Human Reproduction Vol.18, No.2 pp.

399407, 2003

DOI: 10.1093/humrep/deg092

A morphological and chromosomal study of blastocysts developing from morphologically suboptimal human pre-embryos compared with control blastocysts
Thorir Hardarson1, Gunilla Caisander, Anita Sjogren, Charles Hanson, Lars Hamberger and Kersti Lundin
Department of Obstetrics and Gynaecology, Goteborg University, SU/Sahlgrenska, 413 45 Gothenburg, Sweden
1

To whom correspondence should be addressed. E-mail: thorir.hardarson@obgyn.gu.se

BACKGROUND: IVF laboratories performing embryo transfer at day 2 or 3 after fertilization are currently discarding pre-embryos considered suboptimal using morphological criteria. The objective of this study was to investigate whether blastocysts, cultured from such pre-embryos (surplus), were chromosomally and morphologically normal. As a control group we used morphologically good quality embryos (GQE), cultured to the blastocyst stage. METHODS: Human pre-embryos considered suboptimal were cultured to the blastocyst stage. As a control group, frozenthawed pre-embryos of good quality were cultured under identical conditions. The chromosomal status of the blastocysts obtained was studied by multi-colour uorescence in-situ hybridization for chromosomes 13, 16, 18, 21, 22, X and Y. RESULTS: There is, on average, a signicantly higher degree of chromosomal aberrations in blastocysts derived from surplus pre-embryos compared to blastocysts derived from GQE, and the chromosomal aberrations are generally found in a higher number of blastomeres per blastocyst. In addition, blastocysts from surplus pre-embryos had signicantly poorer morphology compared to GQE. Improvement in morphology and/or developmental rate in surplus pre-embryos between day 2 and day 3 did not predict a morphologically/chromosomally normal blastocyst. However, this study shows that close to half of the surplus pre-embryos that reach the blastocyst stage can be considered chromosomally normal when assessed for these seven chromosomes. Furthermore, we found that chromosomal aberrations were more concentrated in a particular cell population within blastocysts derived from GQE, compared with surplus blastocysts. CONCLUSIONS: The study suggests that even if the IVF laboratory is on average making the correct decision about the potential of a pre-embryo, surplus pre-embryos that might become chromosomally normal blastocysts are still being discarded.
Key words: blastocyst/pre-embryo/aneuploidy/FISH/IVF

Introduction A substantial number of human pre-embryos established in the IVF laboratory are discarded, based solely on morphological criteria, since they are not considered to have a reasonable chance of implanting and giving rise to an offspring. The decision to discard these surplus pre-embryos is commonly taken on day 2 or 3 after fertilization, based primarily on criteria such as cell number, fragmentation, cell size and cytoplasmic appearance. This morphological evaluation is furthermore often performed under low magnication. The use of prolonged culture systems has increased during the past decade with the underlying hypothesis that culturing of preembryos to the blastocyst stage will favour `normal' preembryos with a higher potential of implantation over suboptimal pre-embryos and thus aid the selection process of the most optimal pre-embryo for transfer (Gardner et al., 1998b; Huisman et al., 2000). This assumption is partially
European Society of Human Reproduction and Embryology

based on the fact that at the 48-cell stage the pre-embryo is only just beginning to make use of its own genomic expression (Braude et al., 1988) and an aneuploid and/or otherwise suboptimal pre-embryo would therefore not be inuenced until later in development. In addition, delaying embryo transfer to the blastocyst stage has been proposed to yield major benets as regards to obtaining synchronicity between the pre-embryonic development and the uterine physiological state including a reduced uterine wall contractility. Whether or not blastocyst culture will replace cleavage stage pre-embryo transfer is, however, still a matter of debate. It has been suggested that culture of surplus pre-embryos (i.e. pre-embryos of suboptimal quality, not transferred or frozen on day 2 or 3) to the blastocyst stage would enhance the selection of pre-embryos with high implantation potential (Balaban et al., 2001). Genetic studies of blastocysts from such surplus pre-embryos, in particular studies using suitable control 399

T.Hardarson et al.

groups, are limited. Several studies concerning chromosomal aberrations in the human blastocyst have been performed, mainly using the uorescence in-situ hybridization (FISH) technique (Benkhalifa et al., 1993; Evsikov and Verlinsky 1998; Veiga et al., 1999; Magli et al., 2000; Ruangvutilert et al., 2000; Sandalinas et al., 2001; Bielanska et al., 2002). These previous studies include analyses of blastocysts from a variety of patient groups but with no control groups. In contrast, this study was designed to use a clinical approach to study blastocysts, comparing, both morphologically and chromosomally, blastocysts obtained from morphologically low quality pre-embryos with blastocysts that develop from good quality embryos (GQE). In addition, our aim was to nd out if improvement of the morphology and/or developmental state of pre-embryos between day 2 and day 3 is predictive of a chromosomally `normal' blastocyst. Materials and methods
IVF procedure This study was approved by the local ethics committee at Goteborg University, Sweden, and the pre-embryos were donated by patients in our regular IVF programme. Ovarian stimulation was carried out by a desensitizing protocol using a short-acting GnRH agonist preparation in combination with recombinant FSH. Follicular aspiration was performed 3638 h after hCG administration using vaginal ultrasonography. The oocytes either underwent ICSI or were inseminated in our conventional IVF programme, after which they were cultured in IVF-20 (Vitrolife Ltd, Gothenburg, Sweden) for two days. On the day of embryo transfer (day 2), the pre-embryos were scored according to our grading system. This system is modied from Steer et al. (1992) and consists of four main pre-embryo morphological grades, where a developmental rate of 46 cells at day 2 and b6 cells on day 3 is considered normal. Grade I is considered the `optimal' pre-embryo, with no fragments, evensized blastomeres and light, homogeneous cytoplasm. Grade II is divided into three categories: <20% fragments, uneven-sized blastomeres and/or non-homogeneous cytoplasm. Remaining preembryos fall into the third and fourth main groups, mainly due to high cytoplasmic fragmentation (grade III >20%, grade IV >50%). These pre-embryos are only rarely used for transfer and never frozen, and will hereafter be called `surplus pre-embryos'. The surplus pre-embryos, to be discarded on day 2, were instead cultured in separate microdrops of medium until the next day, when their cell number and morphology were again documented as described above. The pre-embryos were then divided into two groups: surplus I: pre-embryos having >5 blastomeres and an improved morphology from day 2; surplus II: pre-embryos with `5 blastomeres and/or that maintained their poor morphology from day 2 (Figure 1). The pre-embryos were transferred on day 3 into rS2 medium (Vitrolife Ltd, Gothenburg, Sweden) for further culture. Pre-embryos which reached the blastocyst stage on day 6, or earlier, were graded morphologically, photographed and their diameter was measured. Grade A blastocysts are expanded or expanding with a distinct trophectoderm and eccentrically located inner cell mass (ICM). Grade B blastocysts are either poorly expanded and/or with less dened trophectoderm and ICM cells but not showing any signs of degenerative foci. Grade C blastocysts exhibit poor morphology, characterized by a number of degenerative foci in the ICM and trophectoderm and a poorly developed blastocyst cavity (Figure 2). The control pre-embryo group was obtained by thawing GQE (i.e.

grade I or II) frozen on day 2 using a slow-freezing method with propanediol as a cryoprotectant (Lassalle et al., 1985). Although the optimal control group would have been fresh GQE cultured to the blastocyst stage, this was not possible due to the ethical implications

Figure 1. Experimental design. The surplus pre-embryos donated on day 2 were cultured until the next day, when their developmental status was assessed (`Improved' or `Non-improved'), and accordingly the pre-embryos were divided into two groups, surplus I and II. The pre-embryos were then cultured further and those which reached the blastocyst stage were xed on day 6. The control group was obtained by thawing good quality pre-embryos, frozen on day 2, and cultured to the blastocyst stage.

Figure 2. Micrographs showing the three different morphology grades (A, B, C) used in this study. Original magnication Q200.

400

Blastocysts from surplus human pre-embryos

involved as all GQE are either transferred or frozen at day 2 in our programme. Only pre-embryos where b75% of the cells survived the freezethaw procedure and resumed cleavage were included. Directly after the thawing procedure, control pre-embryos were cultured in IVF-20 until day 3, when the cleavage status was checked and the preembryos were moved to rS2 for culture to day 6. Only blastocysts meeting the criteria set out above were then included in the FISH analysis. Fixation and FISH analysis The blastocysts were xed on poly-L-lysine-coated glass slides using 0.01 N HCl + 0.1% Tween 20 as rst introduced by (Coonen et al., 1994). Care was taken to remove as much of the cytoplasm as possible from the nuclei to improve the FISH probe penetration. In a few cases, the ICM was isolated during the xation procedure and placed in a separate drop on the same slide. The glass slides were dried and stored at 20C until the FISH analysis was performed. FISH protocol The slides were washed for 5 min in 1Qphosphate-buffered saline (PBS) followed by an ethanol series of 70, 85 and 99.5% for 1 min each. A pepsin treatment was performed in order to increase probe penetration. Briey, the slides were immersed in pre-warmed 0.01 N HCl containing a nal concentration of 0.005% pepsin and incubated for 15 min at 37 T 1C followed by two rinses in water and PBS. The slides were then immersed in Carnoy's xative for 10 min at 4 T 1C and rinsed twice with PBS and distilled water, dehydrated in an ethanol series and air-dried. For simultaneous detection of ve chromosomes, a multi-colour kit was used (MultiVysion PB Multicolour Probe Panel; Vysis, Inc., Downers Grove, IL, USA) which includes chromosomes 13 (spectrum red), 16 (spectrum aqua), 18 (spectrum blue) 21 (spectrum green) and 22 (spectrum gold). A small amount (0.3 ml) of the probe was pipetted onto the nuclei, a cover glass applied and glued and the preparation denatured at 69 T 1C for 7 min on a heating plate. The slides were incubated at 38 T 1C for 4 h in a moist chamber, the cover glasses were then removed and the slides washed at 42 T 1C in 50% formamide + 2Qstandard saline citrate (SSC) (15 min) followed by 2QSSC (10 min) and 2QSSC + 0.001% Igepeal (5 min) to remove non-specic staining. The slides were thereafter air-dried, antifade solution applied and the preparation sealed with a cover glass. The nuclei were observed with a Nikon epiuorescence microscope equipped with appropriate lters and overview pictures (Q100) of the whole blastocyst were taken. These pictures together with drawn maps, allowed us to identify each nucleus and to determine the exact position of each nucleus after the second round of FISH. Before the second FISH round, the slides were washed for 5 min each in 1QPBS, 2QSSC with 0.5% Tween and 2QSSC. The slides were kept on a slow shaking device during this time. After dehydration in an ethanol series, a probe solution for chromosomes 18 (spectrum aqua), X (spectrum green) and Y (spectrum red) was applied (Vysis, Inc.). The FISH procedure was performed as described above except that the incubation time was prolonged to 18 h to increase probe attachment. After the wash, an antifade solution containing [4,6-diamidino-2-phenylindole (DAPI) II] was applied. Signal and data analysis The FISH-scoring criterion was that signals had to be at least one signal's width apart to be scored as two separate signals. A nucleus was considered `normal' when all seven analysed chromosomes were present in correct numbers. The FISH system was tested on normal lymphocytes and two human embryonic stem cell lines with normal karyotypes. FISH was performed in two rounds and the probe for chromosome 18 was

included in both rounds to act as an internal control of nuclear quality and FISH efciency. In the control series, the discordance between the rst and second analysis was never >20%. Hence, this was the limit chosen for acceptance of the FISH analysis of the blastocysts. It was found that the efciency of the chromosome 16 probe was lowered due to repeated freezingthawing of the probe vial, although control slides showed no such effect on the other probes included in the kit. Therefore, in 13 of the blastocysts the results from chromosome 16 were excluded. A sensitivity analysis showed that this had no effect on the overall differences between or within the study groups. Statistical methods Distributions of the variables are given as means, SD, medians and ranges. For comparisons between groups, Fisher's exact test was used for dichotomous variables and the MannWhitney U-test for ordered and continuous variables. Spearman's rank correlations test was used for all correlation analyses. A multivariate analysis of covariance of `% normal cells per blastocyst' between the surplus and control group was performed, with adjustment for maternal age, aspirated oocytes, GQE and IVF/ICSI. Since the distribution of `% normal cells per blastocyst' was not normally distributed, it was rst transformed using the inverse cumulative normal function on the empirical distribution function. All signicance tests were two-tailed and conducted at the 5% signicance level.

Results The demographics for the patients (n = 120) donating preembryos to the study are shown in Table I. The surplus preembryos were collected over a period of 8 months and the frozen pre-embryos were thawed during that same period. No differences were found between the studied groups (surplus group I versus surplus group II, or surplus I + II versus control) in relation to maternal age or number of aspirated oocytes. However, signicantly more GQE per patient were seen in the control group compared with the surplus groups. In Table II the data for all pre-embryos (n = 350) and all blastocysts (n = 78) are shown. There was a trend towards a higher blastocyst rate in the control group but this difference was not statistically signicant (P = 0.14, non-signicant). The proportion of grade A, B and C blastocysts respectively was similar in surplus I and II. However, there was a statistically signicant difference in the proportion of `good quality blastocysts' (grade A + B) between the surplus (I + II) and the control groups (Table II). There was no difference in the grade A + B versus C concerning ICSI and IVF respectively (data not shown). No difference was found between preembryos which reached the blastocyst stage and those that did not regarding maternal age, number of aspirated oocytes or number of GQE (data not shown). A total of 76 blastocysts were successfully xed on day 6. Of these, 74 blastocysts with a total of 7936 cells [mean = 109.0 (SD 19.0), range: 24332] gave detectable FISH signals. In all, 58 blastocysts were included in the analysis of FISH results whereas the remainder (23.7%) did not meet the criteria set out in the internal FISH control (18 versus 18 compliance). The excluded blastocysts (n = 18) were evenly distributed both between and within the experimental groups. The results from 401

T.Hardarson et al.

these 58 blastocysts in relation to morphology are shown in Table III. A primary analysis of morphology grades A and B showed no differences in any of the parameters measured and these grades were therefore combined in Table III (`good morphology') and compared with grade C (`poor morphology'). These results show that the good morphology group had a signicantly higher mean number of cells per blastocyst, mean number of normal cells and number of normal blastocysts (n = 48 versus n = 10) compared with the poor morphology group. There was a signicant correlation between good morphology blastocysts and number of aspirated oocytes

(P = 0.031) and between mean no. of cells per blastocyst and morphology grade (rS = 0.66, P < 0.0001). However, since some blastocysts are derived from the same oocyte cohort, the comparative statistics from these data may be slightly biased. No signicant differences were found between surplus I and surplus II except concerning the number of blastocysts from IVF/ICSI. Maternal age, number of aspirated oocytes, number of GQE, blastocyst rate and morphology grade were similar and these two groups were therefore analysed together as `surplus' group versus the control group in Table IV.

Table I. Demographics of all patients (n = 120) who donated pre-embryos to the study Parameter Mean maternal age (years) Mean (SD) Median (range) Mean aspirated oocytes per patient Mean (SD) Median (range) Mean no. of GQE per patient Mean (SD) Median (range) Mean no of donated pre-embryos per patient Mean (SD) Median (range)
aNo. of patients. NS = non-signicant.

Surplus I (n = 42)a 32.1 (3.4) 33.0 (2539) 14.5 (4.4) 14.0 (827) 4.8 (2.7) 4.0 (212) 3.4 (2.42) 3.0 (112)

Surplus II (n = 49)a 31.8 (3.6) 32.0 (2539) 14.7 (14.0) 14.0 (438) 3.7 (2.6) 3.0 (115) 3.2 (1.98) 3.0 (112)

Surplus (I + II) (n = 91)a 32.0 (3.46) 33.0 (2539) 14.6 (5.3) 14.0 (438) 4.2 (2.7) 3.0 (115) 3.3 (2.18) 3.0 (112)

Control (n = 29)a 30.5 (3.8) 32.0 (2435) 14.5 (4.9) 14.0 (730) 7.6 (3.1) 8.0 (214) 1.8 (0.91) 2.0 (15)

NS NS 0.014 NS

NS NS < 0.0001 < 0.0001

Table II. Data on blastocyst formation in a total of 350 surplus and control pre-embryos Parameter Total no. of blastocysts Embryos derived from conventional IVF Blastocysts from IVF pre-embryos Blastocysts from ICSI pre-embryos Blastocysts Grade A Grade B Grade C Grade A + B Values in parentheses are percentages. of pre-embryos. b versus c: P = 0.020; d versus e: P = 0.33 (NS). f versus g: P = 0.022; h versus i: P = 0.15 (NS). NS = non-signicant.
aNo.

Surplus I (n = 115)a 26 (22.6) 50 (43) 17 (34)b 9 (14)c 12 (10.4) 9 (7.8) 5 (4.3) 21 (18.3)

Surplus II (n = 184)a 36 (19.6) 107 (58) 24 (22)d 12 (16)e 17 (9.2) 11 (6.0) 8 (4.3) 28 (15.2)

P NS 0.018 NS NS NS NS NS NS

Surplus (I + II) (n = 299)a 62 (20.7) 157 (53) 41 (26.1)f 21 (14.8)g 29 20 13 49 (9.7) (6.7) (4.3) (16.4)

Control (n = 51)a 16 (31.4) 37 (73) 9 (24.3)h 7 (50.0)i 7 (13.7) 8 (15.7) 1 (2.0) 15 (29.4)

P NS 0.011 NS 0.008 NS 0.055 NS 0.043

Table III. Comparison of cell number and chromosomal normality between `good' and `poor' morphology blastocysts, compiled from both surplus and control groups Parameter No. of cells per blastocyst Mean (SD) Median (range) Normal cells per blastocyst Mean % (SD) Median (range) Good morphology Grade A + B (n = 48) 121.1 (66.4) 104 (24322) 66.1 (26.1) 76.9 (092.7) Poor morphology Grade C (n = 10) 61.1 (26.7) 65 (24105) 49.3 (23.7) 46.6 (12.591.2) P

0.0008 0.030

402

Blastocysts from surplus human pre-embryos Table IV. Comparison of patient demographics and blastocyst data, between surplus and control blastocysts, presented per patient Parameter Female age (years) Mean (SD) Median (range) No. of aspirated oocytes per patient Mean (SD) Median (range) No. of good quality embryos per patient Mean (SD) Median (range) No. of cells per blastocyst Mean (SD) Median (range) Normal cells per blastocyst Mean % (SD) Median (range) NS = non-signicant. Surplus (I + II) (n = 31) 31.9 (3.2) 32.0 (2538) 14.8 (4.5) 14 (727) 4.2 (2.7) 3 (211) 114.0 (58.1) 105.0 (27282) 59.5 (21.7) 56.4 (0.791.2) Control (n = 9) 31.9 (3.5) 34.0 (2535) 14.2 (3.6) 13 (1022) 7.6 (2.7) 9 (210) 120.0 (87.1) 88.0 (47317) 82.3 (7.9) 85.1 (69.492.7) P

NS NS 0.0074 NS 0.0036

In Table IV the data from the blastocysts analysed with FISH are presented per subject (n = 40), i.e. mean values (regarding number of cells per blastocyst, % normal cells per blastocyst, etc.) are calculated for blastocysts derived from the same subject (oocyte/embryo cohort). Here we found a statistically signicant difference between the surplus (I + II) group and the control group concerning proportion of normal cells (per blastocyst) and proportion of GQE/patient (Table IV). However, there was no difference in mean number of cells per blastocyst (P = 0.78, non-signicant). There were no differences between IVF/ICSI concerning % normal cells or mean number of cells per blastocyst (data not shown). Spearman's rank correlation showed a signicant correlation between number of cells per blastocyst and number of GQE per patient (rS = 0.35, P = 0.030). A multivariate analysis of covariance, based on subjects, between surplus and control group showed that the control group had signicantly (P = 0.0046) higher values of % normal cells per blastocyst after adjustment for maternal age, IVF/ ICSI, aspirated oocytes and number of GQE. The detailed results from the FISH analysis of the different chromosomal aberrations on a cellular level, in relation to developmental and morphology groups, are shown in Tables V and VI respectively. From the 58 blastocysts, 6425 cells were analysed in the rst FISH round and 6164 cells in the second. In general, there is an ~10% overall difference observed between the surplus and the control group when the individual chromosome aberration rate is compared (Table V). By far the most common overall aberration is tetrasomy with a large difference between both developmental and morphological groups (Table VII). As seen in Table VII, 24% (14/58) of the blastocysts were mosaic with one diploid and one tetraploid (2n/4n) cell line. However, the frequency of polyploidy in each blastocyst was rather low [mean = 19.7% (SD 15.2%), range: 762%]. Overall, chromosome Y has a lower aberration rate compared with other chromosomes (Tables V and VI). Interestingly, we found one blastocyst that had extensive monosomy 22 (79/119 cells).

Table V. Overview of the normality rate of the seven different chromosomes analysed by uorescence in-situ hybridization, grouped by the different developmental groups (surplus and control) Group Cells analysed Chromosomea 13 I II Surplus (I + II) Control Total
aValues

16 85 85 85 95 88

18 83 82 83 92 85

21 83 75 78 93 82

22 80 82 81 92 84

X 79 82 81 94 84

Y 91 92 92 98 93

2757 2150 4907 1518 6425

84 83 83 93 85

are percentages.

Table VI. Overview of the normality rate of the seven different chromosomes analysed by uorescence in-situ hybridization, grouped by the different morphological groups Group Cells analysed Chromosome 13 A B C Good (A + B) Poor (C) Total
aValues

16 90 90 74 90 74 88

18 88 88 68 88 68 85

21 85 84 67 84 67 82

22 88 84 69 86 69 84

X 87 85 72 86 72 84

Y 94 94 88 94 88 93

3847 2028 550 5875 550 6425

89 88 68 89 68 85

are percentages.

In a total of 10 blastocysts, we were able to isolate the ICM from the trophectoderm cells. No differences in number of normal cells per blastocyst or individual chromosomal errors were found between cells originating from the ICM versus the trophectoderm.

Discussion The two main ndings of this study are as follows. (i) There is on average a signicantly higher degree of chromosomal aberrations in blastocysts derived from surplus pre-embryos 403

T.Hardarson et al. Table VII. Overview of all the analysed blastocysts categorized according to ploidy and degree of mosaicism Group Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Surplus Control Control Control Control Control Control Control Control Control Control Control Control Control I I I I I I I I I I I I I I I I I I I II II II II II II II II II II II II II II II II II II II II II II II II II II Grade A A A A A A A A A B B B B B B C C C C A A A A A A A A A A A A A A B B B B B B C C C C C C A A A A A A B B B B B B B Cells analysed 116 234 209 157 166 200 142 119 282 85 114 74 27 182 124 66 105 64 83 104 104 78 114 49 108 130 132 183 119 111 132 94 167 72 103 24 101 84 56 32 88 40 39 70 24 98 322 72 97 317 92 88 73 47 97 98 73 44 Sex F M F M M M M F M M F F F M F F F M M M F M M M M M F M F M F F F M M F M F M F F F M M F M F M M F F M F F M F M M Types of cell Dip/Tet Dip Dip/Tet Dip Dip/XXY Dip Dip Dip/Mono X Dip/Tet/Mono X /Mono Y Dip Dip/Chaotic Dip Dip/Tri 22/Tet 21 Dip Dip/Chaotic Dip/Mono 18 Dip Dip/Tet/Mono 13 Dip Dip Dip/Tet Dip/Tet Dip/Mono 21 Dip/Tet Dip/Mono X Dip Dip Dip Dip Dip Dip Dip/Tet/Mono 13 Dip/Hap Dip/Null X/Null Y Dip/Chaotic Dip/Tet 13/Mono 21 Dip Dip Dip/Mono 16/ Mono 21 Dip/Mono 18 Dip/Tet/Tri 22 Dip/Tet/Null 18 Dip/Tet Dip Dip/Tet Dip Dip Dip/Tet Dip Dip Dip Dip Dip/Tet Dip Dip Dip Dip Dip Overall cell type Mosaic polyploidy Normal Mosaic polyploid Normal Mosaic aneuploidy (sex) Normal Normal Mosaic aneuploidy (sex) Mosaic aneuploidy Normal Mosaic chaotic Normal Mosaic aneuploidy Normal Mosaic chaotic Mosaic aneuploidy Normal Mosaic aneuploidy Normal Normal Mosaic polyploidy Mosaic polyploidy Monosomy 21 Mosaic polyploidy Mosaic aneuploidy (sex) Normal Normal Normal Normal Normal Normal Mosaic aneuploidy Mosaic haploidy Mosaic aneuploidy Mosaic chaotic Mosaic aneuploidy Normal Normal Mosaic aneuploidy Mosaic aneuploidy Mosaic polyploidy Mosaic aneuploidy Mosaic polyploidy Normal Mosaic polyploidy Normal Normal Mosaic polyploidy Normal Normal Normal Normal Mosaic polyploidy Normal Normal Normal Normal Normal

The system used is described by Ruangvutilert et al. (2000). For a blastocyst to be classied as not normal, a specic observed chromosomal aberration had to be present in >10% of all cells analysed. This threshold is based on the estimated uorescence in-situ hybridization error rate using these probes (Munne et al., 1998, 1999; Sandalinas et al., 2001). Mono = monosomy; Null = nullisomy; Hap = haploidy; Dip = diploidy; Tri = trisomy; Tet = tetraploidy.

compared to blastocysts derived from GQE, regarding the chromosomes analysed in this study. In addition, surplus blastocysts had a signicantly poorer morphology when compared with blastocysts derived from GQE. (ii) An 404

improvement in morphology and/or developmental rate in surplus pre-embryos between day 2 and 3 does not predict a morphologically/chromosomally normal blastocyst, based on the seven chromosomes analysed.

Blastocysts from surplus human pre-embryos

These results conrm that the morphological and developmental criteria used in the IVF laboratory for pre-embryo selection on day 2 are mainly correct, since this selection results predominantly in apparently chromosomally normal blastocysts (i.e. the control group). The concept of culturing surplus pre-embryos to the blastocyst stage and relying on these surplus blastocysts being chromosomally compatible with blastocysts from GQE is therefore incorrect. Several studies have shown that surplus cleavage stage pre-embryos have a high degree of aneuploidy (Munne et al., 1993, 1994; Marquez et al., 1999) and low pregnancy and implantation rates (Puissant et al., 1987; Steer et al., 1992; Ziebe et al., 1997). Interestingly, no differences were found in mean cell number per blastocysts between the experimental groups and the total cell number was higher than reported in blastocysts derived from similar pre-embryos (Hardy et al., 1989). It is, however, somewhat surprising that the surplus I group, which was dened as morphologically and/or developmentally improved on day 3, did not produce blastocysts of higher quality, since the same primary embryonic parameters as are used to score conventional cleavage-stage pre-embryos were applied. Perhaps morphology and/or synchronicity in developmental speed matters less in the later stages of pre-embryonic development compared to the rst two or three cell divisions. In the human pre-embryo, it is known that the transition between the maternal genome and the pre-embryo's own transcriptional activity occurs at this stage in development (Braude et al., 1988). It is therefore possible that a decline in developmental progress and/or morphological appearance due to suboptimal `genetics' might not occur until later, i.e. at the blastocyst stage, and is therefore not useful as a selection tool between day 2 and day 3. This is supported by ndings that indicate that extended culture until day 3 to select pre-embryo for transfer does not increase the success rate (Laverge et al., 2001). In this paper we have used probes specic for chromosomes X, Y, 13, 16, 18, 21 and 22. It is, however, most probable that other chromosomes are also involved in aneuploidies in blastocysts. Which chromosomes that are most commonly represented in aneuploidies in early blastocysts is still unknown (Bahce et al., 1999; Gianaroli et al., 2001) as chromosomal studies on blastocysts so far have been performed with a limited set of FISH probes. Recently, there have been reports where comparative genomic hybridization has been used to study pre-embryos (Voullaire et al., 2000; Wells and Delhanty 2000; Malmgren et al., 2002). This technique will enable us to draw conclusions about the aneuploidy status for all chromosomes in the pre-embryo. Our ndings conrm the results reported by Sandalinas et al. where surplus pre-embryos which were already diagnosed by PGD were cultured until day 56 and where it was also found that chromosomally abnormal pre-embryos were able to develop into blastocysts but at a lower rate than normal preembryos. What is interesting in the present study is that 42% (19/45) of the surplus blastocysts can be considered chromosomally normal, when assessed for seven chromosomes (Table VII). This indicates that even if the IVF laboratory is on average making the correct decision about the potential of a

pre-embryo, many surplus pre-embryos that might become chromosomally normal blastocysts are still being discarded. However, a chromosomally normal blastocyst does not in itself always lead to implantation and development of a healthy offspring. Nevertheless, it has been shown in several studies that pre-embryonic viability is strongly correlated to chromosomal normality (Gianaroli et al., 1997; Delhanty 2001). In a recent study by Balaban et al. (2001), it was shown that both pregnancy and implantation rates increased when prolonging the culture of surplus pre-embryos to day 5, i.e. transferring blastocysts derived from these poor-quality cleavage stage pre-embryos. Nevertheless, their implantation rates were low as compared with other blastocyst programmes using GQE (Gardner et al., 1998a; Jones et al., 1998), suggesting that a higher proportion of the surplus blastocysts were chromosomally abnormal. Many studies have shown that pre-embryos with serious chromosomal abnormalities are not compatible with blastocyst development (Benkhalifa et al., 1993; Munne et al., 1994; Evsikov and Verlinsky, 1998; Marquez et al., 1999; Veiga et al., 1999; Magli et al., 2000; Ruangvutilert et al., 2000; Sandalinas et al., 2001; Bielanska et al., 2002). But, even if the gravity of the chromosomal imbalance in blastocysts is reduced compared to cleavage stage pre-embryos, there still remains the problem of how to choose the blastocyst able to give rise to a healthy individual, especially if single embryo transfer is preferred. In the absence of a non-invasive method for selecting the chromosomally normal blastocyst, one method may be to perform aneuploidy screening (PGD-AS) at the 8cell stage and transfer only pre-embryos/blastocysts with a normal chromosomal set-up. This method has been shown to be successful by other investigators (Gianaroli et al., 1997; Pergament et al., 2001). The frequency of normal cells, regarding each separate chromosome pair in the surplus group, was higher than the results for each cell, when all the seven chromosomes are combined [7892% (Table V) versus 60% (Table IV)]. In the surplus group, the chromosomal errors are more scattered, i.e. inuencing a higher number of cells, compared with the control group. In the GQE group, on the other hand, the errors are accumulated in a smaller number of cells. This may be less detrimental for the whole blastocyst, possibly enabling it to discard a few chromosomally abnormal cells without loosing developmental capacity. These results are in concordance with our previous ndings where we found that a certain morphological aberration (uneven blastomeres) resulted in a higher incidence and a more widespread distribution of the chromosomal aberrations (Hardarson et al., 2001). The nding that 24% of the blastocysts had two cell lines with two different ploidy stages 2n/4n (Table VII) may not be surprising as it has been suggested that tetraploidization is a part of normal trophoblast development into syncytiotrophoblasts (Drury et al., 1998; Plachot, 1998; Bielanska et al., 2002). Furthermore, as this 2n/4n mosaicism may appear normally as a low degree mosaicism (Clouston et al., 1997) some of these blastocysts may have been regarded as normal due to the 10% FISH error threshold (see Table VII). 405

T.Hardarson et al.

We found that good blastocyst morphology is correlated both to higher cell number per blastocyst and to a higher percentage of chromosomally normal cells, based on the seven chromosomes analysed. Choosing a blastocyst of good morphology would therefore most likely increase the chance of implanting. The degree of blastocyst development in the surplus groups in this study was comparable to that in other studies using surplus pre-embryos (Shoukir et al., 1998; Balaban et al., 2001). In total, we found no differences in blastocyst formation rate between pre-embryos from conventional IVF compared with ICSI pre-embryos, although there were differences within the groups (Table II). Our control group had a numerically higher blastocyst rate than the surplus groups (31.4 versus 22.6%) but this difference was not signicant (P = 0.14). This rate is still low compared with other studies, where between 45% (Marek et al., 1999) and 66% (Gardner et al., 1998a) blastocyst developmental rates have been reported using sequential media, as in the present study. One explanation for this difference may be that we used frozenthawed pre-embryos, which are known to have an impaired developmental capacity and high aneuploidy rates (Iwarsson et al., 1999; Edgar et al., 2002). However, this effect might be partially offset by the fact that many of the patients who donated their frozen pre-embryos had delivered a healthy child from the same pre-embryo cohort, indicating a high pre-embryo capacity. We believe that we are at least not underestimating the difference between an `optimal' control group (i.e. fresh good quality pre-embryos) and a surplus group. In conclusion, we found in this study that blastocysts from morphologically surplus pre-embryos are on average more chromosomally and morphologically abnormal than blastocysts developing from GQE, when assessed for seven chromosomes. Nevertheless, according to the present study, 42% of these surplus blastocysts are chromosomally normal, based on the seven chromosomes analysed, and further studies are needed to nd factors that can help to distinguish between the abnormal and normal ones. The present clinical practice in our IVF laboratory is to discard surplus pre-embryos on day 2 or 3. By prolonged culture combined with PGD-AS, a large number of these pre-embryos may be proved normal, with a theoretically high implantation rate. The benets of such methods must be weighed against the extra costs associated with prolonged pre-embryo culture and PGD-AS and should perhaps only be used for selected patient groups. The possibility of increasing the number of pre-embryos available for IVF patients by prolonged culture surplus of pre-embryos should be taken into due consideration, also in light of the single embryo transfer discussion.

References
Bahce, M., Cohen, J. and Munne, S. (1999) Preimplantation genetic diagnosis of aneuploidy: were we looking at the wrong chromosomes? J. Assist. Reprod. Genet., 16, 176181. Balaban, B., Urman, B., Isiklar, A., Alatas, C., Mercan, R., Aksoy, S. and Nuhoglu, A. (2001) Blastocyst transfer following intracytoplasmic injection of ejaculated, epididymal or testicular spermatozoa. Hum. Reprod., 16, 125129. Benkhalifa, M., Janny, L., Vye, P., Malet, P., Boucher, D. and Menezo, Y. (1993) Assessment of polyploidy in human morulae and blastocysts using co-culture and uorescent in-situ hybridization. Hum. Reprod., 8, 895902. Bielanska, M., Tan, S.L. and Ao, A. (2002) Chromosomal mosaicism throughout human preimplantation development in vitro: incidence, type, and relevance to embryo outcome. Hum. Reprod., 17, 413419. Braude, P., Bolton, V. and Moore, S. (1988) Human gene expression rst occurs between the four- and eight-cell stages of preimplantation development. Nature, 332, 459461. Clouston, H.J., Fenwick, J., Webb, A.L., Herbert, M., Murdoch, A. and Wolstenholme, J. (1997) Detection of mosaic and non-mosaic chromosome abnormalities in 6- to 8-day old human blastocysts. Hum. Genet., 101, 3036. Coonen, E., Dumoulin, J.C., Ramaekers, F.C. and Hopman, A.H. (1994) Optimal preparation of preimplantation embryo interphase nuclei for analysis by uorescence in-situ hybridization. Hum. Reprod., 9, 533537. Delhanty, J.D. (2001) Preimplantation genetics: an explanation for poor human fertility? Ann. Hum. Genet., 65, 331338. Drury, K., Kovalinskaia, R. and Williams, S. (1998) Polyploidy as a normal function of trophoblastic development in human preimplantation embryos observed by in situ hybridization analysis. Fertil. Steril., 70, O-25. Edgar, D.H., Archer, J. and Gook, D.A. (2002) Blastocyst development and cell numbers in thawed human embryos after extended culture. Hum. Reprod., 17, O-236. Evsikov, S. and Verlinsky, Y. (1998) Mosaicism in the inner cell mass of human blastocysts. Hum. Reprod., 13, 31513155. Gardner, D.K., Schoolcraft, W.B., Wagley, L., Schlenker, T., Stevens, J. and Hesla, J. (1998a) A prospective randomized trial of blastocyst culture and transfer in in-vitro fertilization. Hum. Reprod., 13, 34343440. Gardner, D.K., Vella, P., Lane, M., Wagley, L., Schlenker, T. and Schoolcraft, W.B. (1998b) Culture and transfer of human blastocysts increases implantation rates and reduces the need for multiple embryo transfers. Fertil. Steril., 69, 8488. Gianaroli, L., Magli, M.C., Ferraretti, A.P., Fiorentino, A., Garrisi, J. and Munne, S. (1997) Preimplantation genetic diagnosis increases the implantation rate in human in vitro fertilization by avoiding the transfer of chromosomally abnormal embryos. Fertil. Steril., 68, 11281131. Gianaroli, L., Magli, M.C. and Ferraretti, A.P. (2001) The in vivo and in vitro efciency and efcacy of PGD for aneuploidy. Mol. Cell. Endocrinol., 183 (Suppl. 1), S1318. Hardarson, T., Hanson, C., Sjogren, A. and Lundin, K. (2001) Human embryos with unevenly sized blastomeres have lower pregnancy and implantation rates: indications for aneuploidy and multinucleation. Hum. Reprod., 16, 313318. Hardy, K., Handyside, A.H. and Winston, R.M. (1989) The human blastocyst: cell number, death and allocation during late preimplantation development in vitro. Development, 107, 597604. Huisman, G.J., Fauser, B.C., Eijkemans, M.J. and Pieters, M.H. (2000) Implantation rates after in vitro fertilization and transfer of a maximum of two embryos that have undergone three to ve days of culture. Fertil. Steril., 73, 117122. Iwarsson, E., Lundqvist, M., Inzunza, J., Ahrlund-Richter, L., Sjoblom, P., Lundkvist, O., Simberg, N., Nordenskjold, M. and Blennow, E. (1999) A high degree of aneuploidy in frozenthawed human preimplantation embryos. Hum. Genet., 104, 376382. Jones, G.M., Trounson, A.O., Gardner, D.K., Kausche, A., Lolatgis, N. and Wood, C. (1998) Evolution of a culture protocol for successful blastocyst development and pregnancy. Hum. Reprod., 13, 169177. Lassalle, B., Testart, J. and Renard, J.P. (1985) Human embryo features that inuence the success of cryopreservation with the use of 1,2 propanediol. Fertil. Steril., 44, 645651. Laverge, H., De Sutter, P., Van Der Elst, J. and Dhont, M. (2001) A prospective, randomized study comparing day 2 and day 3 embryo transfer in human IVF. Hum. Reprod., 16, 476480. Magli, M.C., Jones, G.M., Gras, L., Gianaroli, L., Korman, I. and Trounson,

Acknowledgements
The authors wish to thank Nils-Gunnar Pehrsson for advice and statistical analyses, patients who donated the pre-embryos and the IVF staff at the Unit of Reproductive Medicine, Sahlgrenska University Hospital, Goteborg, Sweden. The study was supported by grants from Hjalmar Svenssons Foundation, the European Commission (Marie Curie research training grant to T.Hardarson) and the Swedish Medical Research Council (11606).

406

Blastocysts from surplus human pre-embryos A.O. (2000) Chromosome mosaicism in day 3 aneuploid embryos that develop to morphologically normal blastocysts in vitro. Hum. Reprod., 15, 17811786. Malmgren, H., Sahlen, S., Inzunza, J., Aho, M., Rosenlund, B., Fridstrom, M., Hovatta, O., Ahrlund-Richter, L., Nordenskjold, M. and Blennow, E. (2002) Single cell CGH analysis reveals a high degree of mosaicism in human embryos from patients with balanced structural chromosome aberrations. Mol. Hum. Reprod., 8, 502510. Marek, D., Langley, M., Gardner, D.K., Confer, N., Doody, K.M. and Doody, K.J. (1999) Introduction of blastocyst culture and transfer for all patients in an in vitro fertilization program. Fertil. Steril., 72, 10351040. Marquez, C., Sandalinas, M., Bahce, M., Alikani, M. and Munne, S. (1999) Chromosomal abnormalities in 1255 cleavage-stage human embryos. RBM Online, 1, 1726. Munne, S., Lee, A., Rosenwaks, Z., Grifo, J. and Cohen, J. (1993) Diagnosis of major chromosome aneuploidies in human preimplantation embryos. Hum. Reprod., 8, 21852191. Munne, S., Grifo, J., Cohen, J. and Weier, H.U.G. (1994) Chromosome abnormalities in human arrested preimplantation embryos: a multiple probe uorescence in-situ hybridization study. Am. J. Hum. Genet., 55, 150159. Munne, S., Marquez, C., Magli, C., Morton, P. and Morrison, L. (1998) Scoring criteria for preimplantation genetic diagnosis of numerical abnormalities for chromosomes X, Y, 13, 16, 18 and 21. Mol. Hum. Reprod., 4, 863870. Munne, S., Magli, C., Cohen, J., Morton, P., Sadowy, S., Gianaroli, L., Tucker, M., Marquez, C., Sable, D., Ferraretti, A.P., Massey, J.B. and Scott, R. (1999) Positive outcome after preimplantation diagnosis of aneuploidy in human embryos. Hum. Reprod., 14, 21912199. Pergament, E., Fiddler, M., Zhang, J., Meyer, L., Kazer, R.R., Conno, E., Puscheck, E. and Milad, M.P. (2001) Preimplantation genetic diagnosis (PGD) for common aneuploidies: reversing the relationship of PGD and IVF. Am. J. Hum. Genet., 67, 43. Plachot, M. (1998) Aneuploidy rates in blastomeres. In Kempers, R.D., Cohen, J., Haney, A.F. and Youngers, J.B. (ed.), Fertility of Reproductive Medicine. Elsevier, pp. 711720. Puissant, F., Van Rysselberge, M., Barlow, P., Deweze, J. and Leroy, F. (1987) Embryo scoring as a prognostic tool in IVF treatment. Hum. Reprod. 2, 705708. Ruangvutilert, P., Delhanty, J.D., Serhal, P., Simopoulou, M., Rodeck, C.H. and Harper, J.C. (2000) FISH analysis on day 5 post-insemination of human arrested and blastocyst stage embryos. Prenat. Diagn., 20, 552560. Sandalinas, M., Sadowy, S., Alikani, M., Calderon, G., Cohen, J. and Munne, S. (2001) Developmental ability of chromosomally abnormal human embryos to develop to the blastocyst stage. Hum. Reprod., 16, 19541958. Shoukir, Y., Chardonnens, D., Campana, A. and Sakkas, D. (1998) Blastocyst development from supernumerary embryos after intracytoplasmic sperm injection: a paternal inuence? Hum. Reprod., 13, 16321637. Steer, C.V., Mills, C.L., Tan, S.L., Campbell, S. and Edwards, R.G. (1992) The cumulative embryo score: a predictive embryo scoring technique to select the optimal number of embryos to transfer in an in-vitro fertilization and embryo transfer programme. Hum. Reprod., 7, 117119. Veiga, A., Gil, Y., Boada, M., Carrera, M., Vidal, F., Boiso, I., Menezo, Y. and Barri, P.N. (1999) Conrmation of diagnosis in preimplantation genetic diagnosis (PGD) through blastocyst culture: preliminary experience. Prenat. Diagn., 19, 12421247. Wells, D. and Delhanty, J.D. (2000) Comprehensive chromosomal analysis of human preimplantation embryos using whole genome amplication and single cell comparative genomic hybridization. Mol. Hum. Reprod., 6, 10551062. Voullaire, L., Slater, H., Williamson, R. and Wilton, L. (2000) Chromosome analysis of blastomeres from human embryos by using comparative genomic hybridization. Hum. Genet., 106, 210217. Ziebe, S., Petersen, K., Lindenberg, S., Andersen, A.G., Gabrielsen, A. and Andersen, A.N. (1997) Embryo morphology or cleavage stage: how to select the best embryos for transfer after in-vitro fertilization. Hum. Reprod., 12, 15451549. Submitted on July, 2002; resubmitted on October 24, 2002; accepted on November 11, 2002

407