Review Impact of intracytoplasmic sperm injection on the activation and fertilization process of oocytes

Transcription

1 RBMOnline - Vol 3. No Reproductive BioMedicine Online webpaper 2001/194 on web 14 September 2001 Review Impact of intracytoplasmic sperm injection on the activation and fertilization process of oocytes Dr Michael Ludwig was born in 1968 in Germany and concluded university in He is Gynaecologist and Obstetrician at the University Hospital Lübeck and is running the Division of Reproductive Medicine and Gynaecologic Endocrinology at the Department of Gynaecology and Obstetrics. Dr Ludwig has published several papers in different fields of reproductive medicine, especially assisted reproductive techniques. His special interests lie in the field of evidence-based reproductive medicine and epidemiological studies in assisted reproductive treatment. Dr Michael Ludwig Michael Ludwig 1, Annika K Schröder and Klaus Diedrich Department of Gynecology and Obstetrics, University Clinic Hospital Ratzeburger Allee Lübeck Germany 1 Correspondence: Tel. +49 (0) ; fax: +49 (0) ; Ludwig_M@t-online.de Abstract Intracytoplasmic sperm injection (ICSI) is used worldwide to treat preferentially severe cases of male factor infertility. In this review, data regarding the processes of oocyte activation and fertilization in non-assisted conception, conventional IVF and ICSI are discussed. The second messenger calcium shows a typical pattern after ICSI, which is different from that after subzonal insemination (SUZI), which is closer to the conditions of normal fertilization. The onset of calcium spikes is delayed. Sometimes a monophasic calcium pattern, in animals typical for parthenogenetic activation, is observed despite normal subsequent oocyte activation. Furthermore, the frequency of spikes is higher after SUZI, and only one phase instead of two is observed after ICSI for the second onset of calcium release. These alterations may be explained by the differences in oocyte activation after ICSI, since no oolemma-sperm contact is present. Sperm decondensation also follows another pattern after ICSI: as long as residuals of the acrosome are present on the sperm head, no sperm decondensation takes place at that site. Therefore, decondensation is delayed and pronucleus formation, especially that of the male pronucleus, takes longer after ICSI as compared with conventional IVF. Since studies have shown that gonosomes are located preferentially in the apical part of the sperm nucleus, this was proposed to be an explanation for a higher incidence of gonosomal aberrations in offspring after ICSI. However, other explanations, taking clinical data like the background risk of the parents into account, can also be offered for this phenomenon. These alternative theories are more likely to be associated with a slight instead of a frank increase in gonosomal aberrations. The inheritance of paternal mitochondrial DNA seems not to be a problem after ICSI, as shown by different studies. Mitochondrial DNA can be demonstrated in embryos after conventional IVF as after ICSI up to the blastocyst stage but not in children born after ICSI. Finally, lesion of the meiotic spindle by the ICSI procedure seems not to be a problem when data from different studies are taken into account. As assumed also at the beginning of the ICSI era, the meiotic spindle is almost always located in an area of <90 deviation from the polar body axis. Therefore, intrusion of the microinjection needle at the 90 position might not endanger the spindle apparatus. To conclude, several studies using different approaches might show differences in the oocyte activation pattern, the choreography of fertilization and pronucleus formation after ICSI. However, this different pattern does not necessarily mean that ICSI per se is a problem for embryonal development. The different pattern can be explained by the fact that ICSI uses another means of oocyte entry than the normal fertilization process. The clinical data of a high fertilization, cleavage and implantation rate, and especially the data from newborn babies, show that ICSI is a reliable procedure. 230 Keywords: calcium, fertilization, ICSI, meiotic spindle, mitochondrial DNA Introduction Intracytoplasmic sperm injection (ICSI) is performed about 100,000 times per year worldwide. It is estimated that more than 100,000 children have been born following an ICSI procedure so far. Even if no study has shown an increased malformation risk after ICSI up to the present (Ludwig and Diedrich 1999), several arguments have been raised against ICSI because of a possible increased risk of this technique regarding the developmental competence of the treated oocytes. It was proposed that ICSI per se would lead to abnormal development of oocytes, abnormal fertilization, and abnormal embryo cleavage. The slightly increased risk of gonosomal aberrations was said to be a first sign of such abnormalities (Bonduelle et al., 1998). This review was undertaken to collect data from the literature

2 that deal with the fertilization process after ICSI. Theories about abnormal steps during that process will be put against facts, confirmed or compared with data from animal experiments using ICSI. Intracellular calcium concentration: mediator of initial activation of the oocyte During the fertilization process, not only genetic material will be brought into the oocytes by the spermatozoa. A cascade of biochemical and metabolic events is triggered and contributes to activation of the oocyte. The missing or disturbed oocyte activation is discussed as the main cause of fertilization failure following an ICSI procedure (Sousa and Tesarik, 1994). This is different from IVF, since the spermatozoon is brought directly into the cytoplasm, if ICSI is done in the correct way. In our own analysis of cases with fertilization failure after ICSI, it was shown that in cases with fertilizable spermatozoa, an oocyte factor was the most probable cause of this observation (Ludwig et al., 1999b). Receptors of the oolemma, but also intracellular calcium (Ca 2+ ) concentration and phosphorylation/ dephosphorylation of certain enzymes, play a central role in signal transduction during oocyte activation (Tesarik, 1998). Abnormalities of oocyte activation by delayed onset and a disturbed course of concentration changes of the calcium signal will lead to fertilization failure. Some authors even believe these abnormalities to be involved in the occurrence of chromosomal abnormalities or fetal malformations (Tesarik, 1995). Activation of the oocyte during non-assisted fertilization The cell cycle of the maturing oocyte is arrested in metaphase II of the second meiotic division. This is due to high activity of the metaphase promoting factor (MPF) (Ludwig et al., 1998). Inactivation of MPF allows the female nucleus to finish the second meiotic division and protects the male nucleus to enter the metaphase too early, due to MPF of the oocyte. Activation of the oocyte seems to be triggered by interaction of the spermatozoon with the receptors on the oolemma (Tesarik et al., 2000). Animal experiments have shown that these receptors are coupled to G-proteins or tyrosine kinases (Jaffe, 1992; Abassi and Foltz, 1994). This trigger leads to the first, sperm-induced, increase in intracellular Ca 2+ concentration. The development of ICSI, and the observation that ICSI results in a similar calcium response (Taylor et al., 1993; Tesarik, 1994), leads to the assumption, that surface receptors are not indispensable for sperm-induced oocyte activation. A second theory, involving a so-called soluble sperm factor, was proposed. The existence of such an oocyte-activating substance in spermatozoa would explain why oocyte activation can be achieved following ICSI. On the other hand, the absence of contact of the spermatozoon and oolemma would explain slight changes in the oocyte activation pattern after ICSI as compared to SUZI (subzonal insemination), where sperm-oolemma contact is not avoided (Tesarik and Sousa, 1994). Therefore, one can conclude that there must be two methods of oocyte activation, a sperm-oolemma mediated one and a second one, triggered by a soluble sperm factor (Tesarik, 1998). calcium serves as a second messenger during these activation processes. Inactivation of MPF is a main step during this course. Furthermore, oocyte activation leads to the cortical reaction, which leads to the polyspermia blockade. The involvement of calciium as a second messenger during oocyte activation has been described in several species. Ca 2+ comes mainly from intracellular stores including endocytoplasmic reticulum. Two kinds of Ca 2+ channels have been described: the inositol-1,4,5-triphosphate (IP3) receptor and the ryanodine receptor. Both channels open in response to an increase in calcium concentration in the direct neighbourhood. This phenomenon is called Ca 2+ induced-calcium release (CICR). The release of Ca 2+ then leads to a signal which is directed to the oolemma and results in an opening of the Ca 2+ channels in the membrane, resulting in a Ca 2+ influx (capacitative calcium entry) (Tesarik and Sousa, 1996). There is an ongoing discussion about the exact ways in which the spermatozoon leads to intracellular Ca 2+ release. It is possible that the spermatozoon acts directly on calcium stores, induces Ca 2+ release and subsequently CICR. Phosphoinositide of the oolemma is said to be hydrolysed by G-protein mediated activation of the phosphoinositide-specific phospholipase C, which leads to production of IP3. IP3 acts on IP3-sensitive Ca 2+ channels, resulting in calcium release. Those IP3 sensitive calcium stores have been shown in human oocytes (Sousa et al., 1996). Diacylglycerole (DAG), which is also produced during the hydrolysation of phosphoinositide, leads to down regulation of IP3-mediated Ca 2+ release by the action of protein kinase C (PKC). Consequently, the concentration of intracellular free Ca 2+ will decrease again to basal levels. This hypothesis might explain the rapid increase of intracellular Ca 2+ concentration, which is called a Ca 2+ spike. However, other mechanisms such as receptors on the oolemma and its ligands, sperm soluble factors, IP3 and ryanodine sensitive calcium stores may also be involved. The observation that a Ca 2+ spike is induced by injection of sperm extracts into an oocyte (Swann, 1994) does not prove that a sperm soluble factor is the only physiological trigger of oocyte activation. Until now, it has not been possible to calculate the concentration of this factor exactly; therefore, it has been questioned whether, in fact, the necessary dose of sperm factor could be produced by a single spermatozoon. Furthermore, the method of microinjection could possibly change the reaction of the oocyte to the injected sperm factor. These considerations do not exclude the existence of such a factor, but show that simplification and propagation of a single oocyte activating mechanism seems not to be true (Tesarik, 1998). It is not yet clear how the Ca 2+ signal is transmitted to other elements of the oocyte-activating cascade. Changes in protein phosphorylation, such as calmodulin-dependent protein kinase II and PKC, seem to be involved. Phosphorylation of MPF and of a cytostatic factor (CSF), which is responsible for a support of ongoing MPF activity, play a role in MPF regulation (Watanabe et al., 1991; Lorca et al., 1993). In contrast to other species, such as fish and frogs, in which the fertilizing spermatozoon only induces a single Ca 2+ spike, multiple repetitive short Ca 2+ spikes are typical for the Ca

3 232 signal of oocyte activation in the human oocyte. These are called Ca 2+ oscillations (Igusa and Miyazaki, 1986; Sun et al., 1992). Cuthbertson (1985) was the first to describe the Ca 2+ signal in mouse oocytes, which were activated experimentally by incubation using 12-O-tetradecanoyle-phorbolacid (tumour promoting factor, TPA). TPA imitates endogenous DAG and is able to stimulate oocytes by direct activation of PKC. After a period of min following incubation with TPA in a concentration of 200 ng/ml in six out of 12 oocytes, Ca 2+ oscillations were observed. With a higher TPA concentration (1 mg/ml) oscillations started earlier in three out of four oocytes. All oocytes that showed oscillations later developed two pronuclei but no second polar body. The frequency of oscillations was s. A slow increase in extracellular Ca 2+ concentration, by 5-fold, stopped oscillations within a few seconds. Following a decrease in extracellular Ca 2+ concentration, oscillations started again in some but not all oocytes. Furthermore, Ca 2+ concentrations were measured during normal fertilization in four oocytes. Normal fertilization was defined as the occurrence of two pronuclei and a second polar body. The authors observed Ca 2+ oscillations that went on for 4 h. Some Ca 2+ spikes had a duration of min, others a duration of s. The latter had a lower maximum peak level. Cuthbertson (1985) postulated that the Ca 2+ signal following TPA incubation, as well as that following normal fertilization, is able to activate an oocyte. However, the extrusion of a second polar body depends on a certain time schedule and a certain number of oscillations (Cuthbertson, 1985). In human oocytes, the sperm-induced Ca 2+ oscillations can go on for 5 6 h (Taylor et al., 1993). The frequency of single Ca 2+ spikes during these oscillations varies between 2 and 5 min and shows great inter-oocyte variations. Within the same oocyte the frequency is quite constant during the whole duration of the oscillations (Taylor et al., 1993; Tesarik and Sousa, 1994). In most mammals, including humans, the first Ca 2+ signal has a certain shape, which differs from the consecutive spikes. It consists of three to six rapidly occurring Ca 2+ spikes. Thereafter, spikes with a lower frequency follow for the remaining time of the oscillation period (Tesarik and Sousa, 1994; Tesarik and Mendoza, 1999). With regard to the duration of the oscillations, it becomes clear that the spermatozoon is not only the initial trigger (Tesarik and Mendoza, 1999), but also brings in the signal which modifies the involved elements (pumps, channels, stores) for regulation of the consecutive signals. This modification allows an exchange of Ca 2+ between the different cell compartments and the putative signal, subsequently referred to as the oscillator (Tesarik and Mendoza, 1999). The differentiation of a trigger and an oscillator as two distinctive physiological entities is based on the observation that ICSI sometimes leads to a single sperm-dependent Ca 2+ spike or brings the oocyte into a situation in which it is capable of developing oscillations but still needs a further stimulus. Those oocytes can be stimulated using a calcium ionophoreinduced Ca 2+ increase (Tesarik and Testart, 1994). The first situation can be explained by the isolated action of a trigger but without the presence of an oscillator, the second by the presence of an oscillator but not a natural trigger. Cytosolic factors of the spermatozoon, from which only one, oscillin (Parrington et al. 1996), could be identified so far, apparently fulfil the function of such an oscillator. Experimental data showed that Ca 2+ oscillations are coordinated by periodic release and resorption of Ca 2+ from IP3-sensitive calcium stores in the cortex and subcortex of the oocyte, as well as ryanodine-sensitive stores, which are predominantly localized in the ooplasma (Sousa et al. 1996; Tesarik and Sousa 1996). Based on this model, the capacity of an oocyte to keep the oscillations going is due to the different sensitivity of the two calcium stores towards CICR: the initial, limited Ca 2+ release from cortical stores, which are assumed to be IP3 sensitive, causes CICR from cytoplasmic, ryanodine-sensitive stores. Most Ca 2+ is released from the latter. Since the cortical stores have a higher sensitivity towards CICR, the area of increased Ca 2+ concentrations is developing from the initial focus in the cortex towards the periphery of the oocyte, before it migrates to the central ooplasm. In the centre of the ooplasm, the concentration remains increased for a longer time period (Tesarik and Sousa 1995). Another theory only involves the existence of two functionally different IP3 sensitive stores. One should be responsible for the initial Ca 2+ influx and the other for the regenerative Ca 2+ spikes (Berridge, 1996). The function of this model principally is analogous to the other theory; however, for the second one instead of a ryanodine-sensitive store a special IP3-sensitive Ca 2+ store is involved. The sensitivity of the different Ca 2+ stores may be modified by the sperm-induced oscillator, which also could explain the migration of the Ca 2+ wave through the different compartments of the oocyte, which is different from that induced by an artificial stimulus (Tesarik and Sousa, 1995). Oocyte activation by ICSI The experiences with ICSI show that injected spermatozoa can activate an oocyte and can induce Ca 2+ oscillations, which are similar to those after non-assisted fertilization. Therefore, despite the absence of a sperm-oolemma contact, the sperminduced oscillations are sufficient to activate the oocyte. It seems that by injection of a single spermatozoon, the function of the trigger will be fulfilled by the artificial Ca 2+ influx, which is due to the micromanipulation itself (pseudotrigger) (Tesarik et al., 2000). Activation rate of the oocytes can be increased by mechanical manipulation of the spermatozoon before injection, to support the release of the sperm-soluble activation factor (Dozortsev et al., 1995). However, the pseudotrigger alone is not sufficient to induce activation, if a spermatozoon is not injected at the same time (Tesarik and Sousa, 1994). This could be shown by a series of experiments by Tesarik et al. These authors looked for changes in the intracellular Ca2+ concentration in human oocytes, which were injected using medium without a spermatozoon as well as in oocytes which were treated by a sham injection with a needle without a lumen (Tesarik et al., 1994). In a third group, oocytes were injected with a spermatozoon in Ca 2+ -free medium. The initial short (<2 min, maximum after s) increase in Ca 2+ concentration was observed after ICSI directly after insertion of the microinjection needle. The same reaction could be induced by injection of medium without a spermatozoon or by perforation of the oolemma with a non-luminated microinjection needle. If Ca 2+

4 Fluorescence intensity free medium was used, no initial Ca 2+ spike occurred (Tesarik et al., 1994). The extent of the initial Ca 2+ influx seems to depend on the force of oolemma aspiration to break the plasma membrane during ICSI. However, the degree of ooplasma aspiration cannot be precisely quantified not even under experimental conditions (Tesarik and Sousa, 1995). From these experiments, it became clear that the development of the initial spike depends only on the mechanical manipulation of the oocyte in Ca 2+ containing medium, and not from the injection of a spermatozoon. After this initial, sperm-independent spike there were no more oscillations in those oocytes in which only medium was injected. After a normal ICSI procedure, a silent phase was observed over a period of 2 12 h with basal Ca 2+ concentrations, similar to those before the injection. After that, a new increase in Ca 2+ concentration occurred with two different patterns. Either an oscillatory (9/35 oocytes) or a nonoscillatory pattern (26/35) was observed. The non-oscillatory pattern showed a slow increase in Ca 2+ concentration with a plateau after 10 min. This plateau was maintained for a period of min, before Ca 2+ concentrations fell to the initial values (Figure 1). The oscillatory pattern consisted of a series Time after ICSI Figure 1. Pattern of Ca 2+ spikes in the second phase after ICSI (non-oscillatory pattern). Data according to Tesarik et al. (1994). of Ca 2+ spikes each of 20 s duration and with a pause of 1 5 min between two spikes. This pattern was recorded over a period of min (Figure 2). Those oocytes that did not show a new Ca 2+ increase after the initial Ca 2+ spike did not develop to the pronuclear stage. Five out of 13 oocytes and three out of five oocytes with the non-oscillatory and oscillatory pattern respectively did develop two pronuclei and extruded a second polar body. Interestingly, Ozil published results on rabbit oocytes which suggested that only those cells which showed an oscillatory pattern could develop up to the blastocyst stage thereafter. In this study, rabbit oocytes were activated parthenogenetically by an electric stimulus (Ozil, 1990). In another study by Tesarik and Souza, the oscillatory pattern after SUZI was compared with those observed after ICSI (Tesarik and Sousa, 1994). The second Ca 2+ signal following the initial spike after a silent phase, was biphasic after SUZI. The first phase was characterized by three to six consecutive Ca 2+ spikes of s duration. The interval between these spikes was <1 min. Ca 2+ concentrations were above basal concentrations for the whole period. This first phase was followed by a series of Ca 2+ spikes of similar shape and duration, but with a lower frequency (5 15 min). Frequency and amplitude seemed to be higher at the beginning of this second phase and to decrease up to the end. In this second phase, Ca 2+ concentrations decreased to basal levels between the single spikes. The oscillatory phase took 1 3 h (Figure 3). Compared with the data obtained after SUZI, there was a delayed initiation of the initial Ca 2+ signal after ICSI: after 2 12 h after ICSI, and after 50 min to 9 h after SUZI. Compared to SUZI, there were no two phases of oscillations after ICSI. Furthermore, in those oocytes after ICSI with an oscillatory pattern, there was a higher frequency of Ca 2+ spikes during the whole time period as compared to SUZI (1 min). One has to keep in mind that SUZI is supposed to be suitable as a model of normal, non-assisted fertilization, since an oolemma-sperm contact is involved. Accepting this model and comparing the Ca 2+ pattern after SUZI and the data from Ozil (1990) with those after ICSI, the oscillatory pattern has to be called a 'normal pattern and a non-oscillatory pattern a 'parthenogenetic one (Richords and White, 1993; Tesarik et al., 1994). Fluorescence intensity Fluorescence intensity Time after ICSI Figure 2. Pattern of Ca 2+ spikes in the second phase after ICSI (oscillatory pattern). Data according to Tesarik et al. (1994) Time after ICSI Figure 3. Pattern of Ca 2+ spikes in the second phase after SUZI (oscillatory pattern). Data according to Tesarik and Souza (1994). 233

5 234 In other species such as leeches and nematodes, abnormalities of oocyte activation, measured by the amplitude of the Ca 2+ signal, a delayed start of Ca 2+ oscillations and a non-oscillatory pattern instead of an oscillatory one, are associated with arrest of the fertilization process. The delayed start of the oscillations could lead to delayed or incomplete inactivation of MPF and thereby to another course of anaphase II of the meiotic division. Some authors propose that this might be responsible for the development of chromosomal abnormalities after ICSI (Tesarik, 1995). The fact that an oocyte can also be activated by injection of immature stages of spermatogenesis, such as round or elongated spermatids, shows that the oscillator function in these cells is present. Some results have shown that this function is located in the sperm nucleus. Tesarik et al. (2000) observed that oocyte activation was as efficient after injection of spermatid nuclei as after whole spermatids. On the other hand, oocytes which were only injected with cytoplasm of spermatids could not be activated (Tesarik et al., 2000). With these data regarding oocyte activation, it becomes clear, that in theory, genetic alterations may be induced by using immature spermatids. For example, the chromatin of round spermatids is less protected against the oocytes MPF. Therefore, theoretically, if stimulation by Ca 2+ oscillation is not sufficient to inactivate MPF, or if Ca 2+ oscillation starts with a delay, premature condensation of the chromatin and subsequently aneuploidies may occur (Tesarik, 1998). In fact, oscillation-promoting activity of the spermatozoon occurs only as early as in the round spermatid stage (Tesarik, 1998). On the other hand, it should be mentioned that data from animal experiments have raised concerns as to whether round spermatids are actuall capable of fertilizing oocytes. It is well known that in the rhesus monkey ICSI is possible and successful (Hewitson et al., 1998). However, microinjection of elongated spermatids has been shown to lead to fertilization, subsequent cleavage, pregnancy and birth but injection of round spermatids did not (Hewitson et al. 2000a, b). Recently, this was also shown in mice. In no case did injection of round spermatids induced oocyte activation, but elongated spermatid injection induced oocyte activation in almost all cases of (Yazawa et al., 2001). This supports the experience of different laboratories that injection of round spermatids does not lead to ongoing pregnancies in humans either. The number of pregnancies which have been reported to be the result of round spermatid injection, is very low (Ludwig et al., 1999a). In conclusion, the data presented in this section show that during normal fertilization, the combination of an oolemmasperm contact and a soluble sperm factor leads to Ca 2+ oscillations and to oocyte activation. Ca 2+ plays an important role as a second messenger during this process. However, even using ICSI and excluding the oolemma-sperm contact does allow fertilization by induction of a Ca 2+ influx during the microinjection procedure. This pseudotrigger subsequently leads to oocyte activation. The Ca 2+ pattern is similar to that observed under normal conditions, and abnormalities in oocyte activation using ICSI can be explained by a delayed onset of Ca 2+ oscillations and a change in the Ca 2+ release patterns. On the other hand, the high fertilization, cleavage and pregnancy rates using ICSI show that these alterations do not necessarily mean that the fertilization process itself is severely disturbed. It might be difficult to transfer results from animal studies to the fertilization process in humans. All theories about the influence of different activation patterns on the further embryonic development remain speculative. Decondensation of spermatozoa following ICSI Sutovsky et al. showed in 1996 in the rhesus monkey that decondensation of spermatozoa following ICSI is heterogenous, depending on the localization of the DNA in relation to the remaining parts of the acrosome and the sperm membrane. These structures remain at the zona pellucida and the oolemma during conventional IVF and normal fertilization, but could be shown in ooplasm after ICSI (Sutovsky et al., 1996). It was shown in the same year by electron microscopy that in the rhesus monkey, decondensation of spermatozoa occurs as early as 6 h after ICSI. This was the first time point at which oocytes were analysed in that study. The authors also demonstrated that decondensation started in the basal regions and was delayed in the apical part until that time point when the acrosome had totally disappeared (Hewitson et al. 1996). In a very sophisticated study, Bourgain et al. (1998) have analysed 18 human oocytes after ICSI at several time points following an ICSI procedure using electrone microscopy (15, 30, 60, 90 min, 2, 3, 4,, 8 h). Three further oocytes were analysed without ICSI and were used as controls. The authors described alterations, which could be due to the mechanical manipulation of the oolemma, e.g. membrane-bound vesicles and cytoplasmic membrane inclusions, sometimes associated with microvilli or small pinocytotic vesicles. The reaction of membrane-bound cortical granules appeared at about 90 min after the ICSI procedure. Acrosome reaction was observed as early as 15 min after ICSI, and decondensation started at 30 min and lasted until 3 h after ICSI. Since even at the start of male pronucleus formation, 4 h after ICSI, particles of the acrosome were still present at the pronuclear structures, one can conclude a delay in acrosome reaction from these data (Bourgain et al., 1998). Normally, during conventional IVF or normal, non-assisted fertilization, the sperm membrane is uniformly lysed, which leads to a homogenous decondensation process of the chromosomes. The apical part of the sperm nucleus stays condensed until the rest of the acrosome has disappeared. The results of Bourgain et al. (1998) were confirmed by recent studies of Hewitson et al. in the rhesus monkey model. The authors described a clear separation of condensed and decondensed parts until the end of male pronucleus formation. The non-decondensed parts were located in the apical parts of the sperm head, where clearly a perinuclear theca as well as the acrosome could be shown (Hewitson et al., 1999). In another study on human spermatozoa, Luetjens et al. (1999) recently showed, that the chromosomes in human sperm nuclei are not randomly distributed (Figure 4). They used chromosome patenting for chromosomes 18 and X. A study model with Xenopus oocyte extract, which leads to permeabilization and decondensation of spermatozoa, was applied. This is near to the natural situation of the decondensation process. Hybridization was possible in 80% for chromosome 18 (190/234 spermatozoa) and in 40% for

6 a. b. Figure 4. Human spermatozoa after in-vitro decondensation and chromosome painting. a. X chromosome (green), chromosome 18 (red); b. DIC-image overlayed with a fluorescence image stained for DNA with Hoechst (blue). Reprinted with permission from Elsevier Science (Luetjens et al., 1999). chromosome X (230/553 spermatozoa). It was shown that chromosome 18 was preferentially located in the basal part of the sperm nucleus, near the sperm tail (60%), but chromosome X preferentially near the apical part below the acrosome (60%) (Luetjens et al., 1999). Delayed decondensation and the preferential location of gonosomes in the apical part of the sperm nucleus has also been shown recently in the hamster oocyte test using human sperms. Chromosome X and Y were located within the sperm nucleus using chromosome painting. From 12 oocytes the gonosomes were decondensed, but in nine cases even 6 h after ICSI they were still condensed in the apical part (Terada et al., 2000). Also in this study, delayed lysis of the perinuclear theca was confirmed after ICSI (Terada et al. 2000). The group which did these studies proposed that their results might provide an explanation for the slightly increased rate of gonosomal aberrations: delayed decondensation of the apical parts of the sperm head leads to delayed decondensation of the gonosomes, which are preferentially located at this site. The presence of condensed chromosomes during the further steps of the fertilization process might then lead to a disturbed chromosome separation. On the other hand, one might assume that if this was really the case, the rate of gonosomal aberrations would have to be much higher. There are other hypotheses that also might explain this increased risk, e.g. the background risk of the prospective parents with an increased rate of disomic spermatozoa in cases of severe male factor infertility or an increased risk of gonosomal mosaics in mothers and fathers of children conceived after ICSI. These facts are more likely to be associated with the described slightly increased risk of gonosomal aberrations (Luetjens et al., 1999). In fact, the choreography of fertilization (Hewitson et al., 1999) might be another one after ICSI, but this does not necessarily mean that it leads to an increased risk of malformations or chromosomal aberrations. Pronucleus formation after ICSI Light as well as electron microscopy studies on human oocytes after ICSI have shown that the male pronucleus forms before the female one (Nagy et al., 1994; Bourgain et al., 1998). Furthermore, pronucleus formation after ICSI is principally faster as compared to conventional IVF (Nagy et al., 1998). However, when pronucleus formation after IVF and ICSI using bromodioxyuridine administration in the rhesus monkey are compared 16 h after insemination or sperm injection, it can be seen that formation was delayed in ICSI cases. This relates to the ongoing or delayed decondensation of the sperm nucleus, especially in the apical part. In IVF, however, at that time point bromodioxyuridine incorporation could be shown in both pronuclei (Hewitson et al. 1999). Also in the hamster oocyte test, using human spermatozoa, delayed replication of DNA, preferentially in male pronuclei, was observed. Compared to ICSI, oocytes after IVF showed a male and female pronucleus, which was bromodioxyuridine positive in 60% at 6 h post-insemination. On the other hand, in no oocyte after ICSI was a male pronucleus, and in only 40% was the female pronucleus, demonstrated to be bromodioxyuridine positive at that time point (Terada et al., 2000). Therefore, ICSI in an animal model seems not to be comparable to that in the human model, although it introduces quite interesting perspectives. If the results of the cited animal models were directly transferable to the human, there would be no explanation for the observation of Nagy et al. (1998), that not only is pronucleus formation advanced after ICSI, but also the first cleavage. 235

7 30% 24% 20% 27% a. b. IVF 10.8±7.6º 58% 35% 7% ICSI 56±27.5º c. 37% 32% 24% 7% 88% 1% 11% Figure 5. Localization of the meiotic spindle (shown by different shading) in relation to the polar body axis. Data presented from: hamster oocytes according to Silva et al. (1999) (a), rhesus monkey oocytes according to Hewitson et al. (1999) (b), vital human oocytes according to Wang et al. (1999) (c), and according to Hardarson et al. (2000) from in-vitro (d) and in-vivo matured (e) fixed human oocytes. The percentages show the rate of meiotic spindles, which are located in that region of the ooplasm. In (b), only a mean ± SD value for the localization of the meiotic spindle in relation to the 0 position of the polar body axis was given. 236 d. Lesion of the meiotic spindle by the ICSI procedure From the beginning of ICSI, the pioneers of that procedure knew that there might be the possibility of disturbing the spindle apparatus with the injection needle. Therefore, the introduction of the needle was always made at a 90 angle from the first polar body, since the meiotic spindle was assumed to be located in the direct neighbourhood of the polar body. In fact, survival and fertilization rates were the same, independent from microinjection when the polar body was localized at the 6 or 12 o clock position. However, the rate of embryos with <50% fragments was significantly higher in the group with the polar body at the 6 o clock position (83 versus 79%). This study analysed >600 oocytes per study arm (Nagy et al., 1995). In another analysis of 2924 oocytes after ICSI, there was no change in survival, fertilization and developmental rates, independent of whether the polar body was localized at the 6 or 12 o clock position (Dumoulin et al., 2001). However, different groups have published data regarding the correct localization of the meiotic spindle during an ICSI procedure. Silva et al. (1999) studied golden hamster oocytes using polarization microscopy. Compared to all other related studies, the largest deviation from the 0 degree axis (polar body axis) was described. From 30 oocytes the meiotic spindle had a deviation in 30% of up to 45, in 24% from 45 to 90, in 27% from up to 135 and in a further 20% up to 180 (Figure 5) (Silva et al., 1999). Hewitson et al. (1999) studied rhesus monkey oocytes to localize the meiotic spindle in relation to the first polar body. After injection of rhodamin-bound bovine tubuline, a mean deviation of the spindle axis from the polar body axis of 19.8 ± 23.3 (0 68, n = 19) was found. In a control group of human oocytes after in-vitro maturation, the mean deviation was 10.8 ± 7.6 ( , n = 3) and in those after failed fertilization after conventional IVF 10.8 ± 7.6 ( , n = 5). On the other hand, before ICSI in rhesus monkey oocytes, i.e. after mechanical-chemical denudation, the mean deviation was ~56.0 ± 27.5 ( , n = 8) (Hewitson et al., 1999). The authors concluded that by manipulation of the oocytes before e. ICSI, a more significant deviation of the polar body might occur as compared with oocytes which are not manipulated before the fertilization procedure. Another group analysed 112 denuded human oocytes before ICSI with immunohistochemistry by using tubulin antibodies. In 97 oocytes, the deviation of the spindle axis in relation to the polar body axis could be determined. The authors differentiated between in-vitro and in-vivo matured oocytes and described a mean deviation of 26.6 ± 3.3 after in-vitro maturation and of 41.7 ± 4.0 after in-vivo maturation, which was significantly different to each other (P < 0.005), as well as significantly different from a 0 position, i.e. no deviation (P < 0.05). More exact numbers are given in Table 1. The data show that in 100 and 93% of cases in the in-vitro and in-vivo groups respectively, the meiotic spindle was in the hemisphere of the polar body (Hardarson et al., 2000). Finally, Wang et al. published their results using polarization microscopy for the in-vivo localization of the meitoic spindle in human oocytes. In that study, they wrote that 'most' oocytes had the spindle located within 45 from the polar body axis (Wang et al., 2001). However, in an oral presentation during the annual meeting of the American Society of Reproductive Medicine, the same data were demonstrated and rates were shown, which are also included in Figure 1: in 88% the spindle axis was deviated maximally 45, in 11% between 45 and 90, and in only 1% more than 90 (Wang et al., 1999). To conclude, these publications show that by choosing an injection angle of 90, it is still possible to disturb the spindle apparatus, although this is highly unlikely. Only the data from Silva et al. show a higher risk, but it should be remembered that they were collected from a study in another species (hamster) (Silva et al., 1999). It is interesting, however, that the denudation process itself might induce a deviation from the 0 position, a finding that was proposed by one study (Hewitson et al., 1999). However, these data could not be confirmed by others (Hardarson et al., 2000). One can assume that the manipulation should involve as little stress as possible and should only be done by experienced technicians and biologist in clinical practice.

8 Table 1. Localization of the meiotic spindle in relation to the polar body axis in in-vivo and in-vitro matured oocytes before ICSI; data according to Hardarson et al. (2000) Deviation from the axis In-vivo matured oocytes In-vitro matured oocytes polar body (degrees) n % total Cumulative % n % total Cumulative % Total Table 2. Paternal mitochondrial DNA in embryos and blastocysts following IVF or ICSI. The table contain data on the rate of embryos, which show paternal mitochondrial DNA. The rate is calculated in relation to all embryos, which are analysed from that individual patient at the certain stage. Data according to, St John et al. (2000) Patient Embryo Treatment number 4-cell 8-cell Compacted Blastocyst 1 2/2 ICSI 2 3/3 ICSI 3 2/2 ICSI 4 5/5 IVF 5 0/2 0/2 IVF 6 1/1 3/6 IVF 7 0/7 IVF 8 0/2 ICSI Inheritance of mitochondrial DNA by ICSI Another main criticism is directed towards the possibility of transmission of paternal mitochondrial DNA by the ICSI procedure, a process, which is proposed not to be observed under natural conditions. Principally, in cases of natural conception uniparental inheritance of mitochondrial DNA is proposed. This fear might have a realistic background. Chan et al. recently showed that by using ICSI, additional foreign genetic material can be brought into the oocyte. In that study, DNA for the green fluorescent protein was coupled to spermatozoa, which were subsequently injected by ICSI in rhesus monkey oocytes. The marker remained in the ooplasm and was visualized in rising concentrations, beginning with the 4-cell stage up to the blastocyst stage. In contrast, following conventional IVF, the foreign genetic material remained at the oolemma (Chan et al., 2000). Several studies have analysed the possibility of such a transmission (Figure 6). The first study to cover that topic was published by a Swedish group. The authors looked for paternal mitochondrial DNA in six children after ICSI and did not find such transmission in any of them (Houshmand et al., 1997). In a multicentric approach, 27 children of 21 parents were analysed regarding the transmission of paternal mitochondrial DNA. In none of these children could paternal mitochondrial DNA be shown. Therefore, the authors concluded that even in ICSI families, the uniparental inheritance of mitochondrial DNA is the usual case (Danan et al. 1999). A very interesting paper was presented during the annual ESHRE meeting in Tours in 1999 on children after IVF, ICSI and spontaneous conception (Lamb et al., 1999). These data clearly showed that in three families after ICSI, no paternal mitochondrial DNA was found. However, in three children of the seven control families, paternal mitochondrial DNA was found in the placentas (personal communication). Using single cell PCR, Sbracia et al. tried to demonstrate paternal mitochondrial DNA in embryos after ICSI. In all embryos from the 2- up to the 8-cell stage paternal DNA was shown in all three couples. Furthermore, DNA was also shown in three out of seven blastocysts (Sbracia et al., 2000). Finally, St John et al. published data on the presence of paternal mitochondrial DNA in polyploid embryos after IVF and ICSI. Thirty-two embryos were evaluated, and the results are shown in Table 2. Following conventional IVF as well as following ICSI up to the 8-cell stage, but also following conventional IVF up to the blastocyst stage, paternal mitochondrial DNA was present in three cases (St John et al., 2000). Therefore the presence of paternal mitochondrial DNA seems to be a normal finding in embryos after IVF as well as after ICSI up to a certain developmental stage. In children born after ICSI, paternal mitochondrial DNA has not been demonstrated up to now. The arguments of the critics of ICSI do not reflect the reality. The impact of the data from Lamb et al. has to be evaluated further. The thesis of these authors is that perhaps after spontaneous conception, paternal mitochondrial DNA may persist in the peripheral ooplasm, which is later distributed to the trophoblast cells and to the placenta. In ICSI cases, however, the sperm is injected in the centre of the oocyte, which may be able to get rid of that kind of DNA (Lamb et al., 1999). Conclusion The reviewed data clearly show that ICSI provokes some quite special patterns in the process of oocyte activation during the fertilization process. This became apparent from the data of calcium release patterns as well as from those of sperm decondensation and pronuclear formation. 237

9 238 Figure 6. Plasmid transfer by ICSI. Rhodamine-labelled plasmid DNA binds avidly to (A) mouse, (B) bovine, and (C) rhesus monkey spermatozoa. Rhodamine-tagged DNA remains on the surface of micro-injected spermatozoa after ICSI: rhesus monkey spermatozoa micro-injected into (D) a rhesus monkey oocyte or (E) into a bovine oocyte. (A D) Blue = Hoechst DNA imaging. (F1 and F2) Labelled rhesus monkey spermatozoa during IVF. F1, rhodamine-tagged plasmid DNA is lost at the egg surface during IVF (arrows). The partially penetrated spermatozoon demonstrates the loss of exogenous DNA in the penetrated half (green arrow). F2, imaging of the same focal plane. (G) Detection of green fluorescent protein (GFP) expression in a 16-cell stage rhesus monkey embryo using anti-gfp monoclonal antibody (red) and Hoechst DNA staining (blue). (H) Live 4-cell and (I) blastocyst stage rhesus monkey embryos expressing GFP after transgenesis by ICSI using rhodamine-labelled plasmid DNA encoding the GFP gene bound to the injected spermatozoon. (A F) laser scanning confocal microscopy. (A, B and C) produced by overlaying images of 14 labelled spermatozoa and each individual image of a spermatozoon is an overlay of images taken at different focal planes. (G) was collected by digital low light level fluorescence imaging (Princeton CCD, differential interference contrast, Zeiss Axiophot). From Chan et al. (2000). European Society of Human Reproduction and Embryology. Reproduced by permisson of Oxford University Press/Human Reproduction.

10 Damage to the meiotic spindle is unlikely if ICSI is performed at 90 deviation, with the polar body at the 12 or 6 o clock position. The transmission of paternal mitochondrial DNA to the offspring after ICSI seems not to be a problem, despite the sperm mitochondria being transferred to the oocyte during microinjection of whole spermatozoa. The different patterns of oocyte activation and alterations during the fertilization process after ICSI as compared to patterns which are observed after conventional IVF or SUZI can be explained by the fact that (i) ICSI does not involve an oolemma-sperm contact and (ii) that no acrosome reaction takes place. Gerald Schatten s group called these alterations a different choreography of fertilization (Hewitson et al., 1999). This term illustrates that differences do not necessarily mean abnormalities in the subsequent processes like embryo cleavage, implantation and consequences for the children born. However, the processes observed after ICSI have also taught us a lot about the normal fertilization process. We should be aware of the differences but also should carefully watch clinical data of the children born. These data show up to now as a main point a slightly increased risk of gonosomal aberrations. These of course could be the result of this different choreography, but other explanations are also quite likely and can be better associated with an only slightly increased risk: if in fact a delay in the gonosomal decondensation is the cause, a much higher incidence of aberrations should be expected. However, further basic studies would be helpful, to learn more about the process of fertilization in the human and about the consequences of the ICSI procedure. References Abassi YA, Foltz KR 1994 Tyrosine phosphorylation of the egg receptor for sperm at fertilization. Developmental Biology 164, Berridge MJ 1996 Regulation of calcium spikes in mammalian oocytes through a combination of inositol triphosphate-dependent entry and release. Molecular Human Reproduction 2, Bonduelle M, Aytoz A, Wilikens A et al Prospective follow-up study of 1987 children born after intracytoplasmic sperm injection ICSI. In: Filicori M, Flamigni C (eds) Treatment of infertility, the new frontiers. Communications Media for Education, New Jersey, pp Bourgain C, Nagy ZP, De Zutter H et al Ultrastructure of gametes after intracytoplasmic sperm injection. Human Reproduction 13, Chan AWS, Luetjens CM, Dominko T et al Foreign DNA transmission by ICSI: injection of spermatozoa bound with exogenous DNA results in embryonic GFP expression and live Rhesus monkey births. Molecular Human Reproduction 6, Cuthbertson KSRCPH 1985 Phorbol ester and sperm activate mouse oocytes by inducing sustained oscillations in cell Ca 2+. Nature, 316, Danan C, Sternberg D, Van Steirteghem A et al Evaluation of parental mitochondrial inheritance in neonates born after intracytoplasmic sperm injection. American Journal of Human Genetics 65, Dozortsev D, Rybouchkin A, De Sutter P et al Human oocyte activation following intracytoplasmic injection: the role of the sperm cell. Human Reproduction 10, Dumoulin JCM, Coonen E, Bras M et al Embryo development and chromosome anomalies after ICSI: effect of the injection procedure. Human Reproduction 16, Hardarson T, Lundin K, Hamberger L 2000 The position of the metaphase II spindle cannot be predicted by the location of the first polar body in the human oocyte. Human Reproduction 15, Hewitson L, Simerly CR, Tengowski MW et al Microtubule and chromatin configurations during Rhesus intracytoplasmic sperm injection: successes and failures. Biology of Reproduction 55, Hewitson L, Takahashi D, Dominko T et al Fertilization and embryo development to blastocysts after intracytoplasmic sperm injection in the rhesus monkey. Human Reproduction 13, Hewitson L, Dominko T, Takahashi D et al Unique checkpoints during the first cell cycle of fertilization after intracytoplasmic sperm injection in rhesus monkeys. Nature Medicine 5, Hewitson L, Martinovich C, Simerly C et al. 2000a Is round spermatid injection ROSI a therapy for male infertility? ROSI in the rhesus monkey is unsuccessful. Fertility and Sterility 74, S68. Hewitson L, Martinovich C, Takahashi D, Schatten G 2000b Elongated spermatid injection ELSI in the rhesus monkey. Fertility and Sterility 74, S67 Houshmand M, Holme E, Hanson C et al Is paternal mitochondrial DNA transferred to the offspring following intracytoplasmic sperm injection? Journal of Assisted Reproduction and Genetics 14, Igusa Y, Miyazaki S 1986 Periodic increase of cytoplasmic free calcium in fertilized hamster eggs measured with calciumsensitive electrodes. Journal of Physiology 377, Jaffe LA 1992 First messengers at fertilization. Journal of Reproduction and Fertility 42, Lamb VK, Scott-Brown M, Marchington DR et al Does intracytoplasmic sperm injection ICSI result in the inheritance of paternal mitochondrial DNA? Human Reproduction 14, 124. Lorca T, Cruzalegni FH, Fesquet D 1993 Calmodulin-dependent protein kinase II mediates inactivation of MPF and CSF upon fertilization of Xenopus eggs. Nature 366, Ludwig M, Diedrich K 1999 In-vitro-Fertilisation und intrazytoplasmatische Spermieninjektion: Gibt es ein Gesundheitsrisiko für die geborenen Kinder? Deutsches Ärzteblatt, 96, Ludwig M, Küpker W, Diedrich K 1998 Physiologie der reproduktiven Funktionen der Frau. In: Diedrich K (ed.) Endokrinologie und Reproduktionsmedizin III. Urban & Schwarzenberg, München, Wien, Baltimore, pp Ludwig M, Al-Hasani S, Felberbaum R, Diedrich K 1999a New aspects of cryopreservation of oocytes and embryos in assisted reproduction and future perspectives. Human Reproduction 14, Ludwig M, Strik D, Al-Hasani S, Diedrich K 1999b No transfer in a planned ICSI cycle: we cannot overcome some basic rules of human reproduction. European Journal of Obstetrics, Gynecology and Reproductive Biology 87, Luetjens CM, Payne C, Schatten G 1999 Non random chromosome positioning in human sperm and sex chromosome anomalies following intracytoplasmic sperm injection. Lancet, 353, Nagy PZ, Liu J, Joris H et al Time-course of oocyte activation, pronucleus formation and cleavage in human oocytes fertilized by intracytoplasmic sperm injection. Human Reproduction 9, Nagy ZP, Liu J, Joris H et al The influence of the site of sperm deposition and mode of oolemma breakage at intracytoplasmic sperm injection on fertilization and embryo development rates. Human Reproduction 10, Ozil JP 1990 The parthogenetic development of rabbit oocytes after pulsatile electrial stimulation. Development 109, Parrington J, Swann K, Shevchenko VI 1996 Calcium oscillations in mammalian eggs triggered by a soluble sperm protein. Nature 379, Richords LF, White KL 1993 Electroporation of inositol 1,4,5-239