PGD in 47,XXY Klinefelter's syndrome patients

Transcription

1 Human Reproduction Update, Vol.9, No.4 pp. 319±330, 2003 DOI: /humupd/dmg029 PGD in 47,XXY Klinefelter's syndrome patients C.Staessen 1,3, H.Tournaye 1, E.Van Assche 2, A.Michiels 2, L.Van Landuyt 1, P.Devroey 1, I.Liebaers 2 and A.Van Steirteghem 1 1 Centre for Reproductive Medicine and 2 Centre for Medical Genetics, University Hospital, Dutch-speaking Free University of Brussels (Vrije Universiteit Brussel), Brussels, Belgium 3 To whom correspondence should be addressed at: Centre for Reproductive Medicine, Academisch Ziekenhuis, Vrije Universiteit Brussel, Laarbeeklaan 101, B-1090 Brussels-Belgium. catherine.staessen@az.vub.ac.be The use of ICSI has been a major breakthrough in the treatment of male infertility. Even azoospermic patients with focal spermatogenesis in the testis, may bene t from the ICSI technique in order to father a child. As ICSI use has become more common, centres have introduced infertility treatment for Klinefelter patients. To date, 34 healthy children have been born using ICSI without PGD, and the conception of one 47,XXY fetus has been reported. In view of the possible risk of an increased gonosome number in the spermatozoa of Klinefelter patients, a safer approachðoffering these couples ICSI combined with PGDÐhas been used, and has resulted in the birth of three healthy children. Couples in which the male suffered from Klinefelter's syndrome were rst treated in 1995; these patients were offered ICSI + PGD using FISH technology, notably to enumerate the X and Y chromosomes. ICSI + PGD was performed in 32 cycles of 20 couples with spermatozoa originating from a fresh ejaculate (n = 1), testicular biopsy (n = 21) or frozen±thawed testicular biopsy (n = 10). Normal fertilization occurred in % of the successfully injected oocytes. On day 3 of development, 119 embryos from 29 cycles were of suf cient quality to undergo biopsy and subsequent PGD; a positive result was obtained in 113 embryos. Embryos were available for transfer in 26 cycles, with a mean of embryos per transfer. Eight pregnancies were obtained, and ve resulted in a delivery. A total of 113 embryos from couples with Klinefelter's syndrome was compared with 578 embryos from control couples with X-linked disease where PGD was used to determine gender. A signi cant fall occurred in the rate of normal embryos for couples with Klinefelter's syndrome (54.0%) compared with controls (77.2%). Moreover, a signi cantly increased risk of abnormalities was observed for sex chromosomes and autosomes; for each autosome separately, this reached signi cance level for chromosomes 18 and 21 only. Hence, a cautious approach is warranted in advising couples with non-mosaic Klinefelter's syndrome. Moreover, the use of ICSI + PGD or prenatal diagnosis should be carefully considered. Key words: FISH/ICSI/Klinefelter syndrome/pgd/testicular sperm Introduction In 1942, Klinefelter and colleagues rst described a syndrome characterized by gynaecomastia, aspermatogenesis, without aleydigism and increased excretion of FSH (Klinefelter et al., 1942), and some 17 years later this syndrome was related to a 47,XXY karyotype (Jacobs and Strong, 1959). Although most Klinefelter's syndrome patients are considered to have a non-mosaic 47,XXY karyotype, a mosaic 47,XXY/ 46,XY karyotype is found in about 10% of cases (Wilson and Grif n, 1987). Within the general population, the incidence of 47,XXY Klinefelter's syndrome is 0.1% (Nielsen and Wohlert, 1991), and this value reaches 3.1% in the infertile male population (Guichaoua et al., 1993). Although a few cases with proven paternity (Laron et al., 1982; Terzoli et al., 1992) have been reported, non-mosaic 47,XXY Klinefelter patients are in general considered to be sterile. The introduction of ICSI led to a change in the clinical approach towards fertility treatment in these patients. In cases with a 46,XY/47,XXY mosaic karyotype, spermatozoa can usually be recovered from the ejaculate, and fertilization has ensued after ICSI (Harari et al., 1995; Bielanska et al., 2000). Nonmosaic Klinefelter patients (47,XXY) are generally azoospermic owing to primary testicular failure, although in some cases focal spermatogenesis has been present and testicular spermatozoa have been recovered and successfully used for ICSI (Staessen et al., 1996; Tournaye et al., 1996). Investigations on the meiotic products of non-mosaic Klinefelter patients by direct analysis of the ejaculated or testicular spermatozoa with the uorescent in-situ hybridization (FISH) Human Reproduction Update 9(4) ã European Society of Human Reproduction and Embryology 2003; all rights reserved 319

2 C.Staessen et al. Table I. Overview of reported pregnancies obtained after ICSI treatment of patients with apparent non-mosaic Klinefelter's syndrome Reference PGDperformed No. of Outcome of pregnancy Karyotype (n) b cycles (n) a 47,XXY: ejaculated spermatozoa Hinney et al. (1997) No Case report Singleton: miscarriage at 9 weeks Abortion tissue: 46,XX CruÈger et al. (2001) No Case report Singleton pregnancy 46,XX (1) Kitamura et al. (2000) No 2 (2) 1 biochemical pregnancy and 1 miscarriage No information at 8 weeks Tachdjian et al. (2003) No Case report Twin pregnancy 46,XX (1)46,XY (1) 47,XXY: frozen±thawed ejaculated spermatozoa Bourne et al. (1997) No Case report Twin pregnancy 46,XX (1)46,XY (1) 47,XXY: fresh testicular spermatozoa Staessen et al. (1996); Tournaye et al. (1997) Yes 3 (3); 7 (5) 1 biochemical and 2 singleton pregnancies 2 healthy neonates Palermo et al. (1998) No 3 (2) Singleton and twin pregnancy 46,XY (2)46,XX (1) Reubinoff et al. (1998) Yes 5 (4) Singleton pregnancy 46,XY (1) Levron et al. (2000) No Not known (8) 1 triplet, 1 twin and 2 singleton pregnancies 46,XY (4)46,XX (3) Poulakis et al. (2001) No 2 (2) 2 singleton pregnancies 46,XY (1)46,XX (1) Friedler et al. (2001) No 5 (5) 2 singleton and 1 twin c pregnancy 4 healthy neonates Nodar et al. (1999) No Case report Twin pregnancy 46,XY (2) Kitamura et al. (2000) No 1 (1) Singleton pregnancy 46,XX (1) Greco et al. (2001) No Case report Twin pregnancy 46,XY (2) Yamamoto et al. (2002) No 12 (9) 1 twin and 3 singleton pregnancies 46,XY (2)46,XX (3) 47,XXY: frozen±thawed testicular spermatozoa Friedler et al. (2001) No 5 (5) Twin pregnancy Both healthy males Rosenlund et al. (2002) No Case report Singleton pregnancy 46,XY (1) a Number of couples in parentheses. b Number of children in parentheses. c Triplet with 47,XXY fetus reduced in the 14th gestational week. technique have shown varying incidences of normal spermatozoa ranging from 50.0 to 93.7% (Guttenbach et al., 1997; Estop et al., 1998; Levron et al., 2000) and an increased incidence of 24,XX and 24,XY hyperhaploid spermatozoa as compared with a control population (Foresta et al., 1998; Hennebicq et al., 2001; Yamamoto et al., 2002), while the incidence of 24,YY hyperploid spermatozoa was similar (Guttenbach et al., 1997) and disturbed 23,X/23,Y ratios (Guttenbach et al., 1997). These different studies in fact con rm that, in Klinefelter patients, there is an increased incidence of genetically imbalanced spermatozoa, though two different hypotheses have been formulated. The rst hypothesis is that 47,XXY spermatogonia may undergo meiosis to produce hyperhaploid spermatozoa (Guttenbach et al., 1997; Estop et al., 1998; Foresta et al., 1998; Bielanska et al., 2000; Hennebicq et al., 2001; Yamamoto et al., 2002); the second hypothesis is that the rare break-through patches of spermatogenesis in XXY males are due to the presence of normal XY germ cells but, as a result of the compromised testicular environment, these 46,XY germ cells are susceptible to meiotic abnormalities (Mroz et al., 1998). As the use of ICSI has become increasingly widespread, different centres have introduced infertility treatment for Klinefelter patients. Their reportsðthe majority of which are case reportsðrelate the centres' experience with fertility treatment of Klinefelter patients with ICSI but without the use of PGD, and this has resulted in the birth of healthy children (Table I). With the use of freshly ejaculated sperm, the birth of three healthy children has been reported, while two healthy children have been born after the use of frozen±thawed ejaculated sperm, and 26 healthy children have been born after the use of fresh testicular sperm. Finally, three healthy children were born following the use of cryopreserved-thawed testicular spermatozoa. One group (Ron-El et al., 2000a) reported a triplet gestation, one of which (a 47,XXY fetus) was reduced, while others (Kyono et al., 2001) reported a pregnancy following ICSI with fresh testicular sperm from a man with non-mosaic Klinefelter's syndrome and spinal cord injury. The outcome of this pregnancy was not documented however, as at the time of this writing the pregnancy was ongoing. In view of the possible risk of an increased number of gonosomes in their spermatozoa, anotherðmore safeðapproach is to offer these couples ICSI in combination with PGD using FISH technology, in order to enumerate the X and Y chromosomes in particular (Staessen et al., 1996; Reubinoff et al., 1998). To date, by using this approach of combining ICSI with fresh testicular sperm and PGD, three healthy children have been born (Table I). However, since these rst reports were made (Staessen et al., 1996; Tournaye et al., 1997; Reubinoff et al., 1998), a large number of non-mosaic Klinefelter patients have been treated with ICSI in combination with PGD at the authors' centre. The aim of the present review was to describe the fertility treatment in Klinefelter patients, and also to discuss the use of ICSI treatment with and without PGD. Several couples with non-mosaic Klinefelter's syndrome were treated with ICSI combined with PGD, and the genetic constitution of the embryos produced was compared with that of embryos obtained from control couples who 320

3 PGD after ICSI for 47,XXY patients underwent ICSI plus PGD in order to determine fetal gender with relation to sex-linked disease. Assisted reproduction techniques, PGD procedures and results Patients Between November 1994 and March 2002, sperm was obtained from 20 couples of which the husband had a 47,XXY karyotype. The subsequent 32 treatment ICSI cycles which were obtained, together with the data obtained from them, form the basis of the present report. The mean (6 SD) age of the female partners was years (maximum age 38 years), and all women had a normal fertility work-up. The non-mosaic 47,XXY karyotype in the patients was proven by an analysis of 15 to 30 metaphases with G-banding and/or an analysis of 50± 100 interphases with FISH analysis in peripheral blood preparations and/or broblasts. A weak mosaicism of 5±8% was detected after FISH analysis of 200 interphase lymphocytes in three patients; all of these patients consented to undergo PGD by FISH in order to determine the number of sex chromosomes if embryos were subsequently available for transfer. Another group of patients who underwent PGD to determine fetal gender with regard to sex-linked diseases was selected for comparison of the aneuploidy rate. This latter (control) group comprised 43 couples of which the female partner was aged <38 years (mean age years) and included 88 treatment cycles. The results of the latter patient group were published in part previously (Staessen et al., 1999). Sperm retrieval In general, spermatozoa were retrievable in ~50% of Klinefelter patients (Tournaye et al., 1996). The only predictive factor for successful sperm recovery in these patients was the histopathological nding, as for all other parametersð including ultrasonography and frequency of 46,XY in lymphocytes and buccal cellsðno signi cant difference was found between Klinefelter patients with successful sperm recovery and those in whom sperm recovery failed (Westlander et al., 2001). Therefore, sperm recovery is a crucial step in the fertility treatment of Klinefelter patients. Before patients were scheduled for infertility treatment, one or two of their sperm ejaculate samples were carefully evaluated for the presence of sperm cells by using high-speed centrifugation. If the presence of sperm cells was con rmed, the stimulation was planned. However, if no sperm cells were observed in the ejaculate, a testicular biopsy was scheduled (Tournaye et al., 1996; 1997). In cases where no spermatozoa were found after mechanical mincing of the biopsies, enzymatic treatment of the testicular tissue with collagenase type IV was employed (Crabbe et al., 1998) in order to maximize the chance of sperm retrieval. Attempts to freeze testicular sperm cells from non-obstructive males were successful (Verheyen et al., 1997); consequently, it was possible to freeze testicular sperm obtained from a diagnostic biopsy before a treatment cycle could be commenced. Two options exist in scheduling of the testicular biopsy. With the rst approach, the biopsy can be scheduled before starting hyperstimulation treatment, with subsequent freezing of the retrieved spermatozoa. Although this approach prevents the woman from being needlessly hyperstimulated, there is an estimated risk of 19.4% (Verheyen et al., 2002) that the retrieved sperm cells will not survive the freeze±thawing procedure. To counteract this risk, the patient should be scheduled for a back-up fresh biopsy sperm retrieval on the day of oocyte retrieval. The cryopreservation of testicular tissue obtained after diagnostic testicular biopsy is a routine procedure option at the authors' institution. Following an unsuccessful rst ICSI cycle with either fresh or frozen testicular spermatozoa, repeated testicular surgery for subsequent ICSI cycles may be avoided by using frozen specimens. This is important, as repeated biopsy retrieval may be dif cult in Klinefelter patients. The second approach option is to perform an initial testicular biopsy during a hyperstimulation treatment cycle. As sperm is generally recovered in only 50% of Klinefelter patients, this option could be proposed to couples agreeing to the use of donor sperm. ICSI was performed in 32 cycles of 20 couples, with sperm originating from a fresh ejaculate in one cycle, from a testicular biopsy in 21 cycles, and from a frozen±thawed testicular biopsy in 10 cycles (Table II). In two cycles, insuf cient spermatozoa were obtained after thawing of the testicular biopsy specimens in order to inject all the oocytes, and the two couples decided to have the remaining oocytes injected with donor spermatozoa rather than undergo repeat testicular biopsy. The results obtained from oocytes injected with donor sperm were excluded from the data analysis. Oocyte retrieval and ICSI procedure Ovarian stimulation was carried out by pituitary desensitizing with GnRH analogues (Buserelin, Suprefact; Hoechst, Brussels, Belgium) combined with hmg (Humegon; Organon, Oss, The Netherlands) or recombinant FSH (Gonal- F; Serono, Brussels, Belgium or Puregon; Organon) and hcg (Pregnyl, Organon; Profasi, Serono). Oocyte retrieval was carried out using ultrasound-guided puncture at 36 h after hcg administration. The cumulus±corona±oocyte complexes were placed in 25 ml droplets of medium and covered by lightweight paraf n oil. Before ICSI was carried out, the cumulus and corona cells were removed and nuclear maturation was assessed under an inverted microscope. The ICSI procedure was carried out as described previously (Van Steirteghem et al., 1998). The testicular spermatozoa were frequently found to be embedded in Sertoli cell cytoplasm, and had to be carefully extracted. For the injection, a motile sperm cell (if present) was aspirated into the injection pipette and injected into a metaphase II (MII) oocyte which had extruded the rst polar body. Injected oocytes were washed and incubated in 25 ml droplets of medium in a Petri dish at 37 Cin 321

4 C.Staessen et al. Table II. Results of fertility treatment of Klinefelter patients Origin of sperm cells (no. of cycles; no. of transfers) Number of: OCC retrieved (mean 6 SD per cycle) MII injected (mean 6 SD per cycle) 2PN (mean 6 SD per cycle) a Embryos biopsied (with diagnosis) Normal XX or XY (%) Embryos transferred (mean 6 SD per transfer) Positive hcg (with FHB) No. of deliveries Fresh ejaculate (77.8) 5 (4) 3 (75) 2 1 (1) 1 (n =1;1) Fresh TESE (85) 47 (55.3) 31 5 (3) 3 (n = 21; 18) ( ) ( ) ( ) ( ) Frozen TESE (24) 11 (45.8) 8 ( ) 2 (1) 1 c (n = 10; 7) ( ) ( ) b ( ) Total (n = 32; 26) 364 ( ) 310 ( ) 172 ( ) 119 (113) 61 (54.0) 41 ( ) 8 (5) 5 a Mean fertilization rate (%) per cycle. b From one patient only six of 12 oocytes were injected, and from one patient only one of 10 oocytes was injected with sperm from the Klinefelter patient and included in the data analysis. c Child liveborn preterm at 23 weeks, but did not survive. FHB = fetal heart beats; MII = metaphase II; OCC = oocyte±cumulus complex; 2PN = 2 pronuclei; TESE = testicular sperm extraction. an incubator under an atmosphere of 5% CO 2,5%O 2 and 90% N 2. The medium used for oocyte and embryo culture was subject to changes during the period of treatment, though since 1999 a sequential medium (G1-G2; Scandinavian IVF, Vitrolife, GoÈteborg, Sweden) has been used for all patients undergoing ICSI in combination with PGD. In total, 364 oocyte±cumulus complexes (OCCs) were retrieved, with a mean of OCC per cycle (Table II). Among these OCCs, 325 (89.3%) were at MII stage, and 310 of them were injected with sperm from Klinefelter patients. As mentioned earlier, in two patients insuf cient sperm was retrieved to inject all oocytes; in one of these patients only six of 12 MII oocytes were injected, while in the other patient only one of 10 MII oocytes was injected with sperm from the Klinefelter husband. The other oocytes were injected with donor sperm and excluded from further consideration in this report. Although when selecting sperm for microinjection the preference was to use a spermatozoon that had shown signs of motility, suf cient motile spermatozoa could not always be found for the oocytes retrieved; consequently, in some cycles ICSI was carried out with non-motile testicular spermatozoa. In three cycles, all the oocytes were injected with non-motile sperm cells, and in seven cycles some of the oocytes were injected with non-motile sperm cells due to a scarcity of motile sperm cells. In total, 247 (79.7%) oocytes were injected with a motile sperm cell, and 63 (20.3%) with a non-motile sperm cell. Fertilization was checked at 16±18 h after injection (Staessen et al., 1995). Normal fertilization was con rmed by the presence of two distinct pronuclei and polar bodies using microscopy (inverted microscope, original magni cation 3400). A mean fertilization rate of % per injected MII oocyte was obtained in all the cycles. There was no difference 322 in fertilization rate when either fresh sperm or sperm originating from frozen testicular-extraction (TESE) were used (see Table II). In three cycles, only non-motile sperm were injected; of the 26 injected MII oocytes, a total of 15 (57.7%) was fertilized (i.e. a mean of % for the three cycles). Among a total of 64 oocytes in seven cycles, 27 were injected with motile spermatozoa, and 15 (55.6%) were fertilized (i.e. mean % per patient); among 37 oocytes injected with non-motile sperm, 15 (40.5%) were fertilized (mean % per patient). By comparing the fertilization rate of motile versus non-motile sperm cells in sibling oocytes (n = 7 cycles), a tendency towards a higher fertilization rate in the group of oocytes injected with motile sperm cells was observed (Fisher's exact test, P = 0.3; NS). The likely reason for fertilization rates not being signi cantly different in this group was that not enough cases were involved. These data demonstrate that the selection of motile testicular spermatozoa is always preferable for ICSI, although as sperm production in non-mosaic Klinefelter patients is at best scarce, it is worthwhile using non-motile sperm cells for injecting the oocytes. These results were in agreement with fertilization rates obtained after the injection of motile (65%) versus non-motile (45%) testicular sperm cells in patients with complete azoospermia and normal karyotype (Nagy et al., 1998). The same group (Nagy et al., 1998) also showed that embryo quality at 2 days after ICSI was independent of either motile or non-motile testicular spermatozoa being microinjected; this indicated that when normal fertilization has occurred, the quality of the resulting embryos is generally independent of sperm factors (Nagy et al., 1995). Likewise, among 30 fertilized oocytes injected with non-motile sperm cells, 20 developed into embryos suitable for biopsy; a diagnosis was obtained in 15 of these, and four were normal and transferred.

5 PGD after ICSI for 47,XXY patients Table III. Number and quality of embryos for biopsy in patients with Klinefelter syndrome in relation to origin of the sperm, and in control patients Origin of sperm cells (no. of cycles; no. of cycles with biopsy) No. of 2PN (A) No. of embryos for biopsy (% A) <5-cell stage (n) 6- to 8-cell stage (n) Type A-B (%) Type C (%) Type A-B (%) Type C (%) Klinefelter patients Fresh ejaculate (n = 1; 1) (0) 2 (40) 1 (20) 2 (40) Fresh TESE (n = 21; 19) (13.3) 15 (16.7) 54 (60.0) 9 (10.0) Frozen TESE (n = 10; 9) (29.2) 3 (12.5) 12 (50.0) 2 (8.3) Total Klinefelter patients (69.2) * 19 (16.0) 20 (16.8) 67 (56.3) 13 (10.9) (n = 32; 29) Control patients (n = 88) (79.4) * 125 (20.7) 59 (9.8) 368 (60.9) 52 (8.6) * Fisher's exact test: P < As a consequence there was one clinical pregnancy. It was also reported that, for a small number of oocytes, quite similar fertilization rates were achieved regardless of whether motile or non-motile spermatozoa were injected (Ron-El et al., 2000b); among 14 oocytes injected with immotile spermatozoa, seven were fertilized and six cleaved, whereas among the eight oocytes injected with motile spermatozoa, seven were fertilized and cleaved. Both sets of embryos led to the establishment of a pregnancy. Embryo development and biopsy Oocytes with two pronuclei (2PN) were assessed on days 2 and 3 after injection for embryonic development, and embryos of grade A, B or C reaching at least the 5-cell stage on the morning of day 3 were biopsied. Type A embryos were de ned as those in which all the blastomeres were of an approximately equal size, without anucleate fragments. Type B embryos had blastomeres of equal or non-equal size and with anucleate fragments comprising a maximum 20% of the embryo volume, while type C embryos contained anucleate fragments as 20± 50% of their volume. Grade A, B, or C embryos with at least four blastomeres were biopsied on the morning of day 3 after microinjection. The selection criteria for embryo biopsy were similar to those used to decide whether an embryo was transferable on day 3 in the regular ICSI programme without PGD. Those embryos not suitable for biopsy were discarded. The same policy was applied as with embryos not considered for transfer. Before biopsy, the blastomeres were checked for the presence of nuclei. From the 6-cell stage onward, two blastomeres per embryo were removed (Van de Velde et al., 2000; De Vos and Van Steirteghem et al., 2001). Among 172 oocytes showing 2PN, 119 (69.2%) developed into embryos suitable for blastomere biopsy, and the remaining zygotes were arrested or severely delayed on day 3. In three cycles, no embryos suitable for biopsy were obtained, despite fertilization. In two cycles using fresh testicular sperm, ve and two zygotes were obtained respectively, but none of these could be considered for biopsy. In one cycle the only oocyte injected with frozen±thawed testicular sperm was fertilized, but fragmented on day 3 of development. Finally, among 29 of the 32 cycles, embryos were available for biopsy; cell stages and embryo quality for biopsy as de ned by morphological grade are summarized in Table III. Of these embryos, 32.8% had less than six cells, and 67.2% had at least six cells. As mentioned earlier, one group of patients undergoing PGD for sex-linked disease were added retrospectively (as controls) for comparison with the study group. In the control group (see Table III), a signi cantly higher proportion of zygotes (79.4%) developed into embryos suitable for biopsy as compared with the Klinefelter patients (69.2%; Fisher's exact test, P < 0.01), although the distribution of the different grades of embryos for biopsy was not signi cantly different between the control and Klinefelter subjects. One patient with a mosaic Klinefelter karyotype was reported to have a normal fertilization rate after ICSI; however, there was a high incidence of cleavage failure (44%) after pronuclei were observed (Bourne et al., 1995). These results further con rmed that spermatozoa recovered either from the ejaculate or by testicular biopsy from non-mosaic 47,XXY males can induce normal fertilization and cleavage following ICSI, though the cleavage rate may be impaired. FISH procedure The individually biopsied blastomeres were spread onto a Superfrost Plus glass slide (Kindler GmbH, Freiburg, Germany) using 0.01 mol/l HC1/0.1% Tween 20 solution (Coonen et al., 1994; Staessen et al., 1996). Both blastomeres from the same embryo were xed on the same slide, in very close proximity. The FISH procedure was used as described previously (Staessen et al., 1996) and, over a period of time, has included the use of several different probe mixtures. For the rst three PGD cycles, double target FISH was performed using directly labelled DNA probes speci c for chromosomes X (Alpha 323

6 C.Staessen et al. Table IV. Chromosomal abnormalities in embryos of couples with Klinefelter syndrome and control group in relation to the chromosomes analysed Variable Klinefelter syndrome Control Age <38 years Age <38 years No. of embryos analysed for chromosomes XY 4 Not done No. of normal embryos (%) 4 (100) Not done No. of embryos analysed for chromosomes XY, No. of normal embryos (%) 35 (53.0) 241 (76.5) No. of embryos analysed for chromosomes XY, 18, 21 Not done 54 No. of normal embryos (%) Not done 43 (79.6) No. of embryos analysed for chromosomes XY, 13, 18, No. of normal embryos (%) 21 (52.5) 133 (82.6) No. of embryos analysed for chromosomes XY, 13, 16, 18, 21, No. of normal embryos (%) 1 (33.3) 29 (60.4) Total number of embryos analysed Total number of normal embryos (%) 61 (54.0) a 446 (77.2) a a P < 0.01 (contingency table). Satellite DNA probe, Spectrum Green; Vysis, Inc., Downers Grove, IL, USA) and Y (Alpha Satellite DNA probe, Spectrum Orange; Vysis, Inc.) (Staessen et al., 1996). For the next 13 PGD cycles, a triple-target FISH technique was performed using directly labelled DNA probes speci c for chromosomes X (Alpha Satellite DNA, Spectrum Green; Vysis, Inc.), Y [Alpha Satellite III DNA (DYZ1 locus), Spectrum Orange; Vysis, Inc.] and 18 (Alpha Satellite DNA, 1:1 mixture Green/ Orange spectrum; Vysis, Inc.). In the remaining 15 cycles, probes for chromosomes XY, 13, 18 and 21 (DXZ1, Spectrum Blue; DYZ3, Spectrum Gold; LSI13, Spectrum Red; D18Z1, Spectrum Aqua; LSI21 Spectrum Green; Multivision PGT Probe Panel; Vysis, Inc.) were applied. An aliquot (0.2 ml) of the probe solution was added to the nuclei, covered with a round coverslip (4 mm diameter), denatured for 5 min at 73 C and left to hybridize for between 4 h and overnight at 37 C in a moist chamber. After washing in 0.43 standard saline citrate (SSC) at 71 C for 2 min, antifade solution (Vectashield) was added and uorescence signals were evaluated. The nuclei were then examined using a Zeiss Axioskop uorescence microscope with the appropriate lter sets. The FISH images were captured with a computerized system, and the results interpreted by two independent observers. In one cycle, a two-round FISH procedure was performed which allowed the detection of chromosomes XY, 13, 18, 21 (round 1) and 16, 22 (round 2). Following the analysis of the rst set of probes, the coverslips were gently removed and the slides rinsed in 13 phosphate-buffered saline at room temperature, denatured in SSC for 7 min at 75 C, and then dehydrated (70, 90, 100 and 100% ethanol at ±18 C, 40 s each). The second hybridization solution was prepared by mixing a probe for chromosome 16 (Satellite II DNA/D16Z3 probe, Spectrum Orange; Vysis, Inc.) and a probe for chromosome 22 (LSI 22, 22q11.2, Spectrum Green; Vysis, Inc.). The probes were denatured separately in a hot waterbath at 75 C for 5 min. An 324 aliquot (0.2 ml) of the probe solution was then added to the nuclei, covered with a round coverslip (4 mm diameter), sealed with rubber cement and then hybridized overnight in a water bath at 37 C. Finally, the slides were washed for 2 min in 0.43 SSC at 73 C and 23 SSC/0.1% Nonidet P40 for 60 s at room temperature. The washed slides were then mounted with DAPI in antifade solution, and analysed. Following biopsy and FISH procedure, a FISH result was obtained in 113 of 119 (95.0%) embryos available. A normal result was obtained for the chromosomes analysed in 61 of the 113 embryos (54.0%). Three out of four (75%) embryos obtained with fresh ejaculate were normal, 47 out of 85 (55.3%) embryos obtained with fresh testicular sperm were normal, and 11 out of 24 (45.8%) embryos obtained with frozen testicular sperm cells were normal. Although the groups were very small, there was no inter-group difference with regard to the percentage of normal embryos according to the different origins of sperm used. Embryo transfer and outcome During the study period, the policy was changed with regard to the day of transfer, from day 3 to day 5. This allowed the FISH procedure to be performed with a greater number of probes in two consecutive FISH rounds. The number of embryos transferred was also in line with a policy to transfer a maximum of two embryos in patients aged <37 years if the embryo quality was good. Transfer on day 5 also allowed the selection of those embryos which reached the compacting or blastocyst stage on day 5. Spare, genetically normal embryos of suf cient morphological quality were cryopreserved (Van den Abbeel et al., 1997). Implantation was con rmed when two hcg concentrations of >10 U/l were measured at least 10 days after embryo transfer. Clinical pregnancy was de ned by an intrauterine gestational sac with a fetal heartbeat seen by vaginal ultrasound at least 6 weeks after embryo transfer. PGD is a novel technique, and it is recommended that in the case of pregnancy

7 PGD after ICSI for 47,XXY patients Table V. Frequency of abnormalities for each chromosome for patients with Klinefelter syndrome and for controls Abnormality Klinefelter(age <38 years) Control(age <38 years) I. Sex chromosomal abnormalities No. of sex chromosome abnormalities/no. of embryos analysed for this chromosome (%) 15/113 a (13.2) 18/578 a (3.1) Monosomy XO 7 * 10 Monosomy OY 2 0 Trisomy XYY 2 1 Trisomy XXX 1 1 Trisomy XXY 0 4 Mosaic XY/XXY 2 1 Mosaic XY/XYY 1 0 Mosaic XX/XO 0 1 II. Autosomal abnormalities No. of chromosome 13 abnormalities/no. of embryos analysed for this chromosome (%) 3/43 (7.0) 5/209 (2.4) Monosomy Trisomy 13 2 ** 2 No. of chromosome 16 abnormalities/no. of embryos analysed for this chromosome (%) 0/3 4/48 (8.3) Monosomy Trisomy No. of chromosome 18 abnormalities/no. of embryos analysed for this chromosome (%) 8/109 b (7.3) 12/578 b (2.1) Monosomy Trisomy 18 4 ** 1 Mosaic monosomy 18/disomy No. of chromosome 21 abnormalities/no. of embryos analysed for this chromosome (%) 5/43 c (11.6) 8/263 c (3.0) Monosomy 21 3 * 6 Trisomy Mosaic disomy 21/trisomy No. of chromosome 22 abnormalities/no. of embryos analysed for this chromosome (%) 1/3 (33.3) 1/48 (2.1) Monosomy Trisomy III. Ploidy status abnormalities No. of haploid embryos/total no. of embryos analysed (%) 6/113 d (5.3) 4/578 d (0.7) No. of triploid embryos/total no. of embryos analysed (%) 4/113 (3.5) 9/578 (1.6) No. of tetraploid embryos/total no. of embryos analysed (%) 2/113 (1.7) 12/578 (2.1) IV. Combined abnormalities No. of embryos with combined abnormalities/ total no. of embryos analysed (%) 10/113 (8.8) 59/578 (10.2) *Including the double monosomy X-21: in fact counted twice. **Including the double trisomy 13±18: in fact counted twice. a, b, c, d : Fisher's exact test: P < the couple should undergo prenatal diagnosis in order to con rm the PGD diagnosis, either by chorionic villus sampling or by amniocentesis. Among the pregnancies obtained to date, only one couple effectively decided to perform amniocentesis; the fetus was normal and the PGD result con rmed. Clearly, as experience with PGD is increased and accurate estimates of error risks become available, the attitude to prenatal diagnosis will be reconsidered. The introduction of techniques such as comparative genomic hybridization, which enable all chromosomes to be investigated, will remove the risk of all chromosome imbalances. A total of 41 embryos was transferred (see Table II) in 26 transfer procedures (mean embryos per transfer). Embryo transfer could not be performed in three patients, this being due to an absence of normal embryos in two cases, and of poor embryo quality (but genetic normality) in one case. A total of eight cycles with positive hcg was obtained from the 26 transfers; three of these were preclinical abortions and ve were ongoing. This provided an ongoing pregnancy rate of 15.6% per cycle, or 19.2% per transfer. The implantation rate with fetal heart activity was 12.8% per transferred embryo. One patient delivered prematurely at 23 weeks gestation, most likely due to chorioamnionitis; the child was stillborn at delivery. The four other pregnant patients delivered healthy babies (two girls, two boys) at term. To date, among the 20 couples in which sperm were found, four of them delivered (live birth rate 20%). Which genetic abnormalities are found at diagnosis? As mentioned earlier, 54.0% of the embryos from Klinefelter patients were found to be normal, with an almost equal sex ratio (32/61 female; 29/61 male). The numbers of embryos analysed with the different probe mixtures, together with details of embryos 325

8 C.Staessen et al. Table VI. Results of embryos at biopsy and reanalysis of the non-transferred or non-cryopreserved embryos At biopsy FISH results Reanalysed embryos No. of embryos Blastomere 1 Blastomere 2 No. of embryos reanalysed FISH results 32 normal female 4 4: normal female 29 normal male 6 6: normal male 6 monosomy X0 5 3: monosomy X0 1: disomy XX * 1: chaotic 1 double monosomy X : double monosomy X-21 2 monosomy 0Y 2 2: monosomy 0Y 2 trisomy XYY 2 1: trisomy XYY 1: mosaic XYYY/XY/XYY 1 trisomy XXX 1 1: mosaic: monosomy X0 /trisomy XXX 2 XY XXY 1 1: mosaic XY/XXY 1 XY XYY 1 1: mosaic XY1818 / XYY monosomy : monosomy 13 1 trisomy : trisomy 13 4 monosomy : monosomy 18 2: mosaic: disomy 18 / monosomy 18 3 trisomy : double trisomy 18-XYY 1 double trisomy 13 ±18 1 1: double trisomy 13±18 2 monosomy : mosaic monosomy 21/disomy 21/ trisomy 21/ 1 trisomy : trisomy 21 1 disomy 21 monosomy : mosaic monosomy 21/disomy 21 1 monosomy haploid X 4 2: mosaic XX diploid/x haploid; 2: chaotic 2 haploid Y 2 2: chaotic 4 triploid 4 4: triploid 2 tetraploid 1 1: tetraploid 10 combined abnormalities 6 2: normal diploid * ; 4: chaotic * Embryos considered as not con rmed. found to be normal for the chromosomes tested are listed in Table IV. Abnormalities related to the gain or loss of a speci c chromosome, and to the ploidy status of the embryo (haploid, triploid or tetraploid) are detailed in Table V. An abnormality of the gonosomes was apparent in 15 of the 113 (13.2%) embryos tested, and of the autosomes in 17 of 109 (15.6%). Likewise, 12 of 113 (10.6%) embryos showed an abnormality of ploidy status, and 10 of 113 (8.8%) showed combined abnormalities. Are embryos generated by Klinefelter patients at increased genetic risk? In comparison with controls, Klinefelter patients showed a signi cantly higher percentage of abnormal embryos (Fisher's exact test; P < 0.01). The prevalence of sex chromosome abnormalities was only 3.1% in FISH for sex-linked disease group, compared with 13.2% in Klinefelter patients (Fisher's exact test; P < 0.01), while the prevalence of autosomes was 30 of 578 embryos (5.2%) in controls and 17 of 109 (15.6%) in Klinefelter patients (Fisher's exact test; P < 0.01). If data for each autosome chromosome are compared separately, then a signi cantly increased risk was observed for chromosomes 18 and 21 (Fisher's exact test; P < 0.01). Moreover, among the control subjects 25 of 578 embryos (4.3%) had abnormalities of ploidy status compared with 12 of 113 (10.6%) (P < 0.01) in Klinefelter patients. Finally, in the control group 59 of 578 embryos (10.2%) had combined abnormalities compared with 10 of 113 (8.8%) in the Klinefelter group. One group (Magli et al., 2002) reported their experience with PGD in couples with an altered karyotype due to gonosomal mosaicism, and focused on the chromosomal condition of embryos generated. Some 63% of the embryos were chromosomally abnormal, and the incidence of aneuploid mosaics was higher as compared with PGD patients with normal karyotype of the same age. This suggested that constitutional carriers of sex chromosome mosaicism are predisposed to autosomal mosaicism of embryos. Although few data were generated, the incidence of chromosomally abnormal embryos was slightly higher when the woman was the carrier. These authors concluded that their ndings support a predisposition in patients with gonosomal mosaicism to give rise to different cell lines, possibly due to disturbances and malfunctioning of the mechanisms entering cell division. Analysis of non-transferred and non-cryopreserved embryos After PGD and embryo transfer, cells from embryos which were classi ed as genetically normal but morphologically unsuitable for transfer or cryopreservation (n = 10) or as genetically abnormal (n = 52) were spread as a whole as described previously (Staessen and 326

9 Van Steirteghem, 1997), and FISH was carried out using the same probe mixture as was used for the diagnosis. All patients involved consented for further research to be carried out on these embryos. The results were compared with the one-cell diagnosis in order to verify the accuracy of the technique. In total, 62 embryos were spread for reanalysis, of which 527 blastomeres were counted. A FISH result was obtained for 426 blastomeres from 51 embryos (Table VI). The remaining normal embryos (n = 10) which could not be frozen due to poor quality were reanalysed. In all 10 embryos the normal status was con rmed in most of the nuclei. A reanalysis was obtained for ve of the six embryos diagnosed as monosomy X0; in three cases a uniform X0 embryo was con rmed, one embryo was found to be XX diploid (FISH diagnosis not con rmed), and one was chaotic. The diagnosis was con rmed in one embryo with a double monosomy for X and 21. Both 0Y diploid embryos were found to be uniform 0Y diploid embryos. A reanalysis was obtained of both XYY embryos; one embryo was con rmed to be XYY uniformly, while in the other embryo XYY was found to originate from a XYY conception followed by postzygotic non-disjunction resulting in a XYYY and a XY cell line. The trisomic XXX may have arisen from an XX embryo at conception followed by post-zygotic non-disjunction leading to X0 and XXX cell lines. In two embryos, one of the two blastomeres analysed showed a normal XY pattern, and the other an abnormal XXY pattern; both embryos were spread. FISH results were obtained from one embryo only. The XY/XXY mosaic embryo was most likely the result of a XXY conception followed by post-zygotic loss of an X chromosome in one cell line. An analogous situation pertained to the XY/XYY mosaic embryo. Here, a conception of an XYY embryo with subsequent loss of a Y chromosome in one cell might have occurred, or an XY embryo at conception followed by post-zygotic non-disjunction involving a Y chromosome leading to X0/XY/XYY cell lines from which the X0 cell line was lost during the procedure. Intracytoplasmic fertilization of oocytes with a Klinefelter male's 24,XY- or 24,XXbearing spermatozoa, could theoretically produce a 47,XXY or a 47,XXX embryo. These results indicate that indeed, 47,XXY or 47,XXX embryos were formed. The results were given for the autosomal abnormalities and explanations similar to those for the sex chromosome abnormalities are valid. To the present authors' knowledge, only two other groups have reported FISH analysis on embryos of 46,XY/47,XXY or 47,XXY Klinefelter patients. In the rst study (Bielanska et al., 2000), results were reported on FISH analysis with DNA probes speci c for chromosomes X, Y and 18 on 10 spare embryos from fertility treatment of a 46,XY/47,XXY Klinefelter patient. The data revealed that only three of 10 embryos were normal for the chromosomes testedðone embryo with an XX1818 and two embryos with an XY1818 pattern. The abnormalities detected included ve chaotic mosaic embryos and two diploid mosaic embryos with the majority of blastomeres normal for the chromosomes tested. The second study (Reubinoff et al., 1998) reported on PGD with probes for chromosomes X, Y and 18 in three embryos from two patients. One embryo was diagnosed as normal and transferred; this resulted in the birth of a karyotypically normal neonate. The biopsied blastomere from the second embryo had an abnormal XX18 constitution. The blastomere from the third embryo was also found to contain an abnormal pattern, XXY1818. Spreading of the remainder of the cells from the above two embryos revealed nuclei with various X and Y patterns, consistent with chaotic cell division. In contrast with the results of the present study, none of the embryos analysed in either study showed a XXY1818 or XXX1818 pattern, but of course both studies were dealing with a limited number of analysed embryos. The present data and that from both other reports on embryos from Klinefelter patients demonstrate that the incidence of chromosomal mosaicism in in-vitro-fertilized human embryos is substantial. When PGD is attempted, the patients should be consulted regarding the limitations of aneuploidy detection by this experimental technique. In some cases, blastomere analysis at the cleavage stage will not be representative of the whole embryo, due to the high frequency of chromosomal mosaicism in human embryos (Munne and Cohen, 1993; Harper et al., 1995). However, the present data show that on the basis of the results at diagnosis and after PGD reanalysis, the initial diagnosis could not be con rmed in only three of 51 (5.9%) embryos and that, on the basis of the PGD result, the embryos which could have been transferred were actually rejected for transfer. Conclusion PGD after ICSI for 47,XXY patients It is generally accepted that the introduction of ICSI has constituted a breakthrough in the treatment of male infertility. Even patients considered azoospermic but with focal spermatogenesis in the testis, such as those with non-mosaic 47,XXY Klinefelter syndrome, have been able to bene t from the ICSI technique. Furthermore, ICSI of the patient's spermatozoa induced fertilization, produced a high percentage of cleavage-stage embryos, and led to implantation and ongoing pregnancy. All neonates that have been reported as being born after ICSI without PGD using spermatozoa from patients with non-mosaic Klinefelter syndrome are normal. One fetus conceived by spermatozoa recovered from a man with non-mosaic Klinefelter syndrome was of 47,XXY karyotype, and the embryo was reduced at week 14 (Ron-El et al., 2000a). The FISH method with directly labelled DNA probes speci c for the X and Y chromosomes constitutes an ef cient and rapid procedure to determine the exact number of sex chromosomes present in one or two blastomeres of a preimplantation embryo. Since PGD by FISH technology was available at the present authors' institution, a policy was adopted which proposed that patients should combine ICSI treatment with PGD, particularly in order to enumerate the sex chromosomes before transfer (Staessen et al., 1996). The experience of combining ICSI with PGD indicates that there was a signi cant fall in the rate of normal embryos for couples with Klinefelter's syndrome compared with control patients undergoing PGD for gender determination and using ejaculated sperm. Indeed, a signi cantly increased risk of abnormalities for the sex chromosomes as well as for the autosomes was observed; for each autosome separately, this reached signi cance level only for chromosomes 18 and 21. A much higher frequency of disomy 21 was found in the spermatozoa of a Klinefelter patient than in those of control patients (Hennebicq et al., 2001). Any child conceived by ICSI using this sperm will thus have a proportionally higher risk of trisomy 21. Hennebicq explained this by suggesting that, ``The 327

10 C.Staessen et al. most likely reason for the high rate of disomy 21 in our XXY patient is that the abnormal sex chromosome content disturbs the meiotic segregation of some chromosomal pairs, possibly via the presence of an abnormal or absent sex vesicle. Close association of chromosome 21 with one or both sex chromosomes during meiosis could lead to a disturbance of chromosome 21 segregation in the case of an abnormal sex vesicle. The observation provides evidence of the possibility of an interchromosomal effect in patients with Klinefelter's syndrome that results in a higher than normal risk of trisomy 21 in their children?'' The high proportion of normal fetuses that develop from the implantation of embryos generated by fertilization of oocytes with spermatozoa recovered from men with Klinefelter's syndrome may be due to the large percentage of normal (23,X or 23,Y) spermatozoa produced by these Klinefelter patients (Guttenbach et al., 1997; Estop et al., 1998; Levron et al., 2000), though most of the studies investigated only a limited number of chromosomes. Another explanation is the possible reduced implantation potency of the genetically abnormal embryos. It is clear that there is an increased risk of sex and autosomal chromosomal abnormal embryos obtained in Klinefelter patients when compared to patients with normal karyotype and normal spermatogenesis. A possible approach is to transfer all embryos available in Klinefelter patients, assuming that only the best ones will implant. Although this approach is reasonable, it is in contradiction with the general policy of limiting the numbers of embryos for transfer, and in any case does not remove the risk of producing a child with a chromosome abnormality. Without PGD, the chance of selecting and transferring abnormal embryos is high because even embryos with normal morphology have a distinct percentage of abnormalities (Harper et al., 1995; Munne et al., 1995). Moreover, in case of pregnancy there is an indication for prenatal diagnosis with the potential risk of procedure-related miscarriage and, in the case of an abnormal result, termination of the pregnancy. Finally, by transferring genetically abnormal embryos, genetically normal ones will be frozen; as a consequence, since only 50% of embryos will survive thawing, the patients' chances are reduced yet again. With PGD, the identi cation of abnormalities in a cohort of morphologically goodquality embryos prevents the transfer of those embryos that are destined either not to implant or to abort spontaneously. One argument against PGD (Hardy et al., 1990) is that embryo biopsy might affect the viability of these precious embryos which have been generated after ICSI, sometimes with extremely rare sperm. As yet, however, no convincing argument has been produced that would demonstrate a possible dramatic effect of biopsy on the implantation potency of the embryo. The incidence of chromosomal abnormalities in epididymal and testicular sperm retrieved from azoospermic patients undergoing ICSI has been quanti ed (Palermo et al., 2002). The overall aneuploidy rate, analysed by FISH for chromosomes XY, 18 and 21, of 11.4% in men with non-obstructive azoospermia was signi cantly higher than the rate of 1.8% detected in epididymal sperm from men with obstructive azoospermia, and also higher than the rate of 1.5% found in ejaculated sperm. No signi cant difference was found between the epididymal and ejaculated samples. When the chromosomal abnormalities were analysed, gonosomal disomy was the most recurrent abnormality in both obstructive and non-obstructive azoospermic patients, while autosomal disomy was the most frequent in ejaculated sperm. These ndings point to a higher incidence of sperm chromosomal abnormalities in sperm of non-obstructive azoospermic men, where sex chromosome aneuploidy appears to be the most common. The question arising here is whether the increased abnormalities in embryos from Klinefelter patients are related to the 47,XXY karyotype, or to the use of testicular sperm of azoospermic patients. For Klinefelter patients in particular this would imply that the rare break-through patches of spermatogenesis in XXY males may be due to the presence of normal XY germ cells, but as a result of the compromised testicular environment these 46,XY germ cells are susceptible to meiotic abnormalities, including non-disjunction of the sex chromosomes (Mroz et al., 1998) and the autosomes (Hennebicq et al., 2001). A comparative analysis of the incidence and type of chromosomal abnormalities between Klinefelter patients and non-obstructive azoospermic males will provide more information regarding possible mechanisms at the origin of chromosomal abnormalities. In conclusion, at the present authors' institution, broad genetic counselling is offered to couples with non-mosaic Klinefelter's syndrome, and the use of ICSI in combination with PGD is recommended. Ultimately, however, the nal decision must be made by the couples concerned. Acknowledgements The authors thank the clinical, scienti c, nursing, and technical staff of the Centre for Reproductive Medicine, and especially the colleagues of the microinjection and IVF laboratory. They are also very grateful to Mrs Marleen CarleÂ, Sylvie Mertens and Griet Meersdom for technical assistance with the FISH studies, and to Mrs Ines Devolder of the University Language Centre for correcting the English text. Grants from the Belgian Fund for Medical Research are kindly acknowledged. References Bielanska, M., Tan, S.L. and Ao, A. (2000) Fluorescence in-situ hybridization of sex chromosomes in spermatozoa and spare preimplantation embryos of a Klinefelter 46,XY/47,XXY male. Hum. Reprod., 15, 440±444. Bourne, H., Richings, N., Harari, O., Watkins, W., Speirs, A.L., Johnston, H., and Gordon Baker, H.W. (1995) The use of intracytoplasmic sperm injection for the treatment of severe and extreme male infertility. Reprod. Fertil. Dev., 7, 237±245. Bourne, H., Stern, K., Clarke, G., Pertile, M., Speirs, A. and Gordon Baker, H.W. (1997) Delivery of normal twins following the intracytoplasmic injection of spermatozoa from a patient with 47,XXY Klinefelter's syndrome. Hum. Reprod., 12, 2447±2450. Coonen, E., Dumoulin, J.C.M. Ramaekers, F.C.S. and Hopman, A.H.N. (1994) Optimal preparation of preimplantation embryo interphase nuclei for analysis by uorescent in situ hybridization. Hum. Reprod., 9, 533±537. CrabbeÂ, E., Verheyen, G., Silber, S., Tournaye, H., Van de Velde, H., Goossens, A. and Van Steirteghem, A. (1998) Enzymatic digestion of testicular tissue may rescue the intracytoplasmic sperm injection cycle in some patients with non-obstructive azoospermia. Hum. Reprod., 13, 2791±2796. CruÈger, D., Toft, B., Agerholm, I., Fedder, J., Hald, F. and Bruun-Petersen, G. (2001) Birth of a healthy girl after ICSI with ejaculated spermatozoa from a man with non-mosaic Klinefelter's syndrome. Hum. Reprod., 16, 1909± DeVos, A. and Van Steirteghem, A. (2001) Aspects of biopsy procedures prior to preimplantation genetic diagnosis. Prenat. Diagn., 21, 767±780. Estop, A.M., MunneÂ, S., Cieply, K.M., Vandermark, K.K., Lamb, A.N. and Fisch, H. (1998) Meiotic products of a Klinefelter 47,XXY male as determined by sperm uorescence in-situ hybridisation analysis. Hum. Reprod., 13, 124±