Predictive value of cumulus cell apoptosis with regard to blastocyst development of corresponding gametes

Transcription

1 Predictive value of cumulus cell apoptosis with regard to blastocyst development of corresponding gametes Claudia Maria Corn, B.Sc., a Cornelia Hauser-Kronberger, Ph.D., b Marianne Moser, Ph.D., c Gernot Tews, M.D., c and Thomas Ebner, Ph.D. c a University of Salzburg, Institute of Zoology, Salzburg; b General Hospital and Paracelsus University, Institute of Pathology, Salzburg; and c IVF-Unit, Women s General Hospital, Linz, Austria Objective: To test if a high degree of apoptosis in cumulus cells could indicate diminished oocyte quality and developmental competence. Design: Prospective analysis. Setting: Public hospital and university. Patient(s): Thirty seven women who gave written consent to participate in this study. Intervention(s): Cumulus oocyte complexes were denuded separately and the resulting cumulus cell suspensions were analyzed for presence of apoptosis using a terminal deoxynucleotidyl transferase-mediated digoxigenindutp nick-end labeling (TUNEL) assay detection kit. Main Outcome Measure(s): Percentage of apoptotic cumulus cells and blastocyst formation rate. Result(s): Younger patients ( 35 years) showed significantly fewer apoptotic cumulus cells than older ones. In all patients gamete maturity was shown to be highly correlated to the rate of apoptosis in cumulus cells. At zygote and cleavage stages (days 1 to 4) no morphologic features were related to the degree of programmed cell death. However, blastocyst development was predictable taking into account the percentage of apoptosis in associated somatic cells. In addition, there was a trend toward better quality blastocysts from follicles with a lower rate of apoptotic cells. No influence on pregnancy and implantation rate was observed. Conclusion(s): Apoptotic processes within follicles seem to impair oocyte maturation. Though not manifested in the morphologic appearance, gametes and embryos derived from cumulus complexes with no or minor apoptosis have an increased chance of giving rise to optimal blastocysts. (Fertil Steril 2005;84: by American Society for Reproductive Medicine.) Key Words: Apoptosis, blastocyst development, cumulus cells, oocyte maturity, TUNEL assay In contrast to necrosis, which results from injury and affects larger group of cells, apoptosis is a predetermined event in the programmed death of single cells (1). Both forms of cell death can be distinguished morphologically. Although necrosis causes cellular swelling and membrane rupture, apoptotic processes are characterized by nuclear chromatin condensation with subsequent DNA degradation, cellular shrinkage, and the appearance of multiple cytoplasmic fragments. Apoptosis exerts a homeostatic function with regard to tissue dynamics; for example, it is thought to be the underlying mechanism of ovarian follicle degeneration during atresia (2). Palumbo and Yeh (3) were able to show that these events did not only occur in natural cycles but also in gonadotropin-stimulated ones. Taking into account the different inhibitory (E 2, FSH, LH, hcg) and inductive (androgens, GnRH, GnRH analogues) effects of several hormones on apoptosis (4 6), it is not surprising that controlled ovarian hyperstimulation can increase the rate of apoptosis in cumulus cells compared with in vivo cycles (7). Received December 23, 2004; revised and accepted March 17, Reprint requests: Thomas Ebner, Ph.D., Women s General Hospital, IVF- Unit, Lederergasse 47, A-4020 Linz, Austria (FAX: ; thomas.ebner@gespag.at). Vice versa, apoptosis has been used to estimate ovarian reserve in women undergoing IVF (8). Consequently, programmed cell death is likely to be increased in patients with low response (9), advanced maternal age (10, 11), and certain indications for IVF treatment (12). However, because most of these studies either analyzed only a limited number of follicles or actually pooled cumulus cells from all follicles, the predictive value remains questionable. Some investigators (13 15) correlated apoptosis of cumulus cells with gametes and embryos of the corresponding follicle, thus providing first insight into the actual involvement of apoptotic processes in preimplantation development. To our knowledge, all analyses of apoptosis on a per-follicle basis reported in literature were ceased before embryonic genome activation; therefore, this prospective study was set up to track all gametes until day 5 of in vitro development. MATERIALS AND METHODS Patients This prospective study consists of 37 intracytoplasmic sperm injection (ICSI) patients (37 cycles) frequenting the IVF unit at the Women s General Hospital during a 7-month period (February to August 2004). Only those couples were con /05/$30.00 Fertility and Sterility Vol. 84, No. 3, September 2005 doi: /j.fertnstert Copyright 2005 American Society for Reproductive Medicine, Published by Elsevier Inc. 627

2 sidered for the present study, and all gave informed consent. Our institutional review board gave approval, as the cumulus cells would have been disposed anyway. The mean age of the women was 32.9 ( 4.6) years. A total of nine women (24.3%) suffered from secondary infertility, and the remaining patients had no history of pregnancy. In the majority of cases, no (n 22) or only one (n 8) previous cycle could be recorded (mean: cycles). Because this study deals with ICSI cases, the number of exclusive male factor infertility (n 24) is rather high (64.9%). In five couples, female factor infertility was the only logical reason for their childlessness. However, in eight couples a combination of male and tubal factor infertility was observed. Controlled Ovarian Hyperstimulation Two stimulation protocols were applied to harvest a maximum of oocytes without risking ovarian hyperstimulation syndrome. Cycle monitoring consisted of ovarian ultrasonography on cycle day 1 to 3 to exclude functional cysts and further ovarian ultrasonography on and serum E 2, progesterone, and LH measurement from stimulation day 5 onward, with monitoring frequency based on patient response. Approximately half of the patients (n 17) were stimulated using a long protocol. In the long protocol, down-regulation of the pituitary was achieved with the GnRH agonist buserelin (Suprecur; Aventis Pharma, Vienna, Austria). Stimulation was initiated with human menopausal gonadotropin (hmg, Menogon; Ferring Pharmaceutical, Kiel, Germany). In the GnRH-antagonist protocol (n 20), recombinant FSH (Puregon; Organon, Vienna, Austria) was started on day 2 of the cycle. In addition, a GnRH antagonist (Orgalutran; Organon) was administered after 5 to 6 days of stimulation, depending on the presence of a 12- to 13-mm follicle in the ultrasound scan. Immediately after denudation of the gametes, they were removed for further injection, and the cumulus cells were aspirated from the hyaluronidase and transferred into an Eppendorf cup with 7% buffered formalin for fixation. After a 10-minute incubation period, probes were centrifuged (6,000 rpm, 1 minute) and washed twice in PBS. Before storage in the refrigerator until analysis, the pellet of cumulus cells was resuspended. All Eppendorf cups were labeled with both patient and chain-of-custody number to guarantee individual follow-up. The cell suspension was carefully pipetted into a cytospin centrifuge tube which was attached to a silanised glass slide. Probes were concentrated on glass slides by cytospin (Shandon, Pittsburgh, PA) at 6,000 rpm for 7 minutes. Loaded glass slides were air dried for 1 hour. Before staining for apoptosis, cells were digested for 8 minutes using proteinase K (Qiagen, Vienna, Austria), which was to facilitate access of the dye. We used a terminal deoxynucleotidyl transferase-mediated digoxigenin-dutp nick-end labeling (TUNEL) assay (Apop- Tag Fluorescein Detection Kit; Abbott, Wiesbaden, Germany) to detect DNA fragmentation in cumulus cells, known to be a main feature of apoptosis. ApopTag Apoptosis Detection Kit detects apoptotic cells in situ by labeling the free 3=-OH DNA termini with chemically labeled and unlabeled nucleotides. Nucleotides labeled with digoxigenin contained in the reaction buffer were enzymatically added to the DNA by terminal deoxynucleotidyl transferase (TdT) at an incu- FIGURE 1 Apoptosis in cumulus cells analyzed using a TUNEL assay. Apoptotic cells appear yellowish green compared with red nonapoptotic cells. In all patients, ovulation was induced with 5,000 10,000 IU human chorionic gonadotropin (hcg, Pregnyl; Organon), provided that the lead follicle had reached a diameter of 19 mm and the serum E 2 level appeared adequate. Oocyte retrieval was carried out transvaginally under ultrasound guidance 36 hours after hcg administration. Cumulus Cell Removal and Preparation Following follicular puncture, cumulus oocytes complexes (COC) were cultured for at least 3 hours in BM1 medium (NMS Bio-Medical, Praroman, Switzerland). To obtain a maximum of cumulus cells, all COCs were cultured individually in 250- L droplets of hyaluronidase (80 IU/mL; MediCult, Copenhagen, Denmark) for 30 seconds. Final removal of the cumulus cells was performed mechanically using hand-drawn glass pipettes. 628 Corn et al. Cumulus cell apoptosis and development Vol. 84, No. 3, September 2005

3 bation time for 1 hour at 37 C. The DNA fragments labeled with the digoxigenin-nucleotide were then allowed to bind an antidigoxigenin antibody conjugated to a fluorescein reporter molecule for 30 minutes at room temperature. Cells were counterstained with propidium iodine, resulting in red-stained nuclei. The specimens were then analyzed by fluorescent microscopy and a minimum of 75 nuclei of granulosa cells were counted at random. Apoptotic nuclei were defined as positive when showing a green or light green color (Fig. 1); cells without DNA fragmentation show only red nuclear staining. To minimize intraobserver variability, apoptotic cells were counted twice. Intracytoplasmic Sperm Injection The separated sperms (mini swim-up) were incubated in 10- L droplets of fresh BM1 media on the injection dish (Falcon type 1006). Two small droplets of a PVP solution (MediCult, Copenhagen, Denmark) were also prepared. With this constellation of droplets under mineral oil, contamination of the PVP caused by debris carried by the sperm suspension should be avoided. The technique of ICSI has been described in detail elsewhere (16). Briefly, micromanipulation was performed on an inverted microscope ( 200 magnification; Olympus, Vienna, Austria) with Hoffman Modulation Contrast (Modulation Optics Inc., Greenvale, NY) and electronically controlled heat stage and hydraulic micromanipulators (Luigs and Neumann, Ratingen, Germany). Gamete and Embryo Selection At the time of ICSI (day 0), oocyte maturity and quality could be controlled easily. Cytoplasmic anomalies were recorded as well as anomalies of the outer layer (16) as shown in Table 1. All gametes remained individually in the ICSI dish until fertilization control, which was performed 16 to 18 hours after injection (day 1). Zygotes with one or three pronuclei were discarded, and the normal pronuclear patterns were scored according to the criteria of Montag and Van der Ven (17). Additionally, presence of a cytoplasmic halo (18) was controlled. The medium was changed to BlastAssist Medium 1 (MediCult). On day 2, embryo quality was graded, mainly considering the number of blastomeres and the degree of fragmentation (19). In addition, embryos were supplied with fresh medium (BlastAssist Medium 1). According to the number and quality of eight-cell embryos on day 3 (20), transfer was taken into consideration either on day 3 (n 18) or 5(n 19). All embryos that were not transferred to day-3 candidates as well as those concepti that originally were considered for blastocyst transfer had their medium changed to BlastAssist Medium 2 (MediCult). On day 4 of preimplantation development, the degree of compaction was the only criterion analyzed (21). Fresh medium (BlastAssist Medium 2) was supplied once again. At the day of blastocyst transfer (day 5), the degree of expansion and quality of blastocysts was recorded (18, 22). In detail, the cell number and morphology of trophectoderm (TE) as well as inner cell mass (ICM) were scored. Good quality blastocysts derived from patients who TABLE 1 Comparison of controlled ovarian hyperstimulation details and treatment outcome according to women s age. Women s age 35 years 36 years P value No. of patients Basal FSH Dose hmg (IU) 1, , Days of stimulation Endometrium (mm) E 2 at ovulation induction 1, , Apoptotic cumulus cells (%) No. of oocytes PN No. transferred Clinical pregnancies 9 (36.0) 4 (33.3).57 Note: Values in parentheses are percentages. All values are mean SD. FSH follicle stimulating hormone; hmg human menopausal gonadotropin; PN pronuclei. Fertility and Sterility 629

4 had a day-3 transfer and supernumerary blastocysts from day-5 patients were cryopreserved. Embryo Transfer and Luteal Phase Support Taking into consideration all prognostic markers (day 0 to day 5), a maximum of two embryos/blastocysts were chosen for intrauterine transfer. Therefore, all patients were placed in the lithotomy position during embryo replacement, and neither sedation nor anesthesia was used. Embryos were loaded into an Edwards-Wallace Catheter (Smiths Industries, Lancing, United Kingdom) using 10 L of BlastAssist Medium 2 (MediCult) and then expelled approximately 1 cm from the fundus. On days 2, 5, and 8, 3,000 IU hcg (Pregnyl; Organon) were injected to support luteal phase. Seventeen days after intrauterine transfer, the blood concentration of hcg was measured. Clinical pregnancy was determined by visualization of at least one gestational sac with positive heart activity 4 weeks after embryo transfer (ET). Subclinical pregnancy showed no fetal heartbeat. Implantation rate was defined by ultrasound visualization of a gestational sac per transferred embryo (subclinical pregnancies were included). Statistics The apoptosis data are presented as mean standard deviation (SD). Comparison was performed using chi-square and t-tests. P.05 was considered statistically significant %. The group showing empty zonae pellucidae had % apoptotic cumulus cells. A statistically significant correlation was found between the degree of apoptosis in granulosa cells and the maturity of the corresponding gamete. In the group of immature oocytes (GV and MI) the presence of apoptotic granulosa cells was statistically significantly higher (P.0009) compared with mature (MII) oocytes. No such difference could be observed evaluating cumulus cells from follicles not showing any gamete. First polar body morphology could only be evaluated in 94.6% of MII oocytes (polar bodies of 11 gametes were not in focus and thus could not be evaluated). A total of 101 (49.5%) oocytes showed a fragmented polar body, and 92 oocytes (45.1%) had an intact polar body. Fragmented polar bodies showed an apoptosis rate of %, and intact polar bodies an apoptosis rate of % (P.05). A total of 47.1% (96 of 204) MII oocytes showed no morphologic anomaly, but 108 MII oocytes (53.0%) were affected. One hundred and two oocytes had either a cytoplasmic anomaly or an anomaly of the shell, and six had both. As indicated in Table 2, no correlation was found between the percentage of apoptotic granulosa cells and the different anomalies. Zygote Stage (Day 1) Using ICSI, 126 MII oocytes (61.8%) could be fertilized correctly ( % apoptosis). However, 53 zygotes RESULTS A total of 260 oocytes were harvested in 37 patients whose accessory granulosa cells were analyzed with the TUNEL assay. The mean ( SD) number of oocytes aspirated per patient was A mean number of granulosa cells per cumulus complex could be evaluated showing an average apoptosis percentage of % (range from 0 to 62%). However, in seven cases our methodical inclusion criteria ( 75 cells) were not met, resulting in 253 analyzable cumulus complexes (97.3% of all follicles). Pooling all follicles of a patient, it could be demonstrated (see Table 1) that younger patients ( 35 years) showed statistically significantly less (P.03) apoptosis ( %) in their cumulus cells than older ones ( %). Stimulation protocol (antagonist vs. long protocol), cycle day-3 hormonal parameters, and stimulation response turned out to be unrelated to apoptotic processes. Egg Maturity and Quality (Day 0) From these 253 follicles 204 ova (80.6%) were found to be at metaphase II (MII), nine (3.6%) at metaphase I (MI), and 24 (9.5%) at germinal vesicle stage (GV). Empty zonae pellucidae were observed in 16 (6.3%) cases. Apoptosis in MII oocytes showed a mean ( SD) percentage of %, in MI oocytes %, and at prophase 1 (GV) TABLE 2 Morphologic features in MII oocytes and their association with apoptotic processes in corresponding cumulus cells. Type n Apoptosis No anomaly Cytoplasmic anomalies Aggregation of ER Discoloration Central granulation Incorporations Refractile body Bull s eye Vacuole Anomalies of outer layer Granules in PVS Giant oocyte Ovoid shape Note: All values are mean SD; P.05, oocytes with anomalies compared with unaffected gametes. ER endoplasmic reticulum; PVS perivitelline space. 630 Corn et al. Cumulus cell apoptosis and development Vol. 84, No. 3, September 2005

5 TABLE 3 Relation between embryo morphology at cleavage stages and programmed cell death in corresponding cumulus cells. Morphologic feature n Apoptosis Day 2 4 blastomeres blastomeres No fragmentation Moderate fragmentation Severe fragmentation Multinucleation/unequal cells Day 3 6 blastomeres blastomeres No fragmentation Moderate fragmentation Severe fragmentation Day 4 Compacting Not compacting Day 5 Blastocysts a Arrested a Good quality blastocysts Bad quality blastocysts a P.001. (26.0%) had an abnormal number of pronuclei ( %), 42 showed no signs of fertilization ( %), and 25 oocytes (12.3%) were damaged following ICSI procedure ( %). Programmed cell death had no impact on fertilization outcome (P.05). An optimal pronuclear pattern (0A or 0B) was seen in 30 zygotes ( %), and 91 oocytes had patterns 1 to 5 ( %). The pattern of five oocytes could not be evaluated due to pronuclear membrane breakdown before examination. Once again, no correlation could be seen (P.05). Out of these 121 MII zygotes, 98 (81.0%) oocytes showed a halo ( %), whereas 23 oocytes (19%) showed no halo ( % apoptosis). The rate of apoptosis in granulosa cells had no statistically significant influence on halo appearance (P.05). Cleavage Stages (Days 2 to 4) On day 2, 124 (98.4%) out of 126 zygotes showing two pronuclei cleaved. No correlation between apoptosis and developmental rate or embryo morphology could be observed (P.05). The same observation was made for day-3 results. Adequate development on day 4, as assessed by the beginning of compaction, was not related to apoptosis in cumulus cells either (Table 3). Blastocyst Stage (Day 5) Eighty-nine day-3 embryos were considered for blastocyst culture (including surplus embryos from day 3), of which 44 survived until day 5 (49.4%). The apoptosis value in this group ( %) differed in a statistically significant manner (P.0006) from the value of the cohort showing developmental arrest ( %). Optimal blastocysts showed a strong tendency toward a reduced apoptosis rate in corresponding somatic cells compared with blastocysts of fair or bad quality (P.052). Treatment Outcome A mean number of 1.8 ( 0.5) embryos/blastocysts was transferred. In total, a pregnancy rate of 40.5% (15 out of 37) was achieved; however, two subclinical pregnancies reduced the clinical pregnancy rate to 35.1% (13 out of 37). The associated implantation rate was 28.4% (19 out of 67). No difference was observed between embryo and blastocyst transfer (P.14). The multiple pregnancy rate was 26.7% (twins only). Pooling all follicles, the overall mean percentage of apoptosis was inadequate to predict pregnancy (P.68), as was the mean value of those cumulus complexes whose embryos/ blastocysts were transferred (P.38). Considering only those patients showing a 100% implantation rate (n 5) and comparing them with those with no implantation at all (n 22), predictability of implantation improved (P.12) but did not reach statistical significance. DISCUSSION Within growing follicles, oocytes gradually acquire developmental competence to further support early embryonic development. This process is intrinsically linked to folliculogenesis; thus, the fate of the germ cell depends on the health of the developing follicle (e.g., its granulosa cells). In fact, a close association between oocytes and their surrounding cumulus cells is assumed during the whole process of oocyte growth and development. Specialized cumulus cells have transzonal cytoplasmic processes penetrating through the zona pellucida and reaching the oolemma. In this region of contact, gap junctions are present that facilitate the transfer of low-molecular-weight molecules between oocyte and cumulus cells and also between cumulus cells (23). However, the communication axis between germ and somatic cells is not just in one direction but rather is bidirectional (24). This is supported by the work of Depalo et al. (25) who were able to detect apoptosis in both oocytes and cumulus cells of primordial and primary follicles. As a result Fertility and Sterility 631

6 of this mutual dependence, it is very likely that any apoptosisrelated process influencing the cumulus cells will have a similar impact on the oocyte itself; that is, a high degree of apoptosis in cumulus cells would represent diminished oocyte quality and developmental competence. Indeed, Nakahara et al. (13) observed the lowest incidence of apoptotic bodies in granulosa cells of follicles from which the oocytes actually fertilized and developed into good quality embryos. These gametes and embryos from low apoptosis follicles may lead to higher pregnancy rates (25, 26). However, most of this early work on apoptosis was done by analyzing mural cumulus cells from follicular fluid with flow-cytometry. Because differences in the degree of apoptosis between mural and cumulus granulosa cells have been reported (26), cells from follicular fluid may not reflect the actual quality of the respective oocyte or embryo. Because of the high number of cells needed, flow cytometry is not sufficient as a routine assay for analyzing the cumulus cells that might reflect oocyte quality and developmental capacity more accurately (27). Additional problems in terms of prognosis may arise (28) if pooled follicular fluids are used, which will not allow for accurate assignment of apoptotic features to oocytes. Using an in situ apoptosis detection kit, Høst et al. (14 15) evaluated the presence of apoptosis-positive cumulus cells and correlated them with the corresponding oocytes. In short, a remarkable influence of apoptosis on oocyte maturity was detected, as germinal vesicle stage and metaphase I oocytes were separated from cumulus complexes with a high rate of apoptotic cells. This is consistent with our present findings, indicating that apoptotic processes during controlled ovarian hyperstimulation may cause atresia (28) or stop follicle growth in certain instances of diminished quality. No such relationship was documented by Moffatt et al. (29), but they were using different markers (e.g., Fas, Fas ligand, Bcl-X L ) for their immunohistochemical evaluation of programmed cell death. It is interesting that some morphologic features of the oocyte do not seem to be related to the rate of apoptotic cumulus cells, as no increased rate was seen either in gametes with thickened zona pellucida (15) or in eggs with vacuolization or central granulation (29). This is strongly supported by our more detailed data which showed no correlation between individual or pooled cytoplasmic abnormalities and irregularities of the outer shell. Obviously, apoptotic processes gain influence on nuclear (i.e., maturity) rather than on cytoplasmic maturation. There is some evidence that the fertilization rate on day 1 of preimplantation development is higher in oocytes derived from follicles with no or less apoptosis (11, 13 15). However, these studies did not evaluate morphologic criteria at the zygote stage which is known to determine the further fate of the embryo. The present investigation is the first to correlate pronuclear pattern and halo formation with apoptosis, and both parameters turned out not to be related to programmed cell death in cumulus cells. Though a relationship between cleavage stage on day 2 and apoptosis has been suggested (11, 13), our data do not support this suggestion at all because neither the number of blastomeres nor the percentage of fragmentation could be correlated to apoptosis. This developmental stage is before embryonic genome expression, so it would have been of great interest if a suspected relation between cleavage behavior on day 2 and apoptotic phenomena had persisted until day 3. Unfortunately, we are not aware of any data in literature analyzing embryo morphology after day 2 or trying to correlate it with apoptotic processes. Because our strategy was to perform transfer either on day 3 or day 5, prolonged culture allowed for detailed scoring after embryonic genome expression. However, no impact of the degree of apoptotic cumulus cells on corresponding day-3 and day-4 embryos could be seen. This once more emphasizes that day-3 morphology is not an adequate indicator of blastocyst development (30, 31) because survival to blastocyst stage (day 5) was closely correlated with minor apoptosis in corresponding cumulus cells. Together with an observed trend (P.05) toward a better blastocyst quality in follicles with reduced apoptosis, this may indeed explain data from literature indicating that pregnant patients have a significantly decreased mean number of apoptotic cumulus cells compared with nonpregnant ones (11, 25, 26). It is a major problem that any impact on pregnancy rate is only of a theoretical nature because apoptosis values of all follicles are usually pooled; it may indeed be that there is considerable divergence in the number of apoptotic cumulus cells within the follicles of a patient (as indicated by the present data). Thus, analysis would be more accurate if only the apoptosis values of those follicles had been considered from which corresponding embryos or blastocysts had been transferred. Nevertheless, no such relationship could be shown in the present study presumably because not all embryos transferred per patient were implanted. As a consequence (e.g., in case of singleton pregnancies), embryologists are unaware which embryo actually was successful after transfer of more than one embryo. To get the most accurate results, only patients with a 100% implantation rate should be considered and compared with patients with failed implantation. This approach gave promising results in the present study, but the level of statistical significance was not reached, probably due to the rather small patient numbers for this purpose. Our work is the first to accurately correlate most morphologic prognostic parameters (32) throughout the first 5 days of preimplantation development with the degree of apoptosis in corresponding cumulus cells. It could be demonstrated that a higher number of apoptotic cells impairs blastocyst development and quality. Because prediction of implantation at the time of follicle aspiration would allow for selective transfer at the cleavage stage, any theoretical drawback of 632 Corn et al. Cumulus cell apoptosis and development Vol. 84, No. 3, September 2005

7 prolonged culture could be avoided. To support this, a prospective study is currently being conducted that considers for transfer only those day-3 embryos derived from follicles/ cumulus complexes with minor apoptosis. REFERENCES 1. Kerr JFR, Wyllie AH, Currie AR. Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer 1972;26: Tilly JL, Kowalski KI, Johnson AL, Hsueh AL. Involvement of apoptosis in ovarian follicular atresia and postovulatory regression. Endocrinology 1991;129: Palumbo A, Yeh J. In situ localization of apoptosis in the rat ovary during follicle atresis. Biol Reprod 1994;51: Billig H, Furuta I, Hsueh AJ. Estrogens inhibit and androgens enhance ovarian granulosa cell apoptosis. Endocrinology 1993;133: Papadopoulos V, Dharmarajan AM, Li H, Culty M, Lemay M, Sridaran R. Mitochondrial peripheral-type benzodiazepine receptor expression. Correlation with gonadotropin-releasing hormone (GnRH) agonistinduced apoptosis in the corpus luteum. Biochem Pharmacol 1999; 58: Matsubara H, Ikuta K, Ozaki Y, Suzuki Y, Suzuki N, Sato T, et al. Gonadotropins and cytokines affect luteal function through control of apoptosis in human luteinized granulosa cells. J Clin Endocrinol Metab 2000;85: Kaneko T, Saito H, Takahashi T, Ohta N, Saito T, Hiroi M. Effects of controlled ovarian hyperstimulation on oocyte quality in terms of the incidence of apoptotic granulosa cells. J Assist Reprod Genet 2000;17: Seifer DB, Gradiner AC, Ferreira KA, Peluso JJ. Apoptosis as a function of ovarian reserve in women undergoing in vitro fertilization. Fertil Steril 1996;66: Quintana R, Kopcow L, Marconi G, Sueldo C, Speranza G, Baranao RI. Relationship of ovarian stimulation response with vascular endothelial growth factor and degree of granulosa cell apoptosis. Hum Reprod 2001;16: Sadraie SH, Saito H, Kaneko T, Saito T, Hiroi M. Effects of aging on ovarian fecundity in terms of the incidence of apoptotic granulosa cells. J Assist Reprod Genet 2000;17: Lee KS, Joo BS, Na YJ, Yoon MS, Choi OH, Kim WW. Cumulus cell apoptosis as an indicator to predict the quality of oocytes and the outcome of IVF-ET. J Assist Reprod Genet 2001;18: Nakahara K, Saito H, Saito T, Ito M, Ohta N, Takahashi T, et al. Ovarian fecundity in patients with endometriosis can be estimated by the incidence of apoptotic bodies. Fertil Steril 1998;69: Nakahara K, Saito H, Saito T, Ito M, Ohta N, Takahashi T, et al. The incidence of apoptotic bodies in membrana granulosa can predict prognosis of ova from patients participating in in vitro fertilization programs. Fertil Steril 1997;68: Høst E, Mikkelsen AL, Lindenberg S, Smidt-Jensen S. Apoptosis in human cumulus cells in relation to maturation stage and cleavage of the corresponding oocyte. Acta Obstet Gynecol Scand 2000;79: Høst E, Gabrielsen A, Lindenberg S, Smidt-Jensen S. Apoptosis in human cumulus cells in relation to zona pellucida thickness variation, maturation stage, and cleavage of the corresponding oocyte after intracytoplasmic sperm injection. Fertil Steril 2002;77: Ebner T, Yaman C, Moser M, Sommergruber M, Jesacher K, Tews G. A prospective study on oocyte survival rate after ICSI: influence of injection technique and morphological features. J Assist Reprod Genet 2001;18: Montag M, Van der Ven H. Evaluation of pronuclear morphology as the only selection criterion for further embryo culture and transfer: results of a prospective multicentre study. Hum Reprod 2001;16: Ebner T, Moser M, Sommergruber M, Gaiswinkler U, Wiesinger R, Puchner M, et al. Presence of a cytoplasmic halo at zygote stage but not type and extension of the same has a significant influence on preimplantation development and implantation behavior. Hum Reprod 2003; 18: Veeck LL. The morphological assessment of human oocytes and early concepti. In: Keel BA, Webster BW, eds. Handbook of the laboratory diagnosis and treatment of infertility. Boca Raton, FL: CRC, Racowsky C, Jackson KV, Cekleniak NA, Fox JH, Hornstein MD, Ginsburg ES. The number of eight-cell embryos is a key determinant for selecting day 3 or day 5 transfer. Fertil Steril 2000;73: Tao J, Tamis R, Fink K, Williams B, Nelson-White T, Craig R. The neglected morula/compact stage embryo transfer. Hum Reprod 2002; 17: Gardner DK, Lane M, Stevens J, Schlenker T, Schoolcraft WB. Blastocyst score affects implantation and pregnancy outcome: towards a single blastocyst transfer. Fertil Steril 2000;73: Eppig JJ. Intercommunication between mammalian oocytes and companion somatic cells. BioEssays 1991;13: Albertini DF, Combelles CM, Benecchi E, Carabatsos MJ. Cellular basis for paracrine regulation of ovarian follicular development. Reproduction 2001;121: Nakahara K, Saito H, Saito T, Ito M, Ohta N, Sakai N, et al. Incidence of apoptotic bodies in membrana granulosa of the patients participating in an in vitro fertilization program. Fertil Steril 1997;67: Depalo R, Nappi L, Loverro G, Bettocchi S, Caruso ML, Valentini AM, et al. Evidence of apoptosis in human primordial and primary follicles. Hum Reprod 2003;18: Oosterhuis GJE, Michgelsen HW, Lambalk CB, Schoemaker J, Vermes I. Apoptotic cell death in human granulosa-lutein cells: a possible indicator of in vitro fertilization outcome. Fertil Steril 1998;70: Tilly JL. Apoptosis and the ovary: a fashionable trend or food for thought? Fertil Steril 1997;67: Moffatt O, Drury S, Tomlinson M, Afnan M, Sakkas D. The apoptotic profile of human cumulus cells changes with patient age and after exposure to sperm but not in relation to oocyte maturity. Fertil Steril 2002;77: Graham J, Han T, Porter R, Levy M, Stillman R, Tucker J. Day 3 morphology is a poor predictor of blastocyst quality in extended culture. Fertil Steril 2000;74: Milki AA, Hinckley MD, Gebhardt J, Dasig D, Westphal LM, Behr B. Accuracy of day 3 criteria for selecting the best embryo. Fertil Steril 2002;77: Ebner T, Moser M, Sommergruber M, Tews G. Selection based on morphological assessment of oocytes and embryos at different stages of preimplantation development: a review. Hum Reprod Update 2003;3: Fertility and Sterility 633