In vitro maturation of human oocytes for assisted reproduction

MODERN TRENDS Edward E. Wallach, M.D. Associate Editor In vitro maturation of human oocytes for assisted reproduction Marcus W. Jurema, M.D., and Daniela Nogueira, Ph.D. Department of Obstetrics and Gynecology, Division of Reproductive Medicine and Infertility, Women and Infants Hospital, Brown University Medical School, Providence, Rhode Island Objective: To describe and evaluate the current practice of in vitro maturation of oocytes for assisted reproduction. Design: Review of the available and relevant literature regarding in vitro maturation of oocytes. Conclusion(s): In vitro maturation of human oocytes retrieved from antral ovarian follicles is an emerging procedure quickly being incorporated into the realm of assisted reproductive technologies. This new technology has several potential advantages over traditional controlled ovarian hyperstimulation for IVF, such as reduction of costs by minimizing gonadotropin and GnRH analogue use, elimination of ovarian hyperstimulation syndrome, and simplicity of protocol. In vitro maturation of oocytes for assisted reproduction in human beings still is undergoing refinement but currently is providing efficacy and safety outcome comparable to that of traditional IVF in recent selected studies. Implementing in vitro maturation into an established IVF practice is feasible and requires only a few simple adjustments. Crucial to the advancement and optimization of the technology is a better understanding of how to maximize immature oocyte developmental competence and endometrial receptivity. (Fertil Steril 2006;86:1277 91. 2006 by American Society for Reproductive Medicine.) Key Words: In vitro maturation, IVM, in vitro fertilization, IVF, immature oocytes, assisted reproductive technologies In vitro maturation (IVM) of human oocytes, used to assist infertile couples in conceiving a pregnancy, is an emerging technology that has promising potential. Although relatively new in the field of human infertility treatment, IVM technology has been applied successfully in animals and in the livestock industry (1, 2). Early experience with IVM in human beings yielded limited success (3 8) but provided valuable insight into the technology. Subsequently, more recent studies have shown results from IVM that are comparable to those achieved with contemporary IVF cycles (9 12). As a result of the encouraging reports regarding its success, IVM s worldwide clinical use quickly is expanding, whereas its full potential and optimization still are evolving. In vitro maturation has the potential to substitute or at least be an adjuvant to standard IVF protocols for a number of reasons. It does not require the use of large doses of gonadotropins for in vivo follicular growth and oocyte maturation, as currently is practiced with standard controlled ovarian hyperstimulation (COH) for IVF. For patients undergoing IVM, the cost and inconvenience of injectable gonadotropin therapy are avoided, Received December 19, 2005; revised and accepted February 21, 2006. Reprint requests: Marcus W. Jurema, M.D., Women and Infants Hospital, 101 Dudley Street, Providence, Rhode Island 02905 (FAX: 401-453- 7599; E-mail: mjurema@wihri.org). and the risk of ovarian hyperstimulation syndrome (OHSS) virtually is eliminated. In addition, any other potential short- or long-term adverse effects of supraphysiologic concentrations of gonadotropins on sex steroid hormone sensitive tissue (i.e., ovaries, endometrium, and breasts) are minimized. Finally, the need for injections of GnRH analogues is obviated, therefore negating their pituitary and extrapituitary side effects. In vitro maturation provides a patient-friendly approach to assisted reproduction and theoretically could be applied to the general infertility population. It especially is useful for women with polycystic ovaries (PCO), women with previously proven fertility (i.e., after tubal ligation), and couples with a non female infertility issue because they potentially are less likely to tolerate and/or need the superovulation stimulation that is associated with traditional IVF. In this review, the current worldwide clinical experience with IVM will be described. THE CONCEPT OF IVM In 1935, Pincus and Enzmann (13) reported that immature rabbit oocytes removed from their natural ovarian environment were capable of undergoing spontaneous maturation and fertilization in vitro. Similar observations were seen later in human beings by Edwards in 1965 (14). This knowledge was applied 0015-0282/06/$32.00 Fertility and Sterility Vol. 86, No. 5, November 2006 doi:10.1016/j.fertnstert.2006.02.126 Copyright 2006 American Society for Reproductive Medicine, Published by Elsevier Inc. 1277

first in the clinical setting as an effort to rescue and use immature oocytes that were retrieved from stimulated IVF cycles. Veeck et al. (5) allowed immature oocytes that were retrieved from growing follicles exposed to in vivo gonadotropin stimulation to spontaneously mature in the laboratory. It then was shown that these immature oocytes not only were capable of reaching maturation and fertilization in vitro but also of embryonic development and production of live offspring in human beings. However, these rescued immature oocytes frequently are denuded from their cumulus-enclosed environment for evaluation of their maturational status and, as such, are cultured in the absence of surrounding granulosa cells. Studies have shown that this so-called rescue technique produces oocytes with limited fertilization potential and/or embryos with cleavage impairments (15, 16). The suboptimal development of these embryos is not a surprising finding because these immature eggs most likely are dysmature; they have failed to mature in vivo despite exposure to supraphysiologic levels of exogenous gonadotropins. These facts likely account for the high rate of aneuploidy that is seen in embryos that are created from so-called rescue IVM (17 19). Current oocyte IVM technology has moved away from rescue IVM and involves the practice of intentionally retrieving immature oocytes from unstimulated or minimally stimulated small antral ovarian follicles and of culturing these cumulus-enclosed oocytes in appropriate medium. Cha et al. (3) were the first ones to show the success of IVM in human beings using immature oocytes retrieved from antral follicles. Given the improved outcomes, this review will focus on the status of IVM technology practiced in this context. It is important to note that just because retrieved immature oocytes may progress spontaneously to metaphase II (MII) oocytes does not mean that they will achieve oocyte competence. Competent oocyte development requires careful synchronization between nuclear and cytoplasmic maturation. Nuclear maturation consists of germinal vesicle breakdown, resumption of meiosis, extrusion of the first polar body, and arrest at the MII stage. Although these steps are not the sine qua non of oocyte competence, they are essential for fertilization. Unlike nuclear maturation, cytoplasmic maturation is much more difficult to assess microscopically, and its inadequacy tends to present later in development as impaired embryo cleavage or implantation failure (20). Therefore, competent maturation of retrieved immature oocytes requires both nuclear maturation and exposure of these oocytes to the appropriate signals for synchronous cytoplasmic maturation. The ability to obtain a large number of competent oocytes is one of the main goals of assisted reproductive technologies (ARTs) that use extracorporeal fertilization. In standard IVF, exogenous gonadotropins are used to rescue the growth of small follicles that would have been inhibited otherwise by the dominant one. In IVM, the strategy is to retrieve and rescue immature oocytes before they can be adversely affected by the endocrine and paracrine influences of the growing dominant follicle (21, 22). Ideally, retrieval of immature oocytes should occur around the time of selection of the dominant follicle. Studies (2, 23) in the animal model have shown that subordinate oocytes deprived of FSH, either by a short (24 48 hours) exposure to the follicular-dominance phenomenon or withdrawal of exogenous FSH (coasting), undergo morphologic changes similar to those that are observed in preovulatory oocytes after exposure to an LH surge. This process has been termed pseudomaturation because it causes mucification of cumulus granulosa cells and triggers the resumption of meiosis in primary oocytes within small follicles undergoing atresia. It may be during the early stages of this atresia-like process that immature oocytes acquire maturational competence and become suitable for IVM. Therefore, the timing of immature oocyte pickup appears to be a crucial component of the success of IVM. The apparatus necessary for an oocyte to achieve nuclear and cytoplasmic maturation is acquired progressively through its growth phase (24). The human oocyte reaches its mature size (100 120 m) at the antral stage, whereas the follicle itself is only a fraction of its final ovulatory diameter (25). The ability of the oocyte to resume meiosis appears to be acquired when follicular size is only 10% of the presumed ovulatory diameter or approximately 2 mm (26). Therefore, in theory, it appears as though immature oocytes retrieved from very small antral follicles (2 5 mm) already possess the machinery required to undergo full maturation. In practice, it has been shown that the minimum follicular diameter to produce a competent oocyte is around 5 mm (27). Although the actual signal cascade that triggers maturation of these retrieved immature oocytes is unknown, the synchronous activation of nuclear and cytoplasmic maturation for oocyte competence is likely to be highly dependent on the timing and size of the follicle from which the oocytes are retrieved. Much more is known about the mechanisms of animal, compared with human, oocyte maturation in vivo or in vitro. Studies in bovine models have shown that oocytes retrieved in later phases of follicular development have more abundant levels of mitochondrial RNA (mrna) transcripts compared with oocytes from less developed follicles, correlating well with good embryonic developmental quality (28, 29). Maternal antigen that embryos require (Mater), DNA methyltransferase 1 (Dnmt1), and T-cell leukemia/lymphoma 1 (TCL1) are some of the oocyte-derived factors that play an important role in embryonic development, activation, and compaction, respectively (30, 31). Animal data suggest that mrna instability and the absence or abundance of certain transcripts such as LH and FSH receptors, connexin 43, and cyclooxygenase-2 in the cumulus of oocytes from small antral follicles after resumption of meiosis in vitro are predictors of oocyte quality (32 34). The presence of other key components that have been described in rodents, such as growth differentiation factor-9 (GDF-9) and bone morphogenetic protein-15 (BMP-15), mediate important biochemical changes in oocytes required for normal postmeiotic events (30). Identifying mrna expression of such marker 1278 Jurema and Nogueira In vitro maturation of human oocytes Vol. 86, No. 5, November 2006

genes and protein profiles in oocytes before and after IVM can prove very informative with regard to what extent culture conditions induce changes required for favorable oocyte maturation. In vitro maturation requires the maturation of oocytes in the laboratory for 24 hours before insemination. It still is unknown what constitutes a nurturing medium for in vitro oocyte maturation. Many different types of IVM media have been described in the literature, with few studies assessing the contribution or comparing the direct effect of a particular medium on IVM success (4, 35). There are other aspects of IVM that still are in the process of optimization. Among them are patient selection, follicular priming, endometrial preparation and support, method of insemination (IVF vs. intracytoplasmic sperm injection [ICSI]), and timing of ET, all of which will be discussed throughout this review. CLINICAL ASPECTS OF IVM Patient Selection Several publications have described the use of IVM in women with PCO (10, 12, 36 41). These studies have included women with irregular menstrual cycles caused by polycystic ovary syndrome (PCOS) and those with regular cycles but multifollicular ovaries on the basis of ultrasound appearance (PCO). The obvious reasons for choosing women with PCO and PCOS as a target for IVM technology is the greater number of antral follicles available per ovary. Because women with PCO or PCOS are more likely to experience OHSS, they also are the ones most likely to benefit from IVM. In vitro maturation also has been used in women with regular menstrual cycles and normal appearing ovaries. The rationale for using IVM in this group of patients is to attempt a less costly and less demanding treatment for infertility as compared with COH-IVF. For the same reasons, it also appears reasonable to avoid gonadotropin therapy in women whose infertility is based solely on a male factor or in previously fertile women who had a tubal ligation and desire to conceive again. A summary of the current overall worldwide experience is described in Tables 1 (women with PCO and PCOS) and 2 (ovulatory women with normal ovaries). In vitro maturation has been applied, although on a smaller scale, to other specialized groups of patients. One small case series described the successful use of IVM in COH-IVF poor responders (42). In that series, the investigators described the results of eight cycles of IVM performed during treatment with COH-IVF that otherwise would have been canceled because of a poor response. They achieved clinical pregnancy and implantation rates of 37.5% (3/8) and 20% (4/20), respectively. It is important to mention that those patients were on average 30 years of age and that, although one of the pregnant patients had a baseline FSH of 19.1 IU/L, all three pregnant patients had excellent antral follicle count ( 10 follicles sized 8 12 mm) at the time of conversion from IVF to IVM. In vitro maturation has been used in high responders undergoing IVF who were at risk for OHSS (6, 43, 44). In one study (43), 20 cycles of IVF were canceled after an average of 9 days of stimulation (range 5 20 days), and hcg was withheld from patients who had a serum E 2 level ranging from 3,447 to 25,000 pm (939 to 6,812 pg/ml) and had 20 follicles between 11 and 15 mm in diameter. Antral follicles were aspirated to decrease the risk of OHSS without the intention to inseminate them. In these patients, this procedure yielded an average of 8.1 oocytes per retrieval, 18% recovery of oocyte per follicle, and 66% IVM rate within 48 hours. Another study (44) interrupted 56 cycles of COH for IVF because of risk of OHSS and administered hcg when lead follicles were 12 14 mm. The investigators reported retrieving 628 immature oocytes, of which 76% matured in vitro, 82% of those fertilized, and 94% of embryos cleaved, resulting in an overall pregnancy rate (PR) of 46% per transfer. These studies illustrate the potential role of IVM as an alternative to cancellation of IVF in high responders. It is not known how IVM in these circumstances would compare with other options, such as coasting or triggering ovulation with a GnRH agonist. Last, two case reports (45, 46) described successful use of IVM (pregnancies and live births) in couples with severe male-factor infertility that required the use of testicular sperm aspiration or extraction. Taken together, the above mentioned studies illustrate the successful use of IVM in a broad range of infertility patients. However, IVM has not been compared rigorously with traditional IVF, which would be considered the current standard of care for these patients. Until more comparative studies are available, it makes clinical sense to apply IVM to women at risk for side effects from gonadotropins and who have adequate ovarian reserve as assessed by history and/or by standard tests (i.e., antral follicle count, age, basal FSH, and basal inhibin B levels). Follicular Priming One of the major advantages of IVM over standard COH-IVF is that IVM does not require ovarian hyperstimulation. This improves the patient convenience and safety profiles and may decrease the financial burden of infertility treatment with ART. Although gonadotropins, at least in large doses, may not be necessary, the concept of mild ovarian stimulation (so-called follicular priming) in selected patients to increase the efficiency of IVM has been a topic of significant debate. Several studies using IVM in unstimulated patients with regular menstrual cycles showed that IVM can be accomplished without the use of any gonadotropins (47 50). In these studies, the rates of oocyte recovery, maturation, and fertilization were comparable to COH-IVF. However, the resulting IVM embryos appeared to have suboptimal developmental and implantation potential, with PRs of 20% per transfer. Schramm and Bavister (52) showed that mild stimulation with gonadotropins improved the success of IVM in monkeys. Subsequently, Wynn et al. (27) showed the benefit of a short course of FSH in human IVM. Another study has corroborated the findings of improved Fertility and Sterility 1279

1280 Jurema and Nogueira In vitro maturation of human oocytes Vol. 86, No. 5, November 2006 TABLE 1 Clinical outcome of IVM cycles in PCO and PCOS patients. First author and year published (reference citation no.) No. of cycles Priming Average no. of oocytes retrieved % Maturation (duration of culture in h) % Fertilization (type of insemination) % Cleaved embryos Average no. of embryos transferred PR (%) per ET Cha 2000 (41) 94 None 13.6 62.2 (48) 68 (ICSI) 88 4.9 27.1 6.9 20 20 Cha 2005 (51) 203 None 15.5 NA NA NA 5.0 21.9 5.5 24 37 Chian 2000 (38) 13 hcg vs. 7.8 78.2/85.2 (24/48) 90.7 (ICSI) 94.9 2.8 38.5 16.6 3 40 11 none 7.4 4.9 a /68.0 a (24/48) 83.9 (ICSI) 95.7 2.5 27.3 14.8 3 0 Child 2001 (36) 53 (PCO) vs. hcg 10.0 76 (48) 76.3 (ICSI) 94.8 3.3 23.1 8.9 9 40 68 (PCOS) hcg 11.3 77 (48) 79.3 (ICSI) 91.3 3.2 29.9 9.6 10 52.3 Child 2002 (40) 107 hcg 10.3 76 (48) 78 (ICSI) 74 3.2 21.5 9.5 17 26.1 LeDu 2005 (10) 45 hcg 11.4 54.2/63 (24/48) 70.1 (ICSI) 96.3 2.5 22.5 10.9 6 40 Lin 2003 (39) 35 FSH 21.9 43.2/76.5 (24/48) 75.8 (ICSI) 89.4 3.8 31.4 9.7 21 13 hcg vs. 33 hcg 23.1 39.2/71.9 (24/48) 69.5 (ICSI) 88.1 3.8 36.4 11.3 Mikkelsen 2001 (37) 12 None vs. 6.8 44 (24) 69 (ICSI) 64 1.7 0 0 0 0 24 FSH 6.5 59 (24) a 70 (ICSI) 56 1.8 33 a 21.6 3 62.5 Soderstrom-Anttila 2005 (12) IR (%) No. of live births 20 (PCO) vs. None 9.3 54.9 (30 48) 35.0 (IVF, 13) 85.7 1.7 22.2 13.3 2 0 72.4 (ICSI, 7) 61.9 2.0 0 0 0 0 28 (PCOS) None 14.3 58.2 (30 48) 43.8 (IVF, 18) 82.5 1.7 52.9 34.5 6 33.3 78.4 (ICSI, 10) 70.9 1.8 22.2 12.5 1 50 Note: NA not available; PR pregnancy rate; IR implantation rate; SAb spontaneous abortion. a Statistically significant difference compared with the other arm of that study. Jurema. In vitro maturation of human oocytes. Fertil Steril 2006. % SAb

Fertility and Sterility TABLE 2 Clinical outcome of IVM cycles in women with normal ovaries and regular cycles. First author and year published (reference citation no.) No. of cycles Priming Average no. of oocytes retrieved % Maturation (duration of culture in h) % Fertilization (type of insemination) % Cleaved embryos Average no. of embryos transferred PR (%) per ET IR (%) No. of live births % SAb Child 2001 (36) 56 hcg 5.1 78.4 (48) 72.5 (ICSI) 93.1 2.6 4 1.5 1 50 Mikkelsen 1999 (50) 10 None vs. 3.7 76 (36) 62 (ICSI) 54 1.8 33.3 18.8 4 20 10 FSH 3 d 4 85 (36) 65 (ICSI) 62 1.9 22.2 11.8 5 FSH 4.2 71 (48) 61 (ICSI) 48 1.4 20 14.3 1 0 3d,vs. 7 FSH up 2.4 71 (48) 61 (ICSI) 59 1.1 0 0 to6d Mikkelsen 2000 (48) 87 None 6.1 60 (28 36) 77 (ICSI) 87 2.0 17.4 8.8 9 18.9 Mikkelsen 2001 (49) 132 None 3.8 60 (28 36) 73 (ICSI) 87 NA 18 NA 12 20 Soderstrom-Anttila 91 None 6.3 66.9 (30 48) 35.9 (IVF) vs. 84.8 1.4 31 22.6 12 33.3 2005 (12) Yoon 2001 (47) 63 None 9.0 40.7/71.5/74.3 (24/48/56) 100 None 6.5 54.5 (30 48) 67.1 a (ICSI) 85.8 1.5 21 20.0 15 16.7 72.6 (IVF and 89 3.6 17.6 6.5 6 33.3 ICSI) Note: NA not available; PR pregnancy rate; IR implantation rate; SAb spontaneous abortion. a Statistically significant difference compared with the other arm of that study. Jurema. In vitro maturation of human oocytes. Fertil Steril 2006. 1281

oocyte competence from cycles primed with a truncated stimulation of only a few days of FSH followed by a period of coasting (53). Since then, the role of mild gonadotropin stimulation in women undergoing IVM has become the focus of research efforts. In a small, prospective, randomized trial, Mikkelsen et al. (50) investigated whether FSH priming improved IVM outcome in women with regular menstrual cycles, compared with the case of no priming. In this series, FSH (150 IU) was given daily for 3 to 6 days, beginning on the day 3 of the cycle, followed by oocyte retrieval on cycle day 9 10 (24 to 72 hours after a lead follicle of 10 mm in diameter was achieved). Those investigators failed to show any improvement in outcome compared with the case of an unstimulated protocol in these ovulatory women. Cha et al. (41) evaluated the outcome of unstimulated IVM cycles in women with PCO and PCOS. In that report, a PR of 27.1% was achieved; however, the implantation rate was still low (6.9%). Barnes et al. (53) suggested that the developmental competence of oocytes from unstimulated patients with PCO and PCOS, especially in women without pre-ivm P withdrawal bleeding or spontaneous menses, was less than that for women with normal cycles. It therefore was postulated that mild stimulation (follicular priming) for patients with PCO and PCOS after a P-withdrawal bleed may be beneficial. As a result, Suikkari et al. (54), compared the impact of FSH priming in patients with PCO and PCOS with FSH priming in women with regular cycles (both groups began FSH during the preceding late luteal phase) and observed similar number of oocytes retrieved and similar maturation or fertilization rates in both groups. To further analyze whether FSH priming benefits patients with PCO and PCOS, Mikkelsen et al. (37) designed a prospective randomized trial that compared early follicular phase FSH priming (150 IU daily for 3 days) versus no FSH priming. They observed a significant difference in PR that favored priming with FSH (29% vs. 0), along with an implantation rate of 21.6% in women with PCO and PCOS. Taken together, these studies suggest that women with PCO and PCOS may benefit from a short course of FSH priming, more so than women with regular menstrual cycles and normal ovaries. Instead of priming with FSH, Chian et al. (55) began using single injections of 10,000 IU of hcg to prime follicles 36 hours before immature oocyte retrieval and reported the first case series of clinical pregnancies in two patients. Subsequently, in a prospective randomized trial, the same group evaluated the usefulness of hcg priming in 24 patients with PCO and PCOS comparing 13 patients who received hcg with 11 patients who did not (38). The investigators observed a significant increase in the rate of in vitro nuclear maturation (% germinal vesicle breakdown and % MII) that favored the group that was treated with hcg. The remainder of the outcome measures was similar, including PRs (38.5% in the hcg group and 27.5% in the non-hcg group). Other studies by the same investigators have further supported these findings by reporting PRs of 24% and 54% per transfer in hcg-treated IVM cycles in patients with PCO and PCOS (9, 56). Interestingly, in one study, hcg priming in women with regular cycles resulted in lower pregnancy and implantation rates, compared with women with PCO and PCOS, without affecting maturation, fertilization, or embryo-cleavage rates (36). It is possible that such observed differences were a result of the older age of the group of women with regular cycles compared with the case of the younger PCO and PCOS group. These studies suggest that hcg priming may be beneficial in patients with PCO and PCOS. Lin et al. (39) studied the impact of adding FSH stimulation in addition to hcg priming in PCOS patients. In a series of 68 cycles primed with hcg, 35 cycles were pretreated with 75 IU of FSH daily for 6 days, and 33 cycles were not. Their conclusion was that there was no additional benefit of FSH stimulation used concurrently with hcg-primed IVM cycles. One small study comparing FSH priming with hcg priming in PCOS patients revealed superior embryo quality and PRs (33.3% vs. 0) in the hcg-primed group (57). However, the addition of metformin to the PCOS population in conjunction with a truncated course of FSH (150 IU for 3 days) has produced not only excellent but higher PRs (50%), compared with the case of insulin sensitizers alone (26%) or no treatment at all (58). Taken together, these studies suggest an apparent advantage of priming with FSH or hcg over no priming in patients with PCO and PCOS and that the use of insulin-sensitizing agents may improve the effect of FSH priming in this patient group. More studies are needed that compare hcg with FSH priming in women with PCO and PCOS. It is possible that mild stimulation with gonadotropins (hcg or FSH) is necessary to promote the development of oocytes arrested in the androgen-dominant environment of polycystic ovaries. However, a recent study (12) challenged this concept and published acceptable pregnancy and implantation rates in a broad range of unstimulated and unprimed patients. In this study, the group of ovulatory women had a PR per transfer of 31% and an implantation rate of 22.6%. Similarly, despite no stimulation or priming, patients with PCO and PCOS had excellent PRs of 22.2% and 52.9%, respectively, and implantation rates of 15.4% and 34.5%, respectively. Until more studies are available, it still appears that patients with ovulatory dysfunction such as PCOS women may benefit from hcg or FSH priming. Physiologically, FSH priming with or without insulin-sensitizing agents would be a better choice to overcome the androgenic intrafollicular environment that causes ovulatory dysfunction in PCOS patients. Alternatively, oral contraceptives or progestin treatment for withdrawal bleeding before the cycle of IVM may be beneficial in PCOS patients by temporarily suppressing LH secretion. Regarding the use of hcg, it has been shown that bovine cumulus cells from antral follicles as small as 5 mm have mrna transcripts for LH receptors and may respond to hcg stimulation (59). This finding gives more credence to the mechanism by which hcg 1282 Jurema and Nogueira In vitro maturation of human oocytes Vol. 86, No. 5, November 2006

begins the maturation process of small antral oocytes in vivo and facilitates completion of meiosis in vitro. It is, however, unclear whether hastening in vitro nuclear maturation with hcg desynchronizes it from cytoplasmic maturation. The period of final oocyte growth is characterized by both active transcription and translation and, thus, by the accumulation of molecules required for early embryonic cell divisions (60, 61). Oocyte maturation involves a substantial increase in mrna degradation and/or down-regulation, which is essential for embryonic development (62). Isolation of immature oocytes from their antral follicle environment to undergo IVM might disrupt these cytoplasmic processes, possibly compromising mechanisms of activation of the zygotic genome, resulting in the arrest of embryonic development. It has been shown in animals and human beings that protein synthesis in MII IVM oocytes is deficient compared with MII IVF oocytes (35, 63). It is unclear whether this deficiency is a result of a transcriptional inadequacy, and if so, it is unknown whether the addition of a protein source to the medium (i.e., serum) could overcome this potential problem. It also has been shown in vitro that meiosis can be arrested by the addition of germinal vesicle breakdown inhibitors such as phosphodiesterase type 3 inhibitor (64) or 6-diethylaminopurine (65 67) to the medium without interfering with protein synthesis. These agents may delay nuclear maturation but allow for better synchronization of cytoplasmic and nuclear maturation and ultimately improve developmental competence of IVM oocytes. Cycle Monitoring and Outcome Predictors There is no universal or consensus protocol for monitoring IVM cycles. Most published studies include a baseline ultrasound on menstrual cycle day 3 (spontaneous or induced) to rule out the presence of ovarian cysts, dominant follicle, or thickened endometrium. Follow-up ultrasound scans are scheduled according to the individual IVM protocol of each center, and it also depends on the patient population. A repeat ultrasound on cycle day 6 is able to assess the speed in which follicle dominance is taking place. In women with regular cycles, achievement of the desired lead follicle size may be much faster than in women with ovulatory dysfunction; therefore, the timing for subsequent ultrasound follow-up and oocyte retrieval may vary. Besides follicular size, thickness and maturation of the endometrium also are observed and used for determining the timing of oocyte pickup. There are no studies that prospectively evaluate the impact of thickness of endometrium and IVM outcome. Efforts to generate outcome predictors for IVM include only a few retrospective studies. In one of these studies, Mikkelsen et al. (48) observed no differences in number of oocytes aspirated or in maturation, fertilization, cleavage, or pregnancy rates in IVM retrievals occurring when the lead follicle was either 12 mm or 12 mm in women with regular cycles. In the same study, the investigators measured the change in serum E 2 and inhibin A concentrations at baseline compared with time of retrieval. It was concluded that an increase in E 2 of 100% and inhibin A of 80% from baseline to retrieval significantly improved the PR compared with cycles that did not achieve the same increase in serum concentrations of E 2 (19% vs. 0) or inhibin A (24% vs. 0). A follow-up study (49) evaluating the predictive ability of basal levels of E 2 and inhibin A revealed that pregnancy was more likely in cycles starting with a basal E 2 concentration of 200 pmol/l (55 pg/ml) and inhibin A concentration of 10 pg/ml. The investigators did not see an association between PR and basal FSH (up to 15 IU/L) but did notice a clinically but not statistically significant difference in PRs in cycles that started with an antral follicle count of 5 compared with one of 5 (21% vs. 0, respectively). Another group (68) observed an association between basal FSH and number of oocytes retrieved, as well as improved pregnancy chances when basal E 2 was 100 pmol/l (28 pg/ml). Child et al. (40) reported findings of 175 cycles of IVM involving women with normal cycles and women with PCO and PCOS primed with hcg. They found a declining trend in implantation rate with increasing age and a significant improvement in PRs when 6 embryos were produced. Child et al. (69) also found the endometrial thickness at time of ET to be different in pregnancy cycles (10.2 2.0 mm) versus in nonpregnancy cycles (9.4 2.1 mm), whereas there was no difference in thickness of endometrium at time of oocyte retrieval (6.5 1.0 mm vs. 6.6 1.3 mm, respectively). Some investigators have proposed cryopreserving embryos if the endometrial thickness does not reach 7mmonthe day of transfer (10, 38). However, Soderstrom-Anttila et al. (12) observed that the thinnest endometrium at time of retrieval to result in pregnancy was 3 mm. Taken together, studies of IVM outcome predictors suggest that younger patients with adequate ovarian reserve and who produce 6 embryos have a better prognosis. As noted, these studies are limited in number and retrospective in nature, and their results cannot be used to discriminate between poor- and good-prognosis patients accurately. Oocyte Retrieval Commonly, a lead-follicle diameter of between 8 and 12 mm and/or an endometrial thickness of 5 mm are regarded as markers of follicular dominance and are used to trigger immature oocyte retrieval. The underlying principle is to perform the retrieval of the immature oocytes just after induction of atresia but before prolonged exposure to the possibly detrimental endocrine and paracrine effects of the dominant follicle. This concept, although controversial (70), places the ideal timing for retrieval in between the events causing enhancement of oocyte competence (early atresia) and complete atresia. Some investigators have shown a benefit of performing retrieval after the lead follicle reached 10 mm (50), whereas other investigators believe this to be too Fertility and Sterility 1283

late and detrimental and propose canceling the cycle (10, 71). As a result, published studies vary significantly in their ultrasound criteria for oocyte retrieval. Recently, the notion has been challenged that oocyte retrieval for IVM should take place before the development of a dominant follicle. In a bovine study, Chian (72) showed that oocytes retrieved from different phases of the ovarian cycle, from early follicular phase (no dominant follicle) to late follicular phase (dominant follicle present) to luteal phase (corpus luteum present), resulted in no difference in maturation, fertilization, cleavage, or blastulation rates. Those investigators concluded that these immature oocytes exposed to a dominant follicle still possess developmental competence, although the ability to implant or result in a pregnancy or live birth was not tested in that study. Other studies also have suggested that the competence of immature oocytes is independent of the time of retrieval relative to the menstrual cycle (3, 73, 74). In a case series report, Chian et al. (55) described performing natural-cycle IVF of the mature oocyte retrieved in combination with IVM of any immature oocytes retrieved during the same procedure for patients who previously had failed conventional IVF or IUI treatment. In that study, three patients underwent oocyte retrieval 36 hours after hcg of all visible follicles in the presence of a lead follicle of 15 19 mm, and all patients conceived after ET. In one case, only immature oocytes were retrieved and then matured in vitro (the large follicle did not contain an oocyte). The other two pregnancies involved the mixed transfer of embryos that resulted from a combination of in vivo matured and IVM oocytes. From these studies, it is difficult to conclude that the immature oocytes retrieved in this fashion are developmentally competent. The timing of retrieval is dependent upon several factors mentioned. Additional factors, including type of anesthesia, also influence the decision-making process. For example, in centers in which transvaginal oocyte retrievals are performed under local anesthesia (paracervical or paraovarian block) the procedure can be scheduled at a moment s notice because staffing with an anesthesiologist is not required. In this fashion, retrievals can be performed on the day that follicular and endometrial criteria are attained. Centers that are non hospital based and/or that use general anesthesia probably will need 24 hours to arrange coverage by an anesthesiologist. Because most IVF retrievals are scheduled 36 hours in advance, last-minute addition of another retrieval procedure may be disruptive to the operating schedule. This is particularly important because the procedure-room turnover for an IVM retrieval generally takes longer (30 60 minutes) than for IVF (15 30 minutes). In the laboratory, identification and processing of immature oocytes for IVM also takes longer than that for IVF cycles. These differences may become less pronounced with experience. In vitro maturation retrievals best are scheduled early in the morning to allow for insemination the next day during the working hours of the embryologists. Therefore, the incorporation of IVM into an established IVF practice requires some flexibility. The technique of aspirating antral follicles may require certain modifications of the standard IVF retrieval technique. First, special needles have been described for small-follicle aspiration that are more rigid and shorter in total length (to accumulate less aspirate volume) and bevel length (so that the entire bevel is able to fit within the antral follicle) than are IVF-retrieval needles (4). There are no studies in human beings attesting to the superiority of such equipment over standard IVF equipment. A second suggested modification from IVF is the use of lower aspiration-needle suction pressure, in the range of 50 80 mm Hg, compared with 80 100 mm Hg for standard IVF, for fear of denuding the immature oocyte of granulosa cells. Interestingly, IVM retrievals that use standard IVF equipment (consisting of a single-lumen 17-G needle and 80 100 mm Hg of suction pressure) has yielded excellent results (38, 41). The actual technique of aspirating antral follicles differs from that for an IVF aspiration. Unlike an IVF retrieval, the visual collapse of aspirated follicles is not readily seen in IVM retrievals, and one needs to apply multiple passes of the needle throughout the ovary in a back and forth motion while simultaneously rotating the bevel of the needle. The ability to retrieve immature oocytes successfully (especially in non-pcos patients) and the capacity to identify and handle immature oocytes in the laboratory, as with any technical procedure, involves a learning curve. It is suggested that the surgeon gain proficiency in the technique of IVM retrieval from polycystic ovaries before attempting retrieval in a nonpolycystic ovary. Similarly, the embryologist should obtain adequate training, supervision, and experience by performing IVM under the guidance of a person skilled in the technique. Embryo Transfer and Endometrial Preparation Adequate endometrial development is essential in any ART procedure, but even more so in IVM because a dominant follicle or a corpus luteum is not routinely formed, thus possibly compromising both the follicular and luteal sex steroid contribution to the development of the endometrium. Barnes et al. (75) described a successful case in which the dominant follicle was spared at the time of retrieval and the luteal phase was supplemented with hcg. It has also been shown that 4 5 days after retrieval of immature oocytes, another dominant follicle may emerge and adequately supply sex steroids to the endometrium (76); however, this phenomenon would occur only after the ET. Most clinical protocols today rely on the use of exogenous E and P supplementation. Russell et al. (7) compared early versus mid-follicular phase E supplementation therapy and found better results with a midfollicular start, showing that the initiation of E 2 too early in the cycle may be detrimental to the process. However, Trounson et al. (77) did not observe any improve- 1284 Jurema and Nogueira In vitro maturation of human oocytes Vol. 86, No. 5, November 2006

ment in outcome using E supplementation. None of these studies used hcg for priming, which may, in and of itself, improve endometrial receptivity. Indeed, high PRs have been obtained with hcg-primed cycles when E therapy was begun at the time of retrieval and the E 2 dose then was titrated on the basis of the endometrial thickness (10, 38). The standard protocol in our program involves the administration of 6 mg of daily oral E in divided doses starting on the day of retrieval followed by P support (600 mg/d of micronized P per vagina in divided doses) starting the day after retrieval. Because general anesthesia is provided for all our IVM retrievals, the first dose of E is administered vaginally immediately at the conclusion of the retrieval to avoid aggravation of postanesthesia nausea or emesis. Commonly, the transfer of IVM embryos takes place at the cleaved stage (2 3 days after insemination). It may be preferable to transfer IVM embryos at the blastocyst stage to study embryonic competence and development. However, few studies have described successful blastocyst transfer in IVM (75, 78, 79). It has been suggested that oocytes reaching maturation within 24 hours of culture have a better chance of reaching the blastocyst stage and developmental ability (79). It would be counterproductive, however, to attempt blastocyst transfer of IVM embryos in centers that have higher PRs using cleaved IVF embryos as opposed to blastocysts. Therefore, we would suggest that each center perform IVM ETs on the basis of the same criteria currently used for their IVF embryos. The use of assisted hatching has been described and applied successfully in IVM embryos (75). Frozen ETs using IVM oocytes have also been performed successfully (80 82). At this time, the numbers are too low to accurately predict outcomes. Clinical Outcome of IVM Table 1 summarizes selected studies published to date containing pregnancy outcome and live-birth data resulting from fresh IVM-ET cycles in women with PCO and PCOS. Table 2 summarizes the same information in women with normal menstrual cycles. Even though these studies represent a heterogeneous group of experiments, several observations can be made about the overall clinical experience with IVM. As expected, the average number of oocytes retrieved from women with PCO and PCOS is higher than that obtained from women with normal ovaries (12.1 vs. 5.1, respectively). There is a suggestion of a trend toward a higher rate of spontaneous abortions in the PCO and PCOS group that may be related to the diagnosis of PCOS rather than to the IVM technique itself. Last, it is noteworthy that excellent pregnancy and implantation rates can be achieved when transferring an average of 2 embryos in both women with PCO and PCOS and women with normal ovaries (12). LABORATORY ASPECTS OF IVM Identification and Handling of Retrieved Immature Oocytes Laboratory manipulation of immature oocytes for IVM is more time consuming and differs technically compared with traditional IVF. The follicular fluid retrieved in IVM procedures is smaller in volume and contains a greater proportion of whole blood products as a result of ovarian trauma during aspiration. The addition of heparin (2 5 IU/mL) to the collection HEPES-buffered medium minimizes blood clotting. In addition, the aspirates can be filtered through a cell strainer device made of a nylon mesh of 70- m pores (BD Falcon, Franklin Lakes, NJ) to facilitate identification of the oocytes (83). Special attention should be given to ph and temperature settings during the multiple manipulation steps of IVM to ensure appropriate protein synthesis and maintain stable meiotic spindle morphology (84 86). Alterations in these parameters in the bovine (86) and porcine (87) models have shown to negatively impact the capacity of the IVM oocyte to fertilize and develop. Once clean follicular fluid is isolated, the identification of immature oocytes must take into account that the morphology of oocytes retrieved from small antral follicles ( 12 mm) is different from that seen in mature oocytes retrieved from fully grown follicles in standard IVF. Although immature oocytes often are found enclosed within a fully compacted mass of surrounding granulosa cells, the degree of expansion of the cumulus cells can be influenced by the size of the follicle as well as by the amount and duration of exposure to gonadotropins and/or hcg in vivo (Fig. 1). Immature oocytes within a compacted cumulus pattern are harder to identify than are oocytes within a fully expanded cumulus pattern. Expanded cumulus-oocyte complex patterns have a higher expression of LH-receptor mrna, are associated with more efficient oocyte maturation, and result in higher rates of embryo blastulation than compacted patterns (88). In Vitro Maturation of Oocytes Immediately after retrieval, cumulus-enclosed immature oocytes are placed in a specialized IVM medium for 24 48 hours. Son et al. (79) showed that germinal vesicle stage oocytes that matured within 30 hours of culture are developmentally more competent than are oocytes necessitating longer time to mature. Different types of commercially available maturation media have been used in clinical IVM studies. Maturation media usually is supplemented with recombinant FSH and hcg. Patient-inactivated serum or follicular fluid has been added successfully to the maturation media as an exogenous protein source in several studies (12, 38, 40, 48, 71, 80, 89). The likely presence of additional compounds in the serum and follicular fluid such as growth factors, lipids, glycoproteins, steroid hormones, cytokines, and other factors that may be involved in the regulation of oocyte maturation accounts for its preference over synthetic serum substitute or human albumin. After IVM, mature oocytes Fertility and Sterility 1285

FIGURE 1 Morphological classification of cumulus-enclosed oocytes retrieved from small antral follicles. Germinal vesicle stage oocytes are enclosed in 3 (A) to 10 layers (B) of tightly compacted corona cells. (C) Oocytes enclosed in layers of compacted proximal granulosa cells and expanded distal granulosa cells. (D) Oocytes enclosed in expanded cumulus cells (similar to IVF-collected oocytes). (E) Atretic oocytes can be retrieved within fully enclosed cumulus corona cell layers; or partially denuded from cumulus corona cells (as shown here); or completely naked, with or without a degenerative ooplasmic aspect. All panels, scale bar 50 m. Jurema. In vitro maturation of human oocytes. Fertil Steril 2006. may be transported to traditional IVF media for insemination and embryo culture. Insemination of In Vitro Matured Oocytes In the majority of IVM studies, ICSI has been the preferred method of insemination since oocytes are frequently denuded of granulosa cells for evaluation of maturational status. Historically, ICSI has been used to increase the chances of fertilization whether or not a male factor has been detected as a result of the theoretical concern of zona pellucida hardening during the IVM process. A study comparing ICSI versus IVF for the insemination of IVM oocytes reported that ICSI resulted in a higher rate of fertilization than IVF (90). However, the developmental potential of the fertilized egg was similar irrespective of the insemination method. A more recent study (12) reported a higher fertilization rate with ICSI compared with IVF but higher pregnancy and implantation rates with IVF compared with ICSI. Therefore, it is not clear whether ICSI is definitely beneficial or absolutely necessary to effectively inseminate IVM oocytes in the absence of impaired sperm parameters. Further studies still are needed to determine the optimal technique for insemination of IVM oocytes. FOLLOW-UP OF CHILDREN BORN FROM IVM To date, it is estimated that approximately 400 children have been born from IVM technology. Combining all published and unpublished data, there may be information available on 300 of these children (11). The type of published information available is unfortunately scattered, incomplete, and limited by the young age of IVM technology. Most studies have published information regarding obstetric outcome, perinatal outcome, and incidence of congenital malformations such as birth defects or abnormal karyotypes, whereas only one study (91) elaborated on neurological development of children born after IVM (Table 3). Buckett et al. (92) reported an overall healthy obstetric, neonatal, and infant outcome of 48 IVM pregnancies and deliveries, with a multiple-pregnancy rate comparable to their concurrent IVF results. In a more recent update of their data (93), the investigators compared the outcome of 55 IVM cycles with 217 IVF and 169 ICSI cycles that took place during the same period of time (1997 through 2004). The investigators found no differences in perinatal morbidity, birth weight, gestational age at delivery, APGAR scores, umbilical cord ph, or congenital abnormalities among the three groups. 1286 Jurema and Nogueira In vitro maturation of human oocytes Vol. 86, No. 5, November 2006

Fertility and Sterility 1287 TABLE 3 Obstetric, perinatal, and developmental outcome of IVM children. Author/Time Period Children born Single Multiple Cha 2000 (41) 1995 1998 Cha 2005 (51) 1995 2001 Mikkelsen 2005 (11) 1999 2004 Buckett 2004 (91) 1998 2003 Suikkari 2005 (90) 2000 2004 Le Du 2005 (10) 2002 2003 20 (10 boys, 10 girls) 28 (15 boys, 13 girls) 47 (26 boys, 21 girls) Gestational age at delivery Birth Weight Congenital defects at birth 17 3 twins All 37 wks Mean 3.0 Kg None. One child later developed small accessory tissue in left ear 20 8 twins 38.4 weeks for singleton 34.6 weeks for twins 48 27 9 twins 1 triplet 45 2 twins One at 32 weeks, PEC One at 34 weeks, twins Others: median of 40 weeks Median 38 weeks: 8 37 weeks 2 34 weeks Mean 3.2 Kg singleton Mean 2.4 Kg twins Mean 3.7 Kg Mean 3.49 Kg singleton Mean 2.44 Kg twins/triplets 45 39 3 twins Mean 40 weeks Mean 3.5 Kg singletons Mean 2.6 Kg twins 1 hydrops fetalis 1 omphalocele 1 cleft palate 1 soft palate 1 still birth at 42.3 weeks 1 VSD 1 hip dislocation n/a Pregnancy screening and follow up studies Triple screen normal at 16 18 weeks in all pregnancies Omphalocele twin was 45X/46XY by amniocentesis 10 14 week nuchal translucency: all essentially normal 18 22 week U/S in 35 pts: all normal CVS/Amniocentesis karyotype: 11/12 normal. Abnormal one was inherited from father. Cord blood karyotype: 21/21 normal Exam at 24 hours, 1 year, and 5 years: all healthy. Neurologic assessment: At 6 months: 3 minor delays At 12 months: 7 minor, 1 major (glioma found) At 2 years: mean MDI 104 9 ( 85 is adequate) 5 5 None n/a n/a none Oldest child is 1 year: healthy exam. 2 year follow-up planned. Note: PEC Pre-eclampsia; CVS Chorionic Villus Sampling; MDI Mental Developmental Index. Jurema. In vitro maturation of human oocytes. Fertil Steril 2006.