Human Reproduction Vol.20, No.12 pp. 3376–3384, 2005

doi:10.1093/humrep/dei228

Advance Access publication August 25, 2005.

One-step versus two-step culture of mouse preimplantation embryos: is there a difference? J.D.Biggers1,3, L.K.McGinnis1 and J.A.Lawitts2 1

Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, MA, 02115 and 2Beth Israel Deaconess Hospital, Boston, MA 02115, USA

3

To whom correspondence should be addressed. E-mail: [email protected]

BACKGROUND: A comparison has been made of the development of mouse zygotes in either one-step or two-step culture systems. METHODS: Embryo culture, blastocyst cell counts and embryo transfer were done. RESULTS: No significant differences were observed in the proportions of blastocysts, rates of hatching, numbers of cells in the inner cell mass (ICM) and trophectoderm (TE) that developed in protocols: one-step culture in potassium-enriched simplex optimized medium supplemented with glucose and amino acids (KSOMgAA), two-step culture in KSOMgAA/KSOMgAA, and two-step culture in G1.2/G2.2. No gross abnormalities were observed in the fetuses that developed from zygotes in the one-step protocol using KSOMgAA and a two-step protocol using G1.2/G2.2. The body weights of these two groups of fetuses were not significantly different and no developmental abnormalities were observed. No significant differences were observed in the proportions of blastocysts, rates of hatching, numbers of cells in the ICM and TE that developed in protocols: one-step culture in KSOMgAA, two-step culture in KSOMgAA/KSOMgAA, and two-step culture in DM2/DM1. EDTA is not toxic to the initial cleavage stages of development at a concentration of 0.01 mmol/ l in KSOMgAA. CONCLUSIONS: Two-step culture protocols are sufficient for the support of preimplantation mouse development in vitro but they are not necessary.

Introduction Preimplantation mammalian embryos have been cultured from the zygote to the blastocyst either by renewing the medium approximately midway in the culture period (two-step protocol) or not renewing the medium (one-step protocol). One-step and two-step protocols for the culture from the mouse zygote to the expanded blastocyst stage were first described by Whitten and Biggers (1968) and Chatot et al. (1989) respectively. The strong conviction that two-step protocols are absolutely necessary for the successful cultivation of human zygotes to the blastocyst stage has become widespread since the recommendations made by Gardner et al. (1998) (see also Gardner and Lane, 2003). There are several commercially available media sold for this purpose despite the fact that experimental support for this virtual dogma is relatively sparse (Summers and Biggers, 2003). In this paper we present results that compare the development of mouse zygotes to blastocysts using one- and two-step protocols. There are several variants of two-step protocols for the culture of preimplantation embryos. Gardner (1994) first advocated a two-step protocol to remove ammonium and pyrrolidone-5-carboxylic acid, the breakdown products of glutamine (Gln), which arise spontaneously in all media containing Gln (review: Biggers et al., 2004a). Renewing the medium, however, has potentially favourable and unfavourable effects. It may re-supply beneficial compounds that have been significantly depleted during the initial culture period. Conversely, it may remove any beneficial compounds secreted

by the cells during the initial period, such as growth factors. Two-step culture protocols can also involve media in which the second medium differs in composition from the first medium. This can involve the mere addition of a substance, e.g. the addition of glucose to medium CZB, which contains no glucose (Chatot et al., 1989). Alternatively it can involve the omission of a single substance from the first medium, as was done by Lane and Gardner (1995) in which the second medium (DM1) is the first medium (DM2) less EDTA (Table I). The EDTA was considered toxic in the later stages of preimplantation development. More complex two-step procedures can also involve several additions and subtractions of compounds from the first medium to give the second medium, e.g. as is done using medium G1.1 and medium G1.2 (Gardner et al., 1998) (Table I). In these protocols the embryos may be subjected to stress by being transferred to a new chemical environment. The results summarized in this paper describe the effects of renewing potassium-enriched simplex optimized medium supplemented with glucose and amino acids (KSOMgAA), the effects of removing EDTA from KSOMgAA and the effects of replacing medium G1.2 with medium G2.2 and DM2 with DM1.

Materials and methods Nomenclature and preparation of culture media The compositions of the media used in this work are shown in Table I. When two media are used sequentially they will be

3376 © The Author 2005. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: [email protected]

One- and two-step embryo culture protocols

Table I. Compositions of media KSOMgAA, G1, G2, DM1 and DM2 (Mmol/l) Component

KSOMgAA a G1b

NaCl 95 KCl 2.5 KH2PO4 0.35 Na phosphate – CaCl2·2H2O 1.7 MgSO4·7H2O 0.2 NaHCO3 25 Na pyruvate 0.2 Na lactate 10.0 Glucose 5.56 Glutamine 1.0 Alanylglutamine – L-Alanine 0.05 L-Asparagine 0.05 L-Aspartic acid 0.05 L-Glutamic acid 0.05 Glycine 0.05 L-Proline 0.05 L-Arginine 0.3 L-Cystine 0.05 L-Glutamic acid 0.05 L-Histidine 0.1 L-Isoleucine 0.2 L-Leucine 0.2 L-Lysine 0.2 L-Methionine 0.05 L-Phenylalanine 0.1 L-Serine 0.05 Taurine – L-Threonine 0.2 L-Tryptophan 0.025 L-Tyrosine 0.1 L-Valine 0.2 EDTA 0.01 Bovine serum albumin (mg/ml) 1 Streptomycin sulphate (μg/ml) 5 Penicillin G (IU/ml) 100 Choline chloride – Folic acid – Inositol – Nicotinamide – Pantothanate – Pyridoxal – Riboflavin – Thiamine –

90.08 5.5 – 0.25 1.8 1.0 25 0.32 10.5 0.5 – 0.5 0.1 0.1 0.1 0.1 0.1 0.1 – – – – – – – – – 0.1 0.1 – – – – 0.01

– – – – – – – –

G2b

DM2c DM1c

90.08 5.5 – 0.25 1.8 1.0 25 0.1 5.87 3.15 – 1.0 0.1 0.1 0.1 0.1 0.1 0.1 0.6 0.1

98.4 98.4 4.78 4.78 1.19 1.19 – – 1.7 1.7 1.19 1.19 25 25 0.37 0.37 4.79 4.79 3.4 3.4 1.0 1.0 – – 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – 0.1 4 4 5 5 100 100 – – – – – – – – – – – – – – – –

0.2 0.4 0.4 0.4 0.1 0.2 0.1 – 0.4 0.5 0.2 0.4 –

0.0072 0.0023 0.01 0.0082 0.0042 0.0049 0.00027 0.00296

a

Biggers et al. (2000). Gardner and Lane (2002). Media G1 and G2 are often denoted G1.2 and G2.2, being later versions of earlier formulations. c Gardner and Lane (1996). b

denoted generically as X/Y. Two two-step protocols were used: DM2/DM1 (Lane and Gardner, 1995; Gardner and Lane, 1996) was prepared in our laboratory, and G1.2/G2.2 was purchased from Vitro Life (Göteborg, Sweden). At the time our work was done the exact chemical compositions of media G1.2 and G2.2 were a trade secret, but recently their compositions have been published (Lane and Gardner, 2003). KSOMgAA is KSOM (Lawitts and Biggers, 1993) supplemented with amino acids (Biggers et al., 2000), and glucose at 5.56 mmol/l (Summers et al., 2000; Biggers and McGinnis, 2001). A modification of KSOMgAA was also used in which EDTA was omitted and is denoted EDTA-free KSOMgAA. Zygotes were collected in medium FHM (Lawitts and Biggers, 1993) which contained 1 mg/ml polyvinylalcohol (PVA) in place of bovine serum albumin (Biggers et al., 1997). All chemicals and reagents, unless otherwise stated, were from Sigma Corp. (St Louis, MO, USA).

Zygotes Donor females (CF1 mice, 6–8 weeks old; Harlan Sprague Dawley, Indianapolis, IN, USA) were injected i.p. with 5 IU of pregnant mare’s serum gonadotrophin (P.G.600; Intervet, Inc., Millsboro, DE, USA) followed 48 h later with 5 IU HCG. C57BL/6J × DBA/2J (BDF) male mice, 2–11 months old were also obtained from Harlan Sprague Dawley or The Jackson Laboratory, Bar Harbor, ME, USA. Animals were maintained in accordance with the guidelines of the Institutional Animal Care and Use Committee of Harvard Medical School. Culture Embryos were cultured in sets of 12 per micro-drop in a tri-gas atmosphere of 5% O2, 6% CO2 and 89% N2, as reported previously (Biggers and McGinnis, 2001). Embryos cultured using a one-step protocol for 4 days (144 h post-HCG) were cultured in 50 μl drops under oil (Sigma embryo-tested light mineral oil). Embryos cultured using a two-step protocol, where the medium was renewed or changed at 48 h (72 h post-HCG), were incubated in 20 μl drops under oil, following the recommendation of Gardner and Lane (1996). Embryo evaluation In one experiment embryos were graded at 48 h post-HCG at ×40 magnification on a warmed microscope stage at ∼35°C (Wild dissecting microscope), for stage of development including 1-, 2-, 3 – 4-, 8-cell and morula. In all experiments embryos were observed 96, 120 and 144 h post-HCG and graded for the stage of development including compaction, blastocoel formation and hatching. These times correspond approximately to 72, 96 and 120 h in culture. Differential ICM and TE cell counts After 144 h of culture, blastocysts were stained with polynucleotidespecific fluorochromes to differentially stain inner cell mass (ICM) and trophectoderm (TE) cells using a modification of a method originally described by Handyside and Hunter (1984, 1986) and Papaioannou and Ebert (1988). The details of the modified method are described by Biggers et al. (2000). Embryo transfer Pseudopregnant CD1 female mice (Taconic, Albany, NY, USA) were produced by mating CD1 females to vasectomized CD1 males. Cultured blastocysts and compact morulae (4th day of culture) were transferred into day 3 pseudopregnant females. Embryos produced in KSOMgAA were injected into one uterine horn at random and embryos produced in G1.2/G2.2 into the other uterine horn. Fetuses were collected from the pregnant mice on the 12th day after transfer for gross examination and body weight measurements. Experimental design and statistical analysis The results of three randomized block experiments are described in this paper, where each block is a separate replicate of the treatments. Experiment 1 compared four protocols: KSOMgAA, KSOMgAA/KSOMgAA, G1.2 and G1.2/G2.2, cultured until 144 h post-HCG. Experiment 2 was an analogous experiment using four treatments that compared the effects of culturing zygotes for 144 h post-HCG in KSOMgAA, KSOMgAA/EDTA-free KSOMgAA, DM2 and DM2/DM1. Experiment 3 compared the effects of culturing in KSOMgAA and G1.2/G2.2, followed by the transfer of embryos into the uterus of surrogate mothers. In all experiments the experimental unit consisted of 12 zygotes. Two sets of observations were made. Set 1 concerns the morphological development of the zygotes into 2-cell stages, and blastocysts by 96, 120 and 144 h post-HCG, including the numbers of zonaenclosed blastocysts, hatching and completely hatched blastocysts.

3377

J.D.Biggers et al.

These are categorical observations. Before applying exploratory graphical and statistical analyses, the data were re-expressed as the numbers of embryos that at least develop into blastocysts, which at least start to hatch and at least completely hatch. The data on the numbers of blastocysts that developed were examined using the parametric generalized linear model (McCullagh and Nelder, 1989), assuming that the errors follow a binomial distribution. An analysis of deviance partitioned the variation into between the replicates, the treatments and the replicate × treatments interaction, respectively. If the interaction was not significant at the P = 0.05 level the data were pooled over replicates. The blastocysts that developed by 144 h postHCG were then classified into the numbers of blastocysts that remained enclosed in the zona pellucida, those that had partially hatched and those that had completely hatched. The asymptotic or exact Kruskal–Wallis test of significance with one singly ordered variable was used to test whether these distributions differed significantly (Mehta and Patel, 2001). Set 2 concerns the numbers of ICM and TE cells in each blastocyst that had developed 144 h post-HCG in each culture condition. The distributions are displayed using notched box plots. The boxplots show the 10th, 25th, 50th (median), 75th, 90th percentiles; observations outside of the 10th and 90th percentiles are considered outliers. The notches on the box plots are the 95% median confidence limits. Two medians are significantly different if their confidence limits do not overlap. The errors associated with the cell counts are assumed to be formed by a multiplicative process. Thus the data have been submitted to an analysis of deviance assuming lognormally distributed errors (McCullagh and Nelder, 1989), partitioning the variation into between the replicates, the treatments and the replicate × treatments interaction. The analysis is summarized as a two-way analysis of variance (ANOVA) table. If the interaction was not significant at the P = 0.05 level the data were pooled over replicates. The embryo transfer data (embryos that at least implanted and the numbers of fetuses) were treated as two sets of stratified 2 × 2 contingency tables (Mehta and Patel, 2001), each stratum being one mother. The exact probability that the sets were homogeneous was calculated. If P > 0.05 the pooled exact estimate of the common odds ratio and its confidence limits (P = 0.95) was computed. The effects of the two media were considered significantly different if the confidence limits did not include 1. All statistical analyses were done, with one exception, using the S-Plus 5 package (Insightful, Seattle, WA, USA). The exact Kruskal– Wallis test for the analysis of a one-way ordered 2 × n contingency table, and the exact analysis of stratified 2 × 2 contingency tables was the exception analysed using StatXact 6 (Cytel Software Corporation, Cambridge, MA, USA). A difference in all analyses is considered statistically significant if P < 0.05.

Results Development in KSOMgAA, KSOMgAA/KSOMgAA, G1.2 and G1.2/G2.2 (experiment 1) Four protocols were used for the culture of embryos: (i) KSOMgAA until 144 h post-HCG with no renewal of the medium, (ii) KSOMgAA for 48 h followed by culture to 144 h post-HCG in KSOMgAA, (iii) G1.2 until 144 h post-HCG with no renewal of the medium, (iv) G1.2 for 48 h followed by culture to 144 h post-HCG in G2.2. Three replicates were done. Two experimental units were allotted to each treatment in replicate 1 and 3 experimental units in replicates 2 and 3. The results were obtained in two phases. The first phase, lasting until the medium was renewed in two of the protocols, consisted in reality of a comparison of the development of the 3378

zygotes cultured 48 h in two media: KSOMgAA and G1.2. At the end of this period the numbers of zygotes that remained as zygotes and the numbers of embryos that had developed to the 2-cell, 3–4-cell and 8-cell/compacted morula stage were recorded. The second phase, lasting from 48 h from the beginning of culture until 144 h post-HCG, consisted of the comparison of the four protocols. At 96, 120 and 144 h postHCG the numbers of embryos that had developed into zonaenclosed blastocysts, blastocysts that had commenced hatching and completed hatching were recorded. Finally the numbers of ICM cells and TE cells in each blastocyst were counted. Development during the first phase The distributions of the numbers of embryos that remained as 1-cell, developed to 2–4-cell embryos, 5–8-cell embryos and compacted morulae by 72 h post-HCG, when cultured in KSOMgAA and G1.2, are summarized in a 2×4 contingency table (Table II). There was no significant difference between the rates of development in KSOMgAA and G1.2 during this initial period of culture. Morphological development reached by 144 h post-HCG in KSOMgAA, KSOMgAA/KSOMgAA, G1.2 and G1.2/G2.2 The data from the three replicates have been pooled and summarized in Table IIIa. The rates of blastocyst formation were very similar in media KSOMgAA and KSOMgAA/KSOMgAA, reaching a rate of formation of ∼89% by 144 h post-HCG. There was also no significant difference between the rates of blastocyst formation in media G1.2 and G1.2/G2.2, both reaching a rate of ∼82%. Development of blastocysts was slightly slower using the G1.2 and G1.2/G2.2 protocols compared with the KSOMgAA and KSOMgAA/KSOMgAA protocols. In summary, high yields of blastocysts were obtained by 144 h post-HCG using all four protocols. The distributions of the zona-enclosed blastocysts, hatching and hatched blastocysts at 144 h post-HCG observed using the four protocols are shown in Table IIIb. Statistically, the four distributions are significantly different (P = 0.008). The significance is due to variation in the pattern of hatching using the G1.2, G1.2/G2.2 protocols and the KSOMgAA, KSOMgAA/ KSOMgAA protocols. The G1.2 and G1.2/G2.2 protocols supported a majority of blastocysts that started to hatch and completed hatching by 144 h post-HCG. In contrast, a greater percentage of blastocysts cultured using the KSOMgAA, KSOMgAA/KSOMgAA protocols started to hatch but fewer had completed hatching by 144 h post-HCG.

Table II. The distribution of the stages of embryonic development reached after culture for 72 h post-HCG in KSOMgAA and G1.2 Medium

n

1-cell

2–4-cell

5–8-cell

Morulae

KSOMgAA

192 192

1 (0.5) 6 (3.1)

29 (15.1) 31 (16.1)

142 (74.0) 124 (64.6)

20 (10.4) 31 (16.1)

G1.2

Values in parentheses are percentages. Probability that the distributions of the stages of development are not significantly different (exact Kruskal–Wallis test): P = 0.798.

One- and two-step embryo culture protocols

Table III(a). The morphological development of mouse zygotes in KSOMgAA/KSOMgAA, KSOMgAA, G1.2/G2.2 and G1.2 Development to at least the zona-enclosed blastocyst Protocol

Time of culture (h) 96

AA

KSOMg /KSOMg KSOMgAA G1.2/G2.2 G1.2

AA

a

34/60 (56.7) 35/60 (58.3) 18/60 (30) 21/60 (35)

120

144

84/96 (87.5) 85/96 (88.5) 77/96 (80.2) 79/96 (82.3)

84/96 (87.5) 87/96 (90.6) 78/96 (81.3) 79/96 (82.3)

Values in parentheses are percentages. a Only embryos from replicates 1 and 3 were counted at 96 h. Table III(b). The morphological development of mouse zygotes in KSOMgAA/KSOMgAA, KSOMgAA, G1.2/G2.2 and G1.2 Distributions of zona-enclosed blastocysts, hatching and hatched blastocysts 144 h post-HCG Protocol

Zona-enclosed blastocysts

Hatching blastocysts

Hatched blastocysts

Total

KSOMgAA/KSOMgAA KSOMgAA G1.2/G2.2 G1.2 Total

2 (2) 1 (1) 9 (12) 15 (19) 27

43 (51) 55 (63) 14 (18) 16 (20) 128

39 (46) 31 (36) 55 (71) 48 (61) 173

84 (100) 87 (100) 78 (100) 79 (100) 328

Values in parentheses are percentages. Probability that the four distributions not significantly different (exact Kruskal–Wallis test for one ordered variable): P = 0.008.

Numbers of ICM and TE cells in blastocysts that developed by 144 h post-HCG in KSOMgAA, KSOMgAA/KSOMgAA, G1.2 and G1.2/G2.2 The ICM data have been pooled over replicates to generate the box plots shown in Figure 1a. The numbers of ICM cells vary over a wide range (0–73). The notches of the distributions overlap for blastocysts developed in KSOMgAA and KSOMgAA/ KSOMgAA indicating no significant differences between the medians of these two distributions. A t-test based on the ANOVA shows that there is no significant difference between the means of the two distributions (P = 0.597). In contrast, the notches of the distributions for blastocysts that developed throughout in G1.2 and in G1.2/G2.2 respectively just fail to overlap, showing that the number of ICM cells which develop

is slightly less when cultivated in G1.2 compared with G1.2/ G2.2. A t-test shows that the means of the distributions are significantly different (P = 0.008). Only one trophoblastic vesicle was seen in this experiment in an embryo cultivated in G1.2. The analysis of deviance for the TE data shows a significant main effect of replicates (P = 0.008) and protocols (P = 0.0001), and also a significant replicate × protocol interaction (P = 0.0004). The significant interaction is due to a result in one replicate where the number of TE cells produced in G1.2 is greater than in G1.2/G2.2, the opposite of that seen in the other two replicates. The data have been pooled, however, to generate the box plots shown in Figure 1b since the aberrant result does not significantly affect the overall conclusions. The numbers of TE cells vary over a wide range over all protocols (32–212). The notches of the distributions overlap for blastocysts developed in KSOMgAA and KSOMgAA/KSOMgAA indicating no differences in the median TE counts in blastocysts that developed in these two protocols. A t-test shows that there is no significant difference between the means of the two distributions (P = 0.950). The notches for the data cultured in G1.2 and G1.2/G2.2 also overlap showing that the median numbers of TE cells produced in G1.2 and G1.2/G2.2 are not significantly different. A t-test confirms that there is no significant difference between the means of the two distributions (P = 0.597). The results also show that the TE cell counts in blastocysts produced using G1.2/G2.2 are not significantly different from the counts using KSOMgAA and KSOMgAA/KSOMgAA. Significantly fewer TE cells are produced when the blastocysts are produced in G1.2 alone. The distributions of the ratios between the numbers of ICM and TE cells are shown in Figure 1c. Statistical analyses demonstrate no significant differences between these ratios in blastocysts cultured in KSOMgAA, KSOMgAA/KSOMgAA and G1.2/G2.2. The ratio is significantly lower in blastocysts cultured in G1.2, demonstrating that relatively fewer ICM cells developed compared to TE cells. Development in KSOMgAA, KSOMgAA/EDTA-free KSOMgAA, DM2 and DM2/DM1 (experiment 2) The development of mouse zygotes by 144 h post-HCG using four culture protocols was compared. The protocols were culture of embryos in: (i) KSOMgAA until 144 h post-HCG with no renewal of the medium; (ii) KSOMgAA for 48 h

Figure 1. Distributions, summarized as box plots, of the numbers of ICM (a) and TE (b) in blastocysts cultured for 144 h post-hCG administration in KSOMgAA, KSOMgAA/KSOMgAA, G1.2 and G1.2/G2.2. (c) Distributions, summarized as box plots, of the ratio of ICM to TE cells. The box plots show the 10th, 25th, 50th (median), 75th and 90th percentiles, respectively. The values of the median of each box plot are shown to the left of the notches.

3379

J.D.Biggers et al.

followed by culture to 144 h post-HCG in EDTA-free KSOMgAA; (iii) DM2 until 144 h post-HCG with no renewal of the medium; (iv) DM2 for 48 h post-HCG followed by culture to 144 h post-HCG in DM1. Three replicates were done. Eighteen experimental units were allotted to each treatment in replicates 1 and 2, and 15 experimental units in replicate 3. The experimental design was similar to that described in experiment 1, the results being obtained in two phases. Development during the first phase The distributions of the numbers of embryos that remained as 1-cell, developed to 2-cell embryos, 3–4-cell embryos, and 8-cell embryos plus compacted morulae after 48 h of culture in KSOMgAA and DM2 are summarized in a 2×4 contingency table (Table IV). There was a highly significant difference between the rates of development in KSOMgAA and DM2 during this initial period of culture, the embryos cultured in KSOMgAA developing at a faster rate. Morphological development reached by 144 h post-HCG in KSOMgAA, KSOMgAA/EDTA-free KSOMgAA, DM2 and DM2/DM1 The data from the three replicates have been pooled and summarized in Table Va. The rates of blastocyst formation

Table IV. The distributions of the stages of embryonic development reached after culture in KSOMgAA and DM2 for 72 h post-HCG Medium

n

1-cell

2-cell

3–4-cell

8-cell/compacted

KSOMgAA DM2

408 408

10 (2.5) 8 ( 2.0)

40 (9.8) 32 (7.8)

128 (31.4) 235 (57.6)

230 (56.4) 133 (32.6)

Values in parentheses are percentages. Probability that the distributions of the stages of development are not significantly different (exact Kruskal–Wallis test): P < 0.0001.

were very similar in media KSOMgAA, KSOMgAA/EDTA-free KSOMgAA and DM2/DM1 throughout the 144 h period of culture. The rate of blastocyst formation in medium DM2 was similar to the other media for the first 120 h post-HCG but fell off slightly in the next 24 h, causing a significant difference between the yields of blastocysts that developed by 144 h postHCG (P = 0.02). In summary, high yields of blastocysts were obtained by 144 h post-HCG using all four protocols, the yield of blastocysts in medium DM2 (∼73%) being only ∼8% less than development in the other three protocols (∼80%). The distributions of zona-enclosed blastocysts, hatching and hatched blastocysts at 144 h post-HCG observed using the four protocols is shown in Table Vb. Statistically, the four distributions are significantly different (P < 0.001). The difference is due mainly to the very reduced onset of hatching in blastocysts that developed in DM2 (23%). The incidence of hatching and hatched blastocysts was very similar in embryos that developed in KSOMgAA, KSOMgAA/EDTA-free KSOMgAA and DM2/DM1 (82–91%). Overall, ∼36% of the blastocysts completed hatching. Numbers of ICM and TE cells in blastocysts that developed by 144 h post-HCG in KSOMgAA, DM2, KSOMgAA/EDTA-free KSOMgAA and DM2/DM1 The analysis of deviance of the ICM data shows significant main effects of replicates (P = 0.001) and protocols (P < 0.0001) and also a very significant replicate × protocol interaction (P = 0.001). The interaction is due to variation between the magnitudes of the differences between the effects of the four protocols. Nevertheless the patterns of the differences are similar so that the data have been pooled over replicates to generate the box plots shown in Figure 2a. The data show that the numbers of ICM cells vary over a wide range (0–58). Further, the notches of the distributions, which indicate the approximate confidence limits for the medians, overlap for blastocysts

Table V(a). The morphological development of mouse zygotes in KSOMgAA/EDTA-free KSOMgAA, KSOMgAA, DM2/DM1 and DM2 Development to at least the zona-enclosed blastocyst Protocol

AA

KSOMg /EDTA-free KSOMgAA DM2/DM1 DM2

Time of culture (h)

KSOMgAA

96

120

144

45/204 (22.1) 39/204 (19.1) 29/204 (14.2) 27/204 (13.2)

168/204 (82.4) 150/204 (73.5) 156/204 (76.5) 154/204 (75.5)

171/204 (83.8) 161/204 (78.9) 169/204 (82.8) 148/204 (72.6)

Values in parentheses are percentages. Table V(b). The morphological development of mouse zygotes in KSOMgAA/EDTA-free KSOMgAA, KSOMgAA, DM2/DM1 and DM2 Distributions of zona-enclosed blastocysts, hatching and hatched blastocysts 144 h post-HCG Protocol

Zona-enclosed blastocysts

Hatching blastocysts

Hatched blastocysts

Total

KSOMgAA/EDTA-free KSOMgAA KSOMgAA DM2/DM1 DM2 Total

14 (8) 20 (12) 31 (18) 108 (73) 173

81 (47) 86 (53) 86 (51) 34 (23) 287

76 (44) 55 (34) 52 (31) 6 (4) 189

171 (100) 161 (100) 169 (100) 148 (100) 649

Values in parentheses are percentages. Probability that the four distributions not significantly different (exact Kruskal–Wallis test one ordered variable): P < 0.0001.

3380

One- and two-step embryo culture protocols

Figure 2. Distributions, summarized as box plots, of the numbers of ICM (a) TE (b) in blastocysts cultured for 144 h post-hCG administration in KSOMgAA, KSOMgAA/EDTA-free KSOMgAA, DM2 and DM2/DM1. (c) Distributions, summarized as box plots, of the ratio of ICM to TE cells. The box plots show the 10th, 25th, 50th (median), 75th and 90th percentiles, respectively. The values of the median of each box plot are shown to the left of the notches.

developed in KSOMgAA, KSOMgAA/EDTA-free KSOMgAA and DM2/DM1 indicating no significant differences between the effects of these three protocols. The notch of the distribution for blastocysts that developed throughout in DM2 is well below the others indicating that the number of ICM cells that develop under these conditions is much less. A small part of the variation in the ICM count data is due to failure of the ICM to develop, giving rise to trophoblastic vesicles. The numbers of these vesicles observed in embryos that developed in KSOMgAA, KSOMgAA/EDTA-free KSOMgAA , DM2/DM1 and DM2 were one (1/83: 1.2%), two (2/99: 2.0%), two (2/72: 2.8%) and seven (7/69: 10.1%) respectively. The TE data have been pooled to generate the box plots shown in Figure 2b. The numbers of TE cells vary over a wide range (18–181). The notches of the distributions overlap for blastocysts developed in KSOMgAA and KSOMgAA/EDTA-free KSOMgAA indicating no differences in the TE counts in blastocysts that developed in these two protocols. The notch for the data cultured in DM2/DM1 does not overlap with the notches of the KSOMgAA data, and shows that the TE counts in blastocysts produced under this protocol are significantly, but only slightly, less than the counts using KSOMgAA. The notch of the distribution for blastocysts that developed in DM2 is well below the others indicating that the number of TE cells that develop under these conditions is very much less when cultured throughout in DM2. The distributions of the ratios between the numbers of ICM and TE cells are shown in Figure 2c. Statistical analyses (not shown) demonstrate no significant differences between these ratios in blastocysts cultured in KSOMgAA, KSOMgAA/EDTAfree KSOMgAA and DM2/DM1. The ratio is significantly lower in blastocysts cultured in DM2 demonstrating that relatively fewer ICM cells developed compared to TE cells. Implantation and fetal development rates of transferred embryos cultured in either KSOMgAA or G1.2/G2.2 (experiment 3) Blastocysts and compacted morulae for embryo transfer were obtained from zygotes cultured for 96 h post-HCG using either the one-step protocol KSOMgAA or the two-step protocol G1.2/ G2.2. The remainder of the embryos were cultured until 144 h post-HCG. Those embryos grown to 144 h post-HCG in KSOMgAA yielded 89% blastocysts containing median

numbers of 31 and 126 ICM and TE cells respectively, while for those grown in media G1.2/G2.2, 81% of zygotes had developed into blastocysts containing median numbers of 29 and 88 ICM and TE cells respectively. The hatching rates of the blastocysts were similar using the two media. Compact morulae and blastocysts were transferred into 30 recipient females of which 21 became pregnant. Forty-eight and 28 fetuses developed using the KSOMgAA and G1.2/G2.2 protocols respectively. No gross abnormalities, such as exencephaly, were observed in either group of fetuses. Four of the pregnant females had implantations and/or fetuses in only one horn and were excluded from the statistical analysis. These results are summarized in Table VI. The analyses show that the data within each group of mothers formed homogeneous groups and can be pooled. The common odds ratio showed that the numbers of implantations in each group were not significantly different. In contrast significantly fewer fetuses developed using the G1.2/G2.2 protocol compared with the KSOMgAA protocol. The distributions of the fetal body weights are shown in Figure 3. There is no significant difference in body weight of the fetuses that developed using the KSOMgAA and G1.2/G2.2 protocols. Discussion Effect of renewing KSOMgAA The results shown in Table III demonstrate that there are no differences in the percentages of zygotes that develop into blastocysts in KSOMgAA and KSOMgAA/KSOMgAA. Further, the rate of hatching was not significantly different using the

Table VI. The numbers of implantations and fetuses that developed from compacted morulae and blastocysts produced from zygotes cultured in either KSOMgAA or G1.2/G2.2 Medium

Transferred (n)

Implantations (n)

Fetuses (n)

KSOMgAA G1.2/G2.2

102 102

71 (69.6) 60 (58.9)

48 (67.6) 28 (46.6)

Implantations: KSOMgAA versus G1.2/G2.2: common odds ratio 0.59 (95% confidence limits 0.32–1.09). Fetuses: KSOMgAA versus G1.2/G2.2: common odds ratio 0.33 (95% confidence limits 0.14–0.73).

3381

J.D.Biggers et al.

Figure 3. Distributions, summarized as box plots, of the body weights of fetuses 15 and 16 days post-hCG that developed from morulae and blastocysts that had developed in either KSOMgAA or G1.2/G2.2.

two protocols. Also, the results summarized in Figure 1 demonstrate that there were no significant differences between the numbers of ICM cells and TE cells present in the blastocysts that developed in the two protocols. The results comparing development in KSOMgAA and KSOMgAA/EDTA-free KSOMgAA are very similar. The results summarized in Table V demonstrate that there are no differences in the percentages of zygotes that develop into blastocysts nor the rates of hatching, and the results shown in Figure 2 demonstrate that there were no significant differences between the numbers of ICM and TE cells present in the blastocysts that developed in the two protocols. Recently, Gardner and Lane (2003) and Lane and Gardner (2003) have criticized the use of KSOM media because of the accumulation of ammonium in the medium from the spontaneous decomposition of Gln. If this accumulation of ammonium is in fact detrimental then renewal of the medium should be beneficial. Our results show that this is not the case. The fact that we have also found no abnormal fetal development using KSOMgAA in a single-step protocol also suggests that Gardner and Lane’s concern is not supported by this evidence and by other reports published elsewhere (Sinawat et al., 2003; Biggers et al., 2004a,b; Summers et al., 2005). Our results also demonstrate that EDTA in KSOMgAA throughout the culture period did not affect development.

that developed. There have been several independent studies published that confirm the ability of EDTA to overcome the 2-cell block (Hoshi and Toyoda, 1985; Loutradis et al., 1987; Fissore et al., 1989; Nasr-Esfahani et al., 1990). Gardner and Lane (1996), however, reported that the body weights of fetuses that developed from blastocysts produced in vitro in the presence of EDTA and transferred to surrogate mothers were significantly lower than those that developed from blastocysts in DM2/DM1. The practice of not including EDTA in the second phase of culture in two-step protocols stems from this observation. Our results confirm that the presence of EDTA in DM2 during the second phase of culture has deleterious effects, shown by a significant reduction in the rate of hatching and significant reductions of the numbers of ICM and TE cells. This reduction in the numbers of cells in the blastocysts may presage the reduced fetal weights reported by Gardner and Lane (1996). The cause of the inhibitory effects on the development of the mouse blastocyst is unknown. The development of the bovine blastocyst is also inhibited by EDTA, possibly by the inhibition of glycolysis when there is a rapid increase in the utilization of glucose, presumably as an energy source (Gardner et al., 2000). A similar increase in the uptake of glucose begins in the mouse at the morula stage (review: Biggers et al., 1989) whose utilization could be inhibited by a high concentration of EDTA. The reason that EDTA is not deleterious in KSOMgAA but deleterious in DM2 on the development of the blastocyst may well be a concentration effect. The majority of the concentration– response curves, determined independently by Abramczuk et al. (1977), Loutradis et al. (1987) and Fissore et al. (1989), suggested that most of the embryos overcame the 2-cell block when 5–10 μmol/l EDTA is added to media. Higher concentrations were also effective and toxicity did not occur until concentrations >200 μmol/l were used. Chatot et al. (1989) used 100 μmol/l EDTA in a new medium they produced for the culture of mouse preimplantation embryos called CZB, and Gardner and Lane (1996) used 100 μmol/l EDTA in medium DM2. Lawitts and Biggers (1991) exploited the experimental strategy called sequential simplex optimization to determine simultaneously the optimum concentrations of all components in the medium in order to take account of all the possible interactions between the effects of the several components. The result was medium SOM (Lawitts and Biggers, 1992) (the forerunner of KSOM) in which the optimum concentration of EDTA was one order of magnitude less (10 μmol/l) than that used by other investigators. Possibly this lower concentration of EDTA does not lead to the adverse effects that have been observed when 100 μmol/l EDTA is included in the media. Development in G1.2 and G1.2/G2.2

Effect of removing EDTA from medium DM2 Early attempts to culture newly fertilized mouse ova, particularly from outbred and inbred strains, were frustrated by the ‘2-cell block’ (review Biggers, 1993). Abramczuk et al. (1977) overcame this block by adding EDTA to Whitten’s medium (Whitten, 1971). Although the 2-cell block was not absolute in the strains of mice they used, the addition of EDTA spanning a range of concentrations increased the numbers of blastocysts 3382

The proportions of blastocysts that formed in G1.2 and G1.2/G2.2 were not significantly different, as well as those which hatched (P = 0.352). Thus the components and their concentrations in G1.2 are able to support the morphological structure of the blastocyst throughout the culture period without replacement. The changes in the composition of the environment produced by replacing G1.2 with G2.2 do not appear to affect the morphological development. In contrast, the numbers of ICM

One- and two-step embryo culture protocols

cells which formed in blastocysts that developed only in G1.2 is significantly less than those that develop when G1.2 is replaced with G2.2 48 h after the beginning of the culture period. Unfortunately the compositions G1.2 and G2.2 were trade secrets when our work was done. Only recently have the compositions of these media been revealed (Gardner and Lane, 2002). Several differences exist between the compositions of the two media, so it is impossible to pin-point the cause of the effects on ICM numbers. In our hands the numbers of ICM and TE cells produced in the blastocysts were larger using KSOMgAA than G1.2/G2.2 in contrast to the results reported by Lane and Gardner (2003). The importance of such differences should not be overemphasized since, in using commercially prepared media, changes could occur during the inevitable shipping and long storage after preparation. Clinical implications for human IVF The results presented in this paper suggest that a two-step protocol may be sufficient for the culture of mouse zygotes to blastocysts but that it is not necessary. It is important to know whether this conclusion applies to the media used for the culture of the human preimplantation embryo to the blastocyst stage. Two studies suggest that this conclusion may be valid. Biggers and Racowsky (2002) showed that human zygotes can develop into blastocysts in 5 days with high efficiency when cultured using KSOM plus amino acids without renewal of the medium. Normal babies were born using this medium. Macklon et al. (2002) cultured human zygotes in a 17:3 mixture of Earle’s balanced salt solution and Ham’s F10 medium supplemented with albumin which they called the Rotterdam medium. The zygotes developed with equal efficiency when they were cultured either in the Rotterdam medium using a one-step protocol or in G1.2/G2.2 using a two-step protocol. Further, renewing the Rotterdam medium after 48 h culture did not increase the efficiency. Unfortunately, the field is dominated by the strongly held opinion, originating in the writings of Gardner and Lane, that the two-step protocol is absolutely necessary to imitate the environment encountered by preimplantation embryos in situ (see several current reviews: Mehta, 2001; Kolibianakis and Devroey, 2002; Pool, 2002; Quinn, 2004). The popularity of this conviction is due to its intuitive appeal, but critical examination shows that it is not supported by direct experimental evidence (reviews: Biggers, 2003; Summers and Biggers, 2003). Acknowledgements We wish to acknowledge Dr Betsey Williams for helpful criticism of this paper.

References Abramczuk J, Solter D and Koprowski H (1977) The beneficial effect of EDTA on development of mouse one-cell embryos in chemically defined medium. Dev Biol 61,378–383. Biggers JD (1993) The culture of the mammalian preimplantation embryo. In Gianaroli L, Campana A and Trounson AO (eds) Implantation in Mammals. Raven Press, New York, pp 123–136. Biggers JD (2003) Fundamentals of the design of culture media that support human preimplantation development. In Van Blerkom J (ed) Essential IVF. Kluwer, Norwell, MA, pp 291–332.

Biggers JD and McGinnis LK (2001) Evidence that glucose is not always an inhibitor of mouse preimplantation development in vitro. Hum Reprod 16,153–163. Biggers JD and Racowsky C (2002) The development of fertilized human ova to the blastocyst stage in medium KSOMAA: is a two-step protocol necessary? Reprod Biomed Online 5,133–140. Biggers JD, Gardner DK and Leese H (1989) Control of carbohydrate metabolism in preimplantation mammalian embryos. In Rosenblum IY and Heyner S (eds) Regulation of Growth in Development. CRC Press, Boca Raton, FL, pp 20–32. Biggers JD, Summers MC and McGinnis LK (1997) Polyvinyl alcohol and amino acids as substitutes for bovine serum albumin in culture media for mouse preimplantation embryos. Hum Reprod Update 3,125–135. Biggers JD, McGinnis LK and Raffin M (2000) Amino acids and preimplantation development of the mouse in protein-free KSOM. Biol Reprod 63,281–293. Biggers JD, McGinnis LK and Lawitts JA (2004a) Enhanced effect of glycylL-glutamine on mouse preimplantation embryos in vitro. Reprod Biomed Online 9,59–69. Biggers JD, McGinnis LK and Summers MC (2004b) Discrepancies between the effects of glutamine in cultures of preimplantation mouse embryos. Reprod Biomed Online 9,70–73. Chatot CI, Ziomek CA, Bavister BD, Lewis JL and Torres I (1989) An improved culture medium supports development of random-bred 1-cell mouse embryos in vitro. J Reprod Fertil 86,679–688. Fissore RA, Jackson KV and Kiessling AA (1989) Mouse zygote development in culture medium without protein in the presence of ethylenediaminetetraacetic acid. Biol Reprod 41,835–841. Gardner DK (1994) Mammalian embryo culture in the absence of serum or somatic cell support. Cell Biol Int 18,1163–1179. Gardner DK and Lane M (1996) Alleviation of the ‘2-cell block’ and development to the blastocyst of CF1 mouse embryos: role of amino acids, EDTA and physical parameters. Hum Reprod 11,2703–2712. Gardner DK and Lane M (2002) Development of viable mammalian embryos in vitro: evolution of sequential media. In Cibelli I, Lanza R, Campell K and West M (eds) Principles of Cloning. Academic Press, New York, pp 187–213. Gardner DK and Lane M (2003) Towards a single embryo transfer. Reprod Biomed Online 6,470–481. Gardner DK, Schoolcraft WB, Wagley L, Schlenker T, Stevens J and Hesla J (1998) A prospective randomized trial of blastocyst culture and transfer in in-vitro fertilization. Hum Reprod 13,3434–3440. Gardner DK, Lane MW and Lane M (2000) EDTA stimulates cleavage stage bovine embryo development in culture but inhibits blastocyst development and differentiation. Mol Reprod Dev 57,256–261. Handyside AH and Hunter S (1984) A rapid procedure for visualizing the inner cell mass and trophectoderm nuclei of mouse blastocysts in situ using polynucleotide-specific fluorochromes. J Exp Med 231,429–434. Handyside AH and Hunter S (1986) Cell division and death in the mouse blastocyst before implantation. Roux’s Arch Devel Biol 195,519–526. Hoshi M and Toyoda Y (1985) Effect of EDTA on the preimplantation development of mouse embryos fertilized in vitro. Jpn J Zootech Sci 56,931–937. Kolibianakis ES and Devroey P (2002) Blastocyst culture: facts and fiction. Reprod Biomed Online 5,285–293. Lane M and Gardner DK (1995) Removal of embryotoxic ammonium from the culture medium by in situ enzymatic conversion to glutamate. J Exp Zool 271,356–363. Lane M and Gardner DK (2003) Ammonium induces aberrant blastocyst differentiation, metabolism, pH regulation, gene expression and subsequently alters fetal development in the mouse. Biol Reprod 69,1109–1117. Lawitts JA and Biggers JD (1991) Optimization of mouse embryo culture media using simplex methods. J Reprod Fertil 91,543–556. Lawitts JA and Biggers JD (1992) Joint effects of sodium chloride, glutamine, and glucose in mouse preimplantation embryo culture media. Mol Reprod Dev 31,189–194. Lawitts JA and Biggers JD (1993) Culture of preimplantation embryos. In Wassarman PM and Depamphilis ML (eds), Guide to Techniques in Mouse Development. Methods in Enzymology, vol 225. Academic Press, San Diego, CA, pp 153–164. Loutradis D, John D and Keissling AA (1987) Hypoxanthine causes a 2-cell block in random-bred mouse embryos. Biol Reprod 37,111–316. Macklon NS, Pieters MHEC, Hassan MA, Jeucken PHM, Eijkemans MJC and Fauser BCJM (2002) A prospective randomized comparison of sequential versus monoculture systems for in-vitro human blastocyst development. Hum Reprod 17,2700–2705. McCullagh P and Nelder JA (1989) Generalized Linear Models. 2nd edn, Chapman & Hall, London, pp 98–148.

3383

J.D.Biggers et al. Mehta RH (2001) Growth of human preimplantation embryos in vitro. Reprod Biomed Online 2,113–119. Mehta C and Patel N (2001) StatXact 5, Vol 2. Cytel Software Corporation, Cambridge, MA, pp 751–771. Nasr-Esfahani MH, Winston NJ and Johnson MH (1992) Effects of glucose, glutamine, ethylenediaminetetraacetic acid and oxygen tension on the concentration of reactive oxygen species and on development of the mouse preimplantation embryo in vitro. J Reprod Fertil 96,219–231. Papaioannou VE and Ebert KM (1988) The preimplantation pig embryo cell number and allocation to trophectoderm and inner cell mass of the blastocyst in vivo and in vitro. Development 102,793–803. Pool TB (2002) Recent advances in the production of viable human embryos in vitro. Reprod Biomed Online 4,294–302. Quinn P (2004) The development and impact of culture media for assisted reproductive technologies. Fertil Steril 81,27–29. Sinawat S, Hsaio W-C, Flockart JH, Kaufman MH, Keith J and West JD (2003) Fetal abnormalities produced after preimplantation exposure of mouse embryos to ammonium chloride. Hum Reprod 18,2157–2165.

3384

Summers MC and Biggers JD (2003) Chemically-defined media and the culture of mammalian preimplantation embryos: historical perspective and current issues. Hum Reprod Update 9,557–582. Summers MC, McGinnis LK, Lawitts JA, Raffin M and Biggers JD (2000) IVF of mouse ova in a simplex optimized medium supplemented with amino acids. Hum Reprod 15,1791–1801. Summers MC, McGinnis LK, Lawitt, JA and Biggers JD (2005) Development of the mouse following IVF in media containing either L-glutamine or glycyl-L-glutamine. Hum Reprod, 20,1364–1371. Whitten WK (1971) Nutrient requirements for the culture of preimplantation embryos in vitro. Adv Biosci 6,131–141. Whitten WK and Biggers JD (1968) Complete development in vitro of the preimplantation stages of the mouse in a simple chemically defined medium. J Reprod Fertil 17,399–401.

Submitted on April 1, 2005; resubmitted on June 11, 2005; accepted on June 23, 2005