Immature germ cell conception—in vitro germ cell manipulation

BaillieÁre's Clinical Endocrinology and Metabolism Vol. 14, No. 3, pp. 437±452, 2000

doi:10.1053/beem.2000.0090, available online at http://www.idealibrary.com on

8 Immature germ cell conception ± in vitro germ cell manipulation Jan Tesarik

MD, PhD

Laboratoire d'Eylau, 55 rue Saint Didier, 75116 Paris, France Director MAR&Gen, Molecular Assisted Reproduction and Genetics, Gracia 36, 18002 Granada, Spain

Carmen Mendoza

PhD

MAR&Gen, Molecular Assisted Reproduction and Genetics, Gracia 36, 18002 Granada, Spain Professor Department of Biochemistry and Molecular Biology, University of Granada Faculty of Sciences, Campus Universitario Fuentenueva, Granada, Spain

Ermanno Greco

MD

Director Centre of Reproductive Medicine, European Hospital, via Portuense 700, Rome, Italy

Experimental studies in laboratory animals have shown that successful conception can be achieved by fertilizing oocytes with immature male germ cells. This gave rise to the concept that immature germ cells recovered from the testes of azoospermic men with maturation arrest may be used for assisted reproduction. However, in contrast to using germ cells recovered from healthy animals, clinical application to the treatment of male sterility is burdened by inherent defects in germ cells attributable to underlying testicular pathology. The recent introduction of in vitro germ cell culture/manipulation techniques makes it possible, in some cases, to overcome the in vivo maturation arrest by allowing an additional meiotic and post-meiotic di€erentiation and the selective harvesting of cells devoid of apoptosis-related nuclear and cytoplasmic damage. These techniques enabled the ®rst births of normal infants fathered by azoospermic men with maturation arrest at the primary spermatocyte stage and improved the ecacy of assisted reproduction in men with maturation arrest at the round spermatid stage. Key words: germ cell; assisted reproduction; spermatid; secondary spermatocyte; primary spermatocyte; non-obstructive azoospermia; maturation arrest.

The introduction of micromanipulation-based techniques for human in vitro fertilization, and particularly the success of intracytoplasmic sperm injection1,2, has made fertilization independent of sperm motility as well as of the presence of all sperm components that ensure binding to the zona pellucida, zona penetration and sperm ± oocyte membrane fusion during natural fertilization. This had led to questions concerning the feasibility of micromanipulation-assisted fertilization with immature 1521±690X/00/140437‡00 $35.00/00

c 2000 Harcourt Publishers Ltd. *

438 J. Tesarik et al

germ cells and especially with round and elongated spermatids that have already achieved meiosis.3 After the ®rst report, in 1994, of the fertilization of mammalian (mouse) oocytes with round spermatids, leading to the birth of normal o€spring4, and the ®rst demonstration, in 1995, of the successful fertilization of a human oocyte with a round spermatid5, pilot clinical studies followed rapidly. In the same year, the ®rst pregnancy after the fertilization of human oocytes with elongated spermatids6 and the ®rst births after the fertilization of human oocytes with round spermatids7 were reported. In spite of the enthusiasm created by these early reports, immature germ cell conception still remains a challenging technical achievement rather than a practical solution to be routinely applied in clinical practice. One reason for this was the limitation of the technique to post-meiotic maturation arrest, which, although well documented by several authors8,9, is a relatively infrequent condition in the wide spectrum of maturation arrest in human males.10 The diculty in distinguishing round spermatids from somatic cells, experienced by many workers, is another factor contributing to failures of the technique. Finally, spermatogenic arrest is a consequence of di€erent pathological conditions, some of which apparently cause irreparable damage to nuclear DNA or to vital cytoplasmic components that are involved in the triggering of early post-fertilization events in the oocytes.11 Solutions to the problems of immature germ cell conception are, however, beginning to emerge. This chapter reviews the development of novel methods of germ cell in vitro culture and manipulation, analyses the causes of immature germ cell conception failure, discusses ways in which the existing problems may be overcome and suggests practical strategies for the management of infertility caused by male germ cell maturation arrest. IMMATURE GERM CELL CONCEPTION IN ANIMAL MODELS The earliest attempts at fertilizing a mammalian oocyte with a round spermatid worked with the mouse model and used electrofusion12 or direct intracytoplasmic microinjection13 to introduce the spermatid nucleus into the oocyte. The former technique was subsequently used in mouse experiments that resulted in the ®rst birth of normal young after oocyte fertilization with round spermatids.14 The ecacy of this technique was relatively low as only 11% of the embryos transferred to foster mothers developed to term. However, a further development of the technique, including the replacement of electrofusion with a direct intracytoplasmic microinjection15, led to a substantial improvement of the success rate (Table 1). Electrofusion was thus abandoned in this application, all subsequent studies using intracytoplasmic microinjection for spermatid nucleus transfer into the oocyte. This technique was subsequently used for spermatid Table 1. Success rates of conception with immature male germ cells in animal models. Germ cell stage

Species

Technique

Birth rate per embryo

Reference

Round spermatid Round spermatid Round spermatid Spermatocyte II Spermatocyte I

Mouse Mouse Rabbit Mouse Mouse

Electrofusion Microinjection Microinjection Microinjection Microinjection

4/36 (11.1%) 35/131 (26.7%) 14/150 (9.3%) 14/29 (48.2%) 5/333 (1.5%)

14 15 16 17 18, 19

Immature germ cell conception 439

conception in the rabbit16, also leading to the birth of normal o€spring, albeit with a slightly lower ecacy compared with the mouse model (Table 1). Later studies aimed to obtain fertilization by using earlier stages of spermatogenesis for fertilization. These studies were based on the rationale that the oocyte might not detect any di€erence between its own nucleus and a nucleus transferred from a male germ cell when the latter was introduced into it. In fact, the nuclei of mouse secondary spermatocytes were driven to metaphase after their injection into metaphase II mouse oocytes and, after subsequent oocyte activation, underwent the second meiotic division within the oocyte cytoplasm, resulting in spermatocyte nucleus haploidization and the separation of a haploid set of chromatids into a structure resembling the polar body (pseudopolar body).17 Surprisingly, the ecacy of this technique was not inferior to that using round spermatids for conception (Table 1). In contrast, attempts at fertilizing mouse oocytes with primary spermatocytes18,19 were burdened with serious problems, and their ecacy was low (Table 1). The premature separation of primary spermatocyte sister chromatids, leading to the development of chromosomal abnormalities in the resulting embryos, appeared to be the main cause of the poor pre-implantation development and the low birth rate after oocyte fertilization with primary spermatocytes.20 CLINICAL STUDIES WITH HUMAN IMMATURE GERM CELL CONCEPTION BEFORE THE INTRODUCTION OF IN VITRO GERM CELL MANIPULATION Indications The clinical application of immature germ cell conception before the introduction of in vitro germ cell manipulation was limited to cases in which post-meiotic germ cells could be obtained. This applied particularly to cases in which spermatogenesis was arrested at the round spermatid stage or at di€erent stages of spermatid elongation. A recent study shows that, among cases of post-meiotic maturation arrest, blocking usually occurs at the early round spermatid or a late elongated spermatid stage, whereas a block at intermediate stages of spermatid elongation is relatively rare.21 In addition to the con®rmed complete post-meiotic arrest of germ cell di€erentiation, spermatid conception was also applied in cases of non-obstructive azoospermia alternating with severe oligozoospermia in which ejaculated spermatids were used without an attempt to recover more mature forms by testicular biopsy.7,22,23 It is possible, and even probable, that late elongated spermatids and/or spermatozoa could be recovered from the testis in these cases by testicular biopsy, which, however, was not possible, either because of the patient's refusal23 or for other reasons.7,22 Interestingly, spermatids recovered from patients with these relatively less severe disorders of spermatogenesis appear to have a better developmental potential compared with cases of complete spermiogenesis failure because most of the term pregnancies resulting from round spermatid conception reported in the literature were achieved in such cases.7,22,24,27 Ecacy In spite of the encouraging results of the early reports7,22,25, the overall sucess rate, including the results of subsequent studies6,7,21±36, is disappointingly low for round spermatids but acceptable for elongated spermatids (Table 2). This poor ecacy of

440 J. Tesarik et al Table 2. Success rates of human round spermatid injection (ROSI) and elongated spermatid injection (ELSI) before the introduction of in vitro germ cell manipulation.1 Technique

Fertilization rate

Pregnancy rate

Birth rate

ROSI ELSI

571/1875 (30.5%) 370/638 (58.0%)

13/255 (5.1%) 35/97 (36.1%)

5/255 (2.0%) 23/97 (23.7%)

1

Data published from 1995 until 1998 have been taken up from a previously published review.34 More recent published data, compiled from original reports by Ghazzawi et al.35, Sousa et al21 and Gianaroli et al36, as well as our own unpublished data, have been added.

round spermatid conception appears to be partly due to a switch of indication with respect to the original reports. In early studies7,22, spermatid conception was improvised after an unexpected lack of spermatozoa in the ejaculate from men previously shown to have severe oligozoospermia. Later studies deliberately used spermatids from patients in whom the completion of spermatogenesis to spermatozoa had never been documented. These cases have a particularly poor prognosis, and only one term pregnancy was achieved before the introduction of in vitro germ cell manipulation.24 Safety All births resulting from studies summarised in Table 2 gave rise to normal infants. A recent study, however, has reported two cases of congenital malformation after the use of late elongated spermatids for conception.37 In one of these cases, pregnancy was terminated at week 20 of gestation because of a hydrocephalus. The histopathological diagnosis was hydrocephalus, spina bi®da and diaphagmocele, a cytogenetic analysis showing a male with trisomy 9 (47,XY, ‡ 9) of all the amniotic cells. The other case was a boy (46,XY) born at term (2800 g) who had an open lumbosacral myelomeningocele (Arnold ± Chiari syndrome type II), which had not been detected during the pregnancy.37 Even though there is no proven link between these malformations and the use of immature germ cells for assisted conception, and the number of pregnancies and births is too low for statistical analysis, the possibility of an increased risk of chromosomal aberration and birth defect after the use of these techniques must be seriously considered and patients counselled accordingly. MAIN OBSTACLES TO IMMATURE GERM CELL CONCEPTION BEFORE THE INTRODUCTION OF IN VITRO GERM CELL MANIPULATION Limitation to post-meiotic maturation arrest Before the introduction of in vitro germ cell manipulation, only haploid germ cells (spermatids) could be used for assisted reproduction because the use of pre-meiotic and meiotic cells would lead to the introduction of an extra chromosomal material to the oocyte and thus aneuploidy. Spermatogenic arrest at the spermatid stage has been described in certain cases of diabetes.38 This might also be of genetic origin because it was found in patients showing an absence or severe reduction of the expression of testicular cAMP response element modulator protein in the germ line.39 In most cases,

Immature germ cell conception 441

however, the spermatidic arrest is idiopathic. Within the whole population of azoospermic men with maturation arrest, it represents only a relatively small subpopulation, while the arrest is most frequently observed at the end of the ®rst meiotic division (at the primary spermatocyte stage).10 The extension of therapeutic indications for immature germ cell conception to earlier stages of spermatogenesis was thus of great clinical interest. Germ cell apoptosis Programmed cell death (apoptosis) is one of the most important obstacles to the successful use of immature human germ cells for conception. Apoptosis plays an essential role in balancing the Sertoli-to-germ cell ratio during testicular development40 and is involved in the control of cell quality by eliminating abnormal and damaged cells in the adult testis.41,42 It is thus not suprising that a very high frequency of apoptosis, sometimes approaching 100%, was detected among round spermatids from men with a complete arrest of spermatogenesis at the round spermatid stage.43 These ®ndings are compatible with experimental data in transgenic mice with a targeted disruption of the Pp1cg gene, representing an animal model for human spermatogenic arrest.44 Because an apoptotic spermatid cannot be distinguished from a healthy spermatid during the micromanipulation for assisted fertilization, assisted conception attempts with spermatids from men with a complete arrest of post-meiotic development run a high risk of the inadvertent use of apoptotic spermatids. This may lead to fertilization failure or the later developmental arrest of defective embryos.11 A recent study showed that most human embryos developing from oocytes fertilized with round spermatids become arrested later during cleavage, only 8.5% of them developing to the blastocyst stage compared with 46% blastocysts from oocytes fertilized with testicular spermatozoa.45 Nuclear and cytoplasmic immaturity Contrary to earlier fears46, the nuclear and cytoplasmic immaturity of post-meiotic germ cells does not appear to compromise fertilization and embryonic development.15 Most of the cytoplasmic components of the male gamete that develop after meiosis serve to ensure sperm movement and a proper interaction with the oocyte and its vestments. The most important cytoplasmic activity is that responsible for the triggering of the oocyte-activating signal transduction cascade, which is set in motion by a massive release of calcium ions from the oocyte's internal stores.47 The technique of micromanipulation for spermatid intra-ooplasmic injection can induce an in¯ux of calcium from the culture medium, which can mimic the e€ects of endogenous calcium release in the oocyte during normal fertilization.48 Inadequate oocyte activation after intracytoplasmic injection can also be compensated for by producing an arti®cial calcium in¯ux into the injected oocytes with the use of calcium ionophores49,50, and this has been successfully applied in human spermatid conception.24 The complex rearrangement of chromatin structure and nuclear proteins occurring during sperm head condensation in spermiogenesis51 serves principally to achieve a maximal reduction of the nuclear volume in order to facilitate sperm entry into the oocyte. However, these rearrangements, involving a replacement of the basic nuclear proteins histones with sperm-speci®c protamines, also lead to a stabilization of DNA and its increased resistance to modi®cation and degradation.52 The former is important

442 J. Tesarik et al

to predict the rapidity with which spermatid chromatin will react to oocyte cytoplasmic factors after intra-ooplasmic injection. Unlike the chromatin of mature human spermatozoa, which may remain in the original highly condensed state for many hours after injection into metaphase II oocytes that failed to be activated53, the nuclei of round spermatids deposited in the cytoplasm of non-activated human oocytes are rapidly driven to a metaphase con®guration because of the action of the oocyte's metaphase-promoting factor.54 Such premature chromosome condensation is usually disastrous for further development. THEORETICAL AND EXPERIMENTAL BACKGROUND FOR THE CLINICAL APPLICATION OF IN VITRO GERM CELL MANIPULATION In vitro acceleration of meiotic and post-meiotic di€erentiation events Experimental studies performed both in laboratory animals and with human testicular tissue obtained from patients with obstructive azoospermia and normal spermatogenesis converge to indicate that the time schedule of in vivo spermatogenic events can be manipulated by di€erent in vitro treatments (Table 3). Early attempts at the in vitro culture of mammalian testicular tissues suggested that a substantial acceleration of spermatogenesis can be brought about simply by explanting segments of seminiferous tubules and putting them in a standard culture medium without any speci®c hormone supplementation. Most rat primary spermatocytes, for example, completed both meiotic divisions by 2 days in culture and developed into early spermatids, while newly developed late (elongated) spermatids ®rst appeared as early as 4±6 days after the beginning of culture.55 An acceleration of the di€erent phases of germ cell di€erentiation was also noted in early studies of the in vitro culture of human testicular samples.56,57 Although the methodological approaches used in these pioneering studies for the distinction of the individual stages of germ cells before and after culture were subject to some criticism, recent studies, using more sophisticated culture systems and cell identi®cation methods, have con®rmed the trend towards an acceleration of spermatogenesis in vitro in di€erent mammalian species58,59, including the human.60 The mechanism of this general trend towards faster development during in vitro culture is unknown. In the 1990s, several studies aimed to manipulate the timing of spermatogenic events by the exposure of germ cells to speci®c, well-de®ned conditions (Table 3). The most interesting data were generated by two di€erent approaches. One approach was based on a direct exposure of explanted germ cells to okadaic acid, an inhibitor of protein phosphatases modulating the activity of cell cycle-regulating kinases. It was shown that the treatment of mouse primary spermatocytes with okadaic acid accelerated the transition from G2 phase of the ®rst meiotic division to metaphase (G2/M-phase transition), a process that takes several days in vivo, to as little as a few hours.61 Moreover, the okadaic acid-induced acceleration of meiosis in mouse and rat male germ cells involved a correct execution of processes that are unique to meiosis, such as a disassembly of synaptonemal complexes with the resolution of cross-overs as cytologically visible chiasmata, an increase in metaphase-promoting factor activity61 and speci®c modi®cations (phosphorylation) to meiosis-speci®c proteins.62 Interestingly, the action of okadaic acid on meiosis was restricted to primary spermatocytes at

Immature germ cell conception 443 Table 3. Examples of the in vitro acceleration of spermatogenic events in laboratory animals and humans. Species

Accelerated event

Mechanism

Reference

Mouse

G2/M-phase transition

Protein phosphatase inhibition by okadaic acid

61, 63

Rat

G2/M-phase transition

Protein phosphatase inhibition by okadaic acid

62

Mouse

G2/M-phase transition followed by meiosis II

Exposure to oocyte cytoplasm rich in metaphase-promoting factor

18, 20

Human

G2/M-phase transition followed by meiosis II

Unknown, dependent on FSH

64, 65

Human

Post-meiotic di€erentiation

Unknown, dependent on FSH

64, 65

FSH ˆ follicle-stimulating hormone.

pachytene of the ®rst meiotic prophase, whereas no response could be induced for earlier stages (leptotene and zygotene).63 The other approach to arti®cial acceleration of male meiosis was based on the microinjection of mouse primary spermatocyte nuclei to metaphase II mouse oocytes.18,20 The extrusion of a pseudopolar body containing a set of spermatocytederived chromosomes occurred within 6 hours of injection, but this technique was burdened with a high frequency of meiotic errors, supposedly caused by premature sister chromatic separation.20 Recent studies of the in vitro di€erentiation of human male germ cells64,65 have returned to the original, `crude' culture systems using incompletely dissociated Sertoli±germ cell complexes and segments of whole seminiferous tubules, which proved to be more ecient in terms of speed of di€erentiation and the yield of mature cells compared with systems in which individual cell types were dissociated before allowing them to reassociate.66 In addition, the e€ects of the addition of the two hormones known to be involved in human spermatogenesis, follicle-stimulating hormone (FSH) and testosterone, were tested. These studies showed that, in this culture system, the progression of meiosis, postmeiotic morphological di€erentiation and cytoplasmic maturation of post-meiotic cells, monitored by the progression of the assembly of spermatid acrosomal structures marked by immunocytochemistry with a monoclonal antibody against human proacrosin67, was stimulated by FSH in a concentration-dependent manner, with a plateau phase from 50 IU/l FSH.64,65 Moreover, these FSH e€ects were potentiated by testosterone, added at a concentation of 1 mM, whereas the addition of testosterone alone failed to produce any measurable e€ect.65 The potentiating e€ect of testosterone was explained by a protective action of this hormone upon Sertoli cells, which represent the physiological target for FSH action.65 A measurable shift in the proportion of individual stages of meiotic and post-meiotic germ cells cultured in the presence of FSH and testosterone was detected as early as 24 hours after the beginning of culture and was even more accentuated after 48 hours.64,65 No acceleration of meiosis was detected in the absence of FSH, whereas some of the post-meiotic di€erentiation events (condensation, peripheral migration and protrusion of the spermatid nucleus) were FSH dependent, while others (¯agellar growth) were FSH independent.64 Other authors have con®rmed a spectacular acceleration of spermatid ¯agellar growth (several hours) during in vitro culture in hormone-free media68 and

444 J. Tesarik et al

during co-culture on Vero cell monolayers.69 These results (Table 3) formed the basis of experiments aimed at the clinical application of these treatments in order to overcome developmental blocks located at di€erent stages of the spermatogenic process.

Selective pressure against germ cell apoptosis during in vitro culture It was mentioned above that a massive apoptosis of immature germ cells from men with maturation arrest is one of the most important obstacles to their use for assisted reproduction. The few healthy cells that may be present in fresh testicular biopsy samples cannot be distinguished from apoptotic cells without the use of invasive analytical methods. Because apoptotic cells are unlikely to undergo any di€erentiation changes during in vitro culture, one of the ideas in favour of the potential therapeutical interest of human germ cell in vitro culture was the facilitation of the selection of healthy cells to be injected into the oocytes. If, for example, in vivo spermatogenesis is arrested at the round spermatid stage, any elongating and elongated spermatid forms detected after in vitro culture must have resulted from in vitro di€erentiation and can thus be expected to be healthy. Experimental work has indeed con®rmed this hypothesis.70 Overriding apoptosis may not, however, automatically lead to the production of developmentally competent cells. In vivo, apoptosis serves to eliminate cells that have been found to be developmentally incompetent during quality control checks at cell cycle checkpoints. For germ cells, one such known checkpoint is located in late pachytene of the ®rst meiotic prophase and controls the completeness of recombination and synapsis formation. This checkpoint is highly conserved in the evolution of species and can be found in both yeasts and higher eukaryotes, including mammals.71 The prevalence of apoptosis among germ cells downstream of the checkpoint thus depends on both the inherent quality of the precursor cells upstream of the checkpoint and on the eciency of the checkpoint control. This situation is schematically represented in Figure 1. If the checkpoint control functions normally, developmentally competent cells pass undamaged (Figure 1A), whereas those in which abnormalities jeopardizing their developmental potential have been detected become apoptotic (Figure 1B). When the checkpoint control is abrogated or becomes leaky, developmentally incompetent cells do not become apoptotic (Figure 1C), but this does not cure their inherent problem responsible for the germ cell developmental incompetence. From the above perspective, it is important to know what types of checkpoint control are involved in the triggering of apoptosis in germ cells from men with maturation arrest, how they are in¯uenced by di€erent in vitro treatments and whether the developmental competence of the cells that have avoided elimination as a result of the in vitro manipulation can be restored by arti®cial means. Interestingly, the developmental checkpoint controlling recombination and synapsis in mouse pachytene spermatocytes does not appear to be overriden by the meiosis-accelerating treatment with okadaic acid because only those spermatocytes showing complete synapsis can be recruited for accelerated di€erentiation.61 Hence, the authors of that study postulated that, suprisingly, it may not be the genetic events of pachytene that account for the length of time the spermatocyte spends in this phase of its di€erentiation; there may instead be another checkpoint in late pachytene at which the accumulation of

Immature germ cell conception 445

Checkpoint (A)

Developmentally competent cell

Developmentally competent cell

(B)

Developmentally incompetent cell

Apoptotic cell

(C)

Developmentally incompetent cell

Developmentally incompetent cell

Figure 1. Quality control of spermatogenesis. At a certain phase of di€erentiation (checkpoint), the normal progression of developmentally relevant molecular events is checked. Only cells in which no anomaly is found by this intrinsic quality control mechanism are allowed to continue their development, whereas the cells found to be defective are tagged for apoptosis. (A) Developmentally competent cells pass through the checkpoint and carry on with their di€erentiation. (B) Developmentally incompetent cells are arrested at the checkpoint and are eliminated by apoptosis. (C) If the checkpoint control mechanism does not work, or is rendered `leaky' by certain in vitro treatments, developmentally incompetent cells are not recognized and continue to di€erentiate.

precursors for the subsequent transcriptionally silent phase of post-meiotic di€erentiation is controlled.61 Experiments with human in vitro spermatogenesis corroborate this hypothesis because most of the spermatids and spermatozoa ensuing from this process are morphologically atypical64,65, suggesting post-meiotic di€erentiation defects. These morphological abnormalities, although incompatible with natural fertilization, do not compromise the developmental competence of in vitro produced spermatids and spermatozoa when the fertilization is assisted by micromanipulation (see below). In other words, most of the in vitro generated spermatids are actually not `healthy' cells, and their fertilizing ability is entirely dependent on a complex micromanipulation procedure that serves not only to ensure the deposition of the spermatid nucleus in the oocyte cytoplasm, but also to assist oocyte activation events triggering early embryonic development. On the other hand, in vitro culture facilitates the selection of non-apoptotic cells, devoid of irreparable DNA damage. The mechanism by which in vitro cultured germ cells escape apoptosis is not clear but appears to be related to the unusual rapidity of the di€erentiation process in vitro (Figure 2). The apoptotic pathway (Figure 2A) is triggered whan an anomaly is detected at a checkpoint and con®rmed when the cell fails to repair the existing anomaly within the ®xed time frame. The cell is then tagged for death, execution possibly being dependent on additional external stimuli. This is a relatively lengthy

446 J. Tesarik et al

(A) Apoptotic pathway Anomaly detection at a checkpoint

Detection of repair failure Tagging for death

Activation of repair

External death stimulus Death execution

Apoptotic cell (B) In vivo spermatogenesis MI spermatocyte Delayed meiosis

Partial repair

Failed or abnormal spermiogenesis

Apoptotic spermatid

(C) In vitro spermatogenesis MI spermatocyte Accelerated meiosis and spermiogenesis

Healthy spermatid

Figure 2. Model explaining how in vitro cultured germ cells may escape apoptosis. (A) The detection of an anomaly at a checkpoint induces a repair mechanism. If the repair is not achieved, the cell is subsequently tagged for death via apoptosis, and the execution of the sentence may be accelerated by di€erent external stimuli. (B) During in vivo spermatogenesis, the main known checkpoint is localized at the entry of primary spermatocytes to metaphase I (MI). If no repair of the detected anomalies occurs, the further progression of meiosis is delayed or arrested, and the cell undergoes apoptosis. (C) In vitro culture accelerates the di€erentiation events so that the death-triggering mechanism does not have enough time to be activated before the cell reaches the advanced stages of spermatogenesis. Consequently, the most rapidly developing cells can escape apoptosis and develop into healthy spermatids, which, however, still bear the original developmental anomaly (see Figure 1 above).

process, which, when superimposed on the timing of in vivo spermatogenic events (Figure 2B), may ®rst delay the progression of meiosis and subsequently trigger the apoptosis-executing mechanisms, whose e€ect may be fully developed only at relatively late stages of spermatogenesis. In contrast, the highly accelerated progression of spermatogenesis during in vitro culture (Figure 2C) does not leave enough time to the full activation of the death-executing mechanisms before spermatid formation, leading to the prevalence of non-apoptotic spermatids among the culture products. CLINICAL APPLICATION OF IN VITRO GERM CELL MANIPULATION TO IMMATURE GERM CELL CONCEPTION IN MEN WITH MATURATION ARREST Meiotic maturation arrest The ®rst successful assisted reproduction treatment for meiotic maturation arrest was reported in a case of maturation arrest at the secondary spermatocyte stage72 with the use of the technique of secondary spermatocyte haploidization after injection into the oocyte cytoplasm, which had previously proved ecient in the mouse model.17 The ®rst successful assisted reproduction treatment for a patient with maturation arrest at the primary spermatocyte stage73 was, however, realized using a di€erent technique,

Immature germ cell conception 447

that was based on the in vitro culture of testicular biopsy samples in the presence of high concentrations of FSH and testosterone.65 This treatment resulted in a twin pregnancy and the birth of two healthy girls with a normal karyotype.73,74 Another pregnancy, achieved in a similar case of meiotic maturation arrest by using the same technique, is ongoing (Tesarik, work in preparation). Interestingly, as in mouse and rat models62,63, only spermatocytes arrested at pachytene could be recruited for accelerated in vitro meiosis, whereas earlier stages of the ®rst meiotic prophase (leptotene and zygotene) could not.74 If this ®nding is con®rmed with a larger group of patients, the detection of the blocking stage in the ®rst meiosis might be of diagnostic value to predict the chance of successful in vitro maturation in patients with maturation arrest at the primary spermatocyte stage. Taken together, these data suggest that in vitro maturation may be more suitable for human application than the direct injection of immature germ cells followed by maturation within the oocyte cytoplasm because the manipulation is easier, does not impose the need for sacri®cing one extra oocyte for the two-step haploidization treatment, which would be needed in cases in which only primary spermatocytes were available, and does not appear to be burdened with a high frequency of chromosomal abnormalites74 similar to those described after the direct injection of mouse primary spermatocytes into metaphase II oocytes.20 Post-meiotic maturation arrest The same in vitro culture technique as for the in vitro transmeiotic maturation of primary spermatocytes was also used with success in assisted reproduction using in vitro elongated spermatids from a patient with complete in vivo maturation arrest at the round spermatid stage.73,74 Two other births and several ongoing pregnancies have subsequently been achieved for the same indication using this technique, which may thus represent the long-awaited breakthrough in the up until now very unsuccesful treatment of this condition. There is experimental evidence showing that in vitro culture facilitates the selection of non-apoptotic spermatids in men with postmeiotic maturation arrest.70 The mechanism of this selection may be the same as that proposed for avoidance of apoptosis in primary spermatocyte arrest (see Figure 2 above). Recent data suggest that in vitro culture improves the outcome of assisted reproduction in patients with massive germ cell apoptosis even if they produce a very small number of late elongated spermatids or spermatozoa.75 Current limitations and future prospects The very recent results with immature germ cell conception using in vitro spermatogenesis are encouraging but need to be con®rmed by di€erent centres before being proposed to patients on a large-scale basis. In any case, only about half of all patients who possess immature germ cells in their testis can bene®t from this technique because spermatogenesis remains arrested during in vitro culture in the other half. Patients with an only slightly elevated serum FSH level (10±20 IU/l) and those with post-meiotic maturation arrest are most likely to bene®t from this technique compared with patients with a very high FSH (420 IU/l) level and maturation arrest at the primary spermatocyte stage.76 Recent data from our laboratory (unpublished) suggest that, unlike men with normal spermatogenesis, in whom the in vitro e€ect of FSH plateaus at a concentration of 50 IU/ l65, a concentration-dependent increase in the spermatogenesis-promoting activity of

448 J. Tesarik et al

Practice points . the results of immature germ cell conception using freshly obtained germ cells for micromanipulation-assisted fertilization have been disappointing . a major drawback of the use of freshly obtained germ cells for assisted reproduction is the high frequency of apoptosis, which leads to fertilization or implantation failure in the case of ROSI and may be responsible for the reported fetal malformations in the case of intracytoplasmic sperm injection . in vitro culture of testicular biopsy samples in the presence of pharmacological concentrations of FSH and testosterone augments the success rates of immature germ cell conception by facilitating the distinction of germ cells from somatic cells, by decreasing the risk of the inadvertent use of an apoptotic germ cell and by promoting additional nuclear and cytoplasmic germ cell maturation . in vitro culture is the only available option for obtaining potentially fertilizing gametes in azoospermic men with maturation arrest at the primary spermatocyte stage; this technique appears to have more chance of success in cases of spermatocyte arrest in late pachytene compared with leptotene and zygotene of the ®rst meiotic prophase . in vitro culture increases the chance of conception in men with maturation arrest at the round spermatid stage . in cases of non-obstructive azoospermia in which there is some doubt about the possibility of recovering spermatozoa for assisted reproduction by testicular biopsy, it is preferable to postpone oocyte retrieval until 2 days after testicular biopsy and to culture testicular cells for 2 days in a FSH- and testosteronesupplemented medium. If mature spermatozoa are present in the fresh sample, this additional culture period will not decrease their quality for the subsequent assisted reproduction attempt . the in vitro culture of testicular biopsy samples is best carried out using partially mechanically disintegrated testicular tissues in a water bath set at 308C and in a HEPES-bu€ered medium enriched with FSH and testosterone . in vitro di€erentiation can be achieved in about half of men with maturation arrest in whom immature germ cells are present; although a very high serum FSH level indicates a poor prognosis, none of the current clinical and laboratory tests can predict the success or failure of germ cell in vitro di€erentiation

FSH occurs up to a concentration of 500 IU/l in men with maturation arrest and elevated serum FSH. Similarly, a very high concentration of testosterone (10 mM) appears to be superior to the originally used 1 mM65 in this category of patients. We are thus now using these elevated hormone concentrations in all patients with maturation arrest. Nothing of course can be done for patients with complete germinal aplasia (Sertoli cellonly syndrome), from whom no immature germ cells can be recovered. Future research has to focus on these treatment-resistant cases (see below). SUMMARY The feasibility of human conception after fertilization with immature post-meiotic germ cells, round and elongated spermatids, has been demonstrated. The clinical

Immature germ cell conception 449

Research agenda . non-invasive diagnostic methods are needed to predict the presence of immature germ cells in the testes of azoospermic men and of their in vitro di€erentiation capacity . in vitro culture techniques should be optimized to augment the yield, viability, cytoplasmic maturity and biological quality of in vitro matured post-meiotic germ cells . there is a need to develop methods for assisting oocyte activation after the injection of incompletely mature post-meiotic germ cells . long-term culture systems for the in vitro maturation of spermatogonia and prepachytene primary spermatocytes should be devised . research should be carried out into methods of enabling the haploidization of somatic cell nuclei and their use for the formation of arti®cial gametes to be used in cases of complete absence of the germ line ecacy of elongated spermatid conception is acceptable, but a recent report raises concern about the possible increased risk of fetal malformation. In contrast, the ecacy of conception with freshly obtained round spermatids is low, although all the babies born using this technique have so far been normal. An elevated frequency of apoptotic germ cells in men with maturation arrest is likely to be the main cause of conception failure with round spermatids and may be implicated in the developmental abnormalities seen after the use of elongated spermatids. The recent development of in vitro culture techniques for immature male germ cells has made it possible, in some cases, to overcome the existing in vivo maturation arrest and to obtain a highly accelerated meiotic and post-meiotic di€erentiation. This has allowed clinicians, for the ®rst time, to obtain healthy infants by using germ cells from men with meiotic maturation arrest at the primary spermatocyte stage that were prompted to develop to elongated spermatids during in vitro culture. This technique also signi®cantly reduces the risk of an inadvertent use of an apoptotic germ cell for assisted reproduction. These data suggest that the use of freshly obtained immature germ cells for assisted reproduction should be abandoned. Instead, testicular biopsy should be performed 2 days before oocyte recovery in all cases in which the absence of mature spermatozoa in the testis is suspected, allowing for the additional in vitro maturation of germ cells before their injections into oocytes. REFERENCES 1. Palermo G, Joris H, Devroey P & Van Steirteghem AC. Pregnancies after intracytoplasmic injection of a single spermatozoon into an oocyte. Lancet 1992; 340: 17±18. 2. Van Steirteghem AC, Nagy Z, Joris H et al. High fertilization and implantation rates after ICSI. Human Reproduction 1993; 8: 1061±1066. 3. Edwards RG, Tarin JJ, Dean N et al. Are spermatid injections into human oocytes now mandatory? Human Reproduction 1994; 9: 2217±2219. * 4. Ogura A, Matsuda J & Yanagimachi R. Birth of normal young after electrofusion of mouse oocytes with round spermatids. Proceedings of the National Academy of Sciences of the USA 1994; 91: 7460±7462. 5. Vanderzwalmen P, Lejeune B, Nijs M et al. Fertilization of an oocyte microinseminated with a spermatid in an in vitro fertilization programme. Human Reproduction 1995; 10: 502±503. 6. Fishel S, Green S, Bishop M et al. Pregnancy after intracytoplasmic injection of spermatid. Lancet 1995; 345: 1641±1642.

450 J. Tesarik et al * 7. Tesarik J, Mendoza C & Testart J. Viable embryos from injection of round spermatids into oocytes. New England Journal of Medicine 1995; 333: 525. 8. Re M, Carpini-Familari G, Iannitelli M & Vicari A. Ultrastructural characteristics of idiopathic spermatidic arrest. Archives of Andrology 1979; 2: 283±289. 9. AumuÈller G, Fuhrmann W & Krause W. Spermatogenic arrest with inhibition of acrosome and sperm tail development. Andrologia 1987; 19: 9±17. 10. Martin-du Pan R & Campana A. Physiopathology of spermatogenic arrest. Fertility and Sterility 1993; 60: 937±946. 11. Tesarik J, Sousa M, Greco E & Mendoza C. Spermatids as gametes: indications and limitations. Human Reproduction 1998; 13 (supplement 3): 89±107. 12. Ogura A, Yanagimachi R & Usui N. Behaviour of hamster and mouse round spermatid nuclei incorporated into mature oocytes by electrofusion. Zygote 1993; 1: 1±8. 13. Ogura A & Yanagimachi R. Round spermatid nuclei injected into hamster oocytes form pronuclei and participate in syngamy. Biology of Reproduction 1993; 48: 219±225. 14. Ogura A, Matsuda J & Yanagimachi R. Birth of normal young after electrofusion of mouse oocytes with round spermatids. Proceedings of the National Academy of Sciences of the USA 1994; 91: 7460±7462. 15. Kimura Y & Yanagimachi R. Mouse oocytes injected with testicular spermatozoa and round spermatids can develop into normal o€spring. Development 1995; 121: 2397±2405. 16. So®kitis NV, Toda T, Miyagawa I et al. Bene®cial e€ects of electrical stimulation before round spermatid nuclei injections into rabbit oocytes on fertilization and subsequent embryonic development. Fertility and Sterility 1996; 65: 176±185. *17. Kimura Y & Yanagimachi R. Development of normal mice from oocytes injected with secondary spermatocyte nuclei. Biology of Reproduction 1995; 53: 855±862. 18. Sasagawa I, Kuretake S, Eppig JJ & Yanagimachi R. Mouse primary spermatocytes can complete two meiotic divisions within the oocyte cytoplasm. Biology of Reproduction 1998; 58: 248±254. *19. Ogura A, Suzuki O, Tanemura T et al. Development of normal mice from metaphase I oocytes fertilized with primary spermatocytes. Proceedings of the National Academy of Sciences of the USA 1998; 95: 5611±5615. *20. Kimura Y, Tateno H, Handel MA & Yanagimachi R. Factors a€ecting meiotic and developmental competence of primary spermatocyte nuclei injected into mouse oocytes. Biology of Reproduction 1998; 59: 871±877. 21. Sousa M, Barros A, Takahashi K et al. Clinical ecacy of spermatid conception: analysis using a new spermatid classi®cation scheme. Human Reproduction 1999; 14: 1279±1286. 22. Tesarik J, Rolet F, Brami C et al. Spermatid injection into human oocytes. II. Clinical application in the treatment of infertility due to non-obstructive azoospermia. Human Reproduction 1996; 11: 780±783. 23. Barak Y, Kogosowski A, Goldman S et al. Pregnancy and birth after transfer of embryos that developed from single-nucleated zygotes obtained by injection of round spermatids into oocytes. Fertility and Sterility 1998; 70: 67±70. *24. Vanderzwalmen P, Zech H, Birkenfeld A et al. Intracytoplasmic injection of spermatids retrieved from testicular tissue: in¯uence of testicular pathology, type of selected spermatids and oocyte activation. Human Reproduction 1997; 12: 1203±1213. 25. Mansour RT, Aboulghar MA, Serour GI et al. Pregnancy and delivery after intracytoplasmic injection of spermatids into human oocytes. Middle East Fertility Society Journal 1996; 1: 223±225. 26. Antinori S, Versaci C, Dani G et al. Fertilization with human testicular spermatids: four successful pregnancies. Human Reproduction 1997; 12: 286±291. 27. Antinori S, Versaci C, Dani G et al. Successful fertilization and pregnancy after injection of frozen-thawed round spermatids into human oocytes. Human Reproduction 1997; 12: 554±556. 28. Araki Y, Motoyama M, Yoshida A et al. Intracytoplasmic injection with late spermatids: a successful procedure in achieving childbirth for couples in which the male partner su€ers from azoospermia due to de®cient spermatogenesis. Fertility and Sterility 1997; 67: 559±561. 29. Amer M, Soliman E, El-Sadek M et al. Is complete spermiogenesis failure a good indication for spermatid conception? Lancet 1997; 350: 116. 30. Bernabeu R, Cremades N, Takahashi K & Sousa M. Successful pregnancy after spermatid injection. Human Reproduction 1998; 13: 1898±1900. 31. So®kitis N, Yamamoto Y, Miyagawa I et al. Ooplasmic injection of elongating spermatids for the treatment of non-obstructive azoospermia. Human Reproduction 1998; 13: 709±714. 32. Kahraman S, Polat G, Samli M et al. Multiple pregnancies obtained by testicular spermatid injection in combination with intracytoplasmic sperm injection. Human Reproduction 1998; 13: 104±110. 33. Barros A, Bernabeu R, Takahashi K et al. Intracytoplasmic injection of ejaculate and testicle spermatids: report on 35 cycles. Human Reproduction 1998; 13 (supplement 1): 154±155.

Immature germ cell conception 451 34. Al-Hasani S, Ludwig M, Palermo I et al. Intracytoplasmic injection of round and elongated spermatids from azoospermic patients: results and review. Human Reproduction 1999; 14 (supplement 1): 97±107. 35. Ghazzawi IM, Alhasani S, Taher M & Souso S. Reproductive capacity of round spermatids compared with mature spermatozoa in a population of azoospermic men. Human Reproduction 1999; 14: 736±740. 36. Gianaroli L, Selman HA, Magli MC et al. Birth of a healthy infant after conception with round spermatids isolated from cryopreserved testicular tissue. Fertility and Sterility 1999; 72: 539±541. 37. Zech H, Vanderzwalmen P, Prapas Y et al. Congenital malformations after intracytoplasmic injection of spermatids. Human Reproduction 2000; 15: 969±971. 38. Dinulovic D & Radonjic G. Diabetes mellitus/male infertility. Archives of Andrology 1990; 25: 277±293. 39. Weinbauer GF, Behr R, Bergmann M & Nieschlag E. Testicular cAMP responsive element modulator (CREM) protein is expressed in round spermatids but is absent or reduced in men with round spermatid maturation arrest. Molecular Human Reproduction 1998; 4: 9±15. 40. Rodriguez I, Ody C, Araki K et al. An early and massive wave of germinal cell apoptosis is required for the development of functional spermatogenesis. European Molecular Biology Organization Journal 1997; 16: 2262±2270. 41. Odorisio T, Rodriguez TA, Evans EP et al. The meiotic checkpoint monitoring synapsis eliminates spermatocytes via p53-independent apoptosis. Nature Genetics 1998; 18: 257±261. 42. Yin Y, Stahl BC, DeWolf WC & Morgentaler A. p53-Mediated germ cell quality control in spermatogenesis. Developmental Biology 1998; 204: 165±171. *43. Tesarik J, Greco E, Cohen-Bacrie P & Mendoza C. Germ cell apoptosiss in men with complete and incomplete spermiogenesis failure. Molecular Human Reproduction 1998; 4: 757±762. 44. Jurisicova A, Lopes S, Meriano J et al. DNA damage in round spermatids of mice with targeted disruption of the Pp1cg gene and in testicular biopsies of patients with non-obstructive azoospermia. Molecular Human Reproduction 1999; 5: 323±330. 45. Balaban B, Urman B, Isiklar A et al. Progression to the blastocyst stage of embryos derived from testicular round spermatids. Human Reproduction 2000; 15: 1377±1382. 46. Fishel S, Aslam I & Tesarik J. Spermatid conception: a stage too early, or a time too soon? Human Reproduction 1996; 11: 1371±1375. 47. Tesarik J & Mendoza C. In vitro fertilization by intracytoplasmic sperm injection. BioEssays 1999; 21: 791±801. 48. Tesarik J & Mendoza C. Spermatid injection into human oocytes. I. Laboratory techniques and special features of zygote development. Human Reproduction 1996; 11: 772±779. 49. Tesarik J & Testart J. Treatment of sperm-injected human oocytes with Ca2‡ ionophore supports the development of Ca2‡ oscillations. Biology of Reproduction 1994; 51: 385±391. 50. Tesarik J & Sousa M. More than 90% fertilization rates after intracytoplasmic sperm injection and arti®cial oocyte activation with calcium ionophore. Fertility and Sterility 1995; 63: 343±349. 51. de Kretser DM & Kerr JB. The cytology of the testis. In Knobil E & Neill J (eds) The Physiology of Reproduction, pp 837±932. New York: Raven Press, 1988. 52. Gledhill BL, Gledhill MP, Rigler R & Ringertz NR. Changes in deoxyribonucleoprotein during spermatogenesis in the bull. Experimental Cell Research 1966; 41: 652±655. 53. Sousa M & Tesarik J. Ultrastructural analysis of fertilization failure after intracytoplasmic sperm injection. Human Reproduction 1994; 9: 2374±2380. 54. Tesarik J. Oocyte activation after intracytoplasmic injection of mature and immature sperm cells. Human Reproduction 1998; 13 (supplement 1): 117±127. 55. Parvinen M, Wright WW, Phillips DM et al. Spermatogenesis in vitro: completion of meiosis and early spermiogenesis. Endocrinology 1983; 112: 1150±1152. 56. Matte R & Sasaki M. Autoradiographic evidence of human male germ-cell di€erentiation in vitro. Cytologia 1971; 36: 298±303. 57. Ghatnekar R, Lima-de-Faria A, Rubin S & Menander K. Development of human male meiosis in vitro. Hereditas 1974; 78: 265±272. 58. Le Magueresse-Battistoni B, GeÂrard N & JeÂgou B. Pachytene spermatocytes can achieve meiotic process in vitro. Biochemical and Biophysical Research Communications 1991; 179: 1115±1121. 59. Weiss M, Vigier M, Hue D et al. Pre- and post-meiotic expression of male germ cell-speci®c genes throughout 2-week cocultures of rat germinal and Sertoli cells. Biology of Reproduction 1997; 57: 68±76. 60. Tres LL, Smith FF & Kierszenbaum AL. Spermatogenesis in vitro: methodological advances and cellular functional parameters. In Negro-Vilar A & Perez-Palacios G (eds) Reproduction, Growth and Development, vol. 71, pp 115±125. New York: Serono Symposia, Raven Press, 1991. *61. Wiltshire T, Park C, Caldwell KA & Handel MA. Induced premature G2/M-phase transition in pachytene spermatocytes includes events unique to meiosis. Developmental Biology 1995; 169: 557±567.

452 J. Tesarik et al 62. Tarsounas M, Pearlman RE & Moens PB. Meiotic activation of rat pachytene spermatocytes with okadaic acid: the behaviour of synaptonemal complex components SYN1/SCP1 and COR1/SCP3. Journal of Cell Science 1999; 112: 423±434. 63. Cobb J, Cargile B & Handel MA. Acquisition of competence to condense metaphase I chromosomes during spermatogenesis. Developmental Biology 1999; 205: 49±64. 64. Tesarik J, Greco E, Rienzi L et al. Di€erentiation of spermatogenic cells during in-vitro culture of testicular biopsy samples from patients with obstructive azoospermia: e€ect of recombinant follicle stimulating hormone. Human Reproduction 1998; 13: 2772±2781. *65. Tesarik J, Guido M, Mendoza C & Greco E. Human spermatogenesis in vitro: respective e€ects of folliclestimulating hormone and testosterone on meiosis, spermiogenesis, and Sertoli cell apoptosis. Journal of Clinical Endocrinology and Metabolism 1998; 83: 4467±4473. 66. Kierszenbaum AL. Mammalian spermatogenesis in vivo and in vitro: a partnership of spermatogenic and somatic cell lineages. Endocrine Reviews 1994; 15: 116±134. 67. Escalier D, Gallo J-M, Albert M et al. Human acrosome biogenesis: immunodetection of proacrosin in primary spermatocytes and of its partitioning pattern during meiosis. Development 1991; 113: 779±788. 68. Aslam I & Fishel S. Short-term in vitro culture and cryopreservation of spermatogenic cells used for human in-vitro conception. Human Reproduction 1998; 13: 634±638. 69. Cremades N, Bernabeu R, Barros A & Sousa M. In-vitro maturation of round spermatids using co-culture on Vero cells. Human Reproduction 1999; 14: 1287±1293. 70. Tesarik J, Mendoza C & Greco E. In vitro culture facilitates the selection of healthy spermatids for assisted reproduction. Fertility and Sterility 1999; 72: 809±813. 71. Roeder GS. Meiotic chromosomes: it takes two to tango. Genes and Development 1997; 11: 2600±2621. 72. So®kitis N, Mantzavinos T, Loutradis D et al. Ooplasmic injections of secondary spermatocytes for nonobstructive azoospermia. Lancet 1998; 351: 1177±1178. *73. Tesarik J, Bahceci M, OÈzcan C et al. Restoration of fertility by in-vitro spermatogenesis. Lancet 1999; 353: 555±556. 74. Tesarik J, Bahceci M, OÈzcan C. et al. In-vitro spermatogenesis. Lancet 1999; 353: 1708. 75. Tesarik J, Cruz-Navarro N, Moreno E et al. Pregnancy and birth with in vitro cultured spermatids from patient with massive in vivo apoptosis of post-meiotic germ cells. Fertility and Sterility 2000 (in press). 76. Tesarik J, Balaban B, Isiklar A et al. In-vitro spermatogenesis resumption in men with maturation arrest: relationship with in-vivo blocking stage and serum FSH. Human Reproduction 2000; 15: 1350±1354.