality oocytes and poor pregnancy outcome including increased rates of miscarriage. Similarly, several studies in human oocytes have correlated small oocyte size with poorer developmental potential and pregnancy rates (37, 38, 39, 40, 41). As a result of these findings, we speculate that diabetic women who experience uncontrolled or poorly controlled diabetes during ovulation and fertilization may suffer detrimental effects on oocytes/CEOS secondary to hyperglycemia and hypoinsulinemia including developmental delay, up-regulation of cell-death effector pathways, and increased granulosa cell apoptosis. These early insults may then lead to an increased rate of miscarriage and congenital anomalies depending on the signaling and cell-death pathways involved.
Acknowledgments
We appreciate Elizabeth Schlichting and Amanda Wyman for their technical assistance and Dr. Stephen Downs for his helpful scientific advice. We also thank Edward Leiter for his assistance in obtaining the Akita mouse.
References
Greene, MF 1999 Spontaneous abortions and major malformations in women with diabetes mellitus. Semin Reprod Endocrinol 17:127–136
Farrell T, Neale L, Cundy, T 2002 Congential anomalies in the offspring of women with type I, type II and gestational diabetes. Diabet Med 19:322–326
Moley KH, Chi MM, Knudson CM, Korsmeyer SJ, Mueckler MM 1998 Hyperglycemia induces apoptosis in preimplantation embryos via cell death effector pathways. Nat Med 12:1421–1424
Moley KH, Chi MM, Mueckler MM 1998 Maternal hyperglycemia alters glucose transport and utilization in mouse preimplantation embryos. Am J Physiol 275:E38–E47
Moley KH, Vaughn WK, DeCherney AH, Diamond MP 1991 Effect of diabetes mellitus on mouse preimplantation embryo development. J Reprod Fertil 93:325–332
Chi MM, Pingsterhaus J, Carayannopoulos, M, Moley KH 2000 Decreased glucose transporter expression triggers BAX-dependent apoptosis in the murine blastocyst. J Biol Chem 275:40252–40257
Diamond MP, Moley KH, Pellicer A, Vaughn WK, DeCherney AH 1989 Effects of streptozotocin and alloxan-induced diabetes mellitus on mouse follicular and early embryo development. J Reprod Fertil 86:1–10
Colton SA, Pieper GM, Downs SM 2002 Altered meiotic regulation in oocytes from diabetic mice. Biol Reprod 67:220–230
Colton SA, Humpherson PG, Leese HJ, Downs SM 2003 Physiological changes in oocyte-cumulus cell complexes from diabetic mice that potentially influence meiotic regulation. Biol Reprod 69:761–770
Ackert CL, Gittens JEI, O’Brien MJ, Eppig JJ, Kidder GM 2001 Intercellular communication via connexin43 gap junctions is required for ovarian folliculogenesis in the mouse. Dev Biol 233:258–270
Li A, Sato T, Haimovici R, Okamoto T, Roy S 2003 High glucose alters connexin 43 expression and gap junction intercellular communication activity in retinal pericytes. Invest Ophthalmol Vis Sci 44:5376–5382
Krysko DV, Mussche S, Leybaert L, D’Herde K 2004 Gap junctional communication and connexin43 expression in relation to apoptotic cell death and survival of granulosa cells. J Histochem Cytochem 52:1199–1207
Zhang H, Vollmer M, De Geyter M, Litzistorf Y, Ladewig A, Durrenberger M, Guggenheim R, Miny P, Holzgreve W, De Geyter C 2000 Characteristics of an immortalized human granulosa cell line COV 434. Mol Hum Reprod 2:146–153
Idil M, Cepni I, Demirsoy G, Ocal P, Salihoglu F, Senol H, Elibol F, Irez T 2004 Does granulosa cell apoptosis have a role in the etiology of unexplained infertility? Eur J Obstet Gynecol Reprod Biol 112:182–184
Lee KS, Joo BS, Na YJ, Yoon MS, Choi OH, Kim WW 2001 Cumulus cell apoptosis as an indicator to predict the quality of oocytes and the outcome of IVF-ET. J Assist Reprod Genet 18:490–498
Nakahora K, Saito H, Satio T, Ito M, Ohta N, Takahashi T, Hiroi M 1997 The incidence of apoptotic bodies in membrana granulosa can predict prognosis of ova from patients participating in in vitro fertilization programs. Fertil Steril 68:312–317
LeBlanc H, Lawrence D, Varfolomeev E, Totpal K, Morlan J, Schow P, Fong S, Schwall R, Sinicropi D, Ashkenazi A 2002 Tumor-cell resistance to death receptor-induced apoptosis through mutational inactivation of the proapoptotic Bcl-2 homolog Bax. Nat Med 8:274–281
Moley KH 1999 Diabetes and preimplantation events of embryogenesis. Semin Reprod Endocrinol 17:137–151
Izumi T, Yokota-Hashimoto H, Zhao S, Wang J, Halban PA, Takeuchi T 2002 Dominant negative pathogenesis by mutant proinsulin in the Akita diabetic mouse. Diabetes 52:409–416
Wang J, Takeuchi T, Tanaka S, Kubo S, Kayo T, Lu D, Takata K, Koizumi A, Izumi T 1999 A mutation in the insulin 2 gene induces diabetes with severe pancreatic ?-cell dysfunction in the Mody mouse. J Clin Invest 103:27–37
Eppig JJ, Wigglesworth K, Pendola FL 2002 The mammalian oocyte orchestrates the rate of ovarian development. Proc Natl Acad Sci USA 99:2890–2894
Podesta F, Romeo G, Liu WH, Krajewski S, Reed JC, Gerhardinger C, Lorenzi M 2000 Bax is increased in the retina of diabetic subjects and is associated with pericyte apoptosis in vivo and in vitro. Am J Pathol 156:1025–1032
Romeo G, Liu WH, Asnaghi V, Kern TS, Lorenzi M 2002 Activation of nuclear factor-B induced by diabetes and high glucose regulates a proapoptotic program in retinal pericytes. Diabetes 51:2241–2248
Yasui K, Kada K, Hojo M, Lee J-K, Kamiya K, Toyama J, Opthof T, Kodama T 2000 Cell-to-cell interaction prevents cell death in cultured neonatal rat ventricular myocytes. Cardiovasc Res 48:68–76
Generoso WM, Shourbaji AG, Piegorsch WW, Bishop JB 1991 Development response of zygotes exposed to similar mutagens. Mutat Res 250:439–446
Polifka JE, Rutledge JC, Kimmel GL, Dellarco V, Generoso WM 1996 Exposure to ethylene oxide during the early zygotic period induces skeletal anomalies in mouse fetuses. Teratology 53:1–9
Rutledge JC, Generoso WM, Shourbaji A, Cain KT, Gans M, Oliva J 1992 Developmental anomalies derived from exposure of zygotes and first-cleavage embryos to mutagens. Mutat Res 296:167–177
Kimmel CA, Generoso WM, Thomas RD, Bakshi KS 1993 Contemporary issues in toxicology: a new frontier in understanding the mechanisms of developmental abnormalities. Toxicol Appl Pharmacol 119:159–165
Katoh M, Cacheiro NL, Cornett CV, Cain KT, Rutledge JC, Generoso WM 1989 Fetal anomalies produced subsequent to treatment of zygotes with ethylene oxide or ethyl methanesulfonate are not likely due to the usual genetic causes. Mutat Res 210:337–344
Verdu J, Buratovich MA, Wilder EL, Birnbaum MJ 1999 Cell-autonomous regulation of cell and organ growth in Drosophila by Akt/PKB. Nat Cell Biol 1:500–506
Edinger AL, Thompson CB 2002 Akt maintains cell size and survival by increasing mTOR-dependent nutrient uptake. Mol Biol Cell 13:2276–2288
Kang YC, Kim KM, Lee KS, Namkoong S, Lee SJ, Han JA, Jeoung D, Ha KS, Kwon YG, Kim YM 2004 Serum bioactive lysophospholipids prevent TRAIL-induced apoptosis via P13K/Akt-dependent cFLIP expression and Bad phosphorylation. Cell Death Differentiation 11:1287–1298
Larribere L, Khaled M, Tartare-Deckert S, Busca R, Luciano F, Bille K, Valony G, Eychene A, Auberger P, Ortonne JP, Ballotti R, Bertolotto C 2004 P13K mediates protection against TRAIL-induced apoptosis in primary human melanocytes. Cell Death Differ 11:1084–1091
Kezele PR, Nilsson EE, Skinner MK 2002 Insulin but not insulin-like growth factor-1 promotes the primordial to primary follicle transition. Mol Cell Endocrinol 192:37–43
Demeestere I, Gervy C, Centner J, Devreker F, Englert Y, Delbaere A 2004 Effect of insulin-like growth factor-1 during preantral follicular culture on steroidogenesis, in vitro oocyte maturation, and embryo development in mice. Biol Reprod 70:1664–1669
Zhao J, Taverne MA, Van Der Weijden GC, Bevers MM, Van Den Hurk R 2001 Insulin-like growth factor-1 (IGF-1) stimulates the development of cultured rat pre-antral follicles. Mol Reprod Dev 58:287–296
Trounson AO, Gosden RG 2003 Biology and pathology of the oocyte. Cambridge, UK: Cambridge University Press; 307–309
Nataprawira DS, Harada T, Sekijima A, Mio Y, Terakawa N 1992 Assessment of follicular maturity by follicular diameter and fluid volume in a program of in vitro fertilization and embryo transfer. Asia Oceania J Obstet Gynaecol 18:225–230
Wittmaack FM, Kreger DO, Blasco L, Tureck RW, Mastroianni Jr L, Lessey BA 1994 Effect of follicular size on oocyte retrieval, fertilization, cleavage and embryo quality in in vitro fertilization cycles: a 6 year data collection. Fertil Steril 62:1205–1210
Arnot AM, Vanderkerckhove P, DeBono MA, Rutherford AJ 1995 Follicular volume and number during in vitro fertilization: association with oocyte development capacity and pregnancy rate. Hum Reprod 10:256–261
Bergh C, Broden H, Lundin K, Hamberger L 1998 Comparison of fertilization, cleavage and pregnancy rates of oocytes from large and small follicles. Hum Reprod 13:1912–1915(Aimee S. Chang, Alexis N.)
Address all correspondence and requests for reprints to: Dr. Kelle Moley, Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology, Washington University in St. Louis, 4911 Barnes Plaza Road, Box 8064, St. Louis, Missouri 63110-1094. E-mail: moleyk@wustl.edu.
Abstract
Maternal diabetes adversely affects preimplantation embryo development and pregnancy outcomes. The objective of this study was to determine whether diabetes has an impact at an earlier stage of development, the preovulatory oocyte. Models of both acute and chronic insulin-dependent diabetes were used. Acute hyperglycemia was induced by a single streptozotocin injection. Akita mice, which harbor an autosomal dominant mutation causing them to be chronically hypoinsulinemic and hyperglycemic, were used. In both models, preovulatory oocytes were markedly smaller when compared with control animals. A significantly greater number of control oocytes had progressed to meiotic maturation before diabetic oocytes. Both models were found to have smaller, less developed ovarian follicles with a greater number of apoptotic foci by histological evaluation as well as by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling staining. Immunohistochemistry displayed a greater amount of TNF-related apoptosis-inducing ligand (TRAIL) and KILLER, a key murine ligand and receptor involved in the extrinsic pathway, expressed in cumulus cells from hyperglycemic mice compared with controls, suggesting that this apoptotic pathway may be up-regulated under diabetic stress. Elevated KILLER expression was also confirmed through Western blotting. Connexin-43 expression was found to be lower by immunohistochemistry and Western blot analysis in the diabetic samples. Both models of maternal hyperglycemia and hypoinsulinemia may have a detrimental effect on oocyte maturation and development as detailed by the smaller sizes of oocytes and developing ovarian follicles, the lowered percentage reaching germinal vesicle breakdown, and the greater amount of apoptosis. In addition, there may be dysfunctional or decreased communication in diabetic oocytes, as demonstrated by lower expression of connexin-43.
Introduction
IN HUMANS, TYPE 1 OR insulin-dependent diabetes has been found to negatively affect pregnancy by causing poor prenatal outcomes such as an increased risk of congenital anomalies and early miscarriage (1, 2). Likewise, maternal diabetes has been found to adversely affect murine preimplantation embryo development in models of type 1 diabetes (3, 4, 5, 6). Previous studies from our laboratory have suggested that an insult occurring before the one-cell zygote stage may have an impact on the outcome of the resulting embryo. Zygotes removed from streptozotocin- or alloxan-induced hypoinsulinemic and hyperglycemic mice demonstrate retarded in vivo development to a two-cell stage with a lower percentage of two-cell zygotes recovered at 48 h after human chorionic gonadotropin (hCG) compared with nondiabetic controls (7). Similarly, two-cell embryos recovered from diabetic mice and cultured in vitro in control media also experience significant delay in their progression to the blastocyst stage compared with embryos from control mice. As a result of these findings, the oocytes from these diabetic mice were also examined. These oocytes from chemically induced diabetic mice experience significant delays in germinal vesicle breakdown (GVBD) and resumption of meiosis I. This previous study was limited, however, and focused primarily on the effects of hyperglycemia on the embryo, not oocyte. Since those initial studies, another group has confirmed these findings (8, 9). To further investigate the developmental impact of hyperglycemia on oocytes and folliculogenesis, including the effect on maturation, growth, and the surrounding granulosa cells, the following studies were undertaken. We hypothesize that maternal diabetes would have a detrimental effect on oocyte maturation and that the follicular environment of the enclosed cumulus complex and oocyte immediately after hCG administration would be adversely affected.
In the developing oocyte, granulosa cells surround the oocyte, support its growth, and provide hormonal supplementation. They form a multilayered structure around the oocyte. Paracrine signaling and gap junctional communication occur between the granulosa cells and the oocyte. The presence of normal granulosa cell communication and development are critical for differentiation and oocyte growth to occur (10). Because of the importance of this relationship, we also chose to examine whether the diabetic state influenced granulosa cell intercellular communication. It is possible that the maturational delay suspected in oocytes from diabetic mice may be a result of poor paracrine communication between these two compartments because of poor intercellular talk among granulosa cells. Connexin-43 is a key gap junction protein expressed in granulosa cells that is necessary for intercellular communication between granulosa cells and normal folliculogenesis (10). A lack of connexin-43 expression, however, results in abnormal oocytes as well as poor folliculogenesis. Furthermore, diabetic human retinal pericytes display lower levels of connexin-43 when compared with control pericytes (11). Finally, recent studies have shown that connexin-43 expression in granulosa cells is inversely related to apoptotic cell death in the avian follicle (12). Because of these links between diabetes, decreased connexin-43 expression, and apoptosis, one of our objectives in this study was to see whether connexin-43 levels were altered in diabetic cumulus-enclosed oocytes (CEOs).
It is well established that the vertebrate ovarian follicles can undergo apoptosis, via a process called atresia, and this apoptosis is initiated within the granulosa cell layer. In previous studies, we have established that hyperglycemia induces apoptosis in the murine blastocyst. One objective of this study is to examine whether an external stress such as hyperglycemia could cause higher levels of apoptosis in the granulosa cells of developing follicles that might influence oocyte development, maturation (13), and perhaps communication between the granulosa cells and the oocyte. Poor outcomes have previously been associated with granulosa cell apoptosis. Granulosa cell apoptosis has been found to be accelerated in human patients with unexplained infertility (14). Increased apoptosis of surrounding cumulus cells has been correlated with oocyte maturational delay and poor pregnancy outcomes (15, 16).
Two major pathways are involved with apoptosis initiation in vertebrates. They include the intrinsic apoptotic pathway whereby mitochondrial disruption occurs as a result of death-promoting members of the Bcl-2 family, leading to the release of factors that promote caspase-9 activation and downstream apoptosis. The extrinsic apoptotic pathway involves activation of the cell-death receptors by ligands that belong to the TNF superfamily and downstream activation of caspase-8. The two pathways appear to be connected by the protein Bid, which contains Bcl-2 homology domains. Caspase-8, which is usually activated in the extrinsic pathway, can trigger mitochondrial cytochrome c release and thus cause downstream activation of the intrinsic pathway (17). Previously, our laboratory had determined that the intrinsic pathway was active during hyperglycemia-induced apoptosis in the preimplantation blastocyst-stage embryo via the overexpression of Bax, a death-promoting member of the Bcl-2 family (3, 18). It was also discovered that apoptosis was increased in the blastocysts obtained in vivo vs. those cultured in vitro in high-glucose conditions, suggesting that this apoptosis may be occurring under the combined influence of both the extrinsic and intrinsic pathways. In this study, the effects of hyperglycemia on activation of the extrinsic pathway, specifically the TNF-related apoptosis-inducing ligand (TRAIL) (TNF superfamily) in the ovarian follicle, are examined. TRAIL mediates apoptosis through stimulation of DR5 or murine KILLER receptor.
In this study, we find that hyperglycemia, both short and long term, has a significant impact on the developing preovulatory oocyte, causing slower growth, delayed maturation, increased apoptosis in the surrounding granulosa cells, up-regulation of the extrinsic apoptotic pathway, and diminished levels of a key gap junction protein.
Materials and Methods
Oocyte retrieval
All mouse studies were approved by the animal Studies Committee at Washington University School of Medicine and conform to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health. Female immature C57BL/6JXSL/J F1 mice (age 20–24 d) were used for most experiments. Diabetic and age-matched controls received 10 IU equine chorionic gonadotropin, which is also known as pregnant mare’s serum gonadotropin, by ip injection, and 48 h later, they either were killed or given an injection of 5 IU hCG. At the appropriate time points after hCG injection (time 0, 2, and 6 h), the ovaries were removed and placed in a dish containing 1.5 ml culture medium. Oocytes were isolated by puncturing the antral follicles with sterile needles, and then washed through several changes of medium. Before puncturing, the ovaries were stored briefly in an incubator with the settings of 5% CO2, 5% O2, 90% N2 atmosphere at 37 C.
Induction of hyperglycemia
To generate a diabetic model, female mice (age 20–24 d) received a single injection of streptozotocin at a dose of 190 mg/kg (dissolved in sodium citrate buffer, pH 4.4; Sigma Chemical Co., St. Louis, MO). Four days after injection, a tail-blood sample was measured for glucose concentrations via a Hemocue B glucose analyzer (Stockholm, Sweden). If glucose levels were more than 240 mg/dl, these mice were selected and received a priming injection of equine chorionic gonadotropin. A few control mice were also randomly selected; their blood sugar was checked to ensure that it was less than 240 mg/dl.
Akita mice
Akita mice have an autosomal dominant mutation, or Mody mutation, resulting in hyperglycemia and notable pancreatic ?-cell dysfunction (19, 20). Diminished levels of both proinsulin and insulin are produced with resultant hyperglycemia. These mice were maintained at our animal facility, received the standard murine chow diet, and kept on a 12-h light, 12-h dark cycle. These mice were not growth retarded at any stage of development compared with control mice. All Akita mice had glucose levels checked at approximately 5–6 wk through a tail-blood sample by a Hemocue B glucose analyzer. If glucose levels were more than 300 mg/dl, these mice were considered to have the mutation. Age-matched controls were C57BL/6 females; these mice also had glucose levels that were checked by the glucometer.
Oocyte maturation and size
Oocytes from the ruptured antral follicles were observed using Hoffman optics on an inverted Nikon microscope (TMS scope). GVBD was assessed, and the oocytes that retained a germinal vesicle and/or the nucleolus failed to demonstrate maturation. Maturation was expressed as a percentage of GVBD. Oocyte diameters were measured excluding the zona pellucida with an eyepiece graticule (x200 magnification). The volume of each oocyte was calculated based on the formula for the volume of a sphere [(4/3)(r3)]. These measurements were done at 2 and 6 h after hCG.
Ovarian sectioning
Mice were killed, and the ovaries were obtained at specific time points (2 and 6 h). The ovaries were quickly submerged in liquid nitrogen and flash frozen. The ovaries were sectioned using a cryostat machine (Leica 1850, CM) in 11-μm sections. Each ovary was sectioned throughout, and one of every eight ovarian sections was used for immunohistochemical staining.
Evaluation of apoptosis by terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay
The CEOs and cryopreserved ovarian slices were fixed in 3% paraformaldehyde, permeabilized with 0.1% Tween 20, and then incubated in fluorescein-labeled dUTP and terminal transferase in the dark for 1 h at 37 C to label fragmented 3' DNA (TUNEL, Cell Death in Situ Kit; Roche Molecular Biochemicals, Mannheim, Germany) as previously described. Counterstaining of all nuclear DNA was achieved by incubating the tissue in 4 μM To-Pro-3-iodide (Molecular Probes, Eugene, OR) (blue channel) for 20 min. CEOs were visualized using confocal immunofluorescent microscopy with a Nikon C1 laser scanning confocal microscope at x63 magnification. Apoptosis was expressed as the percentage of TUNEL-positive nuclei per CEO or the number of TUNEL-positive follicles per total number of follicles. Student’s t test was additionally used for statistical analysis.
Immunohistochemistry
CEOs were fixed on glass slides in 3% paraformaldehyde for 20 min and permeabilized with 0.1% Tween 20 for 10 min. The CEOs were first incubated in 20% normal donkey serum (Pierce, Rockford, IL) in PBS containing 2% BSA (PBS/BSA) for 1 h to block any nonspecific binding. They were then washed in PBS/BSA and then incubated in the primary antibody [anti-DR5 (Santa Cruz Biotechnology, Santa Cruz, CA), N-19; mouse anti-actin (Chemicon International, Temecula, CA), 1:1000; anti-Trail (Santa Cruz Biotechnology), H-257, 1:200; and rabbit anti-connexin-43 (Zymed Laboratories, San Francisco, CA), 1:50] for 1 h at room temperature. The CEOs were then washed with PBS/BSA and incubated in the appropriate secondary antibody, goat antirabbit Alexa Fluor 488 antibody or goat antimouse Alexa Fluor 488 antibody, for approximately 30 min. The slides were washed, and nuclear staining was performed in 4 μM To-Pro-3-iodide. After extensive washing in PBS/BSA, confocal immunofluorescent microscopy (Nikon C1 laser scanning microscope) was then used to detect fluorescence as described above.
Expression of DR5 (KILLER) receptor and connexin-43 by Western analysis
CEOs were collected at specific time points (0, 2, and 6 h) for DR5 and (t = 0 h, 6 h) for anti-connexin 43 in equivalent groups (n = 70). The pooled samples were washed in several different washes of human tubal fluid and were frozen in human tubal fluid. (HTF) (Irvine Scientific, Irvine, CA). HeLa cell extracts were used as a positive control for the DR5 receptor. The CEOs were placed in Laemmli buffer with ?-mercaptoethanol. All samples were boiled for at least 5 min at 100 C. The samples were subjected to 10% SDS-PAGE and transferred to nitrocellulose. Mouse DR5 or KILLER protein was then detected with a rabbit polyclonal DR5 antibody (Santa Cruz Biotechnology; 1:500) as the primary antibody and a horseradish peroxidase (HRP)-labeled goat antirabbit antibody (Santa Cruz Biotechnology) as the secondary antibody. Connexin-43 protein was detected by a rabbit polyclonal antibody (Zymed Laboratories; 1:1000), and an HRP-labeled donkey antirabbit antibody (Pierce; 1:20,000) was used as the secondary antibody. HRP-labeled bands were quantified using NIH Image (version 1.6). All experiments were performed in triplicate. Mouse anti-actin (Chemicon International; 1:2000) was used as the loading control for all blots.
Ovarian sections
Ovaries were obtained at specific time points and submerged in Bouin’s fixation solution (Sigma). They were then paraffin embedded, cut, and stained with hematoxylin and eosin at the histology core at Washington University in St. Louis Medical School. The sections were then examined and evaluated by a blinded animal pathologist.
Statistics
Differences between the two groups with protein expression and percent TUNEL staining were analyzed using Student’s t test. If multiple groups were being compared, ANOVA was used with Fisher post hoc test. Results are expressed as means ± SE of at least three separate experiments. For each set of oocyte experiments, at least five mice were used for each group, and approximately 15 oocytes were obtained from each animal. All experiments were performed at least two times, and the majority of the experiments were performed three times. The numbers of ovaries sectioned are described below.
Results
Oocytes from streptozotocin-induced diabetic mice were smaller and had a delay in maturation
To test our hypothesis that diabetic mice had delays in oocyte growth and development, oocytes from antral follicles of diabetic mice were collected and examined at different time points. At 2 h after hCG injection, control oocytes (n = 102) were 31% larger than the diabetic oocytes (n = 132). At 6 h, control oocytes (n = 76) were 23% larger than diabetic oocytes (n = 109). These differences are depicted in bar graph in Fig. 1A.
FIG. 1. Oocytes from streptozotocin-induced diabetic mice are smaller, and GVBD is significantly delayed compared with nondiabetic controls. A, At 2 and 6 h after hCG, control oocytes were significantly larger than diabetic oocytes; *, P < 0.01. B, At 2 and 6 h after hCG injection, a significantly lower percentage of oocytes had reached GVBD; *, P < 0.01.
Oocytes from diabetic mice were also found to have a delay in maturation or completion of meiosis I, often referred to as GVBD. Two hours after hCG injection, 27.2% of diabetic oocytes had reached GVBD, whereas 37.2% of nondiabetic oocytes had completed this maturational stage. Similarly, at 6 h, 45.9% of nondiabetic oocytes had reached GVBD, whereas significantly fewer, 36.7% of diabetic oocytes, had matured to this stage as seen in Fig. 1B.
Increased apoptosis was visualized in the hyperglycemic streptozotocin-induced model
At time 0, in animals that had received ovarian priming with pregnant mare serum gonadotropin, but no hCG, there was no significant difference in the number of TUNEL-positive nuclei in the control vs. the diabetic CEOs. The percentage of TUNEL-positive nuclei or apoptotic nuclei was fairly low in both groups with the nondiabetic group at 6.24% and the diabetic group at 9.36%. This difference was not significant. However, at 6 h after hCG injection, there was a trend of higher levels of apoptosis in the diabetic group as predicted. In the diabetic group, 13.81% of the nuclei were TUNEL positive, whereas in the nondiabetic CEOs, approximately 5.2% of the nuclei were TUNEL positive (P = 0.0517). Each study group had approximately 12 animals.
Ovaries from diabetic (n = 4) and nondiabetic mice (n = 4) were also sectioned to assess in situ apoptosis. After TUNEL-stained ovarian sections were examined with confocal microscopy, each of the follicles was examined, and the percentage of TUNEL-positive follicles was counted. At 6 h after hCG, 68.3% of the diabetic follicles were TUNEL positive, whereas only 15.1% of the nondiabetic follicles had TUNEL staining. Similar to what had been visualized in the CEOs, more of the diabetic follicles were TUNEL positive. Representative examples of nondiabetic and diabetic sections are shown in Fig. 2, A and B, respectively. The diabetic follicles had remarkably more TUNEL-positive follicles when compared with the nondiabetic group. These data are demonstrated in bar graph form in Fig. 2C. This difference was statistically significant with a P value of 0.012.
FIG. 2. Ovaries from diabetic mice display increased apoptosis compared with nondiabetic controls. Control ovaries (A) demonstrate decreased TUNEL staining (pink) compared with a streptozotocin-induced diabetic ovaries (B). C, This increase in apoptosis in diabetic, nonatretic follicles is represented graphically at 6 h after hCG; *, P = 0.012. DM, Diabetic; NDM, nondiabetic.
Histology displayed more apoptotic foci in the streptozotocin-induced diabetic model
Ovarian sections were also paraffin embedded and sectioned 6 h after hCG injection. The slides were stained with hematoxylin and eosin. A blinded pathologist then looked at ovaries (n = 4) from streptozotocin-injected mice compared with control mice. The pathologist first examined 16 sections per ovary, picked the five most representative sections, and evaluated these sections. Under light microscopy, the pathologist found that there were more apoptotic-appearing nuclei in the diabetic follicles compared with the nondiabetic follicles. The diabetic follicles, on average, displayed five apoptotic foci per follicle. The nondiabetic follicles, on the other hand, were larger and demonstrated approximately two apoptotic foci per follicle. This correlated with the higher levels of apoptosis that were found in the diabetic follicles.
Increased apoptosis was visualized in the chronic hyperglycemic Akita model
Akita mice harbor an autosomal dominant mutation in the insulin 2 gene that causes them to have chronic insulinopenia and hyperglycemia because of a conformational problem in insulin secretion (19, 20). As a result, these mice have chronic hyperglycemia and hypoinsulinemia with glucose levels in the range of 300 mg/dl and above at the age of 6–9 wk. Ovarian sections from Akita mice (n = 3) and control C57BL/6 mice (n = 3) were obtained 6 h after hCG injection, and TUNEL staining was performed. Approximately 64.07% of the mutant hyperglycemic mice follicles were found to be TUNEL positive, and 36.7% of the control mice follicles demonstrated TUNEL-positive staining. This difference was significant with a P value of 0.035. Examples of control vs. Akita ovarian sections are displayed in Fig. 3, E and D, respectively. This increase in follicle apoptosis confirmed the increased apoptosis seen in the short-term streptozotocin-induced hyperglycemia model.
FIG. 3. Ovaries from Akita hyperglycemic mice demonstrate smaller follicles and increased apoptosis compared with nondiabetic controls. A, Table showing smaller numbers of large, medium/small, and antral follicles and corpora lutea and greater numbers of apoptotic foci in the Akita mice vs. the controls. B and C, Ovarian section, hematoxylin and eosin stained, from an Akita mouse (B) displays several small and midsize follicles, whereas an ovarian section from a control B6 mouse (C) displays a higher number of larger, more well developed follicles. D and E, Ovarian section from an Akita mouse (D) demonstrates increased apoptosis compared with the ovarian section from a control mouse (E).
Oocytes from Akita mice were smaller and had a delay in maturation
To confirm our hypothesis that hyperglycemia may delay growth and maturation, oocytes from the antral follicles of Akita mice were obtained 6 h after hCG injection. In parallel to the streptozotocin-injected mice, the control B6 oocytes (n = 44) were 26.1% significantly larger than the Akita mice (n = 75) (P < 0.05). Similarly, at this time point, only 22.7% of the Akita oocytes had reached GVBD, whereas 34.1% of the control B6 mice had already completed this stage of development. Each experimental group had five animals.
Histology revealed smaller follicles and increased apoptosis in the Akita mice
Paraffin-embedded ovarian sections from Akita mice were visualized 6 h after hCG by a blinded pathologist. Sixteen sections from each ovary (n = 4) were examined. The five most representative sections were recorded, and an average of these five was reported. Comparing the population of follicles from the Akita with those of the control mice, which were strain-matched B6 nondiabetic mice, fewer large follicles were seen in the ovaries from the Akita ovaries. On average, the control mice had seven large follicles per ovary, whereas the Akita mice had 2.5 large follicles per ovary. The average largest diameter of an Akita follicle was 0.44 μm, which was markedly smaller than the average that was seen in a control mouse, 0.62 μm. In addition, there were fewer antral follicles as well as medium and small follicles. An example of the larger, more developed follicles in the control vs. the smaller, less mature follicles from the Akita mice is displayed in Fig. 3, C and B, respectively. Finally, the pathologist noted that the number of apoptotic foci per field was higher in the Akita follicles compared with the strain-matched control ovaries. Again, this was confirmed as seen in the TUNEL experiments on the ovarian sections, showing increased TUNEL-positive, nonatretic follicles in the Akita mice (Fig. 3D) compared with nondiabetic controls (Fig. 3E).
Increased TRAIL and KILLER expression was visualized in the streptozotocin-induced diabetic model
To test the hypothesis that TRAIL/KILLER interaction was involved in diabetes-induced follicular apoptosis, CEOs from the streptozotocin-induced diabetic mice were examined for TRAIL protein expression. The protein appeared to be localized to the plasma membrane as well as the cytoplasm. Higher levels of expression were seen in the diabetic CEOs (n = 4 mice) at 2 and 6 h after hCG, compared with the nondiabetic control CEOs (n = 4 mice) at both time points. A representative example of the increased expression in the diabetic CEO 6 h after hCG administration is displayed in Fig. 4, A and B.
FIG. 4. Increased TRAIL and its receptor, KILLER, protein expression are seen in CEOs from diabetic vs. nondiabetic mice. Significantly greater TRAIL protein is seen by confocal immunofluorescent microscopy in the CEO and oocyte of the diabetic mouse (B) compared with that seen in the nondiabetic mouse (A). Likewise, KILLER protein expression is significantly higher in the diabetic CEO (D) compared with the nondiabetic control (C) with a high expression seen in the surrounding granulosa cells.
Similarly, KILLER protein was examined by confocal immunohistochemistry. This protein appeared to be localized predominantly in the area of the plasma membrane, as predicted by the distribution of receptor expression. More KILLER protein was present in the diabetic mouse CEOs (n = 4 mice) at 2 h after hCG compared with the nondiabetic CEOs (n = 4 mice) with a representative example seen in Fig. 4, D and C. To confirm this finding, protein expression was also quantified by Western immunoblotting using CEOs from diabetic and nondiabetic mice. Approximately 70 CEOs from each different test group were used as samples. The experiment was performed in triplicate at 0, 2, and 6 h. At time 0, no significant difference in the expression levels of KILLER was discerned with a P value of 0.336; however, significant differences were seen at 2 h as displayed in Fig. 5. The values are expressed as X-fold difference over the control value. The difference by ANOVA with the Fisher post hoc test at 2 h was significant with a P value < 0.01; however, at 6 h there was a trend toward significance with a P value of 0.07 by ANOVA with the Fisher post hoc test. As with other TRAIL/KILLER apoptotic signaling, it is possible that this peak in expression at 2 h and decrease at 6 h reflects the initiation and progression of the apoptotic cascade.
FIG. 5. Increased KILLER protein expression in diabetic CEOs by Western immunoblot. The values are expressed as X-fold difference over the control value. The difference at 2 h was significant with a P value < 0.01, whereas at 6 h, the P value was not significant at 0.07 by ANOVA combined with the Fisher post hoc test.
Lower expression of connexin-43 was found in CEOs from streptozotocin-induced diabetic mice
Connexin-43 expression was examined 6 h after hCG in CEOs from diabetic (n = 12) and nondiabetic (n = 12) mice. There appeared to be a greater amount of punctate immunopositive staining surrounding the cumulus cells and the oocytes in CEOs from nondiabetic control samples. An example of increased staining is seen in Fig. 6A with a CEO from a nondiabetic control mouse compared with the immunopositivity seen in Fig. 6B of a CEO from a diabetic mouse. The quantity of immunopositivity was verified using Western blots in triplicate. At 6 h, significantly more connexin-43 protein was present in equal numbers of pooled CEOs from nondiabetic mice compared with the diabetic mice (P = 0.0085). An example of a Western blot is shown in Fig. 6C.
FIG. 6. Connexin-43 expression is decreased in CEOs from diabetic mice. A, In CEOs from nondiabetic mice, connexin-43 protein is seen as punctate immunostaining in the intercellular spaces between granulosa cells. B, CEOs from diabetic mice shows markedly less staining. C, Connexin-43 protein is higher by immunoblotting in the control CEOs compared with the diabetic CEOs.
Discussion
In this study, we demonstrate several changes that occur in the oocyte-cumulus cell complex as a result of exposure to two different types of hyperglycemic and hypoinsulinemic conditions. Specifically, type 1 diabetes in a murine model appears to be associated with 1) oocyte developmental alterations including a decrease in oocyte size and a delay in meiotic maturation; 2) higher levels of granulosa/cumulus cell apoptosis in the CEOs and in the ovarian sections by both TUNEL assay and morphological examination; 3) up-regulation in expression of cell-death signaling proteins such as TRAIL and KILLER, initiated at the period just before the detection of apoptosis; and 4) a decrease in the expression level of key gap junction proteins that are necessary for communication. These are all novel findings that may explain in part the reproductive problems experienced by mice as well as women with hyperglycemia combined with hypoinsulinemia.
The development of ovarian follicles requires a complex set of cell-cell interactions, and the communication between somatic and germ cells involves endocrine, autocrine, paracrine, and gap junction pathways. In order for both the oocyte and granulosa cell components of the follicle to develop successfully, bidirectional communication is essential (21). Colton et al. (9) demonstrated a loss of metabolic communication between granulosa cells and oocytes in CEOs from diabetic mice. It is possible that a decrease in gap junction communication between granulosa cells leads to changes in paracrine communication between oocyte and granulosa cell that is related to the increase in apoptosis. High-glucose conditions have been found to decrease expression of connexin-43 in human retinal pericytes (11) and simultaneously increase apoptosis (22, 23). In addition, studies in rodent myocytes have demonstrated that interrupted cell-cell communication by down-regulation of connexin-43 is associated with accelerated apoptosis (24). Similarly, we have found that hyperglycemia is associated with a decrease in connexin-43 protein expression, which may impact granulosa cell intercellular communication and accelerate granulosa cell apoptosis. A similar inverse correlation between apoptosis and connexin-43 expression has recently been reported in avian granulosa cells (12). It is difficult to discern whether the deficient communication causes differences in cell signaling that then result in higher levels of apoptosis or whether apoptosis acts as a precursor for downstream problems in communication. In either case, this study links the diabetic condition with apoptosis, a decrease in connexin-43 expression, and an up-regulation of the TRAIL-KILLER cell-death pathway, providing some evidence that a relationship between these phenomena may exist in cell-cell communication in the CEO complex.
Delays in oocyte maturation are seen in the oocytes from both types of diabetic mice. We hypothesize that an insult or a preprogramming event may occur at the oocyte stage secondary to maternal hyperglycemia that permanently alters the course of normal development, and this manifests first as a maturational delay. Several prior studies from our laboratory and others support this hypothesis (5, 7, 8, 9). Embryos recovered in vivo at 48 h after fertilization from chemically induced diabetic mice experience an in vivo developmental delay. Only approximately 70% of the diabetic embryos had reached a two-cell stage, whereas 90% of the control embryos had already reached this maturational stage. Similarly, this delay in development persisted in vitro when these same embryos were cultured in control media for 72 h and assessed at each 24-h interval. A significant impairment in development was seen in the embryos from diabetic mice in their rates of progression to blastocyst compared with nondiabetic mice, despite the fact that both were cultured in identical medium conditions (7, 18). We postulate that this later effect on embryo development is related in part to the oocyte maturation delays described and that these changes are also associated with the increase in apoptosis in the CEOs. Similar conclusions have been drawn in experiments involving exposure to certain toxic agents during the preorganogenic period, from the time of sperm entry and zygote development to late blastocyst stage. Generoso and colleagues (25, 26, 27, 28, 29) published a series of papers demonstrating that brief exposure to ethylene oxide resulted in an increased incidence of fetal death and certain types of fetal malformations, such as craniofacial abnormalities, abdominal wall defects, limb defects, and stillbirths. These effects were not associated with induced chromosomal abnormalities or gene mutations (29). Therefore, there is evidence that oocyte-directed insults may be the result of poorly controlled diabetes during folliculogenesis in animal models as well as humans and that these insults may result in later reproductive failures, such as a higher incidence of malformation and miscarriages in this population of patients
In both the streptozotocin-induced and Akita diabetic mice, antral follicle oocyte size is significantly smaller compared with nondiabetic controls. In addition, in the Akita mice, the follicles at all stages of development are smaller compared with the follicles from the control animals. This smaller follicle and oocyte size may reflect abnormal cell growth and survival, which would correlate with the increased apoptosis seen in the CEOs from the diabetic mice. Similar findings of apoptosis and smaller cell size have been described in Drosophila as well as hematopoietic cell lines as a result of decreased Akt expression and activity (30, 31). Akt determines cell size and survival by modulating mammalian target of rapamycin activity and protein synthesis in the leukocytic cell line FL5.12 (31). TRAIL/KILLER interaction has also been linked to decreased Akt/phosphatidylinositol 3-kinase activity (32, 33). Therefore, it is possible that TRAIL/KILLER up-regulation in the CEOs may be modulating the activity of Akt, leading to decreased cell size and apoptosis. Dysregulation of the Akt pathway would also have downstream effects on protein synthesis and transcriptional regulation of genes perhaps involved in further embryonic development. Using this paradigm, up-regulation of TRAIL/KILLER, increased apoptosis, and decreased cell size may all be contributing to the poor oocyte quality and developmental potential of the resulting embryos.
Separating the physiological effects of high serum glucose levels from a lack of insulin is difficult in both of these models of type 1 diabetes, as would be expected in an animal model of human disease. Destruction of the ?-cells of the pancreas by streptozotocin or prevention of insulin secretion by the Ins2 mutation in the Akita mice causes significantly lower insulin and IGF-I levels as well as hyperglycemia. Previous studies from Kezele et al. (34) have shown that insulin acts as a paracrine factor to facilitate transition from primordial to primary follicle at the level of the oocyte. This group has suggested that abnormally low insulin levels as seen in a type 1 diabetic state may inhibit or retard primordial to primary follicle transition. In other studies, Demeestere et al. (35) have shown that IGF-I has a stimulatory effect on follicular steroidogenesis without any effect on oocyte maturation. They did report, however, improved embryo development as a result of exposure to IGF-I during follicular maturation, suggesting that IGF-I improved the quality of the oocyte. Other studies have suggested that IGF-I and insulin enhance granulosa cell proliferation and increase follicle diameter (36). Thus, it is possible that the effects of maternal diabetes seen in this study, specifically granulosa cell apoptosis and decreased oocyte and follicle size, as well as delayed oocyte maturation might be a combined effect of hyperglycemia and hypoinsulinemia. Colton et al. (8) found that the delay in FSH-stimulated oocyte maturation experienced by the streptozotocin-induced diabetic mice could be only partially recreated in vitro with high-glucose conditions, suggesting other factors may be involved. Although the presence of both simultaneous conditions in the animal models may be considered a limitation of this study, these physiological models may more accurately reflect the true follicular milieu caused by this disease, rather than the in vitro experiments examining the two conditions in isolation.
In both the acute chemically induced diabetic mouse model and in the chronically hyperglycemic Akita mouse, detrimental effects are seen in oocyte development and the degree of apoptosis in the surrounding cumulus cells. For these reasons, women with type I diabetes mellitus may be placing their oocytes and offspring at risk. It is known from human studies on patients undergoing in vitro fertilization that increased cumulus cell apoptosis correlates with poorer-quality oocytes and poor pregnancy outcome including increased rates of miscarriage. Similarly, several studies in human oocytes have correlated small oocyte size with poorer developmental potential and pregnancy rates (37, 38, 39, 40, 41). As a result of these findings, we speculate that diabetic women who experience uncontrolled or poorly controlled diabetes during ovulation and fertilization may suffer detrimental effects on oocytes/CEOS secondary to hyperglycemia and hypoinsulinemia including developmental delay, up-regulation of cell-death effector pathways, and increased granulosa cell apoptosis. These early insults may then lead to an increased rate of miscarriage and congenital anomalies depending on the signaling and cell-death pathways involved.
Acknowledgments
We appreciate Elizabeth Schlichting and Amanda Wyman for their technical assistance and Dr. Stephen Downs for his helpful scientific advice. We also thank Edward Leiter for his assistance in obtaining the Akita mouse.
References
Greene, MF 1999 Spontaneous abortions and major malformations in women with diabetes mellitus. Semin Reprod Endocrinol 17:127–136
Farrell T, Neale L, Cundy, T 2002 Congential anomalies in the offspring of women with type I, type II and gestational diabetes. Diabet Med 19:322–326
Moley KH, Chi MM, Knudson CM, Korsmeyer SJ, Mueckler MM 1998 Hyperglycemia induces apoptosis in preimplantation embryos via cell death effector pathways. Nat Med 12:1421–1424
Moley KH, Chi MM, Mueckler MM 1998 Maternal hyperglycemia alters glucose transport and utilization in mouse preimplantation embryos. Am J Physiol 275:E38–E47
Moley KH, Vaughn WK, DeCherney AH, Diamond MP 1991 Effect of diabetes mellitus on mouse preimplantation embryo development. J Reprod Fertil 93:325–332
Chi MM, Pingsterhaus J, Carayannopoulos, M, Moley KH 2000 Decreased glucose transporter expression triggers BAX-dependent apoptosis in the murine blastocyst. J Biol Chem 275:40252–40257
Diamond MP, Moley KH, Pellicer A, Vaughn WK, DeCherney AH 1989 Effects of streptozotocin and alloxan-induced diabetes mellitus on mouse follicular and early embryo development. J Reprod Fertil 86:1–10
Colton SA, Pieper GM, Downs SM 2002 Altered meiotic regulation in oocytes from diabetic mice. Biol Reprod 67:220–230
Colton SA, Humpherson PG, Leese HJ, Downs SM 2003 Physiological changes in oocyte-cumulus cell complexes from diabetic mice that potentially influence meiotic regulation. Biol Reprod 69:761–770
Ackert CL, Gittens JEI, O’Brien MJ, Eppig JJ, Kidder GM 2001 Intercellular communication via connexin43 gap junctions is required for ovarian folliculogenesis in the mouse. Dev Biol 233:258–270
Li A, Sato T, Haimovici R, Okamoto T, Roy S 2003 High glucose alters connexin 43 expression and gap junction intercellular communication activity in retinal pericytes. Invest Ophthalmol Vis Sci 44:5376–5382
Krysko DV, Mussche S, Leybaert L, D’Herde K 2004 Gap junctional communication and connexin43 expression in relation to apoptotic cell death and survival of granulosa cells. J Histochem Cytochem 52:1199–1207
Zhang H, Vollmer M, De Geyter M, Litzistorf Y, Ladewig A, Durrenberger M, Guggenheim R, Miny P, Holzgreve W, De Geyter C 2000 Characteristics of an immortalized human granulosa cell line COV 434. Mol Hum Reprod 2:146–153
Idil M, Cepni I, Demirsoy G, Ocal P, Salihoglu F, Senol H, Elibol F, Irez T 2004 Does granulosa cell apoptosis have a role in the etiology of unexplained infertility? Eur J Obstet Gynecol Reprod Biol 112:182–184
Lee KS, Joo BS, Na YJ, Yoon MS, Choi OH, Kim WW 2001 Cumulus cell apoptosis as an indicator to predict the quality of oocytes and the outcome of IVF-ET. J Assist Reprod Genet 18:490–498
Nakahora K, Saito H, Satio T, Ito M, Ohta N, Takahashi T, Hiroi M 1997 The incidence of apoptotic bodies in membrana granulosa can predict prognosis of ova from patients participating in in vitro fertilization programs. Fertil Steril 68:312–317
LeBlanc H, Lawrence D, Varfolomeev E, Totpal K, Morlan J, Schow P, Fong S, Schwall R, Sinicropi D, Ashkenazi A 2002 Tumor-cell resistance to death receptor-induced apoptosis through mutational inactivation of the proapoptotic Bcl-2 homolog Bax. Nat Med 8:274–281
Moley KH 1999 Diabetes and preimplantation events of embryogenesis. Semin Reprod Endocrinol 17:137–151
Izumi T, Yokota-Hashimoto H, Zhao S, Wang J, Halban PA, Takeuchi T 2002 Dominant negative pathogenesis by mutant proinsulin in the Akita diabetic mouse. Diabetes 52:409–416
Wang J, Takeuchi T, Tanaka S, Kubo S, Kayo T, Lu D, Takata K, Koizumi A, Izumi T 1999 A mutation in the insulin 2 gene induces diabetes with severe pancreatic ?-cell dysfunction in the Mody mouse. J Clin Invest 103:27–37
Eppig JJ, Wigglesworth K, Pendola FL 2002 The mammalian oocyte orchestrates the rate of ovarian development. Proc Natl Acad Sci USA 99:2890–2894
Podesta F, Romeo G, Liu WH, Krajewski S, Reed JC, Gerhardinger C, Lorenzi M 2000 Bax is increased in the retina of diabetic subjects and is associated with pericyte apoptosis in vivo and in vitro. Am J Pathol 156:1025–1032
Romeo G, Liu WH, Asnaghi V, Kern TS, Lorenzi M 2002 Activation of nuclear factor-B induced by diabetes and high glucose regulates a proapoptotic program in retinal pericytes. Diabetes 51:2241–2248
Yasui K, Kada K, Hojo M, Lee J-K, Kamiya K, Toyama J, Opthof T, Kodama T 2000 Cell-to-cell interaction prevents cell death in cultured neonatal rat ventricular myocytes. Cardiovasc Res 48:68–76
Generoso WM, Shourbaji AG, Piegorsch WW, Bishop JB 1991 Development response of zygotes exposed to similar mutagens. Mutat Res 250:439–446
Polifka JE, Rutledge JC, Kimmel GL, Dellarco V, Generoso WM 1996 Exposure to ethylene oxide during the early zygotic period induces skeletal anomalies in mouse fetuses. Teratology 53:1–9
Rutledge JC, Generoso WM, Shourbaji A, Cain KT, Gans M, Oliva J 1992 Developmental anomalies derived from exposure of zygotes and first-cleavage embryos to mutagens. Mutat Res 296:167–177
Kimmel CA, Generoso WM, Thomas RD, Bakshi KS 1993 Contemporary issues in toxicology: a new frontier in understanding the mechanisms of developmental abnormalities. Toxicol Appl Pharmacol 119:159–165
Katoh M, Cacheiro NL, Cornett CV, Cain KT, Rutledge JC, Generoso WM 1989 Fetal anomalies produced subsequent to treatment of zygotes with ethylene oxide or ethyl methanesulfonate are not likely due to the usual genetic causes. Mutat Res 210:337–344
Verdu J, Buratovich MA, Wilder EL, Birnbaum MJ 1999 Cell-autonomous regulation of cell and organ growth in Drosophila by Akt/PKB. Nat Cell Biol 1:500–506
Edinger AL, Thompson CB 2002 Akt maintains cell size and survival by increasing mTOR-dependent nutrient uptake. Mol Biol Cell 13:2276–2288
Kang YC, Kim KM, Lee KS, Namkoong S, Lee SJ, Han JA, Jeoung D, Ha KS, Kwon YG, Kim YM 2004 Serum bioactive lysophospholipids prevent TRAIL-induced apoptosis via P13K/Akt-dependent cFLIP expression and Bad phosphorylation. Cell Death Differentiation 11:1287–1298
Larribere L, Khaled M, Tartare-Deckert S, Busca R, Luciano F, Bille K, Valony G, Eychene A, Auberger P, Ortonne JP, Ballotti R, Bertolotto C 2004 P13K mediates protection against TRAIL-induced apoptosis in primary human melanocytes. Cell Death Differ 11:1084–1091
Kezele PR, Nilsson EE, Skinner MK 2002 Insulin but not insulin-like growth factor-1 promotes the primordial to primary follicle transition. Mol Cell Endocrinol 192:37–43
Demeestere I, Gervy C, Centner J, Devreker F, Englert Y, Delbaere A 2004 Effect of insulin-like growth factor-1 during preantral follicular culture on steroidogenesis, in vitro oocyte maturation, and embryo development in mice. Biol Reprod 70:1664–1669
Zhao J, Taverne MA, Van Der Weijden GC, Bevers MM, Van Den Hurk R 2001 Insulin-like growth factor-1 (IGF-1) stimulates the development of cultured rat pre-antral follicles. Mol Reprod Dev 58:287–296
Trounson AO, Gosden RG 2003 Biology and pathology of the oocyte. Cambridge, UK: Cambridge University Press; 307–309
Nataprawira DS, Harada T, Sekijima A, Mio Y, Terakawa N 1992 Assessment of follicular maturity by follicular diameter and fluid volume in a program of in vitro fertilization and embryo transfer. Asia Oceania J Obstet Gynaecol 18:225–230
Wittmaack FM, Kreger DO, Blasco L, Tureck RW, Mastroianni Jr L, Lessey BA 1994 Effect of follicular size on oocyte retrieval, fertilization, cleavage and embryo quality in in vitro fertilization cycles: a 6 year data collection. Fertil Steril 62:1205–1210
Arnot AM, Vanderkerckhove P, DeBono MA, Rutherford AJ 1995 Follicular volume and number during in vitro fertilization: association with oocyte development capacity and pregnancy rate. Hum Reprod 10:256–261
Bergh C, Broden H, Lundin K, Hamberger L 1998 Comparison of fertilization, cleavage and pregnancy rates of oocytes from large and small follicles. Hum Reprod 13:1912–1915(Aimee S. Chang, Alexis N.)