Medically Assisted Conception: An Agenda for Research : Report of a Study (1989)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

CHAPTER 2 ADVANCES IN 1HE PRACTICE AND SCIENCE BASE OF MEDICALLY ASSISTED CCMCEPTION This chapter is based on papers presented at a workshop en Science Foundations of Efc^v^iiv Assisted Conception sponsored by the Institute of Medicine (ICH) and the Board on Agriculture of the National Research Council. It was held August 21-23, 1988 at the Arnold and Mabel Beckman Center in Irvine, California. For the workshop, the organizing ccnmittee developed a program that explored the recent advances in reproductive and developmental biology that apply to medically assisted conception. This chapter represents a report of science topics selected by the committee. It is therefore not a exupnehensive review of recent advances in reproductive research. The workshop also initiated an interchange of ideas among those involved in patient-related clinical practice, animal IVFET, and those working in basic research as it applies to humans and other animals. Thus, this chapter indicates research areas that promise improvements in the practice of IVFET. First are some developments in human and animal TVFET; subsequently, the processes and recent advances relating to gametogenesis, fertilization, prelmplantation development, and implantation are discussed. This chapter summarizes each talk given at the workshop. The full papers contributed by each author are found in Appendix A. Developments in Human In Vitro Fertilization [1] Eastern Virginia Medical School, one of the premier centers for the clinical practice of in vitro fertilization, has a pregnancy rate of 18.3 percent, based on the number of attempts used to stimulate the ovaries with either follicle-stimulating hormone (FSH) or human menopausal gonadotropin (hfC), and a pregnancy rate of 37.6 percent based on the number of patients. Both patient age and the cause of female infertility have an effect on the outcome of in vitro fertilization. For women infertile because of tubal ligations or endometriosis, the rate of viable pregnancies obtained by XVFET at Eastern Virginia is 16.4 percent. Although the pregnancy rate did not vary according to whether infertility was due to tubal ligation or endometriosis, age has an effect. After the age of 40, it becomes more difficult to stimulate the ovaries to produce mature eggs. Because couples often seek in vitro fertilization to overcome male infertility, the evaluation of semen is important. Currently, this evaluation is descriptive and relatively imprecise. Factors such as number of sperm, sperm motility, and general shape of sperm components(e.g. head shape and tail shape) are important variables in achieving pregnancy. For example, if 14 percent of the sperm are "normal,11 and there are 50,000 sperm per cubic centimeter of ejaculate, in - 25 -

vitro fertilization produces ongoing pregnancies in 30 percent of all attempts. However, even if there are 100,000 sperm per cubic centimeter, but only 4 percent are "normal,1* the ongoing pregnancy rate drops greatly. Abnormal sperm often fail to fertilize an egg because of their inability to penetrate the egg's protective covering, the zona pellucida, therefore workers have tried a technique known as "zona drilling*1 in which holes are produced in the zona pellucida to permit direct access of sperm to the oocyte plasma membrane. Thus far, this technique has not been successful. Another technique, known as zona splitting, has been used at Emory University, Atlanta. By this technique, the zona pellucida is split mechanically. Pregnancies have been reported by use of this method. Basic science questions need to be answered in order to develop better markers for normal and abnormal sperm and to improve the performance of sperm in IVFET. Cryopreservation (freezing) of in vitro fertilized embryos1 is another new and promising technique that will be discussed in greater detail later in this chapter. If embryos are frozen for later use, the stage of their development at the time they are placed in the uterus can be matched with the stage of the uterine wall (endometrium), increasing the likelihood of a successful pregnancy. Before Cryopreservation became an option, embryos had either to be placed in the uterus or disraniRrl. In order to avoid the ethical dilemma of what to do with excess embryos, more than the optimal number of embryos were sometimes transferred into the uterus. By allowing the preservation of embryos for later use, the technique of Cryopreservation reduces not only the chance of multiple pregnancies, but also the number of times a woman's ovaries might be subjected to hormonal stimulation to produce oocytes for additional attempts of IVFET. The use of Cryopreservation has resulted in increased rate of pregnancy at the Eastern Virginia clinic. Nevertheless, basic science research is needed to assess the necessary parameters for successful Cryopreservation and the possible deleterious effects of freezing on the embryo. In cases of ovarian failure, failure of in vitro fertilization, poor quality of eggs, genetic abnormality, or inaccessible ovaries, the only option available to women wanting to bear a child is to use eggs from a donor to perform in vitro fertilization with the husband's sperm. The donated eggs most often come from IVFET patients who have received hormonal stimulation and produced more eggs than necessary for their own use. Sometimes, however, women who are to undergo a tubal ligation agree to ovarian stimulation before the surgery so that eggs can be harvested simultaneously. These eggs are then donated for in vitro fertilization and transfer. Donors are phenotypically matched with recipients and are screened for psychological problems and infectious diseases. Patients who 1 The term embryo in this chapter often refers to in vitro fertilized eggs that developed to two- to four-cell staged embryos. Although there are more technically precise terms for various early stages of development, the more precise terms are used here only when the distinctions are important to the concepts under discussion. - 26 -

conceive through egg donation nay need supplemental hormone therapy to replace the estrogen and progesterone normally released from the ovaries. In such a patient, however, the placenta produces detectable estrogen by week 6 of the pregnancy and progesterone by week 7. Fertilized eggs from oocyte donation, if transferred to the uterus by day 17-19 of the recipient's cycle, resulted in a 36 percent pregnancy rate and 80 percent of those pregnancies going to term. One benefit of oocyte donation, in addition to overcoming the problems mentioned above, is that a tine period of optimum implantation can be achieved such that the stage of the embryo and endometrium are synchronized. This technique has been found to increase the success rate of IVFET. Good quality embryos are critically Important to the suooass of IVFET. Experience at many clinics shows that a number of factors determine the quality of an embryo. Some important factors include the way ovaries are stimulated with hormone supplements and the synchrony between the age of the embryo and the endometrial stage. More research is needed to understand fully these factors and to establish unambiguous and unbiased criteria with which to distinguish poor quality and good quality embryos. Because fertilization is more likely with mature occytes than immature ones, research to learn how to recruit a more synchronous population of follicles is important. Such follicles produce a more synchronous population of occytes capable of responding to the maturational stimuli. Failing the recruitment of mature cocytes, research is needed to develop reliable methods of maturing oocytes in vitro. All of these areas, if improved by increased knowledge of the cell biology of early reproductive events, should greatly increase the ability to identify couples with high probability of success and may increase the success rates of IVFET in human clinical practice. Developments in Assisted Conception in Food-Producing Animals [2] There are critical differences between the goals of assisted conception in animals and human beings. In human beings, the goal is to increase the reproductive ability of those with impaired fertility or, perhaps in the future, to avoid the transmittal of genetic disorders. In contrast, the goal of assisted conception in food-producing animals often is to increase the yield of milk or meat. Because the ethical barriers are lower and the financial stakes higher, a number of advanced technologies are available for use in animals that are not available in the clinical practice of human IVFET. Techniques in commercial or research use include artificial insemination, superovulation, embryo transfer, freezing of embryos, sexing of embryos, multiplication of embryos by bisection and cloning, in vitro fertilization, and the modification of embryos by gene transfer. All western European dairy cows and 60 percent of U.S. dairy cows are impregnated by artificial insemination (AI). AI allows bulls, selected for their transmission of milk producing genes, to achieve up to 500 inseminations from one ejaculate. Thirty years of AI has helped to double the milk production of each cow. - 27 -

Another technique used widely in humans as well as in other animals is superovulation, which involves ovarian stimulation by the administration of hormone supplements. In both animals and humans, supercvulation is unpredictable and often unsuccessful. There has been sane success when cattle are given priming doses of follicle stimulating hormone (FSH) , and when highly purified forms of FSH are used, but the reasons for these effects are unclear. In addition, studies of cattle and other species have found that the oocytes produced by superovulation BonfrtvimRs exhibit abnormal characteristics. Increased understanding of hormonal cycles and the effects of hormones on the cell biology of maturing oocytes would likely have direct application to the solution of some of these problems. The process of embryo transfer in cattle starts with superovulation of a cow, and subsequent mating to a desirable bull. The resultant embryos are collected nonsurgically and transferred to another cow whose estrous cycle has been synchronized with that of the donor cow. Although the success rate with embryo transfer is respectable (about 60 percent of transfers result in pregnancy) , the technique has not produced great increases in milk or meat production. The production of many identical copies of embryos from cattle with highly desirable traits may improve this situation. For this reason, embryo freezing and embryo multiplication procedures are receiving a great deal of attention in the commercial breeding of cattle. Embryo freezing, in concert with embryo transfer, has been generally successful in terms of pregnancy rates. Freezing embryos allows for the storage of rare breeds and preservation of a cattle surplus. However, embryo multiplication has greater potential for the production of large numbers of highly desirable cattle. Two methods of embryo multiplication are used — embryo bisection and nuclear transfer. Embryo bisection is performed at a very early embryonic stage and yields at least two cell masses which are genetically identical. The bisected embryos can then be transferred by normal procedures to a recipient cow. There is a species-specific limit on how many bisections of a given cell mass can be done without compromising viability; with cattle, the maximum yield is four embryos from one. In nuclear transfer, a blastomere (embryonic cell) , or its nucleus, from a valuable embryo is placed into an oocyte from which the nucleus has been removed. The transferred embryo nucleus promotes development of a multicellular mass that can be used to mate a number of copies or clones. There is a great deal of interest in this method both in academic and commercial research. Since these techniques allow more precise selection of desirable traits, both have the potential to effect rapid changes in the prevalence of animals with those traits. While the techniques rtigraraad above are frequently and successfully used, other methods of assisted reproduction are also being investigated. One of the newer approaches is embryo sexing. Sexing of embryos is of particular importance to the dairy industry since only female offspring are needed for milk production. Sexing is done by three methods. The - 28 -

first is karyotyping of embryos by bisecting than and using one half for cytogenetic analysis. The chief disadvantages of this method are the damage that often occurs to the embryo and the variable reliability of the karyotypic identification of sex. The second method employs antibodies to •ale-specific antigens on the embryos to identify the male embryos. The accuracy of this immunological method is greater than that of karyotyping and may hfrrmR widely used, depending on the availability of antisera. Finally, HA hybridization techniques have been y^pt**1 from molecular biology to sex embryos by labeling specific parts of the male Y chromosome obtained by embryo biopsy. There are compelling reasons to develop more fully IVFET procedures in food-producing animals. Such techniques provide a large number of embryos for transfer, research, and embryo multiplication. However, as in human IVFET, the success of in vitro fertilization and embryo transfer in the cattle industry suffers from a lack of basic knowledge of the oetll biology and biochemistry of early gamete maturation and fertilization. Among the areas in which research is lacking is oocyte maturation. As in humans, IVFBT in cattle is more successful if it begins with mature rather than immature oocytes. The biochemistry of sperm capacitation and entry into the eggs also needs to be better understood. A number of chemicals have been identified that are necessary for these processes, but precise identification of the role each chemical plays in fertilization is needed for further progress. Finally, being able to maintain the growth of an embryo in culture for a longer period than is now possible would be helpful. Such an ability would increase the chances of a favorable match between the embryonic and endometrial stages. Gene transfer, if developed further, also shows potential for enhancing the productivity of cattle. With the development of techniques by which genes can be microinjected into embryos, the ability to alter the phenotypic characteristics of food-producing animals has become a possibility. In an early demonstration of transgenic technology, workers injected mouse eggs with the gene for human growth hormone. The resulting mice grew to almost twice the size of normal mice. Such technology could be used in cattle to alter the genes for skeletal muscle in order to produce higher quality meat products, or to alter genes in such a way that biologically active substances would be secreted in milk. There are many areas in which the results of basic research would further the practice of assisted conception and embryo transfer in both human beings and other animals. These include oocyte maturation, sperm morphology, the biochemistry of fertilization, cryopreservation, ovarian stimulation, and molecular genetics. In addition, topics such as membrane biochemistry, hormonal control of testicular and ovarian function, gene expression in early development, and the cell biology of implantation are identified as areas of exploration that would make major contributions to the success of human and animal IVFET. - 29 -

Gametogenesis and Gametes mduction of Gametogenesis and Superovulation [3] Comparative research en ovarian stinulation of non-human primates and domestic animals such as cattle, horses, sheep, and pigs raises interesting issues. Sane work highlights the differences between the normal estrous cycles of animals such as cattle, and the cycles of human and non-human primates. Yet there are problems that are cannon among in the stimulation of follicular development and the induction of mature oocytes. Ovarian stimulation and superovulation are exogenou interventions into a highly regulated physiological process. The normal physiology is not always easily manipulated. Developing strategies to circumvent the differences between normal and induced ovulation is a major goal of research proposed for this area. Non-Hunan Primates Non-human primates exhibit a menstrual cycle that closely approximates that of humans. In monkeys, the hormonal events of the cycle include a geometric increase in serum estradiol, midcycle surges of luteinizing hormone (LH), follicle stimulating hormone (FSH), and a later increase in serum progesterone after IH and FSH decrease in Ttration. These hormone levels are precisely synchronized in the normal cycle and lead to the development of one follicle, the dominant follicle, from which ovulation occurs. One of the purposes of superovulation is to encourage more than one follicle to develop fully, thereby producing multiple oocytes for in vitro fertilization. In monkeys, one of four chemicals is administered in combination with human choricnic gonadotropin to stimulate ovulation. These are human menopausal gonadotropin, human FSH, pregnant mare's serum gonadotropin, or porcine FSH. However, monkeys sometimes produce antibodies that block the activity of hormones obtained from other species activity, and ovarian stimulation fails. Thus, there is a need to develop stores of monkey hormones to be used for ovarian stimulation in this species. The response to the administration of hormones to stimulate ovulation is also not uniform in non-human primates. In general, estradiol levels are higher than in a natural cycle, the IH surge is lacking, IH concentration remains high over a long period, FSH is low, and progesterone remains high in concentration. However, in some subjects there is a premature, spontaneous IH surge that is not synchronized with follicular maturation; this surge stops the follicles and oocytes front maturing. Even if such an IH surge does not occur, there is a wide range of responses to gonadotropin administration. It is impossible to predict the response of any individual animal. The IH surge and the heterogenous responses, however, can often be prevented by additional administration of gonadotropin releasing hormone. Another difficulty often observed with superovulation protocols is an overstimulation of prolactin secretion. These departures from the normal hormonal levels raise important questions relating to the efficacy and unknown effects of ovarian stimulation. - 30 -

Despite the problems of ovarian stimulation in non-human primates, reasonable famous rates have been achieved. Nevertheless, there are additional barriers to progress in this field. The supply of non-human primates is severely limited. In addition, these animals often cannot be used more than once for ovarian stimulation because of the immunological responses to human, equine, and porcine gonadotropins used in the stimulation procedures. Finally, eggs must be recovered from monkeys by either laparotomy or laparoscopy, surgical techniques that are restricted by ethical considerations and formal legal constraints to a low number of repetitions per animal. Although these problems of availability, reuse, and the lack of non-human gonadotropins are more difficult to ow than the biological problems of overriding normal physiology, numerous questions about the basic biological processes remain unanswered. Resolution of the problems associated with the use of non-human primates in this type of research could be facilitated by an increased ability to augment natural hormone and gonadotropin release, by improved resolution of ultrasound imaging to identify mature oocytes and guide their collection by non-surgical means, and by the development of cell lines that could produce larger quantities of non-human primate gonadotropins. Finally, it is important to note that many of the same questions remain unanswered regarding human ovarian stimulation and that particular non-human primate species serve as the best model for human reproductive physiology. Domestic Animals Extensive use of artificial insemination in cattle has increased the genetic contribution of desirable bulls to the overall supply of cattle. However, there is little likelihood that the desirable characteristics of females will be further increased by augmenting the male genetic contributions. Superovulation, on the other hand, combined with embryo freezing and embryo transfer, has the potential to increase directly the gene pool of desirable female traits by increasing the number of offspring from a valuable cow. There are a number of differences between non-human primates and domestic animals in the ways ovarian stimulation is accomplished. Some of these differences arise from differences in the reproductive cycles of the two groups. In dairy and beef cattle, stimulation is accomplished with the administration of either pregnant mare's serum gonadotropin (FMSG) or porcine FSH. The latter is preferred to JMSG because FMSG has a long half-life and often results in asynchronous emulations. Porcine FSH is not without problems either since it is contaminated by variably high levels of luteinizing hormone. Human chorionic gonadotropin, routinely used in non-human primates, is not used in the coranercial practice of ovarian stimulation of cattle. To control the variability and lack of predictability in the estrous cycles of cattle, current protocols often include an injection of a specific type of prostaglandin (PGF-2 alpha) . This treatment causes the regression of the carpus luteum, thereby artificially restarting the cycle. The snrogss rate of superovulation in cattle has been variable. This is caused by a number of factors including seasonal variability, breed differences, dose and timing of gonadotropin administration, and history - 31 -

of previous superovulations. Such differences contribute to difficulty in comparing studies performed on different breeds and, at tines, in comparing studies in the same breeds, but done with varied techniques or in animals of varying age. It has been found that, even if all the factors possible are controlled, there remains a great deal of variability in the success rate of superovulation. A number of strategies to improve the success rate are currently being tried. One is an attempt to purify porcine FSH preparations to exclude the contaminating IH and, further, to determine the FSH to IH ratios that are optimum for successful superovulation. Another strategy is to try to neutralize EMSG with antibodies so that the complications that arise from the long half-life of EMSG can be prevented. Because the lack of normal coordination of endocrine events is a common occurrence in superovulation, investigators are also trying to define the sources of asynchrony and develop ways of normalizing a coordinated series of biological events. Superovulation in domestic animals other than cattle also has had miypri results. In goats, the technique has been tremendously successful. Superovulation of sheep has also been largely successful, and a few innovative strategies have been applied in this species. Sheep are seasonal breeders and attempts have been made, with initial success, to induce ovulation in this species during tunes when they are normally anestrous. In order to bypass the use of exogenous gonadotropins, workers have employed antibodies to inhibin in sheep. Inhibin is a substance made in an ovarian follicle that suppresses FSH secretion. The blocking of inhibin with antibodies raises FSH levels, thereby increasing the natural stimulus for follicular development. Antibodies against steroids have also been tried in sheep to achieve the same end. In horses, superovulation has proven to be difficult, and to require nonstandard protocols. The period of estrous or "heat" in mares is unusually long and variable. In addition, there are piupiujiaimed physiological mechanisms in mares to prevent the development of twins. These two aspects of equine reproductive physiology greatly complicate the successful application of ovarian stimulation techniques. Also of critical importance is the fact that EMSG, even in high doses, does not adequately stimulate the development of follicles in mares. Unlike other animals, where EMSG acts as both FSH and IH receptors, in mares EMSG has only IH activity. Some success has been achieved in mares with administration of porcine FSH, especially if combined with human chorionic gonadotropin. In general, however, progress in successful superovulation in horses lags well behind that in other domestic animals, except for the pig. The most limited of all commercial applications of superovulation in animals has been in the pig. Although possible, the technique does not confer many of the advantages that it holds for other species. One reason is that pigs deliver litters as opposed to one offspring at a time. Therefore, the need to increase the offspring from one individual, while helpful in certain circumstances, is not as compelling in pigs as in other domestic animals. Nevertheless, superovulation and embryo transfer may - 32 -

be able to lower the frequency of ******** by increasing the number of offspring from disease-free or disease-resistant populations. More frequent reasons for the use of embryo transfer in pigs are to obtain disease-free embryos from infected pigs2 and to introduce new genetic material into specific pathogen-free herds. Thus, further research into the normal reproductive physiology of pigs will be useful. To improve IVFET in domestic animals as well as in human and non-human primates, more needs to be known about the normal regulatory events of cvulation, including the physiology and biochemistry of the development of a dominant follicle. Improved ultrasonography has potential for increasing the success of superovulation in all species. As with monkeys, developing ways to bypass the use of exogenous agents like FMSG and porcine FSH would be of enormous use in domestic animals. One way of doing this may be by blocking inhibin as has been done in sheep. In order to improve the success rate of assisted conception in animals, many of these research areas must, nevertheless, be pursued in each species separately with realization of the inter-species differences. It is not always possible to predict when the results obtained from one species can contribute knowledge applicable to other species. Biology and Maturation [4] This section examines three topics relating to the biology and maturation of oocytes: 1) The study of membrane biochemistry, which has relevance to the complex membrane interactions that occur at fertili- zation; 2) the study of molecules that are important in maturation events, is also important to understanding the possible reasons for failure in IVFET; and 3) the biological and physical properties of eggs and embryos that are affected by preservation techniques such as freezing, known as cryopreservation. Membrane Biochemistry The membranes of all cells of the body share certain standard structural characteristics. The basis of a cell membrane is called a bilayer, which is a two-layered collection of phospholipid molecules (lipids containing phosphorus). fhospholipid molecules are polar in that each has one end that is stable in water (hydrophilic), and one end that repels water (hydrophobia). In a bilayer, then, the lipid molecules line up so that the hydrophobia ends meet each other on the inside of the bilayer. The hydrophilic ends are thus oriented on the outside of the bilayer such that one group faces the intracellular space and the other faces the extracellular space. Such an arrangement makes sense when one considers that both the cytoplasm of a cell and the extracellular spaces are oaqpofyri largely of water. It is, in fact, a quite natural response of lipids to form bilayers in an aqueous environment; such a process underlies soap bubble and oil droplet formations. 2 Embryos can be rinsed free of pathogens or treated with enzymes to destroy pathogens making embryos the safest method of moving germ plasm without pathogens around the world. - 33 -

Proteins are part of the membrane also. Some proteins span the thickness of the membrane and act as hydrophilic channels for ions or participate in molecular signal transducticn mechanisms. Other proteins are partly emnerirjeri in the lipid bilayer and partly exposed to the extracellular space. Attached to such proteins are long, often branching, chains of carbohydrate molecules that also extend into the extracellular space. Finally, large extracellular proteins often associate with or stick to the most external parts of the carbohydrate molecules. Thus, the cell membrane, even in its standard form, is a complex structure made up of different types of molecules arranged in regions that each exhibit properties. The possible interactions of membrane molecules are immense since the various molecules are mobile within the bilayer. These interactions underly the equally immense range of biological responses of cells including cell-cell interactions, receptor-mediated signal processing, and transmembrane transport of nutrients or toxins. Reproductive events in which membrane biochemistry is of particular importance include sperm capacitation and fusion of sperm and egg membranes for fertilization. Basic aspects of membrane biochemistry are most often studied using membranes from cells such as red blood cells or membranes manufactured in the laboratory, called liposomes. Such membranes are simpler than the membranes from eggs or sperm and, therefore, easier to study. However, data obtained from these model membranes must be extrapolated to eggs and sperm with great caution. Mammalian Oocyte Mafotration In a normal follicle there is communication between the oocytes and the surrounding nurse cells, called granulosa cells, through specialized junctions (gap junctions) through which small molecules pass from one cell to another. Developing oocytes undergo a type of cell division in which only half the normal number of chranosoroes are retained. This process is known as meiosis and has a number of stages. Until just before maturation, however, the growing oocytes remain in a state of suspended or arrested meiosis. It is thought that the granulosa cells help to maintain this arrest of meiosis by producing substances that enter the cocyte through the gap junctions. It is also possible that substances that induce final maturation of the oocytes are produced in the surrounding cells and are transmitted to the oocytes in the same manner. Since the ability to mature oocytes in vitro would be of enormous value in IVFET, the understanding of the substances that either maintain or abolish meiotic arrest is particularly desirable. It is important to note that many of the studies summarized in this section were done with mouse oocytes and these data were compared to those obtained from other mammals including rabbits and rats. Investigators have also studied frog oocytes, but the mouse is the most common animal model. Eventually, of course, certain critical experiments will have to be replicated using oocytes from human or non-human primates. Several substances are thought to maintain meiotic arrest. One, cyclic adenosine monophosphate (cyclic AMP), is a molecule present ubiquitously in the body. Another, hypoxanthine, is common in the body - 34 -

and belongs to a group of molecules known as purines. Guanosine, another purine, is also a potent inhibitor of neiosis and the purine, adenosine, has been shown to augment the activity of hypoxanthine. By examining the effects of some of these substances on each other, a hypothesis has arisen which suggests that hypoxanthine and other purines may increase the level of cyclic AMP, which then inhibits meiosis. Future research, focused upon the molecules of the cyclic AMP pathway inhibiting meiosis, may elucidate ways in which oocytes can be maintained in meiotic arrest in vitro. While knowledge regarding the inhibition of oocyte maturation is inportant, knowing how to mature oocytes in culture would have many applications that are described later. In vivo, oocyte maturation is produced by the surge of LH immediately prior to emulation. The mechanism of action of IH is unclear, however. One might assume that, if cyclic AMP maintains meiotic arrest, then IH somehow causes a decrease in cyclic AMP in the oocyte or the surrounding nurse cells. It has been shown in vivo that administration of IH in origination with human menopausal gcnadotropin reduces cyclic AMP in the oocyte, but not in the surrounding nurse cells. This effect is puzzling since the gap junctions between the oocytes and the other cells still appear to be functional, so cyclic AMP should pass through these junctions easily. An alternative hypothesis is that IH acts indirectly by causing some maturation-inducing substance to be produced in the surrounding cells that decreases the cyclic AMP in the oocyte. Since an increase in intracellular calcium has been shown to be essential in somtic cell division, it is possible that a maturation-inducing substance operates in the oocytes by causing an increase in intracellular calcium. The search for these substances and their mechanism of action is a major area for future research. The study of oocyte maturation has led to extremely successful methods for maturing mouse oocytes in vitro. Such work has shown that the in vitro culture of oocytes and their surrounding cells can be accomplished in ways that allow subsequent fertilization, implantation, and delivery of normal young to occur. Successful in vitro fertilization of oocytes from other species, however, depends upon continued basic research on oocyte nutrition and metabolism and on the factors or substances that control differentiation and maturation. The potential applications of an increased ability to ma time oocytes in vitro are numerous. As discussed earlier, such an ability would be of enormous value in the production of agriculturally important animals. These techniques also have important applications in attempts to preserve endangered species. For example, it has been demonstrated that oocytes removed from mice up to six hours after death can be matured in vitro (Lazarus effect). Yet another application can be found in basic sci< research into genetic Hiaaagog through the use of transgenic mice. Such mice are produced by introducing known gene fragments into the genome of a mouse. These gene fragments have been inserted by microinjection into eggs. Alternatively, the fragments, incorporated into a harmless virus, have been injected into neonatal female mice. In this way, such fragments are incorporated into the ENA. of developing oocytes in these females. It - 35 -

would be extremely helpful to be able to transfer gene fragments by viral Infection of oocytes, which could then be matured in vitro for subsequent fertilization. Finally, the ability to nature human oocytes nay have some application for human IVFET, since matured oocytes from donors would obviate the need for hormonal stimulation therapy and permit synchronization of the developmental state of the embryo with the state of the uterine lining. One degree to which data obtained from experiments with mouse oocytes can be applied to humans is not known. Therefore, for such applications to be realized in human IVFET, it is probably necessary that basic research be done using oocytes from human ovaries removed for justified clinical reasons and/or oocytes from non-human primates. Cryopreservation of Oocytes and ^bryos Cryobiology is the study of how living tissues can be frozen and stored for later revival. There are a number of uses for cryopreservation of oocytes and embryos in connection with IVFET and other reproductive technologies. Cryopreservation of human and cattle sperm has been a routine and successful procedure for many years. It is thought that, properly stored, frozen embryos can be preserved for many decades before thawing and transfer to the uterine cavity. To date, embryo cryopreservation has been successful in many mammalian species, including mouse, sheep, cow, human, and baboon. Rare mutant animals have also been preserved for later study. The Mouse Embryo Banks at the Jackson Laboratory in Maine contain many examples of cryopreserved mutant mouse embryos. Such mutants are valuable for research into genetic and other diseases. Cryopreservation allows easy transportation of frozen embryos by researchers who move frozen mice embryos from one lab to another, and for the national and international transportation of cattle embryos. If proved safe for use in clinical practice, routine cryopreservation of eggs or embryos could reduce substantially the number of times a woman would have to undergo hormone stimulation, because the eggs harvested could be frozen for later fertilization or fertilized in vitro and the extra embryos stored for future transfer. These stored embryos could be used if the first pregnancy failed or used at a later time so that an infertile couple could have more than one child from one stimnl ation/IVFEr procedure. Thus, such preservation techniques could also minimise the waste of extra eggs and extra embryos produced by superovulation and other IWET procedures. Although the specific methods of cryopreservation vary, the basic steps involved are the same. The first is chemical protection of the tissue against damage from freezing. The tissue is then sometimes cooled in sequential steps to a subzero temperature before being rapidly cooled to -196° C for long-term storage. The procedures used for warming and cooling are variable. After rewarming, the chemical protective solutions are removed by rinsing. Protection from damage by freezing is critical to the survival of an embryo because, without protection, the formation of ice crystals causes tig*"** A*™S*F that can not be repaired. - 36 -

Cryqprotective soluticns, which are made from chemicals such as glycerol or dimethyl sulfcodde, function in a manner similar to antifreeze in a car. They replace the water in the embryo. Because of the importance of this step in cryopreservation, much basic research is being conducted to find appropriate cryoprotective solutions. Certain specific properties of the solutions are important determinants of the survival rate of eggs and embryos. For example, some chemicals, such as sucrose, cannot enter cells and, therefore, do not protect from freezing damage in the same way, or with the same efficiency, as other molecules that can enter cells. Glyoerol, on the other hand, can permeate a cell, partially replace the water, and Impart the cryoprotection necessary. The osmotic characteristics of particular solutions are important also. In the case of glyoerol, if one exposes a cell to a solution of water and glyoerol, glyoerol moves into the rail by osmosis, a process by which molecules in a high concentration will move across a semi-permeable barrier, such as a cell membrane, into a region of low concentration of that molecule. T.ivg&ngg, water moves out of the cell by osmosis into the surrounding fluid. Eventually, the concentration of glyoerol will equilibrate and be the same on both the inside and outside of the cell. Since glyoerol moves into the cell more slowly than water moves out of the cell, the cryoprotection procedure must be carefully controlled so that the cell does not lose water so quickly as to shrink and die. It is especially important to control the movement of molecules across the cell membrane when thawing ova and embryos. After thawing, it is important to remove the glyoerol from the cells and to return the intracellular contents to normal. However, a cell protected with glyoerol, if suddenly exposed to a normal saline solution, will take up a great deal of water very quickly. This influx of water causes the cell to swell and the membrane may burst. Although this problem can be avoided by exposing the cells to solutions of decreasing concentration of glyoerol in a stepwise fashion, this is not always easy to accomplish. Research has shown that adding the impermeable solute, sucrose, to the saline solution results in a more even exchange of glyoerol and water across the membrane and obviates the rapid influx of water into the cell. This one step method has proven to be of particular use in the application of these techniques under field conditions such as in work with endangered species. The method used to freeze and thaw the ova and embryos is also important. Until recently, the tissue was first cooled to a few degrees below zero so that ice formation could be controlled. The tissue was then cooled to some intermediate subzero temperature before being rapidly cooled to the storage temperature of -196° C. Research with mouse ova showed that the cooling procedure used could affect the survival rate in Important ways. For example, it was demonstrated that, if ova lose a considerable amount of their original volume during cooling, the survival rate is greater. Cells that retain most of their original volume usually do not survive. This is probably because a cell is primarily crnfrinnfl of water from which the damaging ice crystals form. The method used to - 37 -

freeze the tissue also determines the subsequent thawing procedure. Thawing is performed stepwise in reverse order to the freezing steps. New freezing protocols have been developed that simplify the multiple step methods. For example, mouse embryos have been successfully frozen in one step after incubation in glycerol and sucrose. This method uses the glycerol to cryoprotect the tissue and the sucrose to dehydrate it before freezing, eliminating the need to slowly cool the tissue before freezing. Variations of this method have also been successful. One novel method is called vitrification, which means freezing without crystal formation. Sonet lines, this can be accomplished by extremely rapid freezing. However, in a variation of this approach, embryos are exposed to a solution of chemicals that form a glassy solid substance when cooled. No ice crystals form in the embryos exposed to this solution so freezing can be rapid and done in one step. These new freezing procedures are still being developed, but they have the potential to increase the success rate of ova and embryo freezing. Improvements in cryobiology would have significant impact. Improved preservation and storage of embryos from valuable cattle and other agriculturally important animals would greatly increase the yield from superovulation and IVFET procedures. The preservation of endangered species and rare genetic breeds of laboratory animals would be enhanced. Finally, successful storage of human embryos that result from IVFET techniques can increase the chances of an infertile couple to have a baby. Intratesticular and Intraovarian Paracrine Control [5] Development of male and female gametes occurs over varied periods of time in special body compartments, the ovaries and the testes. In this section, the biochemical control mechanisms that affect gamete development will be Hisraiggflri. in contrast to hormonal control mechanisms in which bioactive substances are released into the bloodstream and act at distant sites, paracrine control mechanisms involve chemical interactions between neighboring cells. For example, one cell can release an active substance into the extracellular space, the substance then diffuses around other cells in the region and affects the activity of these other cells in some way. These effects most often are mediated by cell surface receptors for the substance released. Control of Tpst^g Function The testes are composed of a number of different cell types, including germ cells in their various developmental stages (spermatogonia, spermatocytes, spermatids, and nature sperm) each of which displays unique characteristics. Other cell types include Sertoli cells, which are nurse cells for the developing spermatogonia and spermatocytes, and Leydig cells, which release the hormone testosterone, required for sperm maturation. Many bioactive hormones and pept ides have also been found in the test is and research has begun to determine the cell uuurces of these chemicals and their function in spermatogenesis. Some of the molecules found in the testis have been shown to affect ™»n cycles and cell differentiation. For example, somatomedin C (also called insulin-like growth factor, IGF) and epidermal - 38 -

growth factor stimulate cells to begin CKA synthesis for mitotic ocTM division. Fibroblast growth factor stimulates quiescent cells to reenter the cell cycle in order to replicate and divide. Various hormones and growth factors act as stimulators to cell differentiation, thus causing the oHIs to exit permanently from the cycle of chromosome replication and mitotic division. Cells are inhibited from exiting the cell cycle by interleukin-1, which is a molecule made by cells of the immune system. The functions of most of the hormones and growth factors in the testis are not known. For a few, however, certain functions are beginning to be described. In the rat, a common animal model for investigations of sperm development, Sertoli cells start to divide and proliferate around the 22nd day of embryonic life, and complete this proliferation by 20 days after birth. Leydig cells, in contrast, do not start to divide until 2 or 3 days after birth and continue to proliferate until 35 days after birth. By the end of these processes, almost equivalent numbers of Sertoli and Leydig cells are found in the testis, despite very different developmental timetables. A hypothesis has been proposed that these cells may regulate each other by complicated interactions during development. Fibroblast growth factor stimulates replication of Sertoli cells by causing an increase in the number of receptors for follicle stimulating hormone (FSH) . As this occurs, the numbers of Sertoli cells increase and release somatomedin C in gr^^t-or and greater amounts. increases the number of receptors for luteinizing hormone (IH) , which causes the Leydig cells to divide and increase in number. The increasing numbers of Leydig cells leads to an increase in beta-endorphin released by the Leydig cells. A negative feedback occurs such that beta-endorphin decreases the proliferation of Sertoli cells. Such paracrine interactions are probably not unusual and may occur between other types of cells in the testis. In addition, non-paracrine mechanisms of coonunication probably occur, including cell to cell adhesion and ocnraunication through gap junctions between cells. Very little is known regarding the mechanisms of interaction between developing sperm cells and their nurse cells, the Sertoli cells. Cell adhesion molecules have been found on spermatccytes that, some suggest, promote continued meiosis of the spermatocytes bound to the Sertoli cells. Experiments done in culture have shown that sperm cells synthesize more ENA and FNA when they are cultured with Sertoli cells (especially if FSH is in the culture medium) , than if the sperm cells are cultured alone, or alone with FSH. Such data suggest that the Sertoli cells are important to sperm nell development. Investigators have looked for proteins that can be related to specific stages of sperm cell development. The stages of sperm cell development are quite complex, but can be generally grouped into four main stages. The primordial germ cells are called spermatogonia, and undergo six mitotic divisions before becoming spermatocytes. Meiosis begins in spermatocytes. Further development forms spermatids, which are immature germ cells in which meiosis is arrested. Spermatogonia, spermatocytes, - 39 -

and early spermatids are in close apposition to Sertoli cells in the long tubular structures of the testis where sperm are formed, called the seminiferous tubules. The seminiferous tubules open into other tubes through which sperm are transported. As spermatids mature and differentiate into early sperm cells, they are released from the Sertoli cells into the seminferous tubule. An amazing feature of the seminiferous tubule in some species is that, in any given segment of the tubule, the germ cells form a similar pattern of stages. While this arrangement argues for some kind of cellular control of developmental stage, it also provides a definable tissue state with which to experiment. Scientists have isolated specific segments of seminiferous tubules and extracted the proteins contained in them. Maps of these proteins have shown that the different stages are associated with different proteins. By isolating some of these proteins and producing monoclonal antibodies to them, it has also been possible to determine which cells contain the protein of interest. One such protein has been found in Sertoli cells, and its synthesis and secretion have been shown to be stage-specific. These experiments have also found that such proteins diffuse into the extracellular spaces of the seminiferous tubules, but because of the slow flow rate within the tubules, remain in spaces near developing sperm cells for quite a long time before being degraded by proteolytic enzymes. Therefore, there seems to be strcirj evidence that Sertoli cell products greatly influence the development of sperm cells. Continued research into the molecular events of male gamete development would fill some of the large gaps in knowledge about the normal development of sperm cells and would further the search for causes and cures of male infertility. Paracrine Control of the Ovary Although it is generally agreed that paracrine control is important to the development of follicles in the ovary, much of the evidence is circumstantial. This state of affairs derives from the nature of autocrine/paracrine regulation. It has proved difficult to manipulate experimentally systems in which the cell of origin of an observed signal is the same as or adjacent to the target cell. Early experiments to examine the importance of intraovarian chemical control involved removing the pituitaries of rats. Without the pituitary hormones (especially LH and FSH), which normally promote follicular development, investigators could then look at the effects of a number of substances on ovarian follicles. Such experiments found that estrogen, which is synthesized and released from cells in the ovary, was nseded for follicular development and that other steroids produced inhibitory effects. Further, it was demonstrated that estrogen could act in concert with or synergistically with IH and FSH to differentiate follicular cells. later experiments used granulosa cells from pig follicles in culture to show various effects of estrogen. Following a brief inhibitory effect, estrogens stimulated progesterone synthesis in the granulosa cells in a manner that was time-dependent and blockable by pharmacological - 40 -

inhibitors to estrogen. Continued research demonstrated that esUujat action was exerted after the generation of cyclic AMP (see earlier section on oocyte maturation for other actions of cyclic AMP) . There are several chemical forms of estrogen that are made by enzyme modification of a basic structure. One type, catecholestrogen, is found in high concentrations in the walls of large ovarian follicles in the pig and probably enhances, by paracrine mechanisms, the effects of FSH in the follicle. This concept arose from culture experiments which found that catecholestrogen with FSH stimulates progesterone synthesis and production of cyclic AMP. However, antiestrogens do not block this effect. Thus, while experiments have shown that estrogen plays an important role in the control of follicular development, its true role in the natural state remains to be Estrogen is only one of the substances under investigation in terms of paracrine regulation in the ovary. Tissue culture experiments have shown a number of peptides and growth factors to cause proliferation of granulosa cells. For some of these factors, it is known that they have effects on follicular cells and that they are present in the ovary in vivo. However, whether or not the factors are necessary for follicular development and normal ovarian physiology is not known. The most research has been done on insulin-like growth factors (IGFs) . IGFs are synthesized by many tissues and have been found to be a powerful local mediator of growth and differentiation. They seem to stimulate all kinds of effects in granulosa cells including increasing enzyme levels, cell numbers, protein synthesis, glucose utilization, and various secretions. Insulin itself and the IGFs seem essential for FSH induction of receptors for IH, for generation of cyclic AMP, and for FSH stimulation of steroids. IGFs also interact with other growth factors in the ovary to increase cell number. Since most of these findings have been from in vitro studies, it is important to demonstrate the presence of IGFs in follicles in vivo. In fact, IGFs have been shown to be present in follicular fluid in in vivo concentrations similar to those that produce effects in culture. Furthermore, the concentration of IGFs increases during follicular growth and under the influence of FSH, IH, and growth hormone. This is important because any substance active in follicular development and local regulation in the ovary would be expected to change in concentration in response to pituitary hormones. Another growth factor found in the ovary is a type of transforming growth factor called TGF-alpha, which is a molecule related to epidermal growth factor. TGF-alpha increases the rate of cell division of ovarian cells in vitro and has a negative effect on the secretion of estrogen front granulosa cells. It also secnc to decrease the effect of FSH on IH receptors, although this effect varies. TGF-alpha has been found in significant concentrations in follicular fluid and is in highest itration in small follicles. Based on these findings, it has been proposed that TGF-alpha promotes cell replication in the follicles and inhibits differentiation. The other type of TGF, TGF-beta, is a molecule very similar to inhibin and Mullerian-inhibiting substance, both of which - 41 -

have effects in the ovary. In general, studies have suggested that TCP-beta augments the actions of FSH to increase estrogen secretion, IH receptor activity, and progesterone synthesis. These effects seem to be dependent on both the concentration of TCP-beta and FSH and seem to vary Fibroblast growth factor is another factor being investigated for its possible role in regulation of ovarian function. This growth factor is also present in many tissues and has been shown to be particularly potent in stimulating the formation of new blood vessels. In vitro, fibroblast growth factor has effects on OR! Is of the corpus luteum and it stimulates granulosa cells to divide. In addition, the factor coomn to support the growth of blood vessels in the corpus luteum. Its effects on synthesis of estrogen and progesterone are probably inhibitory. There are many other psptides that have been found in the ovary, but research on these substances is sparse. Many of them have not even been purified enough to allow analysis of their structure and examination of their effects in vitro or in vivo. This conplex area of research promises to provide useful knowledge about regulation in the ovary and could have far-reaching implications for both IVFET and natural reproductive biology. In order to demonstrate that a particular substance has effects in the ovary, the substance must be shown to be secreted by ovarian cells into the local extracellular space and to produce changes in the activity of neighboring cells. Such minimum criteria have been met for some of the factors discussed above; however, elucidation of the mechanisms involved is critical. Especially important for future research is to determine the biochemical regulation of cell replication and differentiation in the ovary. Also, it is necessary to determine how some of these factors are produced by the cells. For example, are there large precursor molecules produced that are modified inside the cells for specific needs? The interaction of these factors with known hormones is also important and may be of special application in evaluating the effects of superovulation. Strategies to block the activity of these factors could help to determine if any of them are absolutely necessary to normal ovarian function as this is entirely unknown. Finally, it is of utmost importance to relate the findings from in vitro studies to what is actually occurring in the natural state. Assessment of Gametes [6] There are many reasons why IVFET techniques in human clinical practice do not result in ongoing pregnancies one hundred percent of the time. Certainly, some of the failure rate is attributable to such things as the varied skill levels of the practitioners, the physiological problems of the infertile male or female, or the difficulty in synchronizing the stage of the embryo with the stage of the uterine wall. Nevertheless, a significant contribution to the failure of IVFET derives from poorly understood physiological responses of occytes and early - 42 -

embryos. Between 20 and 40 percent of apparently nature eggs fail to be fertilized by IVFET methods. Of the 60 to 80 percent that are fertilized, most of the zygotes produced fail to establish a pregnancy after transfer to the uterus. Such early reproductive failure raises critical questions for scientists interested in both improving IVFET and understanding the normal physiology of early reproduction. Experimental strategies have been developed that are beginning to shed light on some of these problems. One such strategy is to examine the degree of chromosomal abnormality in precvulatoxy oocytes in an attempt to determine whether these abnormalities contribute to either failure to fertilize or failure to develop after fertilization. There has been a variety of studies assessing the number of eggs obtained from women who had undergone ovarian stimulation, that exhibit a particular type of chromosome abnormality called aneuploidy. A normal oocyte contains 23 chromosomes after meiosis, half of the full human complement of 46 chromosomes. Aneuploid oocytes contain less or more than 23 chromosomes. Comparison between the studies is difficult because the results range from 11 to 65 percent of the oocytes being aneuploid. However, the most exhaustive studies have indicated that the best estimates are likely to be between 20 and 25 percent. The interpretation of such studies is also difficult since the oocytes examined were those that failed to be fertilized. Thus, the possibility exists that the chromosome abnormality is simply an indicator of some other developmental problem. In addition, there are no data on the percentages of normal, unstimulated oocytes which are aneuploid, so how much of the observed abnormality is caused by hormonal stimulation is not known. When human oocytes were matured in vitro, one study found only 1.5 percent to exhibit the chromosome abnormality. This suggests that hormonal stimulation may increase the number of aneuploid oocytes. Finally, in a study of 163 women, a very high percentage of abnormal chromosomes was found in oocytes obtained from particular individual women. The possibility exists, therefore, that some women are especially likely to produce oocytes with chromoscme abnormalities. The extent to which chromosomal abnormalities affect fertilization and pre- or early post-implantation failures is important to know. This is especially true when one considers that sometimes aneuploid eggs are fertilized and the resulting embryos develop to term. Also, findings to date suggest that a significant number of embryos produced by IVFET exhibit some type of chrcmosmal abnormality. One way to approach these important questions is to develop a focused, multi-center research program to analyze eggs that either failed to fertilize in vitro, or that were produced in excess of the number needed for IVFET methods. The use of fluorescent stains for chromosomal DMA in such eggs could reveal a variety of abnormalities and answer a variety of questions. Such an approach could determine more precisely the frequency of abnormalities in stimulated cycles. Various stimulation methods could be compared to find ones with the lowest rate of chromosome perturbation. In addition, associations of chromosomal abnormalities with particular patients, or with patients of certain ages, could be made. - 43 -

While the analysis of oocytes and eggs ia important, it is equally necessary to analyze the processes that cause fertilized eggs to develop to varying stages before finally failing to develop. It has been estimated that only 10 to 15 percent of all fertilized eggs develop past a certain stage, and these failures occur at similar frequencies despite differences in ovarian stimulation or in vitro culture conditions. Thus, failures are probably due to processes within the eggs or embryos themselves. In contrast to the analysis of oocytes, analysis of embryos is an extremely difficult endeavor. Between fertilization and implantation, a myriad of cellular and molecular changes take place in an embryo that have to be coordinated precisely and synchronized both spatially and temporally. Research, then, must focus upon individual stages beginning at the earliest stage of fertilization. It has been found that 5 to 10 percent of human eggs that appear to be unfertilized actually do contain one or more sperm in the region just outside the egg membrane when viewed microscopically. At higher magnifications with electron microscopy, such oocytes exhibit an absence of binding between the sperm and the egg. Usually the membranes of the sperm form protrusions, called microvilli, which seem to be important for sperm binding to an egg. Sperm associated with these seemingly unfertilized eggs lack this structural feature. Although the reasons for this failure in sperm/egg binding are unknown, a number of possible explanations are likely. For example, cell surface molecules, which function as cell recognition markers, could be absent. Another possible cause is an abnormal organization of specific proteins, called microfilaments, that form the structural framework necessary for microvilli. This possibility is attractive because eggs have bsen found that lack microvilli except in small restricted regions of their membranes. Since the abnormality has been observed in mature, preovulatory oocytes, it is not likely to be the result of culture conditions. Further, the lack of microvilli is also typical of oocytes from women with a history of failed fertilization by in vitro methods. It may well be necessary to analyze immature but fully grown oocytes from such women to determine if stimulation protocols affect the frequency of this problem. Other research to determine the molecular changes in the egg membrane necessary for fertilization would also be extremely helpful. There are two other conditions associated with early failure at fertilization. In one condition, the CNA of the sperm fails to decondense. Decondensation is a necessary prior condition to the combining of the male and female chromosomes. While it is possible that there is abnormal packaging of the sperm ENA, it is more likely that there is a problem in the egg. Animal studies have suggested that chemicals in the egg cytoplasm are responsible for causing the sperm ENA to decondense. Clearly, more needs to be known about the prooeoc. Another critical stage in fertilization that can go awry is the migration, juxtaposition, and fusion of the male and female pronuclei, each of which contains the respective chromosomal contribution to the - 44 -

resulting embryo. Electron microsoopic analysis has shown that failure here is associated with failure of the membranes between the two pronuclei to dissolve. Therefore, the two pronuclei cannot fuse. Very preliminary data using fluorescent HA stains have shown that, in these cases, HA replication is inccoplete or absent. It is possible that HA replication is required for membrane dissolution, PJHU earch into this problem could pioueeJ in two ways. First, analysis of fertilized eggs that have arrested development at this stage could be done to further knowledge about the mechanisms of such failure. Second, the mouse could provide an animal model with which to study this problem. It has already been demonstrated in the mouse that chemicals, which specifically inhibit HA replication, do not affect the formation of pronuclei, their migration, or their juxtaposition. Such chemicals do, however, prevent the dissolution of membranes between the pronuclei. These results also occur when HA replication is prevented by other means, so probably are not simply a side effect of the inhibiting chemicals. While this discussion has centered upon very early reproductive failures, there remain many other conditions that cause the Aemia* of early embryos. For example, human development is sometimes arrested when the embryo reaches the four-cell stage. This may be due to a failure in the expression of embryonic genes that results in a lack of proteins required for further development. At the 12-16 cell stage, some normally fertilized eggs develop multiple nuclei. Finally, during the first mitotic division of the zygote to form two cells, a chromosome pair sometimes does not separate or disjoin. This results in different numbers of chromosomes in different cell lines of a developing embryo. These are called "mosaics". Research into the causes of these problems can be of benefit to the clinical practice of IVFET as well as to the understanding of human reproduction in general. While animal models, like the mouse, have great usefulness in this research, the use of developmentally arrested zygotes and embryos produced by IVFET is also necessary. Such zygotes and embryos no longer have the potential to develop into a fetus but may, nevertheless, contain information critical to a better understanding of reproductive success and failure. Fertilization [7] Fertilization encompasses sperm and egg maturation as well as the complex cascade of events that occurs once the sperm reaches the egg. Although some aspects of oocyte maturation have been previously discussed, those that directly influence fertilization require further emphasis. The first hurdle for an oocyte is to acquire the ability to continue meiosis. Oocytes from young mice (less than 15 days old) have been shown to be incapable of resuming meiosis if put into tisfaie culture. Therefore, it seems that there is a growth-related biological process that allows for meiotic competence. Following the breakdown of the germinal vesicle, the chromosome pairs in the oocyte separate with emission of the first polar body and arrest at metaphase II; the polar body frequently degenerates. Resumption of meiosis is initiated by fertilization of the egg. - 45 -

Before ovulaticn the oocytes are contained within follicles and, as was previously diBniHRfri, it is thought that the follicular cells inhibit the oocyte from maturing. Many experiments with frog and mouse oocytes have been done to elucidate the molecular mechanisms that regulate maturation. cyclic AMP has been strongly implicated in maintaining meiotic arrest, presumably by activating an enzyme called protein kinase A. Protein kinase A promotes the structural modification of other proteins by adding a phosphorus group; this process is known as protein phosphorylation. Often, proteins are active or inactive biologically, depending upon whether they are phosphorylated or not. In a sense then, modification of proteins by enzymes is one way in which cells turn molecular signals on and off. The proteins that participate in this cascade of reactions associated with oocyte maturation are not known, but are a subject of great research interest. In addition, it must be emphasized that, in addition to cyclic AMP, other types of molecules present in the follicle are likely to participate in maintenance of meiotic arrest. A molecule that is likely to be involved in oocyte maturation has been isolated from frog oocytes and is called maturation promoting factor (MPF). MPF, when injected into immature oocytes of many species, causes chromosome condensation and breakdown of the nuclear membrane. Its biochemical properties are interesting since it behaves like a protein kinase and is only activated when it has been phosphorylated. Oocytes from mice have a similar substance that has not been fully isolated. There have been innovative experiments performed to measure this MPF-like activity. In one experiment, mature mouse oocytes were fused with immature ones. The mixture of the two cytoplasms resulted in the breakdown of the nuclear membrane of the immature cell. Further experiments using similar strategies have shown that the MPF-like factor's biological activity changes according to the developmental stage of the oocyte. It has also been suggested that oocytes that do not develop meiotic competence may contain a substance which blocks the activity of MPF. Greater understanding of oocyte maturation and the molecules that are directly involved in this process might result in improved ability to mature oocytes from humans and other animals in vitro. Research that tests this hypothesis might also further in vitro oocyte maturation. Sperm recovered directly from the testis cannot fertilize eggs; like eggs, these haploid cells must also be matured after formation if fertilization of an ovum is to occur. This maturation takes place during sperm transit through the epididymis, a process that usually takes about 10-14 days. During this time many molecular changes in the sperm, particularly on the surface of the sperm cell, occur. The mechanisms which bring about these surface alterations have not yet been elucidated. Similarly, it is not known which of these modifications are related to the acquisition of fertilizing function by sperm. Identification of the molecules used by sperm for gamete interaction is necessary before the relevant mechanisms can be ascertained. One way that has been used to examine sperm at different stages of maturation in the epididymis is to stain the cells with antibodies that recognize particular sperm antigens. - 46 -

In one case, tan antibody recognizes a sperm protein in the anterior region of the sperm head. The protein recognized by the antibody, termeri M42, is altered subtly during sperm maturation; prior to the protein's alteration, sperm cannot fertilize eggs, whereas after its modification, fertilization can be achieved readily. It is likely that this change in the M42 protein represents only one example of many molecular changes in the sperm during epididymal transit. Continued examination of individual molecules, like M42, may reveal the mechanisms by which the epididymis causes sperm maturation. This information, in turn, could be used productively in two different ways: to promote fertility by learning how to mature sperm in vitro, or to promote cxntraceptive development by preventing sperm maturation specifically. After ejaculation, sperm must undergo a second, and final, maturation phase. This process, termed capacitation, occurs normally in the female reproductive tract, but can be provoked experimentally for many species by incubation in vitro. Like maturation in the epididymis, capacitation is associated with a large number of sperm alterations, the functions of which are largely unknown. But at least two changes in the sperm's surface membrane are now recognized for essentially all mammlian sperm as a function of capacitation: changes in lipid composition and the loss of many surface-associated components. Capacitation is a prerequisite for fertilization, and these alterations apparently permit sperm to penetrate through the physical barriers that surround the egg at the time of fertilization. Clearly, the ability to control this process would assist fertility regulation considerably. In the ideal circumstance, sperm and egg meet in the fallopian tube and fertilization takes place. However, fertilization is an exceedingly complex process in which the sequence of events is of the utmost importance. At the time of fertilization, the egg is surrounded by two kinds of barriers. Enveloping the egg directly is a jelly-like covering, called the zona pellucida, which is surrounded itself by a cellular layer, called the cumulus. The zona plays a major role during fertilization and many of its molecular features have been defined. There are three important proteins in the mouse zona — ZP1, ZP2, and ZP3. Each protein has sugar sidechains attached to it and is, thus, a glycoprotein. The structure of the extra-cellular zona pellucida is of polymers (compounds produced by the combination of two like molecules) of ZP2 and ZP3 formed into long filaments which are cross-linked by ZP1. ZP2 and ZP3 both function as receptors for sperm. ZP3 also induces the acrosome membrane of the sperm to break down. Both ZP2 and ZP3 are chemically modified following fertilization and play important roles in keeping extra sperm from penetrating the egg. The genes coding for mouse ZP2 and ZP3 have been isolated. Each is a single-copy gene in the mouse genome and each is highly conserved among mammalian species. The zona genes are expressed only in occytes where their expression is developmentally restricted to the growth phase of - 47 -

oogenesis that occurs just prior to ovulation. The determination of the structure and the nucleic acid sequence of the ZF2 and ZP3 messenger RNAs has led to the prediction of the amino acid sequence of the ZP2 and ZP3 proteins. Using the cloned genes, the zona proteins can now be expressed by tissue culture cells. By modifying the cloned zona genes, scientists will be able to produce mutant zona proteins for further analysis of how the zona proteins interact with one another as well as the biological function of individual zona proteins. Antibodies have been produced which bind either to ZP2 or to ZF3. After injection into female mice, these antibodies localize to the zonae pellucidae surrounding growing oocytes. The presence of these antibodies causes highly effective, long-term, but ultimately reversible contraception. Recent studies have demonstrated that immunization of female mice with a synthetic ZP3 peptide (based on the ZP3 DNA sequence) induces the formation of circulating anti-ZP3 antibodies that bind to eggs and effectively preclude fertilization. Because the zona genes are highly conserved, the development of this contraceptive strategy using homologous (same species) peptide as a vaccine may have wide spread application. This method of contraception is desirable because it would prevent fertilization, not implantation, and would avoid the dangers and side effects of the hormones administered in birth control pills. Nevertheless, the reversibility of the strategy remains to be rigorously The surface changes that sperm undergo during capacitation permit them to navigate through the cumulus cells. A secretory organelle that is essentially an enzyme bomb located at the apex of the sperm head must not be disrupted, however. The sperm will use this bomb, called the acrosome, to digest a path through the zona. If the acrosane breaks down prematurely, sperm stick to the cumulus cells and cannot reach the zona surface. The sperm's recognition site for the zona is located on its surface membrane and binds to the ZP3 protein of the zona. The ZP3 molecule then triggers sperm to release its acrosonal enzymes. Particular components within the acrosome bind to ZP2 and the acrosomal enzymes, together with the motility of the sperm cell, allow the sperm to penetrate the zona and reach the egg membrane. This process is remarkably similar across mammalian species. It is likely that factors from the egg itself are not involved in fertilization up to this point since these events occur even if the zona contains no egg, a so-called zona ghost. Like the antibodies against zona proteins diBnisBfirl above, antibodies directed against specific sperm proteins can also inhibit fertilization. The sperm protein M42, described earlier, that is altered during epididymal maturation is one of several candidate sperm proteins for this purpose. Sperm incubated with an anti-M42 antibody cannot fertilize eggs in vitro since the sperm are blocked from losing their acrosomes. In a more natural experimental situation, injection of female mice with the same anti-M42 antibody decreases pregnancy levels markedly. Strong evidence suggests that some forms of infertility in women may be caused by the presence of antisperm antibodies. Perhaps the continued examination - 48 -

of this experimental system—the induction of infertility with specific anti-sperm antibodies—will assist in explaining the mechanisms underlying such naturally occurring types of infertility. Once the sperm has penetrated through the zona, it associates with the egg surface rapidly. Membrane fusion between the two rpmat-o* is initiated at a specific location, near the middle of the sperm head, and pmueaJh around the OR!] so that the entire sperm is incorporated into the egg. Once the sperm has fused with the egg, mnll packages called cortical granules, which contain enzymes that modify the zona pellucida, move to the egg membrane, fuse with it, and release their contents into the space between the egg and the zona pellucida. The released enzymes modify the ZP2 protein so that acxosome-reacted sperm can no longer bind to ZF2. Modification of ZP3 also occurs that affects sperm binding and the induction of an acrosome reaction. In brief, it seems as though nature has devised backup systems to ensure that only one sperm enters the Very little is known regarding the precise biochemical events that take place during the cortical granule reaction. In vitro experiments have shown that activated protein kinase C inhibits fertilization by modifying ZP2 in the same way as occurs in the natural state. Protein kinase C also modifies ZP3 but it only prevents ZP3 from inducing a complete acrosome reaction. Sperm binding to ZP3 is unaffected. Other experiments have implicated other enzymes and co-factors (inositol phosphate, G proteins, and phospholipase C), all of which participate in cascades of biochemical reactions. Much research needs to be done before any of the actual biochemical events can be known. For example, it is not known what chemicals are released by the granules. The precise changes in the zona proteins are not clear. Since some of these chemical cascades produce changes in calcium storage and release, the role of calcium is not clear. Knowledge of these events could contribute greatly to the ability to fertilize eggs in vitro. Finally, while the block to polyspermy is taking place, changes occur in the egg that allow the male and female pronuclei to combine. When the sperm first penetrates the egg, its chromosomes are condensed, held together by chromatin-associated proteins called protamines. For the chromosomes of the male to combine with the female, the chromosomes must deoondense. It is thought that a chemical change in certain molecules of the protamines is necessary for the chromatin to deoondense. Such a chemical change can be induced by glutathione, which has been found in high concentration in mature eggs. Furthermore, depending upon the developmental state of the oocyte, glutathione can cause decondensation of oocyte chromatin when applied in vitro. It is possible that the substance that causes sperm chromatin decondensation is contained in the germinal vesicle, since sperm injected into oocytes before germinal vesicle breakdown retain their onrienaeri chromatin. The substance could be a co-factor required for another reaction, could be an activator of some other molecule, could be solely responsible for the effect, or could actually act as an inhibitor of a molecule whose action is to prevent - 49 -

chronatin decondensation. Further information about these pmcccecn may suggest strategies with which IVFET could be enhanced. To summarize, the sequence of events for gamete interaction in leading to fertilization is: 1. The capacitated speno penetrates between the cunulus cells. 2. The acxosane-intact sperm binds to ZP3 in the zona 3. ZP3 triggers the loss of the acrosone (the acxoscne reaction) , causing release of digestive enzymes from the sperm head. 4. One aerosols-reacted sperm binds to ZP2 and penetrates through the zona matrix. 5. The sperm and egg fuse with the eventual formation of two pronuclei and restoration of the diploid state. 6. At the time of sperm fusion with the egg, a reaction is initiated that prevents any other sperm from entering (the block to polyspermy). Preinplantation Development [8] In this section, a variety of events that occur between fertilization and implantation are discussed. These include the special metabolic requirements of early embryos and the first expression of embryonic genes. Finally, the results of experiments involving micrcoanipulation of animal embryos and embryo splitting will be described. The short description that follows reviews the main events of preimplantation development in a general way. It is also offered as an explanation of terms that will be used throughout this section. The major steps in preimplantation development occur in the fallopian tubes. The fertilized egg, or zygote, undergoes mitotic cell divisions called cleavage. The first cleavage results in the formation of two cells, each of which divide in the second cleavage to form four cells, and so on through 8-cell and 16-cell stages until there are about 50 or 60 cells, at which *-i"«» blastulaticn occurs. During the cleavage divisions, the QellB are called blastomeres and all the blastomeres are encapsulated by the zona pellucida. Blastulation begins about the fourth day after fertilization, just as the developing embryo reaches the uterus. At blastulaticn, cavities form in the cell mass and the cells separate into two regions, an inner cell mass, from which the fetus and some extra embryonic membranes will form, and an outer cell layer, called the tzophectoderm. coincident with these changes, the zona pellucida degenerates. The blastocyst, as the embryo is called at this tine, can be considered to be a polarized structure in that the inner onll mass is attached to the trcphectcderm layer at one side of the blastocyst. If cut - 50 -

through the middle, the blastocyst would look lite a ring of nells with a clunp of other cells stuck to one part of the inside of that ring. mis polarization is important for implantation later since the pole of the blastocyst containing the inner cell mass will first touch the lining of the uterus in preparation for implantation in sane species (in others, the opposite is true). Metabolic Substrates and Pathways Metabolism encompasses the processes by which QellB extract energy from their environment and synthesize the necessary building blocks to form large molecules. Thus, metabolism is an exceedingly complex network of biochemical reactions that can be viewed as intersecting sets of interactions. The sets overlap at key points that regulate which set of reactions will be more or less active, depending on the needs of the cell at the time. The regulators at these key points are enzyme proteins. In most cascades of biochemical reactions, one enzyme's activity will be more important than any other's. Such an enzyme is called a rate-limiting enzyme and is often positioned at the key points of overlap between different pathways of metabolism. In organisms dependent upon oxygen, glucose is used as a basic fuel for metabolism. There are three linked systems that subserve glucose metabolism—glycolysis, the citric acid cycle, and the oxidative respiratory chain. In glycolysis, glucose is chemically changed in a series of enzyme-catalyzed reactions to pyruvate. In the process, molecules of adenosine triphosphate (ATP) are produced. ATP is the energy source for vast numbers of biochemical reactions. The citric acid cycle uses pyruvate, formed by glycolysis, as a substrate for another series of reactions to produce carbon dioxide and molecules (NAEH and FAEH) that are required for a particular type of chemical reaction also important in biological systems. Other products of the citric acid cycle are precursors for the synthesis of other molecules. For example, certain amino acids, which are the building blocks for proteins, are formed from the citric acid cycle. Finally, the NADH and FAEH from the citric acid cycle are used in the oxidative respiratory chain to transfer electrons to molecular oxygen and form additional molecules of ATP. Glucose is also necessary in two other systems that synthesize important biological chemicals. Both of these systems intersect with the glycolytic pathway. The first reaction of glycolysis is the conversion of glucose into gluoose-6-phosphate (G-6-P) by the enzyme, hexokinase. The G-6-P produced can be diverted when necessary to enter the pentose-phosphate path or to be used to synthesize glycogen. In the pentose-phosphate path, G-6-P is converted to another kind of sugar by a series of reactions and, in the process, molecules of NADFH are formed. NADfH functions as a hydrogen and electron donor in biosynthethic - 51 -

reactions. In glyoogen synthesis, the G-6-P is converted to glycogen, which is the way cells store glucose for later use. When cells need ATP, for example, glyoogen can be broken down to form G-6-P, which can then continue in the glyoolytic pathway. Eventually, through the citric acid cycle and oxidative respiratory chain, every G-6-P molecule leads to the production of 37 ATP molecules. For many years, glucose metabolism was not studied in eggs and early embryos, because it was generally assumed that results of studies from other cells would also apply to eggs. With the advent of tissue culture, it became apparent that eggs and early embryos required unexpected nutrients in their culture media, which nnrmnl to violate this assumption. Specifically, it was shown that, for 2-cell mouse embryos, glucose alone would not promote survival in culture. Rather, lactate was critical for these embryos. Lactate is formed from pyruvate when there is little oxygen in the tissue. Exercised muscles, for example, produce a great deal of lactate from pyruvate. Lactate can also be converted to pyruvate, which may then be used in the citric acid cycle. Thus, the finding of a lactate requirement of these embryos suggested that certain pathways of metabolism were restricted, especially the glyoolytic pathway. It was later found that maturing occytes and embryos up to the 8-cell stage needed pyruvate and other citric acid cycle precursors to survive. While the exact requirements have been shown to differ from one animal species to another, it is interesting that pyruvate is often used in culture mndia for human and monkey eggs and early embryos despite the lack of experimental data to justify its use. The nutrient studies described above are not the only evidence that metabolism is restricted in oocytes and early embryos. Morphological studies have demonstrated that the intracellular organelles in which the oxidative respiratory chain takes place are not the same in early cells and embryos of 8 or more cells. These organelles are called mitochondria and are small membrane bound packages of enzymes and other molecules. Inside each mitcchondricn is a complex labyrinth of membranes called aristae. It has been known for a long time that the arrangement of the aristae differ among cell types depending on the function of the cell and the metabolic state of the cell. Studies of eggs and early embryos showed a pattern of aristae that included a concentric and transverse arrangement. However, the concentric aristae disappeared by the 8-cell stage in most species. This change in structure correlated well with an observed increase in oxygen consumption by these 8-cell embryos. Other studies have explored which substrates were taken up by cells. These studies also showed that eggs and early embryos took up pyruvate rather than glucose until they reached the 8-cell stage. To assess whether the cells actually used the pyruvate, studies were done to measure the carbon dioxide produced by metabolism of pyruvate compared to glucose. This is done by labeling the pyruvate or glucose with radioactive tags which, if the pyruvate is metabolized to carbon dicodde, results in the formation of radioactively tagged carbon dioxide. As might be expected, most carbon dioxide was produced from breakdown of pyruvate in early (1 to 2 cell) embryos. At 8 cells, the carbon dioxide was mostly from glucose metabolism. - 52 -

While the results of these studies seem clear, it is important to note some possible confounding factors in these experiments. The most important of these is that, since much of the work has been done with cells in culture, the differences between the amount of oxygen available in culture and the amount available in the natural state may have affected the results. The amount of oxygen available can, by itself, cause metabolism to be switched from one pathway to another. It is also important to determine if the enzymes responsible for glycolysis are present, and if glucose is being used in a pathway other than glycolysis. Various studies have compared the metabolic processes at play by using the strategy of starvation and refeeding of cells with the substances of interest. It was found that, if one starves embryos of glucose at any stage, the level of G-6-P decreases dramatically. Upon refeeding with glucose, the G-6-P levels increase in al l stages also. This suggests that of glucose into cells and the first enzymatic reaction by hexokinase to form G-6-P are both normal across different stages of embryonic development. In contrast, the levels of fructose 1,6, biphosphate (an intermediate molecule formed in glycolysis) do not change with starvation of glucose at any embryonic stage. However, the levels of this intermediate increase substantially with refeeding of glucose in late embryos, but not early ones. Such data provide strong evidence that, like all cells, the rate-limiting enzyme for glycolysis (phosphofructokinase) is the one that catalyzes the formation of fructose 1,6, biphosphate. Moreover, the results suggest that this enzyme operates at a very low level in early embryos and does not reach its normal activity level until after the 8-oell stage. A similar starvation/refeeding experiment has suggested that the presence of pyruvate inhibits the formation of fructose 1,6, biphosphate from glucose and that the addition of glucose resulted in no increase in the metabolites of the citric acid cycle, whereas pyruvate did result in an increase in these metabolites at all stages of embryonic development. There is much that should be known regarding the control of glycolysis in oocytes and early embryos. It is important to continue to investigate the rate-limiting enzymes at the intersection points between different metabolic pathways. Since the energy charge of a cell determines to a great extent the pathway chosen, it is also necessary to consider changes in the total energy charge of oocytes as they proceed through maturation and embryonic development. The energy charge can be expressed as the amount of ATP molecules available. Often, it is measured by determining the ratio of ATP to adenosine diphosphate, ADP, which is a lower energy state form. It is well known that if the energy charge of a cell is high, the pentose-phosphate and glycogen synthesis pathways are favored. This is because more energy in the form of ATP molecules is not needed. If there are not enough ATP molecules, the energy level is low and glycolysis, citric acid cycle, and oxidative respiratory chains are the favored routes. It is known that, in the mouse, zygotes and early embryos have a high energy charge that begins to fall as development procseds. Such a finding is consistent with the evidence that these early DS exhibit a restricted metabolism in the glycolytic and citric acid - 53 -

pathways. It is important to know the metabolic requirements of oocytes and early embryos because these requirements are critical to successful maintenance of these cells in culture. Gene Expression in Early Embryonic Development The determination of when embryonic genes begin to be expressed for the synthesis of new proteins is important in the detection of genetic diiwises in embryos. Of the general population, two percent are at risk to produce children with genetically transmitted diiwiRes. For some the risk is as high as one chance in two. With the development of IVFET techniques, it may become possible to fertilize eggs from high risk couples in vitro to allow for early diagnoses of genetic diseases. In theory, one cell from a multicelled embryo could be removed, analyzed for disease and, if found to be affected, a decision could be made not to implant that embryo. As an alternative, naturally fertilized eggs could be removed from the uterus by flushing, biopsied, and then returned via embryo transfer. For such diagnostic procedures to be realized, however, certain prerequisites are necessary. First, there must be a realistic chance of maintaining a pregnancy after embryo transfer. Embryo biopsy may adversely affect the chance of a ssful transfer. Second, the embryo should not be damaged by biopsy. In cattle, embryo biopsy is a relatively ouuuoi technique and has been done with embryos at various stages. The optimum stage for biopsy is, nevertheless, important to determine. For example, techniques using 8-cell embryos involve removing the zona pellucida and removing one cell. An advantage of biopsy at this stage is that the embryo can be transferred to the mother when it is well synchronized with the state of the uterine wall. There are some disadvantages at this stage, however. These include the fact that there is only one cell with which to work, and the fact that embryos lacking zonae can stick together during culture or transfer, and can develop into a mosaic or chimera that, at best, is undesirable and, at worst, is fatal to the embryo. In addition to these problems, research has suggested that zona-free embryos do not exhibit normal cleavage and often exhibit cells that vary greatly in size. In later embryonic stages, it is possible to nick the zona and slice off the 5 to 10 cells that nerniate out the nick. This approach will likely result in a lower implantation rate, because the optimum fci™» for implantation would be lost. However, the embryos could be frozen for later implantation. Clearly, much more needs to be known about these changes and their effects. The third prerequisite for preimplantation diagnoses is to have reliable probes with which to identify the presence of genetic diseases. Ideally these probes would make the identification at the level of the genome or ENA, as is possible with new technologies, such as Polymerase Chain Reaction techniques. In situ hybridization is a technique whereby a strand of ENA is made that matches with the strand of ENA one is looking for. To use in situ hybridization, however, the sequence of the gene in - 54 -

question must be known. Some of the sequences for human genetic are known, especially the sex-linked disuses. The problem, however, is that many in situ probes are linked to the Y sex chromosome. This causes an unacceptable rate of false negative readings. A procedure with greater potential is to look at the defective gene products, like hemoglobin, as an indication of genetic blood diseases, or like hypoxanthine phosphoribosyl transferase (HPRT) as an indication of Lesch-Nyhan disease. The last prerequisite is absolutely critical, and that is that the probes must assess the activity of embryonic genes not maternal genes. Studies in the mouse have suggested that the embryonic genome turns on between the 2- and 4-cell stage. Before this stage, intracellular proteins are determined by FNA which was stored in the occyte. Investigators at the Medical Research Council Unit of Mammalian Development in London have developed a microassay for the presence of HEKT in embryos that, applied in mouse embryos, has begun to answer some of the questions about the time of appearance of embryonic gene activity. HFRT is an enzyme that is lacking in children born with the sex-linked Lesch-Nyhan disease. The hallmark of the disease is self-mutilation and such children also exhibit strange motor behavior, are often mentally retarded, and usually die by the age of 10. Using the microassay for HPRT, workers have found that the enzyme increases in concentration in 8-cell mouse embryos suggesting that the embryonic genes have become active. So, it seems fairly clear that, in the mouse, gene activity begins at this point. However, before preimplantation diftgnmnen can be achieved in humans, one must know the timing of human embryonic gene activity. In Great Britain, research on human embryos is allowed under the control of the Voluntary Licensing Authority. In the Cambridge University clinic, all patients are asked to donate their excess eggs and embryos for research purposes. Of 300 patients, only two refused permission for the research, and these two were still afforded all the clinical services available. Investigators at this clinic have thus conducted a number of studies on human embryos. In a series of experiments, radioactive methionine was added to the culture media of cultured embryos at various stages of development. Methionine is taken up by cells and incorporated into newly synthesized proteins and so is a qualitative measure of protein synthesis. Researchers found new proteins that contained the radioactively tagged methionine were synthesized for the first time between at the 4- and 8-cell stage. In another experiment, when chemicals, which blocked transcription of messenger FNA, were added to the culture medium, they had no effect on unfertilized oocytes or early embryos and development proceeded normally. However, the blockade of FNA, or protein synthesis, from the 4-cell stage caused the embryos to stop developing. While the results suggested that embryonic gene activity, which directs synthesis of new proteins, is important at the 4- to 8-cell transition period, the data are complicated by the fact that embryos have a tendency in vitro to stop development spontaneously at this stage. There is, however, further support for the idea that the embryonic genes are active at this stage. Since earlier experiments had looked at the - 55 -

pattern of proteins preocnt in early and late embryos and found that the pattern changed quite clearly, the protein patterns were measured with and without the blockade of messenger RNA. These results demonstrated that, when FNA synthesis was blocked, the changes in protein pattern characteristic of the 8-oell stage and beyond did not occur, but the pattern remained in the early state. Because the objective of such research was to develop preimplantaticn diagnostic techniques, the presence of HHRT was also assayed from human embryos. Unlike the findings in the mouse, there was a large variation in HPRT levels in human embryos, and the massive rise in HFRT at the 8-oell stage characteristic of the onset of gene activity in the mouse was not seen in the human. The reasons for these differences and for discrepancies between these data and that of others are not clear. It could be that HFKT is synthesized by the human embryo later than in the mouse embryo, or that the HFRT from the maternal genes is being broken down as rapidly as new HFRT is produced so that no net increase can bee seen. It is also possible that there are simply unexplained differences in storage, culture conditions, or other variables. Whatever the reasons, the application of HPRT assays to preimplantation diagnosis of Lesch-Nyhan Syndrome in the human has been disappointing. The situation does point out, however, the dangers of extrapolating from animal studies and underscores the need to conduct some research with human embryos that, otherwise, would be wasted. Regulative Potential of Micromanipulated Embryos Many of the classic studies on embryonic development involved manipulation of certain parts of animal embryos and subsequent analysis of the effects of that manipulation on final development. Such manipulations included removal of certain parts (limb buds, for example, to study regenerative capacity), transplantation or exchange of parts (as in exchanging upper spinal cord regions with lower spinal cord to see if nerve outgrowth would follow the limb buds in the normal way), and grafting of parts of one embryo to another. Many new technologies allow mioTMnanipulation of embryos at the level of cells and, sometimes, even individual molecules. It is clear that the developmental potential of cells gradually narrows as development proceeds and that embryos have remarkable capabilities to readjust to disturbance. Yet little is known about what governs the point at which readjustment can no longer occur because the potential of a given embryonic region has been irreversibly determined. Further, it is expected that that point of determination will vary according to the specific region in question and according to the species involved. It has long been thought that prior to blastulation an embryo is totipotent, meaning that no cell of the developing embryo is committed to any particular developmental fate. In other words, the cells are - 56 -

undifferentiated and uncxxnoitted. However, it is becoming evident that embryos of different species, even though exhibiting the sane morphological state and the sane number of cells, nay differ in the degree of determination and potential of each cell. Such findings call into question the long-held assumptions of totipotentiality. Since an embryonic cell is totipotent only if it can develop into an entire organism, experiments on embryo splitting can answer questions regarding the real potential of embryonic cells at different stages and in different species. If a 2-cell stage embryo of a laboratory mouse is bisected, development proceeds normally. up to the 8-cell stage, individual blastomeres can be dissected from an embryo and aggregated with another embryo successfully. However, such isolated blastomeres cannot develop or organize into viable fetuses on their own. The findings in the rat are similar. In contrast, rabbit blastomeres isolated from 4- and 8-cell stage embryos have been shown to be totipotent. Other experiments have examined the effects of bisecting mouse embryos exactly in half. While 65 percent of half embryos survive from the 2-cell stage, only about 45 percent survive from later stages. The success at later stages may not mean that all the cells are totipotent; rather, it may mean that each half embryo contained the required number and type of cell to complete development. Information on domestic species is derived mostly from experiments with sheep and cattle embryos. In sheep, isolated blastomeres have been shown to develop normally from embryos up to the 8-cell stage. However, for unknown reasons, only some blastomeres are capable of such development. One case was reported in which an 8-cell embryo was divided into four equal parts. Transfer of these split embryos into ewes resulted in the birth of four lambs. Sheep embryos, which have been halved, seem to survive well from the 2-cell to blastocyst stage. In cattle, similar success has been achieved with quartered and halved embryos from the 8-cell stage to the blastocyst stage. Compared to the mouse, sheep and cattle exhibit blastulation at a later cleavage stage. Therefore, the number of cells is higher at blastulation for sheep and cattle. It is possible that a higher total cell number may allow for greater flexibility in embryonic adjustment to manipulations such as splitting. Beyond questions of immediate survival of ndcromanipulated embryos, there are other questions that have been addressed by investigators. Some studies have assessed the size of embryos after splitting and the birth weight of resulting offspring. While one study in mice found no differences in these measurements between control and halved embryos, others have reported differences at various times of development. An examination was made of blastocyst formation following 2-cell stage embryos that had been bisected. One half of the embryo was disnaniRd and the other was cultured, returned to a synchronized female, or placed in an immature oviduct. Those placed into culture conditions exhibited delayed - 57 -

development. A higher percentage of the half embryos developed normally when returned to a synchronized female or an immature oviduct. Such findings indicate that placing the split embryos in vitro somehow compromises their development. However, it is possible that the number of degenerated half embryos were underestimated in the in vivo conditions, since it is more difficult to recover and identify degenerated embryos from the uterus or oviducts. These studies further showed that viability was correlated with cell number such that the lower the cell number achieved in the first few days after bisection, the lower the chances of survival. Many of the half embryos failed after implantation, indicating that the bisection did not result in a failure to implant. One of the most clear suggestions from these studies is that the cell number at blastulation (which occurs at a particular time, independent of the cell number) is critical and sometimes blastulation results in an embryo with a low number of inner cell mass cells. A reduction in the inner cell mass may then cause the A»nHg«> of some half embryos. This idea was tested directly by another group of studies that used differential stains to compare the number of cells in the inner cell mass to the total number of cells. Embryos with a low cell number at blastulation had a lower inner cell mass ratio and exhibited a lower viability than embryos with a higher cell number at blastulation. These experiments further showed that maintaining the half embryos in vivo, even in a nonpregnant uterus, increased the formation of the inner cell mass, resulting in a higher inner cell mass ratio and higher viability than that achieved in vitro. In a study examining the long-term effects of embryo splitting, genetically identical embryos were either bisected and immediately placed into foster mothers or were left intact and placed into foster mothers. The half embryos were generally less successful and those born had a higher rate of neonatal mortality than did the controls. However, some of the neonatal mortality might have been because the half embryos were born from smaller litters than the controls. Small litters usually result in longer pregnancies, which are also associated with higher neonatal mortality rates. Comparison of other phenotypic characteristics of the neonates born revealed no significant differences between half and whole embryos in measures such as the age at eye opening or growth curves. Thus, by birth, some type of regulation has taken place. The studies taken together argue strongly for the existence of crisis points for half embryos including blastulation and implantation. Further study of the mechanisms of regulation in micramanipulated embryos not only can add to our basic knowledge of reproductive biology, but can also elucidate important requirements for successful splitting of embryos in different species. Implantation [9] Implantation, one of the most poorly understood processes in reproductive biology, involves a complex interaction between the embryo and the cells of the uterus. There are also significant species differences in the process of implantation that can be generally divided - 58 -

into three types. The type of implantation found in rabbits is called fusion; whereas the type found in rodents is called displacement. Implantation in humans and other primates is called intrusive and is the most important to understand for the purposes of this report. The general model for implantation in the human begins with the attachment of the blastocyst to the endometrial vail. At this point, the outer layer of trophectoderm of the embryo proliferate and differentiate into two types of trophcblast cells. The cells close to the embryo cytotrophcblasts and the cells in direct contact with the endometrium fuse with each other to henrme the syncytiotropncblast. A syncytium is a large, multinucleated, mass of cytoplasm formed by the fusion of many separate cytotrophoblastic cells into one. The syncytiotrophoblast acts to erode a path into the endometrial tissue that, in turn, allows the embryo to burrow into the wall of the uterus. Some cells of the endometrium are induced to build up large stores of nutrient molecules, which are then released into the extracellular spaces (close to the embryo) as these cells degenerate. The changes in endometrial cells can be visualized with a number of experimental techniques and are called the deridiM] reaction. Eventually, the trophoblastic cells penetrate deeply enough into the uterine wall to contact maternal blood vessels. The interactions between the trophoblast and these blood vessels results in the formation of the vascular supply of the placenta. The study of implantation involves the search for and identification of factors controlling uterine receptivity and maternal recognition of pregnancy. Some of these factors treom to be produced and released by the cells of the blastocyst prior to implantation. The understanding of the biology of implantation has far-reaching implications for human and non-human reproduction. Both in the natural reproductive process and in the practice of medically assisted conception, there is a huge gap between the number of successfully fertilized eggs and the number of offspring born. A large part of this gap can be accounted for by the loss of embryos at or around the point of implantation. For example, in human IVFET practice, 60 to 70 percent of the eggs are fertilized, but only twenty to thirty percent of the embryos placed in the uterus result in an ongoing pregnancy. Moreover, it has been estimated that, in couples without fertility problems and not practicing birth control, a pregnancy occurs in one out of three menstrual cycles in which a fertilized egg is present. In sheep and cattle, twenty to forty percent of the fertilized eggs do not survive and, in the pig, the loss rate is thirty to forty percent. Early embryonic loss is cannon across species. While some early embryonic loss can be explained by factors such as heat stress, nutritional deficiencies, and genetic abnormalities, it has been proposed that much of this loss results from three other possible causes. First, the uterine environment is probably only narrowly permissive to implantation by an embryo. Second, embryos may fail to signal their presence and, consequently, fail to induce the necessary hormonal and uterine changes necessary to maintain pregnancy. Third, embryos may be rejected as a foreign body by the mother's ianune system. - 59 -

The idea that the uterus allows inplantaticn only under precise conditions almost certainly underlies the problems Aianiaa*A elsewhere in this chapter relating to synchrony between the embryonic stage and the uterine stage. Lack of synchrony has been demonstrated to be a problem in all species examined. It has been shown that if precise synchrony cannot be achieved, it is generally better to transfer embryos that are more advanced than the uterine environment. This makes sense if one considers that, following ovulation, the corpus luteum in the ovary begins to produce progesterone, which acts to prepare the uterus for pregnancy.If the corpus luteum degenerates, the drop in progesterone causes the uterine lining to be shed resulting in menstruation. A hormone secreted by the txophectcderm, chorionic gonadotropin, prevents the degeneration of the corpus luteum and, thus, maintains the progesterone secretion. It is reasonable then that an embryo that is transferred to a uterus more advanced than the embryo may not produce enough chorionic gonadotropin to be able to rescue the corpus luteum from degeneration. There have been attempts to control the state of the uterine lining through the administration of exogenous hormones. Attempts to advance the uterine environment of pigs by administration of progesterone have been unsuccessful. However, because pig embryos secrete estrogen which also affects the uterine lining, exogenous estrogen was tried. This approach worked, but only during a narrow time window and, if given too early, was actually toxic to the embryos. Other research suggests that it is too simplistic to assume administration of one type of hormone would be sufficient to control the uterine environment. Analysis of proteins synthesized and secreted by embryos indicates that there are a number of different chemicals made by embryos that can affect the uterus or the corpus luteum. It is nevertheless probable that some of the entoryo-producjed proteins cause changes in the secretions of uterine cells that are necessary and/or supportive of an implanting embryo. A number of proteins are also secreted by endcmetrial cells, and much research has focused upon isolation of these proteins, analysis of their functions, and mapping the changing patterns of protein synthesis associated with implantation. One experimental strategy has been to label implantation sites in mouse uterus with a dye, pontamine blue. This dye, when injected into the veins of a pregnant mouse, causes implantation sites to be colored blue without staining the rest of the endometrium. After this labeling, investigators can then remove the uterus and maintain explants of it in culture. Using precursors for protein synthesis that have been radioactively labeled, investigators can label the new proteins synthesized and compare the pattern of proteins from implantation and non-implantation sites. This strategy has shown that the rate of protein accumulation in implantation sites is up to forty percent greater than in other sites. Much of the increase is accounted for by increases in proteins desH"*** to be secreted from endometrial cells. Further work red the pattern of protein synthesis from natural implantation sites to those that were mechanically induced to look like implantation sites. Although there was an increase in the synthesis of some proteins by the mechanical method, some proteins were not increased. Such a finding - 60 -

argues for there being entoryo-dependent protein synthesis in the enckmetrium at implantation sites. As will be rtiRniiwri in a later part of this chapter, it is known that embryos release certain factors and proteins that directly affect the cells of the endometrium. At this point, however, it is useful to consider some of the proteins found to be released by endometrial cells. Erdoroetrial protein 15 has been isolated in humans. This protein is only present in the secretory phase and first trimester of pregnancy. The presence of this protein has been shown to change depending upon the hormonal state, but so far no function has been ascribed to this protein. Endonetrial protein 14 has been found in human and mouse endcnetria. This protein is identical in structure to an IGF, which is present in numerous sites of the body. In rabbits, a protein called uteroglobin has been shewn to be induced by progesterone and present in early pregnancy or in pseudopregnancy. This protein has a variety of functions including an anti-inflammatory action. Factors isolated from the mouse include epidermal growth factor and a type of colony stimulating factor, both of which increase in response to progesterone. It is surmised that these, and probably other, growth factors may function to control the proliferation of the placenta. Studies of endometrial proteins in the pig are of special interest since this species does not exhibit the intrusive type of implantation. In fact, the maternal and embryonic blood supplies never cone close to each other in the pig as they do in humans and other primates. Such a situation sets up potential problems in bringing nutrients to the developing embryo. Nutrients must be released from the endometrium and diffuse to the embryo. Proteins have been found in the pig that seem to help in this respect. For example, uteroferrin is a protein that carries iron to the embryo. Another type of protein, which increases in response to progesterone, transports water insoluble molecules to the embryo. These are called retinol-binding proteins. Two other proteins — plasmin and trypsin inhibitors — protect the uterine cells from destruction by embryonic enzymes, and lysozyme protects against infection. In summary, many classes of proteins are produced and secreted by endometrial cells in response to estrogen, progesterone, or eridryD-prcduced factors. While some of these proteins serve nutritive functions in species in which the embryo either invades the endcnetrium late or not at all, other functions of these proteins are not clear. It is probable that early embryonic loss is, in some cases, due to abnormal expression of uterine proteins. A poor quality environment for the embryo could result from either excessively low levels of necessary proteins or excessively high levels of proteins, which could be toxic to the embryo. research has been devoted to examining the incidence or mechanisms of immunological rejection of embryos. The uterus is not isolated from the immune system, especially in species in which the maternal and fetal blood are hardly separated. An embryo is like a transplanted organ, which must be transplanted to a site that is protected - 61 -

fron imnune cells or an iiiiuuiiubuppressive drug mist be given. In sane species, an implanting entoryo causes what looks very much like an allergic inflammatory reaction in the endonetrium, that is, dilatation of blood vessels, proliferation of blood capillaries, and fluid accumulation. In addition, it is certain that cell-surface antigens capable of eliciting an innune reaction on the part of the mother are eventually present on enbryos. Thus, the reasons why all embryos are not rejected immunologically are interesting and important areas of research. It has been found that enbryos release substances, especially interferon, that, in other tissues, act to suppress or modulate immune cell activity. Therefore, it is possible that the embryo acts to control locally the innune response of the mother by secretion of immunomodulating substances. However, more research is needed before such a mechanism can be established. The study of implantation in human beings is particularly difficult because there is no in vitro model available. Such a barrier has special importance because of the wide species differences that exist in regard to implantation. Findings from other species cannot be assumed to be true for human beings. However, it is possible to examine some aspects of human implantation by using in vitro methods recently developed. One of these methods involves the isolation of cytotropncblast cells. It has been found that cytotropnoblast cells placed into culture under certain conditions will proliferate and differentiate into syncytiotrophoblast. Such studies have shown conclusively that syncytiotrophoblast is derived from cytotrophoblasts. In a culture dish, this differentiation takes place in two steps. First, the trophoblast cells aggregate. Then they fuse to become a syncytium. It is likely that this process is mediated by cell-adhesion molecules or CAMS, which have important roles in cell-cell interactions during development of the nervous system and other tissues. The CAMs are produced by the trophoblast cells because blockade of proteinsynthesis prevents aggregation of the cells. It is also known that the process depends on calcium, since it does not occur in culture medium lacking calcium. In contrast to the aggregation step, the mechanisms responsible for fusion of the trophoblast are entirely unknown. The interactions of cultured trophoblastic cells with various oorrponents of extracellular matrix have also been studied. Extracellular matrix is simply the intercellular space and its component molecules, which are usually synthesized by the surrounding cells and secreted into the extracellular space. The exact composition of the matrix varies from tissue to tissue, but it is important in the uterus since interactions between it and the blastocyst occur during implantation. For trophoblast aggregation and membrane fusion to occur, there must be serum proteins in the culture media and/or the di«heg must be coated with extracellular matrix proteins. It is thought that the serum is required because it contains matrix proteins, (e.g., fibrcnectin). If these proteins are added, the serum is no longer required. Collagen, fibronectin, and laminin are structural proteins that are plentiful in many extracellular spaces. - 62 -

The importance of extracellular matrix was underscored by additional experiments that used endcraetrial explants to co-culture the trophoblast cells. This method is more like the natural situation, since the trophoblast differentiates into syncytiotrophoblast, as before, but now interacts with endometrial tissue. These experiments showed that trophoblast from first trimester and term placentas bind to epithelial oel Is of secretory endcnetrium. The cells also bind to cut surfaces of the explants, areas where extracellular matrix was exposed. After 24 to 48 hours, a zone of tissue necrosis developed in the endometrial explants where the syncytiotrophoblast associated with the tissue. Moreover, trophoblastic cells bind to and invade nests of cultured endometrial gland cells. The trophoblastic cells dislodge the endometrial cells and penetrate beneath them in a process resembling intrusive implantation. These experiments suggest that the extracellular matrix always permits attachment and differentiation of the syncytiotrophoblast. That fact has important implications in dispnse states where the epithelium is eroded and, especially, when the lining of the fallopian tubes is eroded. Such states may cause implantation to occur in either less desirable sites or in totally undesirable sites such as the fallopian tubes. The mechanisms for trophoblast invasion of the endometrium are not known. There are a variety of proteases (enzymes which break down proteins) that have been implicated, such as plasminogen activator. In some strains of mice, blastocysts are less invasive of the endometrium apparently because they produce less plasminogen activator. Cultured human trophoblasts produce plasminogen activator (urokinase). Urokinase may degrade f ibronectin and activate other enzymes (e.g., oollagenase). It is possible that the action of proteases like urokinase could be controlled by the presence of cell surface receptors for the enzyme, which localize its actions, and by specific plasminogen activator inhibitors,two of which are known to be generated by trophoblastic cells. This control would be important since the invasion of the endometrium must have an endpoint since embryos do not burrow all the way through the endometrial wall. It is expected that many of the biochemical interactions between the trophoblast and the endometrium would occur through paracrine mechanisms. Substances released by the syncytiotrophoblast and cytotrophoblast could reach a high local concentration in areas of the endcnetrium. Substances known to have powerful paracrine effects in other tissues may play a role in implantation including protein and steroid hormones. Substances released from both the trophoblast and the endometrial cells, could account for the inductive and interdependent changes in both these tissues. Further research, however, is clearly necessary to answer these questions. On the basis of the above studies, investigators have proposed a working model for implantation in the human. The first step of this model holds that the trophoblast binds to specific endometrial cell-adhesion molecules. After binding, the trophoblast penetrates the endometrium and - 63 -

attaches to the extracellular matrix by mechanisms involving proteases and controlled by specific inhibitors. This model, then, forms an adequate starting point for further it?aciari:h. Knowledge of the biochemistry is particularly deficient and could be furthered by the application of technologies such as monoclonal antibodies and ENA probes for particular cell—adhesion molecules. MErr£ifirsf<t'iT< of the molecular F**"fo"vi«mg of placenta formation may determine the factors that regulate the proteases. Such findings could be of great use in understanding rtififwre states like tmcpmiii in which abnormal invasion of trophoblastic cells is suspected to be a cause. Finally, investigators see a real need for the institution of a national or international registry of IVFET programs to track the incidence of abnormal implantation and correlate those data with the types of ovarian stimulation protocols used and hormone replacement therapy given. In this way, the disturbances of implantation caused by various aspects of the procedures of IVFET could be separated from those which occur naturally. - 64 -

BUDDIES 1. Based en talk by Dr. Zev Rosenwaks. 2. Sunmazy of talk given by Dr. Neal First. 3. Summary of talks given by Dr. Robert F. Williams and Dr. Joanne Fortune. 4. Suranary of talks given by Dr. Ray Hammerstedt, Dr. John Eppig, and Dr. Stanley Leibo. For additional information on cryobiology, see S.P. Leibo, Cryobiology: Preservation of Mammalian Embryos. In Genetic Engineering of Anjm^^sf j.w. Evans and A. Hollaender (Eds.), Plenum Press, New York, 1985. 5. Summary of talks given by Dr. William W. Wright and Dr. James M. Hammond. 6. Summary of talk by Dr. Jonathan Van Blerkom. 7. Combined summary of talks by Dr. Patricia N. Saling, Dr. Richard Schultz, and Dr. Jurrien Dean. 8. Summaries of talks by Dr. John D. Diggers, Dr. Peter Braude, and Dr. Virginia Papaioannou; summary of talk by Dr. Harry M. Weitlauf, which was given in the fertilization section of the workshop, is combined in this chapter with talks included under implantation. For additional information, see Braude, P.R., Bolton, V.N., and Moore, S., Human gene expression first occurs between the four- and eight-cell stages of pre-implantation development. Nature 332:459-461, 1988; and Braude, P.R., Bolton, V.N., and Johnson, M.H., The use of human pre-embryos in infertility research. In Embryo Research; Yes or No. Ciba Foundation Study Group, G. Bock and M. O'Connor (Eds.), Tavistock Press, United Kingdom, pp. 63-82, 1986. 9. Combined summaries of talks given by Dr. Harry M. Weitlauf, Dr. Jerome Strauss, and Dr. R. Michael Roberts. For additional detail see Weitlauf, H.M. and Suda-Hartman, M., Changes in secreted uterine proteins associated with embryo implantation in the mouse. Journal of Reproduction and Fertility. 84: 539-549, 1988. - 65 -