Spermatid injection into human oocytes. I. Laboratory techniques and special features of zygote development

Human Reproduction vol.11 no.4 pp.772-779, 1996 Spermatid injection into human oocytes. I. Laboratory techniques and special features of zygote development Jan Tesarik 1 and Carmen Mendoza American Hospital of Paris, 63 Boulevard Victor Hugo, 92202 Neuilly sur Seine, France, and Division of Reproductive Biology, University of Granada Faculty of Sciences, Campus Universitario Fuentenueva, 18071 Granada, Spain 'To whom correspondence should be addressed at: 50 Rue des Carrieres, 92150 Suresnes, France Spermatid injection into the oocyte cytoplasm has been shown recently to yield viable human embryos developing to term after transfer to the mother. This study provides details of the laboratory techniques related to spermatid injection (ROSI) and spermatid injection (ELSI) and focuses on some special features of zygote development associated with the use of these types of sperm precursor cells for fertilization. A spermatidenriched fraction was obtained by centrifugation of cells from azoospermic ejaculates through a discontinuous Percoll gradient column. Individual or spermatids were identified in this fraction and injected deep into oocytes. Oocyte activation was boosted by a vigorous aspiration of the ooplasm at the time of injection. The fertilization rates after ROSI and ELSI were 45 and 44% respectively. A single large syngamy nucleus was detected in 36% of the zygotes that previously showed two normalsized pronaclel. This condition did not appear to delay the first cleavage division. These observations underscore the importance of distinguishing the syngamy nucleus of diploid zygotes from the female pronucleus of haploid, parthenogenetically activated eggs. Key words: non-obstructive azoospennia/oocyte activation/ pronuclear development/spennatid injection into oocytes/ syngamy Introduction Since the introduction of intracytoplasmic sperm injection (ICSI) into the treatment of human male infertility (Palermo et al., 1992), the ICSI technique has been perfected to give >50% pregnancy rates even in cases of severe oligoasthenoteratozoospermia in some centres (Tesarik and Sousa, 1995a). Moreover, ICSI has been successfully combined with microsurgical epididymal sperm aspiration or testicular biopsy to treat cases of obstructive azoospermia (reviewed in Silber, 1994). However, men with non-obstructive azoospermia, due to defects of spermatogenesis, were still excluded from reproduction. The observations that incorporation of spermatid 772 nuclei into hamster and mouse oocytes can lead to fertilization and the beginning of embryonic development (Ogura and Yanagimachi, 1993; Ogura et al, 1993) and that normal young can be obtained, at least in the mouse, after transfer of such embryos into recipient females (Ogura et al, 1994; Kimura and Yanagimachi, 1995) suggested diat the use of spermatids for conception might be applicable to the treatment of nonobstructive azoospermia in humans (Edwards et al., 1994). This expectation has been reinforced by the successful application of the method in rabbits (Sofikitis et al, 1994). hi the past 2 years, attempts at obtaining viable human embryos by injecting spermatids or spermatid nuclei into oocytes have been reported from several laboratories. The feasibility of human conception with the use of spermatids instead of spermatozoa has been suggested by the observation of apparently normal fertilization and cleavage up to the fourcell stage of human oocytes injected with spermatids (Vanderzwalmen et al, 1995) and confirmed by the first two childbirths resulting from injection of spermatids into human oocytes (Tesarik et al., 1995,1996). Another continuing pregnancy resulting from injection of an spermatid to a human oocyte has been reported (Fishel et al, 1995), whereas four other pregnancies, established after injection of isolated spermatid nuclei, were terminated by spontaneous abortion (Hannay, 1995). Spermatids can be obtained from the testicular tissue by pulling them off Sertoli cells with the use of a micromanipulator (Silber and Lenahan, 1995). However, some spermatids can also detach spontaneously from Sertoli cells and appear in the ejaculate of azoospermic men. The first reported births after spermatid injection into human oocytes (Tesarik et al, 1995, 1996) have been achieved with the use of such 'ejaculated' spermatids. In this study we present a detailed description of the laboratory technique, for both spermatid injection (ROSI) and spermatid injection (ELSI). We also describe some special features of zygote development related to the use of spermatids for conception. The knowledge of these features is crucial for a correct evaluation of fertilization outcomes after ROSI and ELSI. Materials and methods Source and preparation of spermatids Spermatids were prepared from ejaculates of 15 azoospermic men. After centrifugation (1200,10 min), pellets were resuspended in 1 ml Tyrode's salt solution (TSS; Sigma, St Louis, MO, USA), loaded on top of a discontinuous Percoll (Pharmacia, Uppsala, Sweden) gradient column consisting, from top to bottom, of 2 ml 40%, 1 ml 60%, 1 ml 70%, 1 ml 80%, and 1 ml 100% Percoll solutions diluted European Society for Human Reproduction and Embryology

Laboratory aspects of spermatic! injection spermatids in all cases, but no exact quantitation was performed. After centrifugation of whole ejaculates, the identification of spermatids was problematic because of accumulation of cellular debris and abundant amorphous material. Bacteria were also frequent in these ejaculates, and their density was considerably reduced in the spermatid-enriched fraction. Figure 1. (A) A and (B) an spermatid (arrow) just before the aspiration into the microinjection needle (7 u.m, inner diameter at the needle tip). in Ham's F 10 medium (ICN Biomedicals, Costa Mesa, CA, USA) and centrifuged at 330 g for 20 min. Preliminary experiments showed that most living spermatids can be recovered from the 70% Percoll fraction. This fraction was carefully removed from the Percoll column without mixing it with the other fractions and diluted 1:10 with TSS. Cells from the 70% Percoll fraction were then washed by two cycles of centrifugation (600 Xg, 10 min) followed by resuspension in 10 ml TSS and pelleted again by centrifugation at 600Xg for 10 min. The resulting cell pellet was resuspended in 0.5 ml sperm preparation medium (Medi-Cult, Copenhagen, Denmark) and maintained at 37 C for a period ranging from 10 min to 2 h during which oocytes were being prepared. Spermatids were distinguished from other cell types according to their shape, size, and the form and size of the nucleus. Living spermatids had a smooth outline, a nucleus and a regular zone of cytoplasm suring the nucleus (Figure 1A). They did not possess a tail. A developing acrosomal granule could be recognized in some spermatids as a bright spot adjacent to the spermatid nucleus. All spermatids injected in this study presented this distinctive feature. As to the other cells present in the ejaculate, monocytes, polymorphonuclear leucocytes and spermatocytes were distinguished from spermatids by their larger size, whereas lymphocytes have a larger nucleocytoplasmic ratio so that a continuous cytoplasmic zone a the nucleus cannot be seen. These differences are important because a vacuole can occasionally occur adjacent to the nucleus of these cells and mimic the developing acrosomal granule of spermatids. The presence of the acrosome is thus a useful but not the only criterion for the recognition of spermatids from the other types of cells. In some ejaculates, only spermatids were present, whereas some spermatids (Figure IB) were also identified in others. The developing tail was identified exceptionally in the spermatids. This may have been due to a block of spermiogenesis at very early stages of spermatid elongation, to degeneration and lysis of more advanced stages of spermatids during their passage from the testis to the ejaculate or to the loss of the tail during the preparation of the spermatid-enriched fraction. Even in the enriched fraction, the percentage of spermatids was always very low. It was estimated as 1-5% of all nucleated cells present in this fraction. It has to be noted, however, that anuclear cytoplasmic fragments and cellular debris prevailed over intact cells in this fraction although the relative contribution of this material was significantly lower as compared to sediments after centrifugation of whole ejaculates. There appeared to be more spermatids than Source and preparation of oocytes Oocytes were recovered from 15 wives of azoospermic men by ultrasound-guided follicular aspiration after pituitary desensitization with a growth hormone-releasing hormone (GnRH) agonist and ovarian stimulation with gonadotrophins. Cumulus oophorus and corona radiata cells were removed, respectively, by a brief incubation with 80 nj/ml hyaluronidase in sperm preparation medium. This ready-to-use hyaluronidase solution was purchased from Medi-Cult. As soon as the cumulus intercellular matrix was dissolved by the action of hyaluronidase, oocytes were transferred to fresh sperm preparation medium, and the adhering corona radiata cells were stripped off by repeated aspiration of the oocytes in and out of a finely drawn Pasteur pipette. Denuded oocytes were then examined for meiotic maturity. Only metaphase II oocytes (showing one compact or fragmented polar body) were used in this study. The oocytes were maintained at 37 C in IVF medium (Medi-Cult) equilibrated with 5% CO2 in air for 1-6 h before injection. Injection technique Basic features of the injection technique used for deposition of spermatids in oocyte cytoplasm conformed to our standard ICSI technique as described previously (Tesarik and Sousa, 1995a). However, in view of the different size and form of spermatids as compared with spermatozoa, some modifications were necessary to adapt the technique for the injection of this particular type of cell. Injection needles used for spermatid injection (Figure 1) had a less sharply bevelled tip than those used for standard ICSI. This was particularly important for injection of spermatids which were slightly larger than the internal diameter of the injection needles. With sharp-bevelled tips, these cells do not enter the needle easily and spin at the needle tip as a result of turbulence created by medium aspiration. The bevel angle of the needles was 45, and the inner and outer diameters of the injection needles (at the tip) were 7 u.m and 9 ]ira respectively. Spermatid-enriched cell fractions (see above) were centrifuged (600 #,10 min) and resuspended in 10% polyvinylpyrrolidone solution in sperm preparation medium. A microdrop (5 H-l) of this suspension was placed on a Falcon 3001 plastic culture dish (Becton Dickinson, Plymouth, UK). A drop of the 10% polyvinylpyrrolidone solution alone and drops of sperm preparation medium containing oocytes to be injected were added to the same dish. All drops were covered by a layer of embryo-tested light mineral oil (Sigma) previously incubated overnight with an equal volume of FVF medium under a gas phase of 5% CO2 in air, and the dish was located on the heated stage (37 C) of an inverted research microscope equipped with Hoffman modulation contrast optics. The microscope used was either an Olympus IMT-2 or a Leica DMTRB. Two Narishige (Tokyo, Japan) manipulators, MN-151 mechanical joystick manipulator and MMO- 202D hydraulic remote-control manipulator, were mounted on the microscope, the former on the left side and the latter on the right side of the microscope stage. The left-hand and the right-hand manipulators moved the oocyte-holding pipette and the injection needle respectively. Both instruments were prepared from Narishige G-l capillary glass tubes as described previously (Tesarik, 1993) but 773

J.Tesarik and CMendoza for the above modifications of the injection needle bevel angle and diameter. As soon as the dish with spermatids and oocytes was placed on the microscope stage, the tip of the micromjection needle was introduced to the drop containing 10% polyvinylpyrrolidone, and a small amount (~5 pi) of this solution was aspirated into the needle. The needle was then moved to the drop containing spermatids. If spermatids were present, they were used preferentially because, as cells developmentally closer to mature spermatozoa, they were expected to give a better chance of normal fertilization than spermatids. Only in cases in which spermatids were not found, spermatids were injected. The technique used for the injection of spermatids differed slightly from that for injection of spermatids. Once a spermatid was chosen as a potential candidate for injection, an attempt was made to aspirate it into the microinjection needle. The behaviour of the cell during this step was decisive for its final acceptance or rejection. Round spermands were slightly larger than the internal diameter of the microinjection needle and were thus deformed as they entered the needle. However, some spermatids did not disintegrate upon aspiration, whereas others did. The cells that deformed but did not disintegrate were assumed to be viable and selected for injection. Each spennatid having passed through this preliminary selection step was subsequently transferred to the adjacent drop with 10% polyvinylpyrrolidone solution to be washed. When expelled from the needle, living cells tended to resume their initial form. Cells that disintegrated during aspiration into the microinjection needle or remained deformed after expulsion to the polyvinylpyrrolidone drop were not injected. This viability assessment checkpoint was not possible for injection of spermatids, whose smaller width allowed a smooth passage through the microinjection needle without any deformation. After being washed in the polyvinylpyrrolidone drop, the chosen spermatid was aspirated into the microinjection needle again and moved to one of the drops containing oocytes (Figure 2A). Before injection into the oocyte, the spennatid was positioned as close as possible to the needle tip. However, a direct contact between the spermatid and the zona pellucida surface is to be avoided because it often leads to a premature expulsion of the spermatid from the needle during subsequent zona penetration. When the optimal position of the spennatid in the needle was achieved, the needle was pushed rapidly but without excessive haste through the zona pellucida into the oocyte cytoplasm. This step demanded great caution because the needles for spermatid injection are less sharp-bevelled than those used for sperm injection. Thus, hesitation during the injection increases the degree and duration of oocyte deformation, whereas too a hasty movement may lead to a transpiercing of the oocyte which is usually lethal. When the tip of the microinjection needle approached as closely as possible the oocyte pole opposite to that at which the oolemma was penetrated, the oocyte cytoplasm was aspirated vigorously into the needle so that the spennatid within the needle moved ~100 (im backwards. This movement was accompanied by a surge of ooplasm up the injection needle in all oocytes injected in this Figure 2. Injection of a spermatid into a metaphase II oocyte. (A) Preparation for injection. The oocyte is immobilized on the holding pipette (left) with the first polar body in the 12-o'clock position. The tip of the microinjection needle (right) touches the zona pellucida and produces a slight impression reflex on the oolemma. This reflex is used to verify whether the needle is positioned correctly, i.e. in the oocyte equatorial plane. A spennatid (arrow), deformed by aspiration, is seen in the needle. (B) Spennatid deposition in the oocyte cytoplasm. The spennatid (anow), just expulsed from the injection needle, is still deformed. (C) Freshly injected oocyte with the site of needle entry still visible (short arrows) and with the injected spennatid (arrow) at the blunt end of the partially resorbed puncture canal. The injected spennatid has already resumed its original shape. 774

Laboratory aspects of spermatid injection study. The vigorous ooplasmic aspiration has been shown to boost oocyte activation after ICSI (Tesarik and Sousa, 1995a). The spermatid, together with the aspirated ooplasm, was then expelled slowly into the oocyte, and the needle was withdrawn. Extreme caution is necessary at the moment at which the spermatid leaves the needle tip because the pressure within the needle increases as the needle lumen is filled up by the spermatid leading to the risk of massive medium outflow when the free passage is re-established after spermatid expulsion. During the needle withdrawal, as much as possible of the polyvinylpyrrolidone solution previously injected into the oocyte was re-aspirated from the injection canal. Freshly injected spermatids were still deformed by aspiration into and out of the injection needle (Figure 2B), but this deformation disappeared within 2 min after injection, and the injected cells resumed their shape (Figure 2C). The technique of spermatid injection was essentially the same as for spermatid injection. However, the risk of medium overload was lower because spermatids usually did not fill up the needle lumen. larger than the normal size of a pronucleus, was subsequently detected in many of the zygotes that previously showed two equal-sized pronuclei (Figures 3A and B). This nucleus may have resulted from fusion of both pronuclei and is referred to as syngamy nucleus throughout this paper. In addition to Assessment of fertilization, embryo culture and transfer After injection, oocytes were treated as any other injected oocytes in a standard ICSI programme. FVF medium equilibrated with 5% CO2 in air was used for all subsequent culture as well as for embryo transfer. Oocytes from each patient were cultured together in 0.5 ml medium in Nunclon four-well tissue culture dishes without mineral oil protection. In most cases, fertilization was assessed by three consecutive inspections of the oocytes performed at 10-12 h, at 16-18 h, and at 22-24 h after injection. In some cases at the beginning of this experimental series, only two inspections were done at 16-18 h and at 22 24 h after injection. The presence of two pronuclei was the only fertilization criterion applied. Only oocytes in which two pronuclei were detected during at least one of the three inspections were considered further for transfer. Embryonic cleavage was assessed two days after injection. Only embryos uiat had undergone at least one cleavage division were transferred to the patient's uterus. Results Normal fertilization (development of two pronuclei) was achieved in 43% (20/47) of the oocytes treated by ROSI and in 44% (17/39) of the oocytes treated by ELSI. The variability of fertilization outcomes among individual patients and the frequency of abnormal fertilization are described in detail elsewhere (Tesarik et ai, 1996). With the exception of two zygotes (see below), both pronuclei were of similar size and did not differ from those observed in zygotes fertilized by mature spermatozoa. Fully developed pronuclei were observed during the first inspection at 10-12 h after spermatid injection in 15 out of 62 oocytes that were examined at that time (the remaining 24 oocytes were inspected only at the 10-12 h and 22-24 h time points). However, a single nucleus, slightly Figure 3. Post-fertilization development of a zygote obtained by ROSL (A) Zygote appearance during the first inspection performed at 10 h after ROSI. Two pronuclei in close apposition and two polar bodies in the perivitelline space are evident (B) Zygote appearance during subsequent inspection performed at 17 h after ROSI. Two polar bodies are still visible in the same locality, but the two pronuclei have fused giving rise to a single large syngamy nucleus. (C) When observed at 34 h after ROSI, the zygote has developed into a two-cell embryo. 775

J.Tesarik and CMendoza Figure 4. Abnormal pronuclear development after ROSI. This zygote shows only one normal-sized pronucleus (arrow), whereas the other pronucleus (short arrow) is of considerably smaller size. The two polar bodies are out of focus in this picture. their larger size, syngamy nuclei also differed from pronuclei by the number of nucleolar precursors. When evaluated at 16-18 h after injection, the number of nucleolar precursors per pronucleus was 4.7 ±0.1 (mean ± SEM, n = 40), whereas the corresponding value for the syngamy nucleus was 9.1 ± 0.3 (mean ± SEM, n = 7). This difference was significant (f-test, P < 0.01). Table I summarizes the changes in the nuclear status observed at three different time points after spennatid injection in 28 of those zygotes in which two pronuclei were detected on at least one of these occasions. The remaining 9 zygotes, for which only two inspections were done, are not included in Table I. The data in Table I show that two pronuclei were present in 54% (15/28) of the zygotes as early as 10-12 h after spermatid injection. In 53% (8/15) of these zygotes, however, the two pronuclei formed a single syngamy nucleus by the time of the subsequent inspection at 16-18 h after injection. In all these eight zygotes, the syngamy nucleus disappeared by 22-24 h after injection (Table I), and cleavage was observed on the following day (Figure 3C). In two other zygotes, two pronuclei were first detected at 16-18. h after injection, and the syngamy nucleus was seen no earlier than 22-24 h after injection. These two zygotes also cleaved by the next day. The overall cleavage rate of two-pronucleated zygotes was 92% (34/37). Table I also shows the type of spermatid ( versus ) with which individual oocytes giving rise to each of these zygotes were fertilized. The development of two pronuclei followed by the appearance of a syngamy nucleus was observed both in zygotes resulting from ROSI and in those from ELSI. The number of cases available for analysis was too small to allow a comparison of the frequency of this phenomenon between these two techniques. In two cases (zygote numbers 2 and 23 in Table I), only one of the two pronuclei had a normal size, whereas the other one was much smaller and denser (Figure 4). These two zygotes also underwent cleavage. 776 Discussion Different techniques were employed in previous attempts at obtaining mammalian embryos with the use of spermatids instead of spermatozoa. The ability of spermatid nuclei to form male pronuclei and to participate in syngamy was first demostrated by Ogura and Yanagimachi (1993), who injected isolated spermatid nuclei into the cytoplasm of hamster oocytes. Because none of the zygotes produced by this method developed beyond the two-cell stage, these authors subsequently abandoned the direct injection technique and gave preference to a less invasive one, based on electrofusion of oocytes with whole spermatids previously deposited in the perivitelline space (Ogura et al, 1993). The latter technique was tested in the mouse and hamster models; however, as with the direct nuclear injection technique, none of the resulting zygotes cleaved more than once. Notwithstanding, some of the two-cell embryos produced by electrofusion of mouse oocytes with spermatids were viable because normal young were bom after transfer of such embryos into the oviduct of recipient mice (Ogura et al, 1994). Birth of normal progeny has also been reported in the rabbit with the use of a method based on direct intra-ooplasmic injection of isolated spermatid nuclei (Sofikdtis et al, 1994), quite similar to that originally used by Ogura and Yanagimachi (1993) in the hamster. This method has reportedly been applied to humans in 93 infertile couples at several American and Greek clinics and resulted in four pregnancies of which, however, all were terminated by spontaneous abortion (Hannay, 1995). The method described in this study is quite different from the above ones and is based on a recent modification of the technique of ICSI using mature spermatozoa (Tesarik and Sousa, 1995a). The direct intracytoplasmic injection of spermatids was given preference over electrofusion because the current experience with ICSI does not suggest any significant adverse effect of the injection trauma on the viability of oocytes and future embryos. However, the electrical pulse delivered to mouse oocytes to trigger oocyte fusion with the spermatid, applied in the presence of micromolar concentrations of Ca 2+ and Mg 2+ ions, has also been shown to be essential for mouse oocyte activation (Ogura et al, 1993). Thus, preliminary experiments with aged human oocytes from unsuccessful FVF attempts were carried out in our laboratories to find out how to activate human oocytes after spermatid injection. Oocyte activation after ICSI has been shown to be essentially dependent on the release of an as yet unidentified cytosolic oocyte-activating factor from the injected spermatozoon (Tesarik and Testart, 1994; Tesarik et al, 1994). During normal fertilization, this factor probably acts in concert with another oocyte-activating mechanism which involves an interaction between the cell surfaces of both gametes (Tesarik and Sousa, 1994). However, in the absence of the contact between the gamete cell surfaces, the oocyte activation process relies exclusively on the cytosolic sperm factor and needs to be boosted by a massive influx of extracellular Ca 2+ into the oocyte; this massive Ca 2+ influx can be mediated either by the action of a Ca 2+ ionophore (Tesarik and Testart, 1994;

Laboratory aspects of spermatid Injection Table I. The development of different types of nuclei in human oocytes fertilized by spermatid injection Zygote no. Type of spermatid Number and type of nuclei detected at different times after injection 10-12 h 16-18 h 22-24 h Cleavage* 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 no Cleavage was assessed 2 days after injection. Tesarik and Sousa, 1995b) or by the hydrodynamical sucking effect produced by vigorous aspiration of oocyte cytoplasm (Tesarik and Sousa, 1995a). Both of these methods were also equally effective at promoting oocyte activation after spermatid injection (Tesarik, unpublished observation), and the latter method was given preference to the former for the therapeutic ROSI and ELSI application. The injection of whole spermatids rather than isolated spermatid nuclei was considered preferrable for three reasons. Firstly, the sperm-derived oocyte-activating factor is supposed to be present in the cytoplasmic cell compartment rather than within the nucleus (Swann and Ozil, 1994) although it was reported to associate subsequently with blastomere nuclei (Kono et ai, 1995). We are now examining the possibility that this factor also associates with human spermatid nuclei. If this is not the case, the removal of most of the spermatid cytoplasm is likely to lead either to a lack or to abnormalities of oocyte activation. Such abnormalities may still be compatible with fertilization but may jeopardize further development of the embryo. Secondly, the procedure of spermatid nuclei isolation may lead to the loss of spermatid centrosomal material. In humans, as in most other animal species with the exception of mouse and some other rodents, paternally inherited centrosomes are required for normal nucleation of zygote microtubuli to form an aster which is involved in pronuclear co-localization and organization of subsequent mitotic spindles of the early embryo (Sathananthan et at, 1991; Palermo et al, 1994; Schatten, 1994; Van Blerkom and Davis, 1995). Thirdly, the procedure of nucleus isolation may cause damage to the nuclear envelope and thus facilitate partial loss of intranuclear material. Some of these factors may have been at the origin of the spontaneous abortions reported after injection of isolated spermatid nuclei into human oocytes (Hannay, 1995). Injection of whole spermatids into oocyte cytoplasm has also been used by Fishel et al (1995), who achieved an ongoing pregnancy which now approaches term (Fishel, personal communication). The time course and morphology of the fertilization process after spermatid injection showed slight differences from what is normally seen after IVF or after ICSI with mature spermatozoa. The development of pronuclei did not appear to be delayed as compared with injection of mature spermatozoa (Nagy et al., 1994). Interestingly, a single syngamy nucleus was observed in most of these initially two-pronucleated zygotes when they were inspected again at 16-18 h after spermatid injection. The larger size of the syngamy nuclei and the number of the nucleolar precursors, which was nearly double that observed in a pronucleus, was consistent with the notion that the syngamy nuclei arose from fusion between the two pronuclei. The presence of diploid syngamy nuclei in some human zygotes developing after ICSI was reported recently (Levron et al, 1995). It is possible that the fusion between both pronuclei is a normal feature of human development, but the resulting syngamy nucleus may have a very short lifespan and disappear rapidly as the zygote enters the mitotic phase of the first cleavage division. Accordingly, the finding of a syngamy nucleus is infrequent when oocytes have been fertilized with mature spermatozoa. On the other hand, 777

J.Tesarik and C-Meudoza the use of sperm precursor cells for fertilization may lead to a relative insufficiency of some yet unknown controlling elements needed for the syngamy nucleus to pass through the mitosis entry checkpoint Anyway, the prolongation of syngamy does not delay development because all of the zygotes in which a syngamy nucleus was observed at 16-18 h after spermatid injection cleaved by the next day. If oocytes are not examined for the presence of pronuclei as early as 10-12 h after spermatid injection, the stage of two pronuclei may be easily missed and only one syngamy nucleus may be found. Thus, care must be taken not to confound this diploid nucleus with a haploid female pronucleus, which could lead to elimination of zygotes with potentially normal developmental competence but misinterpreted as parthenogenetically activated oocytes. Sequential appearance and breakdown of both pronuclei followed by cell division, a condition described to be associated with a history of recurrent hydatidiform mole and supposed to lead to the development of androgenetic diploids (Edwards et al., 1992), is also to be distinguished from the prolonged syngamy stage. The syngamy nucleus is slightly larger than a pronucleus, but this difference is not striking and may be overlooked. The number of nucleolar precursors is thus an additional useful element of distinction between the two types of nucleus. On average, nine big nucleolar precursor bodies (similar in size to fully developed nucleoli of embryo blastomeres) were usually observed in a syngamy nucleus and only four to five in a pronucleus. It has to be stressed, however, that small nucleolar precursor bodies should not be taken into account because they are numerous at early stages of pronuclear development and their fusion subsequently gives rise to the big nucleolar precursors (Tesarik and Kopecny, 1989). The finding of one normal-sized and one small condensed pronucleus in one ROSI and one ELSI zygote was similar to the observation by Ogura et al. (1993) after electrofusion of spermatids with oocytes in the mouse. It is possible that this abnormality can result from an injection by error of a non-spermatogenic cell into the oocyte. In this study, however, this condition developed in two zygotes, one of which was injected with a spermatid and the other with an one. At least in the latter case such an error in cell identification is highly improbable because the distinction of an spermatid does not pose a major problem. In conclusion, this is a pilot study describing the technique of spermatid injection into oocytes by which the first term pregnancies have been achieved in humans. The fertilization process showed some special features which, however, did not appear to compromise further development. The knowledge of these features is important for a correct evaluation of fertilization outcomes after spermatid injection. Studies are in progress to determine whether some developmental stages of spermatids are more suitable for use in assisted conception than others and whether the technique can be further improved. For the time being, many questions concerning the unusual variability of the success rates and the potential hazards of this novel technique of assisted reproduction remain open. Until more information about these issues is available, this technique should be applied with extreme caution and only after making sure that the patients treated are fully aware of its experimental nature. Acknowledgements Data presented in this study are derived from treatment cycles performed at the American Hospital of Paris (Neuilly sur Seine, France), at the Instituto Valenciano de Infertilidad (Valencia, Spain) and at the Centro Hispalense de Procreacion Asistida (Sevilla, Spain). We are grateful to all workers of these centres who contributed to the success of this collaborative work. References Edwards, R.G., Crow, J., Dak, S., Macnamee, M.C., Hartshorne, CM. and Brinsden, P. (1992) Pronuclear, cleavage and blastocyst histories in the attempted preimplantation diagnosis of the human hydatidiform mole. Hum. Reprod, 7, 994-998. Edwards, R.G., Tarin, JJ., Dean, N., Hirsch, A. and Tan, S.C. (1994) Are spermatid injections into human oocytes now mandatory? Hum. Reprod, 9, 2217-2219. Fishel, S., Green, S., Bishop, M. et al (1995) Pregnancy after intracytoplasmic injection of spermatid. Lancet, 345, 1641-1642. Hannay, T. (1995) New Japanese IVF method finally made available in Japan. Nature Med, 1, 289-290. Kimura, Y. and Yanagimachi, R. (1995) Mouse oocytes injected with testicular spermatozoa and spermatids can develop into normal offspring. Development, 121, 2397-2405. Kono, T., Carroll, J., Swann, K. and Whittingham, D.G. (1995) Nuclei from fertilized mouse embryos have calcium-releasing activity. Development, 121, 1123-1128. Levron, J., Munne, S., Willadsen, S., Rosenwaks, Z. and Cohen, J. (1995) Male and female genomes associated in a single pronucleus in human zygotes. Biol. Reprod., 52, 653-657. Nagy, Z.P., Liu, J., Joris, H., Devroey, P. and Van Steirteghem, A.C. (1994) Time-course of oocyte activation, pronucleus formation and cleavage in human oocytes fertilized by intracytoplasmic sperm injection. Hum. Reprod, 9, 1743-1748. Ogura, A. and Yanagimachi, R. (1993) Round spermatid nuclei injected into hamster oocytes form pronuclei and participate in syngamy. Biol. Reprod., 48, 219-225. Ogura, A., Yanagimachi, R. and Usui, N. (1993) Behaviour of hamster and mouse spermatid nuclei incorporated into mature oocytes by electrofusion. Zygote, 1, 1-8. Ogura, A., Matsuda, J. and Yanagimachi, R. (1994) Birth of normal young after electrofusion of mouse oocytes with spermatids. Proc. NatL Acad Set USA, 91, 7460-7462. Palermo, G., Joris, H., Devroey, P. and Van Steirteghem, A.C. (1992) Pregnancies after intracytoplasmic injection of a single spermatozoon into an oocyte. Lancet, 340, 17-18. Palermo, G., Munne\ S. and Cohen, J. (1994) The human zygote inherits its mitotic potential from the male gamete. Hum. Reprod, 9, 1220-1225. Sathananthan, A.H., Kola, I., Osbome, J. et al (1991) Centrioles in the beginning of human development. Proc. NatL Acad. Set. USA, 88, 4806-4810. Schatten, G. (1994) The centrosome and its mode of inheritance: the reduction of the centrosome during gametogenesis and its restoration during fertilization. Dev. Biol, 165, 299-335. Silber, SJ. (1994) The use of epididymal sperm in assisted reproduction. In Tesarik, J. (ed.), Male Factor in Human Infertility. Ares-Serono Symposia, Rome, pp. 335-368. Silber, SJ. and Lenahan, K. (1995) Sertoli cell surgery: spermatid retrieval and ICSI in azoospermic patients with maturation arrest. Hum. Reprod. Update, 1, No. 6, CD-ROM, item 26, (video). Sofikitis, N.V., Miyagawa, I., Agapitos, E. et al. (1994) Reproductive capacity of the nucleus of the male gamete after completion of meiosis. J. Assist. Reprod Genet., 11, 335-341. Swann, K. and Ozil, J.P. (1994) The dynamics of the calcium signal that triggers mammalian egg activation. InL Rev. Cytol, 152, 183-222. Tesarik, J. (1993) Subzonal sperm insertion with aged human oocytes from failed in-vitro fertilization attempts: fertilization results and some applications. Hum. Reprod, 8, 1245-1250. 778

Laboratory aspects of spermatid injection Tesarik, J. and Kopecny, V. (1989) Development of human male pronucleus: ultrastructure and timing. Gamete Res., 24, 135-149. Tesarik, J. and Sousa, M. (1994) Comparison of Ca 2+ responses in human oocytes fertilized by subzonal insemination and by intracytoplasmic sperm injection. Fertil. SteriL, 62, 1197-1204. Tesarik, J. and Sousa M. (1995a) Key elements of a highly efficient intracytoplasmic sperm injection technique: Ca 2+ fluxes and oocyte cytopiasmic dislocation. FertiL SteriL, 64, 770-776. Tesarik, J. and Sousa, M. (1995b) More than 90* fertilization rates after intracytoplasmic sperm injection and artificial induction of oocyte activation with calcium ionophore. FemL SteriL, 63, 343 349. Tesarik, J. and Testart, J. (1994) Treatment of sperm-injected human oocytes with Ca 2+ ionophore supports the development of Ca 2+ oscillations. BioL Reprod., 51,385-391. Tesarik, J., Sousa, M. and Testart, J. (1994) Human oocyte activation after intracytoplasmic sperm injection. Hum. Reprod, 9, 511-518. Tesarik, J., Mendoza, C. and Testart J. (1995) Viable embryos from injection of spermatids into oocytes. N. Eng. J. Med., 333, 525. Tesarik, J., Rolet, F., Brami, C. et al. (1996) Spermatid injection into human oocytes: n. Clinical application in the treatment of non-obstructive azoospermia. Hum. Reprod., 11, 780-783. Van Blerkom, J. and Davis, P. (1995) Evolution of the sperm aster after microinjection of isolated human sperm centrosomes into meiotically mature human oocytes. Hum. Reprod., 10, 2179-2182. Vanderzwalmen, P., Lejeune, B., Nijs, M. et al. (1995) Fertilization of an oocyte microinseminated with a spermatid in an in-vitro fertilization programme. Hum. Reprod, 10, 502-503. Received on October 27, 1995; accepted on January 24, 1996 779