Evidence that glucose is not always an inhibitor of mouse preimplantation development in vitro

John D. Biggers1 and Lynda K. McGinnis

Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, MA 02115, USA


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
A factorial experimental design was used to examine the effects of 16 combinations of four concentrations of glucose (0.20, 0.60, 1.8, 5.4 mmol/l) and four concentrations of potassium dihydrogen phosphate (KH2PO4; 0.05, 0.15, 0.45, 1.35 mmol/l) on the development in vitro of outbred CF1 mouse zygotes. Three responses were measured: (i) the number of zona-enclosed blastocysts; (ii) the number of blastocysts that started to hatch; and (iii) the total cell counts in the blastocysts. General linear modelling was used to estimate the most parsimonious two-dimensional concentration–response surfaces that represent the three responses to the different concentrations of glucose and KH2PO4. There were no significant interactions between the effects of glucose and KH2PO4 in all cases. Thus, the effects of glucose and phosphate are independent. No significant effects of glucose on blastocyst formation and the initiation of hatching were observed. Increasing the concentration of KH2PO4 inhibited slightly (<=20%) the development of zygotes into blastocysts and the initiation of hatching. The slight inhibitory effects of KH2PO4 appeared to be due to the inhibition of the development of a few sensitive embryos. No significant effects of glucose and KH2PO4 were observed on the total cell counts.

Key words: embryo culture/glucose/mouse preimplantation development/phosphate


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The finding during the late 1980s (Schini and Bavister, 1988Go) that physiological concentrations of glucose inhibit the development of the 2-cell hamster embryo in vitro was soon observed using embryos of other species: mouse (Chatot et al., 1989Go, 1990Go; Aoki et al., 1990Go; Diamond et al., 1991Go; Lawitts and Biggers 1991aGo; Brown and Whittingham, 1992Go; Scott et al., 1993Go; Scott and Whittingham, 1996Go; Haraguchi et al., 1996Go), rat (de Hertogh et al., 1991Go; Reed et al., 1992aGo; Miyoshi et al., 1994Go), sheep (Thompson et al., 1992Go), cow (Takahashi and First, 1992Go) and human (Conaghan et al., 1993Go; Hardy, 1994Go; Quinn et al., 1995Go; Coates et al., 1999Go). This seemingly overwhelming evidence has tended to discount the significance of the fact that glucose does not always inhibit the initial stages of preimplantation development in vitro. Thus, in 1985 it was reported that cleavage of the sheep zygote was stimulated by glucose (Betterbed and Wright, 1985Go). Subsequent investigations on the preimplantation development in vitro of sheep failed to show any inhibitory effect of glucose (Thompson et al., 1989Go; McGinnis and Youngs, 1992Go). No inhibitory effect on preimplantation development in vitro has been reported in the pig (Petters et al., 1990Go; Hagen et al., 1991Go), except in the case of one experiment out of several reported later (Reed et al., 1992bGo). The claim by others (Seshagiri and Bavister, 1991Go) that glucose inhibits the preimplantation development of the pig appears to be due to their misinterpretation of the published analyses of variance (Petters et al., 1990Go). The latter authors found a statistically significant glucosexphosphate interaction due to low responses when glucose and phosphate were both omitted from the medium. Soon after this it was reported (Lawitts and Biggers, 1992Go) that the addition of glucose to a medium called simplex optimised medium (SOM), in a concentration as high as 5 mmol/l, did not inhibit the development of outbred CF1 mouse zygotes to blastocysts. Medium SOM was superseded by a modified medium called KSOM (Lawitts and Biggers, 1993Go). Using KSOM, other workers (Summers et al., 1995Go, 2000Go; Biggers et al., 1997Go) have obtained further independent evidence showing that glucose does not inhibit the development of the outbred CF1 mouse zygote into hatching blastocysts when added in concentrations of 5.56 mmol/l.

The data presented previously (Schini and Bavister, 1988Go) on the joint effects of phosphate and glucose in overcoming the 2-cell block in hamsters suggested that glucose and phosphate inhibit development additively, but independently. In contrast, it was shown later (Reed et al., 1992aGo) that the development of 8-cell rat embryos was inhibited only when glucose and phosphate were present in the medium together. A similar observation was made subsequently (Scott and Whittingham, 1996Go) using two outbred strains of mice (CF1 and CD1). The results of a factorial experiment on the joint effects of glucose and KH2PO4 on the development of CF1 zygotes in a modified human tubal fluid medium have been graphically sketched (Quinn and Horstman, 1998Go). These authors' illustration suggests that glucose inhibits development in the absence of phosphate but is barely inhibitory in the presence of 1.85 mmol/l phosphate. The possibility of the existence of a glucosexphosphate interaction was not tested statistically. No studies have been reported on the effects of varying the phosphate concentration in KSOM. Therefore, the effects of different combinations of KH2PO4 and glucose concentrations (in a factorial experimental design) on the development of outbred CF1 mouse zygotes cultured in KSOM were examined. A detailed statistical analysis of results from this type of experiment was necessary to determine whether the effects of glucose on the development of mouse preimplantation embryos cultured in KSOM are influenced by the concentration of KH2PO4. A multivariate approach was used to assess the joint effects of glucose and KH2PO4 on preimplantation development, by measuring: (i) the percentages of blastocysts that develop from zygotes; (ii) the percentages of blastocysts that initiate hatching; and (iii) the total cell counts in the blastocysts that developed after 144 h after human chorionic gonadotrophin (HCG) administration.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Reagents
All chemicals and hormones were obtained from Sigma Chemical Corporation, St Louis, MO, USA.

Donors
Zygotes were obtained by mating CF-1 female mice, 6–8 weeks old (Charles River Laboratories, Willmington, MA, USA) and BDF1 males, 2–11 months old (Charles River Laboratories, Raleigh, NC, USA). Females were stimulated with 5 IU pregnant mare's serum gonadotrophin, and superovulated 48 h later with 5 IU HCG. The oviducts were flushed 22–26 h after HCG administration with a glucose/KH2PO4-free HEPES-buffered modification of KSOM, denoted FHM (Lawitts and Biggers, 1993Go), and embryos with two pronuclei were selected for culture. The medium used for embryo collection and holding prior to culture was also FHM.

Culture media
All culture media were formulated from KSOM (Table IGo). The medium, omitting glucose and KH2PO4, was prepared as a 2x solution (complete KSOM without the CaCl2), frozen in 50 ml culture tubes at –70°C for up to 3 months. On the day before embryo collection, 50 ml of KSOM was thawed and supplemented with CaCl2 (also stored as a frozen stock), glucose, KH2PO4 and H2O. Eighteen different media were prepared, each containing a specific combination of the concentrations of glucose and KH2PO4 shown in Table IIGo.


View this table:
[in this window]
[in a new window]
 
Table I. Composition of KSOM (Lawitts and Biggers, 1993Go)
 

View this table:
[in this window]
[in a new window]
 
Table II. Osmolality (mosmols) and K+ concentration (mmol/l K+, in parentheses) of media used in the 4x4 factorial experiment
 
Embryo culture
Embryos were cultured in groups of 12 per 50 µl droplet of medium overlaid with embryo-tested mineral oil, in modular incubator chambers (Billups Rothenberg, Inc., Del Mar, CA, USA) which were gassed with a mixture of 5% O2, 6% CO2 and 89% N2 (Lawitts and Biggers, 1993Go). Culture plates (60 mm suspension culture plates, Corning Inc., Corning, NY, USA) were prepared one day before embryo collection and equilibrated in the module overnight. Embryos were cultured for 144 h post-HCG administration.

Embryo evaluation
Embryos were observed at 96, 120 and 144 h post-HCG administration, on a Wild dissecting microscope fitted with a warmed stage at ~35°C using 40x magnification. They were graded for stage of development including compaction, blastocoele formation and hatching.

Total blastocyst cell counts
Embryos were fixed 144 h post-HCG administration in 3% formaldehyde for 15 min at 37°C, after which the nuclei were stained with the fluorochrome, Hoechst 33258 (1 µg/ml) in Dulbecco's phosphate-buffered saline (PBS) for 15 min at room temperature (Ebert et al., 1985Go). Blastocysts were mounted on glass slides, in groups of one to four, and covered with a mounting medium (50% glycerol, 50% PBS, 5 mg/ml sodium azide and 1 µg/ml Hoechst 33258). A glass coverslip was placed over the embryos and sealed in place with clear nail polish. Nuclei were counted at a magnification of 40x using an inverted Zeiss epifluorescence microscope with a 365 nm band pass excitation filter and a 420 nm long pass barrier filter.

Biometrical considerations
Concentration–response surfaces as models for chemically defined media
It was proposed previously (Biggers et al., 1957Go) that the responses of cells and organs cultured in chemically defined media can be modelled by a multidimensional concentration–response surface. This model is a generalization of the dose–response line familiar in hormone biological assays. Later, it was suggested that a similar model be used to represent the response of preimplantation embryos in vitro in chemically defined media (Biggers et al., 1972Go). An 11-dimensional concentration–response surface was assumed when sequential simplex optimization was used determine the concentrations of 10 components that were adopted in SOM (Lawitts and Biggers, 1991bGo). In the present paper, it is assumed that the response of the zygote to joint changes in the concentrations of glucose and KH2PO4 can be represented by a two-dimensional surface. An estimate of the shape of part of this surface was obtained by using a 4x4 factorial design in which four concentrations each of glucose and KH2PO4 were used in all 16 combinations to span the region of interest.

Multivariate characteristics of the responses
Each embryo was observed at three different times, so that the responses were repeat (longitudinal) measurements (Diggle et al., 1994Go). At each time, the numbers of embryos that developed into three developmental stages: zona-intact blastocysts, partially and completely hatched blastocysts were counted, so that the data were also ordinal categorical responses (Clogg and Shihadeh, 1994Go). Thus, the multivariate observations were serially correlated both longitudinally and ordinally. The raw data were re-expressed as the numbers of zygotes that developed at least into blastocysts and at least begin to hatch by 120 and 144 h post-HCG in each experimental unit. The response in each drop was then expressed as the percentage of blastocysts that developed from zygotes and the percentage of these blastocysts that initiated hatching. The mean percentage of the percentages in a series of drops was estimated as a weighted mean, weighting each observation in proportion to the number of blastocysts that formed in each drop. Total cell counts were made on all the blastocysts that developed.

General linear regression (McCullagh and Nelder, 1989Go)
All categorical responses just described were analysed using general linear regression by fitting the model:



where g({eta}) is the link function; b0 is the intercept; r is the replicate number; g is the coded logarithm of the glucose concentration; p is the coded logarithm of the KH2PO4; b1, b2, b3, b12, b13, b23, b123 are regression coefficients, and eijk is the random error. It is also assumed that the data have binomial errors and the link function is the logit (natural logarithm of the odds ratio) transformation. The dependent variables were treated as factors in the first analysis of each set of data. The non-significant main effects and interactions were eliminated to find the most parsimonious model needed to adequately describe the data. In these regression analyses the glucose and phosphate concentrations were treated as real numbers. The cell counts have also been analysed by general linear regression using the same model (1), assuming that the errors arise from a multiplicative process with lognormally distributed errors where the link function is 1. The goodness of fit of each model was tested by the residual deviance which is distributed as a {chi}2. All statistical analyses were summarized as analysis of deviance tables. All computations were performed using S-Plus 2000 (MathSoft, Cambridge, MA, USA).

Boxplots (McGill et al., 1978Go)
All distributions were summarized using notched boxplots. The boxplots show the 10th, 25th, 50th (median), 75th and 90th percentiles and outliers. The notches on the boxplots were the median confidence limits. Two medians were significantly different if their confidence limits did not overlap. All computations were performed using S-Plus 2000.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Experimental design
A randomized block experimental design was used in which 16 treatment combinations defined by four concentrations each of glucose and KH2PO4 in the medium were simultaneously compared. The four concentrations of glucose were 0.2, 0.6, 1.8 and 5.4 mmol/l, and the four concentrations of KH2PO4 were 0.05, 0.15, 0.45 and 1.35 mmol/l. Both sets of concentrations form geometric progressions to estimate efficiently the concentration–response surface. The concentration of glucose ranged from 0.2 mmol/l to 5.4 mmol/l, the highest concentration being that normally present in blood. This range spans the concentration of glucose reported to occur in the mouse oviduct (3.40 mmol/l) (Gardner and Leese, 1990Go). The KH2PO4 concentration ranged from 0.05 mmol/l to 1.35 mmol/l; this range spans the concentration of phosphate (0.66 mmol/l) reported in the sheep oviduct (Restall and Wales, 1966Go). This value was obtained after converting the published value in mequiv/l to mmol/l using an effective equivalent weight of 1.8 (personal communication, R.G.Wales, Murdoch University, Perth, Western Australia). The range also spans the concentration of phosphate (1.26 mmol/l) in the human oviduct, converted from the published value expressed as phosphorus in mg % (Lippes et al., 1972Go). [Yamada and Nishikimi (1999) have indirectly referred to these two papers on phosphate concentrations in the sheep and human, but have imprecisely reported a range of 0.37 to 1.19 mmol/l]. Also, the upper concentration of phosphate used in the current experiment was <2 mmol/l, the concentration reported to be optimal in media for the culture of several cell lines (Waymouth, 1954Go). The osmolalities of the 16 variants of KSOM differed only slightly (Table IIGo). The experimental unit randomized to the treatment combinations was a set of 12 zygotes. Four replicates were done; two of the replicates had one experimental unit per treatment and the other two replicates had two experimental units per cell. External control groups were also simultaneously cultured in standard KSOM as well as in KSOM from which glucose and KH2PO4 had been omitted.

Effect of KSOM and glucose/KH2PO4-free KSOM on blastocyst development and initiation of hatching
The percentages of zygotes that developed in KSOM at least to the blastocyst stage and at least starting to hatch by 120 and 144 h post-HCG are summarized in Table IIIGo. These levels of response are comparable with independent results using KSOM described elsewhere (Erbach et al., 1995Go; Biggers et al., 2000Go). The percentages of zygotes that developed in glucose/KH2PO4-free KSOM at least to the blastocyst stage and at least starting to hatch by 120 and 144 h post-HCG administration are also summarized in Table IIIGo. The omission of glucose and KH2PO4 from KSOM had no effect on the percentages of zygotes that developed at least to blastocysts, or at least began to hatch.


View this table:
[in this window]
[in a new window]
 
Table III. Percentages of zygotes developing at least to the blastocyst stage and at least starting to hatch in KSOM, and glucose/KH2PO4-free KSOM (external controls)
 
The distributions of the numbers of cells in blastocysts cultured from zygotes in KSOM and glucose/KH2PO4-free KSOM are summarized by the notched box plots shown in Figure 1Go. The median cell counts in the blastocysts produced in KSOM and glucose/KH2PO4-free KSOM were 72 and 56 respectively. The fact that the notches of the boxplots overlap suggests that this reduction in cell counts is not significant. Failure to reach statistical significance, however, was due in part to the differences between the levels of responses in the two replicates. An analysis of deviance (not shown), in which the variation between replicates is eliminated, showed that the reduction in cell number was statistically significant (P = 0.016).



View larger version (14K):
[in this window]
[in a new window]
 
Figure 1. Boxplots showing the distributions of total cell counts in blastocysts that developed from zygotes into blastocysts by 144 h post-HCG administration in KSOM and glucose/KH2PO4-free KSOM. For explanation of boxplots, see text.

 
Effect of varying the content of glucose and KH2PO4 in KSOM on blastocyst development and the initiation of hatching
The percentage of embryos that developed at least to the blastocyst stage after culture in the 16 media for 120 h post-HCG are shown in Table IVGo. There was a high percentage (range 64–92%) of zygotes that developed into blastocysts over all the 16 combinations of glucose and KH2PO4 concentrations. The responses observed in the external control media (Table IIIGo) fell into this range. The analysis of deviance, using model 1 (Table VGo) shows that there were highly significant differences between the mean responses of the four replicates. However, all the replicatexfactors interactions were not significant, with the exception of the replicatexglucose interaction, the significance of which was marginal (P = 0.049). The patterns of responses of the embryos to the 16 combinations of glucose and KH2PO4 concentration combinations were therefore assumed to be comparable in all four replicates. The analysis of deviance also showed that there was no significant effect of glucose, but an effect of KH2PO4 (P = 0.002) was shown. Further, the effect of KH2PO4 was independent of the concentration of glucose, since the glucosexKH2PO4 interaction was not significant. Thus, the effects of varying the concentrations of glucose and KH2PO4 could be assessed separately by their main effects. The distributions of these effects are summarized in the form of boxplots in Figure 2a,eGo. A linear regression of the logit of proportion of blastocysts on the logarithm of the concentration of KH2PO4, pooled over all glucose concentrations, adequately fitted the data, and its slope was estimated to be b = –0.47, SEM = 0.13, df = 89. The non-inhibitory effect of increasing the concentration of glucose, and the minor inhibitory effect of increasing the concentration of KH2PO4 in the medium, were clear.


View this table:
[in this window]
[in a new window]
 
Table IV. Percentages of zygotes developing at least to the blastocyst stage after culture for 120 and 144 h post-HCG administration respectively in different combinations of glucose and KH2PO4 concentrations
 

View this table:
[in this window]
[in a new window]
 
Table V. Analyses of deviance of the percentages of zygotes that developed at least into blastocysts by 120 and 144 h post-HCG administration, and the percentages of blastocysts at least starting to hatch by 120 and 144 h post-HCG administration
 


View larger version (32K):
[in this window]
[in a new window]
 
Figure 2. A matrix of boxplots showing the main effects of glucose and KH2PO4 on the development of zygotes in a factorial combination of different concentrations of glucose and KH2PO4. (a, e) Percentage of zygotes that develop at least to the zona-enclosed blastocyst stage by 120 h post-HCG administration; (b, f) percentage of zygotes that develop at least to the zona-enclosed blastocyst stage by 144 h post-HCG; (c, g) percentage of blastocysts that at least partially hatch by 120 h post-HCG; (d, h) percentage of blastocysts that at least partially hatch by 144 h post-HCG.

 
The results obtained on blastocyst formation 144 h post-HCG are also shown in Tables IV and VIGoGo (see also Figure 2b,fGo). The responses observed in the external control media (Table IIIGo) fell into this range. There was no significant effect of glucose concentration on the percentages of blastocysts that developed, and a marginally significant effect of KH2PO4 concentration (P = 0.03). The logit of the proportion of blastocysts was again linearly related to the logarithm of the concentration of KH2PO4 and was negative (b = –0.41, SEM = 0.15, df = 89). The results were very similar to those described for the data obtained at 120 h post-HCG, except that the responses were slightly higher and the inhibitory effect of KH2PO4 was less. The statistical significance of these differences was not tested as the results obtained at 120 and 144 h were serially correlated and therefore not independent.


View this table:
[in this window]
[in a new window]
 
Table VI. The percentages of blastocysts at least starting to hatch after culture for 120 and 144 h post-HCG administration respectively in different combinations of glucose and KH2PO4 concentrations
 
The results obtained 120 h post-HCG administration on the effects of different concentrations of glucose and KH2PO4 on the initiation of hatching are shown in Tables V and VIGoGo (see also Figure 2c,gGo). The initiation of hatching was very variable at 120 h post-HCG, ranging from 9 to 38%. The responses observed in the external control media (Table IIIGo) fell into this range. The analysis of deviance showed a significant difference between replicates (P = 0.0004), and also a significant glucosexphosphate interaction (P = 0.029). There was no significant effect of glucose and KH2PO4 concentrations on the percentages of blastocysts that began to hatch at 120 h post-HCG, although the slight inhibitory effect of KH2PO4 was close to significance at the P = 0.05 level (P = 0.09). All the other interactions were not significant. By 144 h post-HCG, the initiation of hatching had increased, ranging from 21 to 53% (Table VIGo, Figure 2d,hGo). The responses observed in the external control media (Table IIIGo) fell into this range. The analysis of deviance (Table VGo) showed a significant difference between replicates (P = 0.0013), and a significant inhibitory effect of increasing the concentration of KH2PO4 (P = 0.012). The effects of glucose and all the interactions were not significant. A linear regression of the logit of the blastocysts that began to hatch on the logarithm of the concentration of KH2PO4 144 h post-HCG administration did not fit the data. The significance of the effect of KH2PO4 was due to a slight fall in the initiation of hatching (~10%) when the concentration of KH2PO4 was increased from 0.15 to 0.45 mmol/l.

The total cell counts in blastocysts that developed from zygotes in 144 h in two of the replicates were counted. The mean total cell counts observed in each medium are shown in Table VIIaGo. The analysis of deviance (Table VIIbGo) showed that the only significant effect was between replicates. The effects of glucose, KH2PO4 and their interaction were not statistically significant. The mean number of cells over all groups was 66. The median cell count in blastocysts produced in standard KSOM was similar (Figure 1Go). In contrast, the median cell counts in blastocysts produced in glucose/KH2PO4-free KSOM were lower (Figure 1Go).


View this table:
[in this window]
[in a new window]
 
Table VIIa. Mean cell counts in blastocysts that developed from zygotes in KSOM containing different concentrations of glucose and KH2PO4
 

View this table:
[in this window]
[in a new window]
 
Table VIIb. Analysis of deviance of log10(cell counts)
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Effects of simultaneously varying the concentrations of glucose and KH2PO4 in KSOM
A factorial experimental design was adopted to investigate the joint effects of glucose and KH2PO4 on the development of outbred CF1 mouse zygotes into blastocysts in KSOM. Figure 3Go shows a parsimonious two-dimensional concentration–response surface calculated from the incidence of blastocyst formation 144 h post-HCG administration. Its shape is convex, curving slightly downwards only along the KH2PO4 axis as the concentration of this compound is increased. The number of cells that developed in those embryos that formed blastocysts 144 h post-HCG was unaffected by the concentrations of glucose and KH2PO4 in the media, and thus the data were fitted by a horizontal plane through a mean cell number of 66. To attain the mean cell count of 66 cells the zygotes would have undergone about six cell divisions.



View larger version (42K):
[in this window]
[in a new window]
 
Figure 3. Estimated concentration–response surface showing the regression of the percentage of zygotes that developed into at least zona-enclosed blastocysts on the concentrations of glucose and KH2PO4 after culture for 144 h post-HCG administration in KSOM. The data are adequately fitted by the equation y = logit(p) = 1.409–0.157x1–1.011x2, where x1, x2 are the log(glucose) and log(KH2PO4) concentrations, respectively (residual deviance 88.32, df 88, P = 0.47). The graph has been produced by transforming the logits back to proportions.

 
The results confirm previous reports that glucose, in a concentration as high as that found normally in blood, does not inhibit the development of outbred CF1 mouse zygotes at least to blastocysts and of blastocysts that begin to hatch when cultured in KSOM (Lawitts and Biggers, 1991bGo, 1992Go, 1993Go; Summers et al., 1995Go, 2000Go; Biggers et al., 1997Go, 2000Go). Thus, the acceptance of the view that glucose in concentrations within normal physiological ranges universally inhibit mammalian preimplantation development in vitro is not valid. Neither is it justified to imply, as previously (Barnett and Bavister, 1996Go), that all culture media for the culture of preimplantation embryos should contain low concentrations of glucose.

No evidence was obtained to show that the slight inhibitory effect of KH2PO4 on the percentages of zygotes that at least develop into blastocysts and that begin to hatch is dependent on the concentration of glucose in the medium. The effect of KH2PO4 is the same over all concentrations of glucose. A useful model of the distribution of the effects of an inhibitor on a population is a tolerance distribution (Finney, 1952Go). Specifically, let there be a threshold concentration of KH2PO4 for each zygote that uniquely inhibits blastocyst formation, and that these threshold concentrations are distributed say as a lognormal distribution. Then the slight inhibitory effect of KH2PO4 on blastocyst formation is due to the arrest of a small group of zygotes that are particularly sensitive to the compound. Presumably, concentrations of KH2PO4 >1.35 mmol/l would have blocked the development of more zygotes. Results from other studies (Haraguchi et al., 1996Go, 1999Go) have raised the possibility that there may be considerable differences between the sensitivities of different mouse strains to KH2PO4. The development of zygotes of the AKR/N strain were completely blocked at the morula stage by 1 mmol/l KH2PO4. The difference in sensitivity to KH2PO4 needs to be confirmed, however, by comparing the responses of the two strains using the same medium [Haraguchi et al. (1996) used a modified Whitten's medium, while KSOM was used in the current study]. Differences in sensitivity to KH2PO4 may be even greater between species. For example, the rat and hamster preimplantation embryo during the initial stages of cleavage seems to be particularly sensitive to phosphate (rat: Miyoshi et al., 1994; Miyoshi and Niwa, 1997; Yamada and Nishikimi, 1999; hamster: Lane et al., 1999).

Media dependency of the effects of glucose
It has been suggested (Biggers, 1993Go) that the effects of a compound in a medium may be differentially affected by the concentrations of other components. The combined results of several investigators on the effects of glucose and KH2PO4 in media for the culture of a single strain of mouse preimplantation embryos provide a good example of this possibility. Table VIIIGo shows the background composition of four media in which the effects of glucose and KH2PO4 on the development of outbred mouse CF1 zygotes have been studied. It has been shown in the present paper that the addition of glucose to one of these media (KSOM) has no inhibitory effect on development, and the addition of KH2PO4 has only a slight inhibitory effect. In contrast, the addition of both compounds in similar concentrations are inhibitory when added to the other three media [M16 (Lawitts and Biggers, 1991a; EMGP (Scott et al., 1993Go); modified human tubal fluid (HTF; Quinn and Horstman, 1998)]. These observations on a single strain of mouse provide strong evidence that the effects of glucose and KH2PO4 may depend on the background composition of the media in which they are studied. The variable effects of glucose on preimplantation development of the sheep may well be due to the differences in the media used: Betterbed and Wright (1985) used Brinster's medium, Thompson et al. (1989) used medium SOF, McGinnis and Youngs (1992) used medium CZB, and Thompson et al. (1992) used medium SOF supplemented with amino acids.


View this table:
[in this window]
[in a new window]
 
Table VIII. Background compositions of four media (KSOM, M16, EMGP and HTF) in which the effects of glucose and phosphate on early outbred mouse CF1 development have been studied
 
At present, it is impossible to identify the cause(s) of these differences between the effects of glucose and phosphate that seem to occur when outbred CF1 zygotes are cultured in different media. Whether the inhibition is associated with higher concentrations of KCl, pyruvate and lactate is unknown. In studies of the behaviour of the 8-cell hamster embryo in vitro, evidence was produced for the Crabtree effect (Seshagiri and Bavister, 1991Go), a phenomenon found in metabolically active cells (e.g. cancer and ascites cells) in which glucose stimulates glycolysis at the expense of mitochondrial respiration (for a review, see Wojtczak, 1996). Others (Thompson et al., 1992Go) have also supported the view that the Crabtree effect explains the inhibitory effect of glucose they observed using sheep embryos in an amino acid-supplemented medium SOF. It would be of interest to know whether the Crabtree effect occurs in early mouse embryos cultured in M16, EMGP and modified HTF, and not in KSOM. Presently, the mechanisms involved in the Crabtree effect are controversial (for a review, see Wojtczak, 1996). Explanations based on the competition between respiration and glycolysis for precursors of ATP (ADP and inorganic phosphate), shifts in intracellular pH, changes in the permeability of the inner mitochondrial membrane, the specific pattern of isoenzymes and their regulation in the glycolytic pathway and the topography of enzymes in rapidly growing cells (hexokinase) are far from satisfactory. Very recently, evidence has been published (Wojtczak et al., 1999Go) that the Crabtree effect is produced by the release of bound intracellular Ca2+ which is subsequently taken up by mitochondria where it inhibits F1F0-ATP synthase. The possibility that the Crabtree effect is mediated in preimplantation embryos by the perturbation of intracellular free Ca2+ has become a reality due to recent studies (Lane et al., 1999Go) which showed that culturing hamster zygotes increases the concentration of intracellular free Ca2+ by the 2-cell stage. It is well known that the release of bound intracellular Ca2+ plays a very important role in functions necessary for development following sperm binding (for reviews, see Stachecki and Armant, 1996; Lane et al., 1998). The functions of Ca2+ seem to depend on reaching a critical threshold concentration at which oscillatory fluctuations in concentration are induced. It is possible that some media perturb the release of bound intracellular Ca2+ in physiological concentrations sufficiently to block further development.

Media formulation
KSOM was formulated using the sequential simplex optimization technique to find a medium in which the 2-cell block did not occur (Lawitts and Biggers, 1991bGo, 1992Go, 1993Go). Optimun concentrations of glucose and KH2PO4 in which this block was minimal were found to be 0.20 mmol/l glucose and 0.35 mmol/l KH2PO4. These values, however, are likely to be ill-determined optima because the results described in this paper show that the concentration–response surfaces for glucose and KH2PO4 in the neighbourhood of these concentrations are, for all practical purposes, horizontal planes. When these conditions exist, the resolution of optimization strategies falls, since the sequential estimates of an optimum do not converge but tend to wander randomly about the surface (Walters et al., 1991Go). What, in the light of the results reported in the present paper, should be the concentrations of glucose and KH2PO4 in a recommended version of KSOM?

It is suggested that in the case of glucose the `back to nature' principle enunciated earlier (Leese, 1998Go) be used, and a glucose concentration of 3.40 mmol/l be selected, which has been found in the oviduct of the mouse (Gardner and Leese, 1990Go).

Lawitts and Biggers (1991b) described a medium denoted aKH2PO4 that contained no KH2PO4, and which supported a high rate of passage through the 2-cell block. This result was discussed by others (Bavister, 1995Go), and has been used to question the inclusion of 0.35 mmol/l KH2PO4 in SOM and its subsequent modification KSOM (Quinn, 1997Go; Quinn and Horstman, 1998Go). It has also been pointed out (Lawitts and Biggers, 1991bGo) that this medium was obtained after only four cycles of the ongoing simplex optimization procedure. After a further 16 cycles, however, the medium converged to one referred to as SOM that contained 0.35 mmol/l KH2PO4 (Lawitts and Biggers, 1992Go). The results reported in the present paper have demonstrated a statistically significant inhibitory effect of KH2PO4 in KSOM on development at least to the blastocyst stage and the initiation of hatching at 120 and 144 h post-HCG respectively (Tables IV and VIGoGo; Figure 3Go). Nevertheless, the effect of including KH2PO4 in concentrations ranging from 0.05 to 1.35 mmol/l still allows the development of >=76% blastocysts by 144 h post-HCG administration (Table IVGo, Figure 3Go). The fitted concentration–response surface shows that raising the concentration of KH2PO4 from 0.05 to 0.35 mmol/l reduces the percentage of blastocysts formed from 87 to 83%. Similar findings were obtained when the observed response was the initiation of hatching. Thus, the effect of including 0.35 mmol/l KH2PO4 in KSOM is very small and may be of no biological significance. In many cells a high intracellular phosphate concentration is maintained by a balance between phosphate transporters in the cell membrane that import and export phosphate across the cell membrane. The nature of these transporters is unknown in the preimplantation embryo. At present, the concentration of phosphate in the mouse oviduct is not known, so the `back to nature' principle cannot be used as in the case of glucose. Nevertheless, it is reasonable to assume that phosphate is present in the mouse oviduct since its presence has been measured in the sheep (Restall and Wales, 1966Go) and human (Lippes et al., 1972Go). We believe it is important not to exacerbate the natural phosphate gradient across the cell membranes by eliminating KH2PO4 completely from the medium. It is recommended that the empirically determined concentration of 0.35 mmol/l KH2PO4 in KSOM, found by sequential simplex optimization, be retained until further information on the regulation of phosphate transport in preimplantation embryos becomes available.

The belief that glucose is physiologically inhibitory to early preimplantation embryos of all species, including man, has now attained dogmatic status. Recent publications recommend that low concentrations of glucose be used in media for the culture of human preimplantation embryos (Bavister, 1999Go; Coates et al., 1999Go), and the putative beneficial effects of media containing low concentrations of these compounds are touted in advertisements of commercially available media for human IVF and embryo transfer (e.g. Scandinavian IVF Science AB, medium IVF-50TM; Irvine Scientific, Inc.). There is no doubt that glucose and KH2PO4 inhibit or adversely affect the early development of the human embryo in most of the media currently being used. The results of the present study suggest, however, that media could be found in which these compounds are not inhibitory in concentrations similar to those present in the natural in-vivo environment. In fact, an indication has been provided (Coates et al., 1999Go) that such conditions could be found. A clinical study was carried out in which human preimplantation embryos were cultured in Earle's balanced salt solution with and without 5.5 mmol/l glucose. No significant effects of the glucose on pregnancy rates were observed. Adverse effects of the glucose on both the cleavage rate and grade of embryo were reported, which led the authors to suggest that `... a reduction of the glucose concentration of the medium used for embryo culture from the pronucleate stage to embryo transfer on day 2 or 3 is prudent'. Although the tests of significance show that the effects of glucose on cell number and embryo grade are both statistically significant, some caution should be exercised in suggesting that the effects are biologically significant, since the differences were only 5% based on the observation of two large group sizes (1283 exposed to glucose versus 1471 not exposed to glucose). A minimal difference of little physiological interest can be statistically significant if the sample size is sufficiently large (Lindsey, 1999Go).


    Acknowledgments
 
This work has been supported as part of the National Cooperative Program on Non-Human In Vitro Fertilization and Preimplantation Development and was funded by the National Institute of Child Health and Human Development, NIH, through cooperative agreement HD21988. Animals used in this study were maintained in accordance with the guidelines of the Committee on Care and Use of Laboratory Animal Resources, National Research Council. We thank Dr Betsey S.Williams for helpful criticism of the manuscript.


    Notes
 
1 To whom correspondence should be addressed. E-mail: john_biggers{at}hms.harvard.edu Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Aoki, K., Nakamura, M., Namiki, H. et al. (1990) The effect of glucose and phosphate on mouse 2-cell embryos to develop in vitro. Zool. Sci., 7, 973–977.[ISI]

Barnett, D.K. and Bavister, B.D. (1996) What is the relationship between the metabolism of preimplantation embryos and their developmental competence? Mol. Reprod. Dev., 43, 105–133.[ISI][Medline]

Bavister, B.D. (1995) Culture of preimplantation embryos: facts and artefacts. Hum. Reprod. Update, 1, 91–148.[ISI]

Bavister, B.D. (1999) Glucose and the culture of human embryos. Fertil. Steril., 72, 233–234.[ISI][Medline]

Betterbed, B. and Wright, R.W. (1985) Development of one-cell ovine embryos in two culture media under two gas atmospheres. Theriogenology, 23, 547–553.[ISI]

Biggers, J.D. (1993) The culture of the mammalian preimplantation embryo. In Gianaroli, L., Campana, A. and Trounson, A. (eds), Implantation in Mammals. Raven Press, New York, pp. 123–136.

Biggers, J.D., Rinaldini, R. and Webb, M. (1957) The study of growth factors in tissue culture. Symp. Soc. Exp. Biol., 11, 264–297.

Biggers, J.D., Whitten, W.K. and Whittingham, D.G. (1972) The culture of mouse embryos in vitro. In Daniel, J.C. (ed.), Methods in Mammalian Embryology. Freeman, San Francisco, pp. 86–116.

Biggers, J.D., Summers, M.C. and McGinnis, L.K. (1997) Polyvinyl alcohol and amino acids as substitutes for bovine serum albumin in mouse preimplantation embryo culture media. Hum. Reprod. Update, 3, 125–135.[Abstract/Free Full Text]

Biggers, J.D., McGinnis, L.K. and Raffin, M. (2000) Amino acids and preimplantation development of the mouse in protein-free KSOM. Biol. Reprod., 63, 281–293.[Abstract/Free Full Text]

Brown, J.J.G. and Whittingham, D.G. (1992) The dynamic provision of different energy substrates improves development of one-cell random-bred mouse embryos in vitro. J. Reprod. Fertil., 95, 503–511.[Abstract]

Chatot, C.I., Ziomek, C.A., Bavister, B.D. et al. (1989) An improved culture medium supports development of random-bred 1-cell mouse embryos in vitro. J. Reprod. Fertil., 86, 679–688.[Abstract]

Chatot, C.L., Lewis, J.L., Torres, I. et al. (1990) Development of 1-cell embryos from different strains of mice in CZB medium. Biol. Reprod., 42, 432–440.[Abstract]

Clogg, C.C. and Shihadeh, E.S. (1994) Statistical Methods for Ordinal Variables. Sage Publications, Thousand Oaks, CA, pp. 1–3.

Coates, A., Rutherford, A.J., Hunter, H. et al. (1999) Glucose-free medium in human in vitro fertilization and embryo transfer: a large-scale, prospective, randomized clinical trial. Fertil. Steril., 72, 229–232.[ISI][Medline]

Conaghan, J., Handyside, A.H., Winston, R.M.L. et al. (1993) Effects of pyruvate and glucose on the development of human preimplantation embryo in vitro. J. Reprod. Fertil., 99, 87–95.[Abstract]

De Hertogh, R., Vanderheyden, I., Pampfer, S. et al. (1991) Stimulatory and inhibitory effects of glucose and insulin on rat blastocyst development in vitro. Diabetes, 40, 641–647.[Abstract]

Diamond, M.P., Pettway, Z.Y., Logan, J. et al. (1991) Dose-response effects of glucose, insulin, and glucagon on mouse pre-embryo development. Metabolism, 40, 566–570.[ISI][Medline]

Diggle, P.J., Liang, K. and Zeger, S.L. (1994) Analysis of Longitudinal Data. Clarendon Press, Oxford, pp. 117–130.

Ebert, K., Hammer, R.E. and Papaioannou, V.E. (1985) A simple method of counting nuclei in the preimplantation mouse embryo. Experientia, 41, 1207–1209.[ISI][Medline]

Erbach, G.T., Lawitts, J.A., Papaioannou, V.E. et al. (1995) Differential growth of the mouse preimplantation embryo in chemically defined media. Biol. Reprod., 50, 1027–1033.[Abstract]

Finney, D.J. (1952) Probit Analysis. 2nd edn. Cambridge University Press, pp. 8–12.

Gardner, D.K. and Leese, H.J. (1990) Concentrations of nutrients in mouse oviduct fluid and their effects on embryo development and metabolism in vitro. J. Reprod. Fertil., 88, 361–368.[Abstract]

Hagen, D.R., Prather, R.S., Sims, M.M. et al. (1991) Development of one-cell porcine embryos to the blastocyst stage in simple media. J. Anim. Sci., 69, 1147–1150.[Abstract/Free Full Text]

Haraguchi, S., Naito K., Azuma, S. et al. (1996) Effects of phosphate on in vitro 2-cell block of AKR/N mouse embryos based on changes in cdc2 kinase activity and phosphorylation states. Biol. Reprod., 55, 598–603.[Abstract]

Haraguchi, S., Naito, K. and Sato, E. (1999) Phosphate exposure during the late 1-cell and early 2-cell stages induces a time-specific decrease in cyclin B and cdc25B mRNAs in AKR/N mouse embryos in vitro. Zygote, 7, 87–93.[ISI][Medline]

Hardy, K. (1994) Effects of culture conditions on early embryonic development (Abstract). Hum. Reprod., 9 (Suppl. 4), 95.

Lane, M., Boatman, D.E., Albrecht, R.M. et al. (1998) Intracellular divalent cation homeostasis and developmental competence in the hamster preimplantation embryo. Mol. Reprod. Dev., 50, 443–450.[ISI][Medline]

Lane, M., Ludwig, T.E and Bavister, B.D. (1999) Phosphate induced developmental arrest of hamster two-cell embryos is associated with disrupted ionic homeostasis. Mol. Reprod. Dev., 54, 410–417.[ISI][Medline]

Lawitts, J.A. and Biggers, J.D. (1991a) Overcoming the 2-cell block by modifying standard components in a mouse embryo culture medium. Biol. Reprod., 45, 245–251.[Abstract]

Lawitts, J.A. and Biggers, J.D. (1991b) Optimization of mouse embryo culture media using simplex methods. J. Reprod. Fertil., 91, 543–556.[Abstract]

Lawitts, J.A. and Biggers, J.D. (1992) Joint effects of sodium chloride, glutamine, and glucose in preimplantation embryo culture media. Mol. Reprod. Dev., 31, 189–194.[ISI][Medline]

Lawitts, J.A. and Biggers, J.D. (1993) Culture of preimplantation embryos. Methods Enzymol., 225, 153–164.[ISI][Medline]

Leese, H.J. (1998) Human embryo culture: back to nature. J. Assist. Reprod. Genet., 15, 466–468.[ISI][Medline]

Lindsey, J.K. (1999) Revealing Statistical Principles. Hodder Headline Group, London, p. 170.

Lippes, J., Enders, R.G., Pragay, D.A. et al. (1972) The collection and analysis of human fallopian tubal fluid. Contraception, 5, 85–103.[ISI][Medline]

McCullagh, P. and Nelder, J.A. (1989) Generalized Linear Models. 2nd edn. Chapman & Hall, London.

McGill, R., Tukey, J.W. and Larson, W.A. (1978) Variations of box plots. Am. Statist., 32, 12–16.[ISI]

McGinnis, L.K. and Youngs, C.R. (1992) In vitro development of ovine embryos in CZB medium. Theriogenology, 37, 559–569.[ISI]

Miyoshi, K. and Niwa, K. (1997) Stage-specific requirement of phosphate for development of rat 1-cell embryos in a chemically defined medium. Zygote, 5, 67–73.[ISI][Medline]

Miyoshi, K., Funahashi, H., Okuda, K. et al. (1994) Development of rat one-cell embryos in a chemically defined medium: effects of glucose, phosphate and osmolarity. J. Reprod. Fertil., 100, 21–26.[Abstract]

Petters, R.M., Johnson, B.H., Reed, M.L. et al. (1990) Glucose, glutamine and inorganic phosphate in early development of the pig embryo in vitro. J. Reprod. Fertil., 89, 269–275.[Abstract]

Quinn, P. (1995) Enhanced results in mouse and human embryo culture using a modified human tubal fluid medium lacking glucose and phosphate. J. Assist. Reprod. Genet., 12, 97–105.[ISI][Medline]

Quinn, P. (1997) Comments. In Elliott, T. and Elder, K. (eds), Blastocyst Culture, Transfer and Freezing. Ladybrook Publishing, West Leederville, W. Australia. p. 43.

Quinn, P. and Horstman, F.C. (1998) Is the mouse a good model for the human with respect to the development of the preimplantation embryo in vitro. Hum. Reprod., 13 (Suppl. 4), 173–183.[Medline]

Quinn, P., Monipanah, R., Steinberg, J.M. et al. (1995) Successful human in vitro fertilization using a modified human tubal fluid medium lacking glucose and phosphate ions. Fertil. Steril., 63, 922–924.[ISI][Medline]

Reed, M.L., Jin, D.I. and Petters, R.M. (1992a) Glucose and inorganic phosphate inhibits rat 8-cell embryo development in vitro. Theriogenology, 37, 282.

Reed, M.L., Illera, M.J. and Petters, R.M. (1992b) In vitro culture of pig embryos. Theriogenology, 37, 95–109.[ISI]

Restall, B.J. and Wales, R.G. (1966) The fallopian tube of the sheep. III. The chemical composition of the fluid from the fallopian tube. Aust. J. Biol. Sci., 19, 687–698.[ISI][Medline]

Schini, S.A. and Bavister, B.D. (1988) Two-cell block to development of cultured hamster embryos is caused by phosphate and glucose. Biol. Reprod., 39, 1183–1192.[Abstract]

Scott, L. and Whittingham, D.G. (1996) Influence of genetic background and media components on the development of mouse embryos in vitro. Mol. Reprod. Dev., 43, 336–346.[ISI][Medline]

Scott, L.F., Centaurium, S.G. and Smith, S. (1993) The relevance and use of mouse embryo bioassays for quality control in an assisted reproductive technology program. Fertil. Steril., 60, 559–568.[ISI][Medline]

Seshagiri, P.B. and Bavister, B.D. (1991) Glucose and phosphate inhibit respiration and oxidative metabolism in cultured hamster eight-cell embryos: evidence for the `Crabtree effect'. Mol. Reprod. Dev., 30, 105–111.[ISI][Medline]

Stachecki, J.J. and Armant, D.R. (1996) Transient release of calcium from inositol 1,4,5-trisphospate-specific stores regulates mouse preimplantation development. Development, 122, 2485–2496.[Abstract/Free Full Text]

Summers, M.C., Bhatnagar, P.R., Lawitts, J.A. et al. (1995) Fertilization in vitro of mouse ova from inbred and outbred strains: complete preimplantation embryo development in glucose-supplemented KSOM. Biol. Reprod., 53, 431–437.[Abstract]

Summers, M.C., McGinnis, L.K., Lawitts, J.A. et al. (2000) IVF of mouse ova in a simplex optimized medium supplemented with amino acids. Hum. Reprod., 15, 1791–1801.[Abstract/Free Full Text]

Takahashi, Y. and First, N.L. (1992) In vitro development of bovine one-cell embryos: influence of glucose, lactate, pyruvate, amino acids and vitamins. Theriogenology, 37, 963–978.[ISI]

Thompson, J.G.E., Parton, G.A.J., Cruikshank, G.W. et al. (1989) Development of sheep preimplantation embryos in media supplemented with glucose and acetate. Theriogenology, 32, 323–330.[ISI]

Thompson, J.G.E., Simpson, A.C., Pugh, P.A. et al. (1992) Requirements for glucose during in vitro culture of sheep preimplantation embryos. Mol. Reprod. Dev., 31, 253–257.[ISI][Medline]

Walters, F.H., Parker, L.R., Jr, Morgan, S.L. et al. (1991) Sequential Simplex Optimization. CRC Press, Florida.

Waymouth, C. (1954) Some effects of inorganic phosphate and bicarbonate on cell survival and proliferation in chemically defined nutrient media. Biochem. J., iv–v.

Whittingham, D.G. (1971) Culture of mouse ova. J. Reprod. Fertil. (Suppl.), 14, 7–21.[Medline]

Wojtczak, L. (1996) The Crabtree effect: a new look at the old problem. Acta Biochim. Pol., 43, 361–368.[ISI][Medline]

Wojtczak, L., Teplova, V.V., Bogucka, K. et al. (1999) Effect of glucose and deoxyglucose on the redistribution of calcium in Ehrlich ascites tumour and Zajdela hepatoma cells and its consequences for mitochondrial energetics. Eur. J. Biochem., 263, 495–501.[Abstract/Free Full Text]

Yamada, M. and Nishikimi, A. (1999) Inhibitory effect of phosphate on in vitro development of 2-cell rat embryos is overcome by a factor(s) in oviductal extracts. FEBS Lett., 462, 71–74.[ISI][Medline]

Submitted on July 7, 2000; accepted on September 21, 2000.