Abstract

Mitochondria are cellular organelles regulating metabolism and cell death pathways. This study examined changes in mitochondrial membrane potential (ΔΨm) throughout the stages of preimplantation development in mouse embryos conceived either in vivo or in vitro and human embryos donated to research from IVF. Embryos stained with the ΔΨm‐sensitive dye (JC‐1) were quantified for the ratio of high‐ to low‐polarized mitochondria using a deconvolution microscope. Overall, mouse zygotes and early embryos contain a subset of high‐polarized mitochondria with a progressive increase in the ratio of ΔΨm observed with increasing cleavage. A transient increase in the ratio of high to low ΔΨm was observed in in vivo fertilized 2‐cell stage embryos, coincident with embryonic genome activation in the mouse, but not in 2‐cell embryos obtained through IVF. We further observed that arrested mouse 2‐cell embryos possessed an increased ratio of ΔΨm compared with non‐arrested embryos. In human 8‐cell embryos we observed an increased ratio of high‐ to low‐polarized mitochondria with increasing degrees of embryo fragmentation. We concluded that the pattern of mitochondrial membrane potential progressively changes throughout preimplantation development, and that an aberrant shift in ΔΨm could contribute to, or is associated with, decreased developmental potential.

Introduction

Progression of fertilized mammalian oocytes through cleavage, blastocyst formation and implantation is dependent on the successful implementation of specific genetic and developmental programmes. Successful interaction of paternal and maternal gametes is required for normal embryonic development. The oocyte controls several important aspects of meiosis (Polanski, 1997), fertilization and regulation of early cleavage, and modulates epigenetic development of the embryonic genome manifested later in embryogenesis (Latham and Sapienza, 1998). Thus, variability in the quality of oocytes is reflected in embryonic developmental competence and in the high incidence of embryonic wastage observed in human IVF programmes (Bolton et al., 1989). Maternal factors that could contribute to abnormal preimplantation development include cytoplasmic maternal mRNA, proteins, antioxidants and organelles, including mitochondria.

Mitochondria are the major source of energy in all eukaryotic cells, producing ATP through oxidative phosphorylation and the citric acid cycle. They regulate calcium homeostasis and modulate apoptosis through release of several cell death‐inducing molecules (for general reviews see Lenaz et al., 2002; Ravagnan et al., 2002). Almost all knowledge of the behaviour and function of mitochondria have been elucidated through the study of somatic cells.

The pattern of mitochondrial inheritance has a profound impact on development and reproductive performance (Smith and Alcivar, 1993). In most mammalian species, mitochondria are exclusively maternally derived since the few sperm mitochondria that enter the oocyte at fertilization appear to be degraded (Sutovsky et al., 2000) or diluted out during sequential cleavage and cannot be detected at the blastocyst stage (Cummins et al., 1998). Oocyte mitochondria display several unique features suggesting an undifferentiated phenotype. In cross‐section, they appear spherical, contain a very dense matrix with a low number of cristae, have one haploid DNA molecule per organelle and possess low metabolic activity (Jansen and de Boer, 1998). Upon fertilization and embryonic development, mitochondria undergo a maturation process characterized by a less dense matrix, and a slow change towards an elongated shape (Motta et al., 1988). As this maturation process occurs in each individual mitochondrion, a mixed population of differentiated and undifferentiated mitochondria is present within the cells of the blastocyst (Plante and King, 1994).

Each blastomere is entirely dependent on the energy produced from oocyte‐inherited mitochondria, as embryonic mitochondria only resume transcription and gain replicative ability following the blastocyst stage (Smith and Alcivar, 1993; Cummins, 1998). This dependence on maternal mitochondria therefore suggests that the number of mitochondria found in the oocyte at the time of ovulation could be critical to ensuing embryo development. Several studies on mtDNA copy number in both mouse (Piko and Taylor, 1987) and human (Steuerwald et al., 2000; Reynier et al., 2001) oocytes have indicated that a highly variable number of mtDNA molecules can be found in individual oocytes, with the average copy number being ∼100 000 for mouse and ∼220 000 for human, with the range in human oocytes being more highly variable. One conclusion that can be drawn is that not all oocytes, be they of mouse or human origin, will contain the same number of mtDNA molecules, and that this variable could affect reproductive performance and embryo quality.

As women age, the success rate of IVF falls, and recent evidence suggests that mitochondrial defects may in part be responsible for this decline in fertility. Increased maternal age is accompanied not only by mitochondrial mutations (Keefe et al., 1995) but it also affects mitochondrial function (Van Blerkom et al., 1995), numerical density (Muller‐Hocker et al., 1996; Steuerwald et al., 2000) and metabolic activity (Wilding et al., 2001). These observations are further supported by data showing that infusion of mitochondria‐enriched cytoplasts to mouse oocytes resulted in apparent increases in ATP production (Van Blerkom et al., 1998). In addition, Cohen et al. (1998) demonstrated that transfer of donor ooplasm to recipient human oocytes improves in vitro development of embryos for patients whose previous attempts at IVF had failed and results in mitochondrial heteroplasmy in the offspring (Brenner et al., 2000). The results of these studies therefore suggest that the microinjected mitochondria remain viable and could in fact be contributing to ATP production of the embryo (Steuerwald et al., 2000). There has been an increasing interest towards improving the understanding of the role of mitochondria in successful preimplantation development with a particular emphasis on clinical applications, and the ethical issues surrounding this form of treatment have been extensively debated (for review see Thorburn and Dahl, 2001; Hawes et al., 2002; St John, 2002).

Mitochondrial membrane potential is a key indicator of cellular viability, as it reflects the pumping of hydrogen ions across the inner membrane during the process of electron transport and oxidative phosphorylation, the driving force behind ATP production. Through use of the lipophilic mitochondrial probe JC‐1, which can be used to estimate changes in membrane potential (ΔΨm), heterogeneity between ΔΨm of individual mitochondria within a cell have been found in a wide range of somatic cell types (Diaz et al., 1999) and more recently in mouse and human female germ cells (Wilding et al., 2001; Van Blerkom et al., 2002). The goal of the current study was to quantitatively evaluate ΔΨm heterogenity and to determine changes in ΔΨm throughout preimplantation development in both mouse and human embryos. Furthermore, we studied whether the mode of fertilization and the growth environment alters ΔΨm in mouse embryos and assessed embryos with aberrant developmental potential to determine whether ΔΨm and mitochondrial distribution could be implicated in the regulation of blastomere arrest or fragmentation.

Materials and methods

Embryo collection

In vivo fertilized embryos

The Mount Sinai Hospital Animal Care Committee approved all animal protocols and protocols were in compliance with standards for the ethical treatment of animals. ICR female mice 6–8 weeks old (n = 70; Harlan, USA) were superovulated using 5 IU of equine pregnant mare’s serum gonadotrophin (obtained from NHPP, NIDDK and Dr A.F.Parlow) followed 48 h later by 5 IU of hCG (Wyeth, Canada). The mice were subsequently mated with ICR males of proven fertility and plugs were verified the next morning.

In vivo fertilized embryos at the various preimplantation stages were retrieved from mated females immediately prior to analysis [at day 0.5, 1.5, 2, 2.5, 3 and 3.5 post‐coitum (p.c.)]. For the 2‐cell arrest experiments, zygotes were flushed at day 0.5 p.c. in modified human tubal fluid medium (mHTF; Irvine Scientific, USA) supplemented with 0.5% bovine serum albumin (BSA; Sigma, Canada). A subset of these zygotes was placed into HTF supplemented with 0.5% BSA while the remaining zygotes were washed and transferred to synthetic oviductal medium enriched with potassium (KSOM) with amino acids (Specialty Media, USA). Both groups of zygotes were cultured in a humidified incubator at 37°C, 5% CO2 for up to 48 h to assess the incidence of 2‐cell arrest.

In vitro fertilized embryos

An ICR male of proven fertility was killed by cervical dislocation and the epididymides and vasa deferentia were removed and placed in 1 ml of HTF supplemented with 0.5% BSA. The epididymides were punctured with a needle and the sperm removed from the vas deferens. The suspension of sperm was placed in a humidified incubator at 37°C, 5% CO2 for 1 h in order to capacitate. The oocyte–cumulus cell complexes were retrieved from 6–8 week old ICR females (superovulated as above) 14 h post‐hCG. The oocyte–cumulus cell complexes were placed in 500 µl HTF supplemented with 0.5% BSA to which 40 µl of capacitated sperm suspension was added. The fertilization dishes were incubated in a humidified environment at 37°C, 5% CO2 for 5 h, at which point the oocytes were washed free of sperm through several HTF and KSOM microdrops and were transferred to KSOM. The oocytes were cultured overnight and assessed for 2‐cell cleavage the following morning. A subset of these embryos was analysed at the 2‐cell stage and the remaining embryos were cultured in KSOM medium and analysed at subsequent preimplantation stages.

In order to control for the rate of parthenogenetic activation, the sperm suspension was heated to 56°C in order to disrupt the acrosome (Cozzi et al., 2001). Treated sperm were found to have no motility and disrupted plasma membrane upon dual staining with 4,6‐diamidino‐2‐phenylindole (DAPI; Sigma, Canada) and propidium iodide (Sigma) (data not shown). The heated sperm suspension was then used in the same manner as untreated sperm described above.

Human embryos

Spare human embryos donated to research were obtained from the Mount Sinai Hospital IVF Program with approval from the Mount Sinai Hospital Research Ethics Board. A total of 140 embryos was received over an 8 month period, and these embryos were successfully imaged and analysed using the ΔΨm staining protocol described below. Samples were received between days 1 and 7 post‐retrieval, where day 0 was the day of retrieval. Samples were cultured in HTF for 72 h until the 6–8‐cell stage, at which point they were switched to G2.2 (Vitrolife; Meditech, Canada). Culture conditions were 37°C, 5% CO2, 90% N2 and 5% O2 until the time of analysis.

Mitochondrial membrane potential analysis

JC‐1 (5,5′,6,6′‐tetrachloro‐1,1′,3,3′‐tetraethylbenzimidazoyl carbocyanine iodide, DePsipher Kit TA700; R&D Systems Inc., USA) is a lipophilic cationic dye that enters the inner mitochondrial matrix in its monomeric form when the mitochondrial membrane is polarized (Reers et al., 1991, 1995). When the mitochondrion has a high ΔΨm, the dye crosses the membrane and forms J‐aggregates, which appear red under UV light. If ΔΨm is low, the dye remains in its monomeric form and fluoresces green. There are two different forms of monomeric green staining: one is where the dye has crossed the mitochondrial membrane, which has a low potential, and the staining is punctate; when the mitochondria are dead, and hence have no ΔΨm, the dye is excluded entirely from the mitochondria and results in diffuse green cytoplasmic staining.

The membrane‐sensitive dye JC‐1 was prepared at half the manufacturer’s recommended concentration (0.5 µl DePsipher, 900 µl ultrapure water, 100 µl reaction buffer and 10 µl stabilizer) from stock solutions directly prior to use. The dye was prewarmed in a 37°C water bath prior to addition to embryos, which were stained for 25 min in a humidified incubator at 37°C, 5% CO2. Following staining, the embryos were individually washed and placed in 3 µl drops of mHTF supplemented with 0.5% BSA under light mineral oil (Specialty Media) in a Delta TPG Culture Dish (Bioptechs, USA). To verify the specificity of JC‐1 staining, mouse oocytes were imaged once, treated with the mitochondrial uncoupler carbonyl cyanide p‐trifluoromethoxyphenylhydrazone (40 µmol/l final concentration, FCCP; Sigma) and subsequently imaged at 2 min intervals for a total of 14 min elapsed time. Samples were imaged on a deconvolution microscope (Olympus IX70; Applied Precision Inc., USA), using the ×20 objective under the fluorescein isothiocyanate (FITC) and rhodamine isothiocyanate (RITC) filters. Optically sectioned images of the samples, where each section had a thickness of 2 µm, were obtained with 10 total sections visualized for each mouse embryo and a larger number of sections (up to 15) analysed for human embryos. Following imaging, mouse embryos were individually transferred to 10 µl drops of KSOM under light mineral oil and cultured in a humidified incubator at 37°C, 5% CO2 until the blastocyst stage. Between the time of imaging and the blastocyst stage, the embryos were evaluated at least once every 24 h to assess their developmental progress.

Acquired images were analysed using DeltaVision Software (Applied Precision Inc.), which allows for quantification of signal intensity of the JC‐1 stained embryo. The ratio of RITC (J‐aggregate) to FITC (J‐monomer) staining was determined for all sections of the embryo, from which an average ratio of J‐aggregate to J‐monomer staining for the entire embryo was determined.

Mitochondrial distribution in 2‐cell embryos

Computer‐based morphometry was used to identify precisely the image properties of 2‐cell embryos imaged using deconvolution microscopy. Two main steps were involved in this process; image segmentation, a process that identifies objects within the image, and image analysis, a process that determines the features of the identified objects. Image segmentation was performed by first applying an edge‐oriented segmentation, followed by region‐oriented segmentation, to the embryo image. Image analysis was focused on extracting image properties that could be used to differentiate between the groups of embryos. Since the goal was to identify possible differences in distribution patterns among 2‐cell embryos, we implemented an image‐feature extraction system using Matlab R12 software (MathWorks Inc., USA). The intensity of RITC and FITC channels was normalized in order to handle differential illumination of RITC and FITC images. The analysis was performed in steps; an edge detection algorithm (Jurisica et al., 2000) was used to identify the outline of the embryo, 32 morphometric features (such as boundary coordinates, perimeter, area, convex hull, Euler number, and eccentricity) were extracted and sorted based on their power to differentiate, and the distribution of pixels within the embryo was analysed.

We have applied the case‐based reasoning system TA3 (Jurisica et al., 1998) to quantitatively analyse differences in mitochondrial distribution. The analysis is based on assessing similarity among embryos using image features extracted from the embryo sections. Case‐based reasoning is a process of remembering and re‐using experience in the form of cases stored in a database. In our application, cases represented individual embryo sections, which were described by 32 extracted image features. We have analysed each embryo using the three middle sections most closely surrounding the nuclei using the convex hull attribute. During the analysis, we have measured similarity among image sections based on their representation using TA3 (Jurisica and Glasgow, 2000). In addition, we have confirmed these results with a Binary Tree‐Structured Vector Quantization (BTSVQ) method (Sultan et al., 2002). We have applied BTSVQ to cluster the embryos based on all 32 features extracted from individual embryo sections.

TUNEL

Terminal deoxynucleotidyl transferase‐mediated dUDP nick‐end labelling (TUNEL) assays were performed on embryos that reached the blastocyst stage. The blastocysts were fixed in 4% paraformaldehyde and stored at –20°C until analysis. Frozen slides were defrosted, washed in PBS and subjected to TUNEL analysis as previously described (Jurisicova et al., 1998). The TUNEL‐stained blastocysts were imaged on the deconvolution microscope using the RITC and DAPI filters. Total cell counts as well as cells positive for TUNEL or displaying condensed chromatin as evidenced by DAPI staining were counted on captured images.

Statistics

JC‐1 staining intensity data were subjected to Kruskal–Wallis one‐way analysis of variance (ANOVA) on ranks followed by Dunn’s test to evaluate the significance of the results within pairs of groups. Two‐way ANOVA was used to determine significance between the TUNEL counts, percentage condensed chromatin and cell counts between treatment groups. The rate of 2‐cell arrest was tested for significance using the χ2‐test, while statistical differences in observations of mitochondrial distribution were tested with Fisher’s exact test and McNemar’s test (SigmaStat 1.0; Jandel Scientific Corp., USA).

Results

Mouse embryo results

Mitochondrial membrane potential

Specificity of mitochondrial staining by JC‐1 was verified through FCCP treatment of stained oocytes, resulting in marked decreases in ΔΨm and the intensity of RITC staining. The level of FITC staining remained fairly constant throughout the imaging time, indicating that the RITC JC‐1 staining was specific to mitochondria with a polarized membrane potential (data not shown) (Van Blerkom et al., 2000; Wilding et al., 2001). This dye was used to stain two profiles of mouse preimplantation development, in vivo and in vitro fertilized, and spare human embryos donated to research (Figure 1A–L).

In vivo fertilized embryos

Experiments performed with the in vivo fertilized and developed group of embryos (see Figure 2A) revealed dramatic changes in ΔΨm associated with individual developmental stages. A substantial increase (P < 0.05) in ΔΨm was observed between the zygote (n = 27) and 2‐cell stage (n = 26), followed by a sudden decrease at the 4‐cell stage (n = 19, P < 0.05). A gradual increase was seen between the 4‐cell stage and blastocyst stage (n = 20), which showed significantly higher potential (P < 0.05) when compared with zygotes, 4‐cell and 8‐cell stages (n = 21), although the increase was not significant compared with the 2‐cell stage or the compacted morula stage (n = 28).

In vitro fertilized

The fertilization in vitro was successful, as on average 59% (283/473) of the oocytes cleaved to the 2‐cell stage 21 h following insemination. Only 2.2% (3/136) of oocytes that were inseminated with heated sperm cleaved to the 2‐cell stage, indicating that a very small percentage of the embryos studied could have been parthenogenetically activated. A subset of IVF embryos was not manipulated but followed as controls in culture, where 59% (47/80) reached the blastocyst stage.

In order to overcome variability in ΔΨm ratios caused by IVF, we increased the sample size in several of the early developmental stages. Analysis of the overall profile of ΔΨm obtained from in vitro fertilized embryos (Figure 2B) showed differences compared to the in vivo fertilized embryos, especially in the early cleavage stages. There was no change in ΔΨm between the zygote (n = 19) and 4‐cell stage (n = 48), followed by a rise at the 8‐cell stage (n = 40), reaching its maximum value at the blastocyst stage (n = 25), similar to the profile described for the late preimplantation stages in the in vivo fertilized embryos. Blastocysts had a significantly (P < 0.05) higher ΔΨm than all other stages, with the only other significant increase in ΔΨm being the compacted morula (n = 19) versus the 4‐cell stage. Interestingly, the increase in ΔΨm seen at the 2‐cell stage in the in vivo fertilized embryos, which coincides with activation of the embryonic genome and mtDNA transcription (Piko and Taylor, 1987), was not seen in the in vitro fertilized 2‐cell embryos (n = 44).

Comparison of stages across groups

Comparison of JC‐1 levels at all embryo stages between the two modes of fertilization yielded interesting differences (Figure 2C). Those embryos that were in vitro fertilized and assessed at the zygote stage did show a significant (P < 0.05) 25% elevation in ΔΨm when compared with zygotes fertilized in vivo. Interestingly, 2‐cell in vitro fertilized embryos showed a significant 40% deficit in ΔΨm when compared with in vivo fertilized 2‐cell embryos, indicating that the increase in ΔΨm found in in vitro fertilized zygotes is lost by the 2‐cell stage. At the 8‐cell stage, there was a significant (P < 0.05) 30% increase in ΔΨm in the in vivo fertilized embryos. By the compacted morula and blastocyst stages, there were no statistically significant differences in ΔΨm between the two groups, probably due to the large variance within the samples.

Mitochondrial distribution

Mitochondrial distribution was also assessed at the 2‐cell stage for the in vivo and in vitro fertilized groups. Two researchers independently performed qualitative analysis of JC‐1 staining where the pattern of J‐aggregate staining was diffuse in both groups of embryos. A pattern of perinuclear clustering of low polarized mitochondria (J‐monomer staining) was observed in 10/44 in vitro fertilized, but only in 1/26 in vivo fertilized, 2‐cell embryos suggesting that differences exist in the subcellular localization of mitochondria between these two groups of embryos. We performed computer‐aided image analysis using the convex hull parameter as the determinant of mitochondrial distribution. High correlation was observed in staining within each group at respective wavelengths (∼86%) while the overall correlation between the groups was low (∼14%). The image analysis therefore confirmed that differences in the mitochondrial distribution pattern do exist. This was further confirmed when we used case‐based reasoning to analyse differences in the distribution of mitochondria using graphical representation of similarities, which clearly partitioned the two sub‐groups of embryos (data not shown). These results suggest that the mode of fertilization and/or the environment in which the early embryo is cultured directly affects ΔΨm at certain stages and the distribution of the differentially polarized mitochondria at the 2‐cell stage.

TUNEL

This assay was used for determination of possible apoptotic cells through incorporation of dUTP at the site of DNA strand breaks, a hallmark of apoptosis. The fixed blastocysts were also counterstained with DAPI to determine the DNA pattern in the nucleus, where condensed chromatin is another indicator of cell death. Through these two means, the total cell numbers and incidence of cell death were determined for embryos that had been both in vivo and in vitro fertilized (Table I).

The cell numbers for in vivo fertilized embryos, retrieved at the blastocyst stage at day 3.5 p.c., was lower (31± 2) than the average cell number of blastocysts that were in vitro fertilized (62 ± 7) and cultured until day 4.25 p.c. due to a 12–18 h developmental delay in reaching the blastocyst stage in in vitro culture.

All embryos analysed by JC‐1 were returned to culture and assessed to determine whether the use of JC‐1 had a detrimental effect on subsequent embryo development. 62% of in vitro fertilized embryos and 57% of in vivo fertilized embryos exposed to JC‐1 reached the blastocyst stage. There was no correlation between ΔΨm at the pronuclear or 2‐cell stage and developmental potential (data not shown). Furthermore, embryos that were analysed with JC‐1 and cultured following analysis were fixed at the blastocyst stage and subjected to DAPI/TUNEL analysis. Although the percentage of TUNEL‐positive cells from in vivo fertilized blastocysts was not different, blastocysts that were in vitro fertilized and cultured had a higher frequency of cells with condensed chromatin, although this difference did not reach statistical significance. This increase in cell death incidence with fertilization in vitro supports our previous findings using the same protocol (Jurisicova et al., 1998).

While there was no observable detrimental effect of JC‐1 on the frequency of blastocyst formation in in vitro cultured embryos, there were negative effects on cell viability. Both in vivo and in vitro fertilized embryos that were stained with JC‐1 had a higher percentage of DAPI‐positive cells with condensed chromatin than non‐manipulated cultured embryos, but this increase was not statistically significant. These results suggest that while JC‐1‐stained embryos are competent to develop to the blastocyst stage under these analysis conditions, this assay results in an increased incidence of cell death.

In vitro arrest

A cohort of ICR zygotes was flushed on day 0.5 p.c. and cultured in HTF medium, which is known to cause 2‐cell arrest in this particular strain of mice. The frequency of 2‐cell arrest under these conditions was 42.2% (n = 96), in comparison with an incidence of 18% arrest (n = 230) when zygotes are cultured in KSOM medium (Figure 3A). Subsets of 2‐cell embryos from each of these culture conditions were analysed with JC‐1 44 h after hCG. We considered embryos that failed to cleave within 64 h post‐hCG to be arrested at the 2‐cell stage. These embryos were analysed at ∼68 h post‐hCG, to determine whether differences in ΔΨm exist between those embryos that arrest and those that carry on through cleavage (Figure 3B). It was found that the embryos cultured in KSOM (n = 44) had a significantly (P < 0.05) lower level of mitochondrial membrane potential when compared with 2‐cell embryos cultured in HTF. While there was a trend towards increased mitochondrial membrane potential in the arrested 2‐cell embryos at 68 h post‐hCG (n = 20), this trend did not reach significance against the HTF group analysed at the earlier time point (n = 22), suggesting that changes in mitochondrial activity precede developmental arrest. It is important to note that embryos cultured in HTF (at 44 h after hCG) consisted of a mixed population of non‐arrested and arrested embryos, since it was not possible to distinguish between the two populations at the time of analysis.

Moreover, subcellular distribution of JC‐1 staining of 2‐cell embryos was also altered. We specifically looked for differences between those 2‐cell embryos that arrested and those that continued through development. Two researchers independently performed qualitative analysis of the embryos (Table II). The pattern of FITC staining was different between the 2‐cell embryos that were developmentally competent and those that arrested, with an increased prevalence of perinuclear clustering of low polarized mitochondria in the arrested embryos. Staining of high‐polarized mitochondria was not qualitatively different between the three groups, as all exhibited combinations of pericortical and diffuse cytoplasmic staining. These observations were more comprehensively assessed using an image analysis program that had we developed (Figure 4A–F). Similarities between the developmentally competent and arrested 2‐cell embryos were determined and have been presented both graphically and numerically. Graphically, as seen in Figure 4G, there was a minor correlation between the distribution of mitochondria in the HTF cultured 2‐cells using all 32 extracted image features, independent of developmental competence. The most significant correlation between groups was present in the HTF and KSOM 2‐cell embryos, suggesting that developmentally competent embryos have similar mitochondrial distribution patterns irrespective of culture media. Analysis using our BTSVQ clustering algorithm (data not shown) also supported similarity of the KSOM and HTF cultured embryos, with the arrested 2‐cell embryos not fitting the same profile as the developmentally competent KSOM and HTF cultured embryos. Similarity in patterns was further determined for both J‐aggregate and J‐monomer staining between the three groups by using mitochondrial distribution as determined by the convex hull parameter. The mitochondrial distribution pattern of HTF arrested embryos was used for comparison against both the HTF and KSOM cultured groups. The percentage correlation of the J‐monomers (FITC) was high within the HTF arrested 2‐cell population (76.8%) and was lower when compared to both the HTF (18.8%) and KSOM (4.5%) cultured embryos. On the J‐aggregate (RITC) channel, similarities were again present within the arrested 2‐cell embryos (65.9%) and lower in both the HTF (31.0%) and KSOM cultured embryos (3.2%). The high self‐correlation of mitochondrial distribution in HTF arrested embryos suggests a common distribution pattern in developmentally incompetent embryos. Results from the computer analysis indicate that large differences exist in J‐aggregate and J‐monomer mitochondrial distribution between those 2‐cell embryos that are developmentally competent and those that arrest.

Human embryo results

Mitochondrial membrane potential analysis

All human samples were analysed for ΔΨm using JC‐1 (Figure 1M–R). Unfortunately, all human embryos were arrested or fragmented poor quality embryos (n = 140), while the mouse experiments allowed analysis of good quality samples. Abnormally fertilized pronuclear stage embryos (n = 21, 18.6 ± 4.54% degree of fragmentation) were found to have ratios of RITC to FITC fluorescence approximately equal to the arrested 1–3‐cell embryos (n = 17, 30.8 ± 7.33% degree of fragmentation) (Figure 5A). In the 4–6‐cell stage embryos (n = 19, 40.6 ± 5.9% degree of fragmentation) and 8‐cell embryos (n = 64, 45.9 ± 2.3% degree of fragmentation), there was an extremely high ratio of mitochondria with elevated ΔΨm. Since we had a large number of 8‐cell embryos for analysis, the quantified ΔΨm in this group was further broken down by the percentage fragmentation (Figure 5B) as follows: 20% (n = 15), 40% (n = 23), 60% (n = 13), and 80% (n = 7). There was a trend towards increased ΔΨm with increased levels of fragmentation, suggesting that cellular fragmentation is associated with elevated ΔΨm.

Discussion

Mitochondrial membrane potential

ΔΨm is used as an indicator of mitochondrial health, since this measure of ion transport reflects metabolic activity and integrity of the mitochondrial membrane. Changes in ΔΨm with embryo development may therefore be indicative of necessary changes in mitochondrial metabolic activity with embryo cleavage. Our observations of ΔΨm heterogeneity within the population of mitochondria in preimplantation stage embryos are in agreement with several other studies (Wilding et al., 2001; Van Blerkom et al., 2002). Interestingly, using a mouse model we observed that changes in ΔΨm associated with progression of embryos through preimplantation development coincide with important developmental milestones.

Oocyte mitochondria and pre‐compaction stage embryos have been shown to have low levels of metabolic activity (Van Blerkom et al., 1995), low respiratory rates and O2 consumption (Houghton et al., 1996; Thompson et al., 1996; Trimarchi et al., 2000) and limited glucose metabolism (Biggers et al., 1967; Leese and Barton, 1984). These metabolic parameters change as embryo division proceeds, with post‐compaction embryos displaying a marked increase in O2 consumption (Trimarchi et al., 2000) and a switch towards glucose utilization (Gardner, 1998). Since there is no mitochondrial replication until the blastocyst stage (Piko and Taylor, 1987; Jansen and de Boer, 1998), the initial population of mitochondria must be partitioned and the metabolic activity of the smaller population of mitochondria per cell must increase to meet the increasing demands of cellular activity. This in turn would require recruitment of some quiescent mitochondria into an active subset leading to an overall increase in the ratio of high to low ΔΨm as found here. Although there is no definitive correlation between ΔΨm and mitochondrial activity in somatic cells (Richter et al., 1996; Diaz et al., 1999), many mitochondrial functions, including protein import, ATP generation and lipid biogenesis, depend on the maintenance of ΔΨm (Voisine et al., 1999). It is therefore plausible that the trend towards an increase in ΔΨm may be necessary to support the changing metabolic demands of the preimplantation embryo.

Other studies of mitochondrial morphology in human oocytes and preimplantation embryos indirectly support our findings of increasing ΔΨm with embryo cleavage. Throughout oocyte development, from the primordial germ cell to the mature oocyte, mitochondria undergo many spatial and morphological changes (Motta et al., 1988), with all changes leading towards a more differentiated mitochondrial phenotype. Further, post‐fertilization alterations involve a gradual transition to a more elongated shape and an increase in the number of cristae (Dvorak and Tesarik, 1985; Jansen and de Boer, 1998). Not all mitochondria mature at the same rate, suggesting that individual mitochondria may respond to intracellular signals differentially (Jansen and de Boer, 1998), further supporting the concept of mitochondrial heterogeneity. Marked changes in mitochondrial morphology occur at the 8‐cell stage in humans, coincident with embryonic genome activation (Dvorak and Tesarik, 1985; Sathananthan and Trounson, 2000). These variations in structure not only coincide with changes in embryonic genome activation but also with a switch in the metabolic substrates from pyruvate and lactate to glucose (Braude et al., 1988), suggesting that the morphological changes may be a result of a change in metabolic activity and energy substrates. The altered mitochondrial morphology throughout preimplantation development is therefore consistent with our findings of increased ΔΨm with advanced development, concomitant with the differentiation of mitochondria to an amplified number of cristae and a boost in metabolic activity.

While a gradual increase in ΔΨm with advancing development is to be expected, the degree to which changes in the mitochondrial membrane potential occurs could also be indicative of physiological disturbances in development. In examining the ΔΨm changes in in vivo fertilized mouse embryos (Figure 2A), two features are striking: the high ratio ΔΨm at the 2‐cell stage, and the subsequent gradual increase in ΔΨm from the 4‐cell stage through to the blastocyst stage. As the 2‐cell stage is the time of activation of the mouse embryonic genome (Prather and First, 1988), it could possibly explain the need for a transient increase in ΔΨm in order to satisfy the metabolic demands of the embryo. Additionally, it is important to note that this increase is not present in in vitro fertilized embryos, despite the fact that they were cultured in the media most agreeable with mouse embryo development (Ho et al., 1995). Perhaps this lack of increase in ΔΨm reflects the artificial environment, compared with in vivo fertilization, which is associated with decreased developmental potential of IVF in human embryos (Quinn et al., 1984).

While the transient increase in ΔΨm at the 2‐cell stage in mouse embryos may be necessary for concurrent activation of the embryonic genome, other results suggest that aberrant increases in the ratio of mitochondria with high to low membrane potential are detrimental. In examining the effects of increased degrees of fragmentation in human 8‐cell embryos, it is obvious that an increase in ΔΨm accompanied the increase in degree of embryo fragmentation. Similar results were also observed in fragmented mouse oocytes and embryos (B.M.Acton and A.Jurisicova, unpublished observations). While measurement of ΔΨm has been used to study somatic cells under a number of pathological conditions, including hypoxia, apoptosis and anoxia (Diaz et al., 1999), the connection between apoptosis and ΔΨm remains unclear. Experiments have provided mixed results as to whether opening of the mitochondrial permeability transition pore during apoptosis causes a loss of mitochondrial membrane potential (Marchetti et al., 1996; Zamzami et al., 1996) or whether ΔΨm (Ankarcrona et al., 1995) or a functional electron transport chain (Jia et al., 1997) is maintained for ATP requirements by the cell undergoing apoptosis. While the mitochondrial membrane potential is sustained in somatic cells undergoing induced apoptosis, the remaining high‐polarized mitochondria of late apoptotic cells represent only a fraction of the mitochondria originally found in the cell, the majority of which have been degraded (Diaz et al., 1999). These last results support experiments performed in fragmented human embryos, which were shown to have decreased ATP (Van Blerkom et al., 1995) despite an increase in ΔΨm as shown in this study. Since we are measuring the ratio between high and low potential mitochondria, increases in this ratio could come about through either an increase in the number of high‐polarized mitochondria or a decrease in the number of mitochondria with low polarity through degradation resulting from apoptosis. A further complicating factor is that the human embryos were examined following their cellular deterioration; therefore at the present time, we cannot determine whether changes in ΔΨm precede or are a consequence of fragmentation.

The mitochondrial membrane potential developmental profiles created in both the mouse and human suggest that there are optimal ratios of ΔΨm at each developmental stage. Embryos, or even individual blastomeres, with ratios that do not fall within a so‐called metabolically acceptable range may experience developmental difficulties. Aberrant increases in the ratio could disrupt the normal developmental pathway and perhaps result in fragmentation. In addition, the lack of healthy human 8‐cell embryos for study does not allow direct confirmation of this hypothesis. Conversely, embryos or blastomeres that are lacking in mitochondria with high ΔΨm could have deficiencies in mitochondria‐specific functions, such as oxidative phosphorylation, electron transport or mitochondrial membrane integrity that could impair their developmental competence. This may be consistent with the observation that fragmented embryos generate more reactive oxygen species (Yang et al., 1998), which could be a consequence of an increased metabolic activity or an underlying reason for the generation of mtDNA mutations, which would in turn trigger cell death manifested by hyperpolarization of mitochondria.

Mitochondrial distribution

Localization of mitochondria is also of key importance in embryo development as it is altered in embryos that are developmentally compromised. Reports of mitochondrial distribution at the 2‐cell stage of mouse and rat embryos suggest that these organelles are homogeneously distributed throughout the cytoplasm (Batten et al., 1987; Muggleton‐Harris and Brown, 1988; Matsumoto et al., 1998). Embryos from some strains of mice are prone to in vitro arrest, however, and they have been shown to exhibit nuclear mitochondrial clustering at the 2‐cell stage (Muggleton‐Harris and Brown, 1988). At the present time, the molecular pathways resulting in 2‐cell arrest have not been elucidated and therefore mitochondrial clustering may not necessarily be related to the inability to develop in vitro (Bavister and Squirrell, 2000). Extensive studies in hamster embryos have shown that mitochondrial relocalization is critical to developmental competence of the embryo and that substances which block development of hamster embryos disturb the normal pattern of mitochondrial clustering at the 2‐cell stage, yet the physiological explanation for relocation of mitochondria remains unknown (Bavister and Squirrell, 2000). The mitochondrial distribution studies in this paper support the previous findings in both the mouse and hamster that localization of mitochondria is of importance to development.

Comparing the mitochondrial distribution patterns qualitatively revealed some differences between the developmentally competent and arrested 2‐cell embryos. Systematically extracting image features and using these to analytically evaluate the same group of embryos using case‐based reasoning and self‐organizing maps indicated extensive differences in the distribution of both low and high polarized mitochondria, and provided us with three orthogonal approaches and statistical analysis evidence to support our qualitative findings. The arrested 2‐cell embryos were found to have different mitochondrial distributions compared with the KSOM and HTF cultured embryos. An increase in ΔΨm was also seen in arrested 2‐cell embryos, following the same trend as increases seen in fragmented human embryos. There were also large differences in mitochondrial distribution when comparing in vivo and in vitro fertilized mouse embryos using our image analysis system. In terms of mitochondrial distribution, our study confirms previous findings in which the distribution pattern of mitochondria was altered between arresting and non‐arresting embryos in both the mouse and hamster (Muggleton‐Harris and Brown, 1988; Bavister and Squirrell, 2000) and further suggests that the mode of fertilization and/or the culture environment contributes to alterations in mitochondrial distribution.

In conclusion, this study shows regulated changes in the ratio of high‐ to low‐polarized mitochondria throughout both mouse and human preimplantation development. There seems to be a propensity towards an increased ratio of high‐polarized mitochondria in instances of compromised developmental competence, through either developmental arrest or fragmentation. These results corroborate previous studies based on mitochondrial morphology and metabolic activity and show that mitochondria play a decisive role in successful preimplantation development. Further studies are required to directly confirm whether changes in ΔΨm are the cause or the consequence of morphological development of the germ cell mitochondria and the changing metabolic demands of the developing embryo.

Acknowledgements

The authors would like to thank Dr T.J.Brown for help with statistical analysis and the Mount Sinai Hospital IVF team for assistance with the human embryos. This study was funded by the Canadian Institutes of Health Research (Grant Number MOP14058). Computational analysis was supported in part by the National Science and Engineering Research Council of Canada (Grant Number 203833‐98). BMA is supported by a CIHR Doctoral Research Award.

Figure 1. Representative images of mouse and human embryos stained with JC‐1 (AF). In vivo fertilized mouse embryos: (A) zygote, (B) 2‐cell, (C) 4‐cell, (D) 8‐cell, (E) compacted morula and (F) blastocyst. (GL) In vitro fertilized mouse embryos: (G) zygote, (H) 2‐cell, (I) 4‐cell, (J) 8‐cell, (K) compacted morula and (L) blastocyst. (MR) Human embryos donated to research: (M) pronuclear, (N) 2‐cell, (O) 5‐cell, (P) 8‐cell, (Q) compacted morula and (R) blastocyst.

Figure 1. Representative images of mouse and human embryos stained with JC‐1 (AF). In vivo fertilized mouse embryos: (A) zygote, (B) 2‐cell, (C) 4‐cell, (D) 8‐cell, (E) compacted morula and (F) blastocyst. (GL) In vitro fertilized mouse embryos: (G) zygote, (H) 2‐cell, (I) 4‐cell, (J) 8‐cell, (K) compacted morula and (L) blastocyst. (MR) Human embryos donated to research: (M) pronuclear, (N) 2‐cell, (O) 5‐cell, (P) 8‐cell, (Q) compacted morula and (R) blastocyst.

Figure 2. Ratio of JC‐1 staining during mouse preimplantation development. (A) In vivo fertilized embryos. (B) In vitro fertilized embryos. (C) Direct comparison of JC‐1 staining in in vitro and in vivo fertilized embryos at the specific preimplantation stages of development. Above ratios were computed through obtaining a ratio of J‐aggregate to J‐monomer staining for each individual embryo section, averaged per embryo and then averaged among embryos from the same stage of development. Bars represent the mean ± SEM of individual embryos. Within each graph, data with different letters indicate a statistically significant difference (P < 0.05) by Kruskal–Wallis ANOVA on ranks followed by Dunn’s test, except in (C) where data were compared within and not between the same embryonic stage.

Figure 2. Ratio of JC‐1 staining during mouse preimplantation development. (A) In vivo fertilized embryos. (B) In vitro fertilized embryos. (C) Direct comparison of JC‐1 staining in in vitro and in vivo fertilized embryos at the specific preimplantation stages of development. Above ratios were computed through obtaining a ratio of J‐aggregate to J‐monomer staining for each individual embryo section, averaged per embryo and then averaged among embryos from the same stage of development. Bars represent the mean ± SEM of individual embryos. Within each graph, data with different letters indicate a statistically significant difference (P < 0.05) by Kruskal–Wallis ANOVA on ranks followed by Dunn’s test, except in (C) where data were compared within and not between the same embryonic stage.

Figure 3.In vitro arrest experiment. (A) Incidence of in vitro arrest for ICR 2‐cells in potassium‐enriched synthetic oviductal medium (KSOM) and human tubal fluid (HTF) media. Bars represent the mean ± SEM of in vitro arrest. Within the graph, data with different letters indicate a statistically significant difference (P < 0.05) by χ2 analysis. (B) Ratio of JC‐1 staining of 2‐cell embryos in KSOM and HTF, and arrested 2‐cell embryos in HTF. Above ratios were computed through obtaining a ratio of J‐aggregate to J‐monomer staining for each individual embryo section, averaged per embryo and then averaged among embryos from the same stage of development. Bars represent the mean ± SEM of individual embryos. Within each graph, data with different letters indicate a statistically significant difference (P < 0.05) by Kruskal–Wallis ANOVA on ranks followed by Dunn’s test.

Figure 3.In vitro arrest experiment. (A) Incidence of in vitro arrest for ICR 2‐cells in potassium‐enriched synthetic oviductal medium (KSOM) and human tubal fluid (HTF) media. Bars represent the mean ± SEM of in vitro arrest. Within the graph, data with different letters indicate a statistically significant difference (P < 0.05) by χ2 analysis. (B) Ratio of JC‐1 staining of 2‐cell embryos in KSOM and HTF, and arrested 2‐cell embryos in HTF. Above ratios were computed through obtaining a ratio of J‐aggregate to J‐monomer staining for each individual embryo section, averaged per embryo and then averaged among embryos from the same stage of development. Bars represent the mean ± SEM of individual embryos. Within each graph, data with different letters indicate a statistically significant difference (P < 0.05) by Kruskal–Wallis ANOVA on ranks followed by Dunn’s test.

Figure 4. Qualitative and quantitative analysis of mitochondrial distribution in 2‐cell embryos using JC‐1 staining. (A) Potassium‐enriched synthetic oviductal medium (KSOM) cultured 2‐cell embryo, (B) human tubal fluid (HTF) cultured 2‐cell embryo, (C) HTF cultured arrested 2‐cell embryo, (D) HTF cultured arrested 2‐cell embryo from which images (E) and (F) were extracted. (E) Computer analysis of mitochondrial distribution from the fluorescein isothiocyanate channel for embryo (D). (F) Computer analysis of mitochondrial distribution from the rhodamine isothiocyanate channel for embryo (D). (G) Pseudocolour representation of a correlation matrix, showing a relationship between JC‐1 staining in the KSOM, HTF and HTF arrested 2‐cell embryos. Embryos 1–20 are HTF cultured, 21–39 are HTF arrested, and 40–61 are KSOM cultured. The colour map corresponds to the scale of correlation coefficients; non‐correlated data display a coefficient of zero (light blue), negative correlation dark blue, and positive correlation ranging from yellow to red. The diagonal of the symmetric correlation matrix represents self‐correlation and thus is equal to 1 (dark red).

Figure 4. Qualitative and quantitative analysis of mitochondrial distribution in 2‐cell embryos using JC‐1 staining. (A) Potassium‐enriched synthetic oviductal medium (KSOM) cultured 2‐cell embryo, (B) human tubal fluid (HTF) cultured 2‐cell embryo, (C) HTF cultured arrested 2‐cell embryo, (D) HTF cultured arrested 2‐cell embryo from which images (E) and (F) were extracted. (E) Computer analysis of mitochondrial distribution from the fluorescein isothiocyanate channel for embryo (D). (F) Computer analysis of mitochondrial distribution from the rhodamine isothiocyanate channel for embryo (D). (G) Pseudocolour representation of a correlation matrix, showing a relationship between JC‐1 staining in the KSOM, HTF and HTF arrested 2‐cell embryos. Embryos 1–20 are HTF cultured, 21–39 are HTF arrested, and 40–61 are KSOM cultured. The colour map corresponds to the scale of correlation coefficients; non‐correlated data display a coefficient of zero (light blue), negative correlation dark blue, and positive correlation ranging from yellow to red. The diagonal of the symmetric correlation matrix represents self‐correlation and thus is equal to 1 (dark red).

Figure 5. Analysis of human embryos. (A) Ratio of JC‐1 staining at the human preimplantation stages of development. (B) Ratio of JC‐1 staining in human 8‐cell embryos broken down by degree of fragmentation. Above ratios were computed through obtaining a ratio of J‐aggregate to J‐monomer staining for each individual embryo section, averaged per embryo and then averaged among embryos from the same stage of development. Bars represent the mean ± SEM of individual embryos. Each data set was tested for significance using the Kruskal–Wallis ANOVA on ranks followed by Dunn’s test; there were no significant differences between the groups.

Figure 5. Analysis of human embryos. (A) Ratio of JC‐1 staining at the human preimplantation stages of development. (B) Ratio of JC‐1 staining in human 8‐cell embryos broken down by degree of fragmentation. Above ratios were computed through obtaining a ratio of J‐aggregate to J‐monomer staining for each individual embryo section, averaged per embryo and then averaged among embryos from the same stage of development. Bars represent the mean ± SEM of individual embryos. Each data set was tested for significance using the Kruskal–Wallis ANOVA on ranks followed by Dunn’s test; there were no significant differences between the groups.

Table I.

Cell death frequency in mouse blastocysts was analysed using both condensed chromatin and TUNEL for the incidence of cell death

Mode of fertilization Analysis Total cell number % condensed chromatin % TUNEL positive 
In vivo Unmanipulated (n = 17) 30.8 ± 2.0  1.49 ± 1.05  1.31 ± 0.7  
 JC‐1‐stained (n = 36) 61.62 ± 7.39  4.43 ± 0.92  0.55 ± 0.22  
In vitro Unmanipulated (n = 9) 47.6 ± 2.6  5.55 ± 1.34  0.69 ± 0.51  
 JC‐1‐stained (n = 20) 41.0 ± 0.18  7.14 ± 4.53  1.35 ± 0.62  
Mode of fertilization Analysis Total cell number % condensed chromatin % TUNEL positive 
In vivo Unmanipulated (n = 17) 30.8 ± 2.0  1.49 ± 1.05  1.31 ± 0.7  
 JC‐1‐stained (n = 36) 61.62 ± 7.39  4.43 ± 0.92  0.55 ± 0.22  
In vitro Unmanipulated (n = 9) 47.6 ± 2.6  5.55 ± 1.34  0.69 ± 0.51  
 JC‐1‐stained (n = 20) 41.0 ± 0.18  7.14 ± 4.53  1.35 ± 0.62  

No statistically significant differences were found using two‐way ANOVA.

Table II.

Pattern of J‐monomer and J‐aggregate staining in 2‐cell embryos

2‐Cell culture milieu Total Analysed FITC (J‐monomers) RITC (J‐aggregates) 
  Diffuse Diffuse and perinuclear Diffuse Pericortical Diffuse and pericortical 
KSOM 25 22a 3a 13a 2a 10a 
HTF 21 17a,b 4a,b 9a 9a 3a 
HTF arrested 20 10b 10b 10a 3a 7a 
2‐Cell culture milieu Total Analysed FITC (J‐monomers) RITC (J‐aggregates) 
  Diffuse Diffuse and perinuclear Diffuse Pericortical Diffuse and pericortical 
KSOM 25 22a 3a 13a 2a 10a 
HTF 21 17a,b 4a,b 9a 9a 3a 
HTF arrested 20 10b 10b 10a 3a 7a 

Qualitative analysis of mitochondrial distribution by JC‐1 staining of mouse 2‐cell embryos. 2‐Cell embryos were assessed for the localization of low‐ (FITC) and high‐ (RITC) polarized mitochondria. Evaluation of FITC and RITC staining patterns indicated that perinuclear staining of low‐polarized mitochondria occurs more frequently in arrested embryos. Distribution of RITC‐stained mitochondria appeared to be similar in all three embryo groups, and no statistically significant differences were observed. Statistical analysis using Fisher’s exact test indicated a significant difference (P = 0.008) between the potassium‐enriched synthetic oviductal medium (KSOM) and human tubal fluid (HTF) arrested groups on the FITC channel as indicated by the different superscript letters.

References

1
Ankarcrona M, Dypbukt JM, Bonfoco E, Zhivotovsky B, Orrenius S, Lipton SA and Nicotera P (
1995
) Glutamate‐induced neuronal death: a succession of necrosis or apoptosis depending on mitochondrial function.
Neuron
 
15
,
961
–973.
2
Batten BE, Albertini DF and Ducibella T (
1987
) Patterns of organelle distribution in mouse embryos during preimplantation development.
Am J Anat
 
178
,
204
–213.
3
Bavister BD and Squirrell JM (
2000
) Mitochondrial distribution and function in oocytes and early embryos.
Hum Reprod
 
15
(Suppl 2),
189
–198.
4
Biggers JD, Whittingham DG and Donahue RP (
1967
) The pattern of energy metabolism in the mouse oocyte and zygote.
Proc Natl Acad Sci USA
 
58
,
560
–567.
5
Bolton VN, Hawes SM, Taylor CT and Parsons JH (
1989
) Development of spare human preimplantation embryos in vitro: an analysis of the correlations among gross morphology cleavage rates and development to the blastocyst.
J In Vitro Fertil Embryo Transfer
 
6
,
30
–35.
6
Braude P, Bolton V and Moore S (
1988
) Human gene expression first occurs between the four‐ and eight‐cell stages of preimplantation development.
Nature
 
332
,
459
–461.
7
Brenner CA, Barritt JA, Willadsen S and Cohen J (
2000
) Mitochondrial DNA heteroplasmy after human ooplasmic transplantation.
Fertil Steril
 
74
,
573
–578.
8
Cohen J, Scott R, Alikani M, Schimmel T, Munné S, Levron J, Wu L, Brenner C, Warner C and Willadsen S (
1998
) Ooplasmic transfer in mature human oocytes.
Mol Hum Reprod
 
4
,
269
–280.
9
Cozzi J, Monier‐Gavelle F, Lievre N, Bomsel M and Wolf JP (
2001
) Mouse offspring after microinjection of heated spermatozoa.
Biol Reprod
 
65
,
1518
–1521.
Cummins J (
1998
) Mitochondrial DNA in mammalian reproduction.
Rev Reprod
 
3
,
172
–182.
Cummins JM, Wakayama T and Yanagimachi R (
1998
) Fate of microinjected spermatid mitochondria in the mouse oocyte and embryo.
Zygote
 
6
,
213
–222.
Diaz G, Setzu MD, Zucca A, Isola R, Diana A, Murru R, Sogos V and Gremo F (
1999
) Subcellular heterogeneity of mitochondrial membrane potential: relationship with organelle distribution and intercellular contacts in normal hypoxic and apoptotic cells.
J Cell Sci
 
112
,
1077
–1084.
Dvorak M and Tesarik J (
1985
) Differentiation of mitochondria in the human preimplantation embryo grown in vitro.
Scr Med
 
3
,
161
–169.
Gardner DK (
1998
) Changes in requirements and utilization of nutrients during mammalian preimplantation embryo development and their significance in embryo culture.
Theriogenology
 
49
,
83
–102.
Hawes SM, Sapienza C and Latham KE (
2002
) Ooplasmic donation in humans: the potential for epigenic modifications.
Hum Reprod
 
17
,
850
–852.
Ho Y, Wigglesworth K, Eppig JJ and Schultz RM (
1995
) Preimplantation development of mouse embryos in KSOM: augmentation by amino acids and analysis of gene expression.
Mol Reprod Dev
 
41
,
232
–238.
Houghton FD, Thompson JG, Kennedy CJ and Leese HJ (
1996
) Oxygen consumption and energy metabolism of the early mouse embryo.
Mol Reprod Dev
 
44
,
476
–485.
Jansen RP and de Boer K (
1998
) The bottleneck: mitochondrial imperatives in oogenesis and ovarian follicular fate.
Mol Cell Endocrinol
 
145
,
81
–88.
Jia L, Allen PD, Macey MG, Grahn MF, Newland AC and Kelsey SM (
1997
) Mitochondrial electron transport chain activity but not ATP synthesis is required for drug‐induced apoptosis in human leukaemic cells: a possible novel mechanism of regulating drug resistance.
Br J Haematol
 
98
,
686
–698.
Jurisica I and Glasgow J (
2000
) Extending case‐based reasoning by discovering and using image features in IVF. ACM Symposium on Applied Computing (SAC 2000), ACM Press, Villa Olmo, Como, Italy, pp.
52
–59.
Jurisica I, Mylopoulos J, Glasgow J, Shapiro H and Casper RF (
1998
) Case‐based reasoning in IVF: prediction and knowledge mining.
Artif Intellig Med
 
12
,
1
–24.
Jurisica I, Glasgow J and Mylopoulos J (
2000
) Incremental Iterative Retrieval and Browsing for Efficient Conversational CBR Systems.
Int J Appl Intellig
 
12
,
251
–268.
Jurisicova A, Latham KE, Casper RF and Varmuza SL (
1998
) Expression and regulation of genes associated with cell death during murine preimplantation embryo development.
Mol Reprod Dev
 
51
,
243
–253.
Keefe DL, Niven‐Fairchild T, Powell S and Buradagunta S (
1995
) Mitochondrial deoxyribonucleic acid deletions in oocytes and reproductive aging in women.
Fertil Steril
 
64
,
577
–583.
Latham KE and Sapienza C (
1998
) Localization of genes encoding egg modifiers of paternal genome function to mouse chromosomes one and two.
Development
 
125
,
929
–935.
Leese HJ and Barton AM (
1984
) Pyruvate and glucose uptake by mouse ova and preimplantation embryos.
J Reprod Fertil
 
72
,
9
–13.
Lenaz G, Bovina C, D’Aurelio M, Fato R, Formiggini G, Genova ML, Giuliano G, Pich MM, Paolucci U, Castelli GP et al. (
2002
) Role of mitochondria in oxidative stress and aging.
Ann NY Acad Sci
 
959
,
199
–213.
Marchetti P, Castedo M, Susin SA, Zamzami N, Hirsch T, Macho A, Haeffner A, Hirsch F, Geuskens M and Kroemer G (
1996
) Mitochondrial permeability transition is a central coordinating event of apoptosis.
J Exp Med
 
184
,
1155
–1160.
Matsumoto H, Shoji N, Sugawara S, Umezu M and Sato E (
1998
) Microscopic analysis of enzyme activity mitochondrial distribution and hydrogen peroxide in two‐cell rat embryos.
J Reprod Fertil
 
113
,
231
–238.
Motta PM, Nottola SA, Micara G and Familiari G (
1988
) Ultrastructure of human unfertilized oocytes and polyspermic embryos in an IVF‐ET program.
Ann N Y Acad Sci
 
541
,
367
–383.
Muggleton‐Harris AL and Brown JJ (
1988
) Cytoplasmic factors influence mitochondrial reorganization and resumption of cleavage during culture of early mouse embryos.
Hum Reprod
 
3
,
1020
–1028.
Muller‐Hocker J, Schafer S, Weis S, Munscher C and Strowitzki T (
1996
) Morphological–cytochemical and molecular genetic analyses of mitochondria in isolated human oocytes in the reproductive age.
Mol Hum Reprod
 
2
,
951
–958.
Piko L and Taylor KD (
1987
) Amounts of mitochondrial DNA and abundance of some mitochondrial gene transcripts in early mouse embryos.
Dev Biol
 
123
,
364
–374.
Plante L and King WA (
1994
) Light and electron microscopic analysis of bovine embryos derived by in vitro and in vivo fertilization.
J Assist Reprod Genet
 
11
,
515
–529.
Polanski Z (
1997
) Genetic background of the differences in timing of meiotic maturation in mouse oocytes: a study using recombinant inbred strains.
J Reprod Fertil
 
109
,
109
–114.
Prather RS and First NL (
1988
) A review of early mouse embryogenesis and its applications to domestic species.
J Anim Sci
 
66
,
2626
–2635.
Quinn P, Warnes GM, Kerin JF and Kirby C (
1984
) Culture factors in relation to the success of human in vitro fertilization and embryo transfer.
Fertil Steril
 
41
,
202
–209.
Ravagnan L, Roumier T and Kroemer G (
2002
) Mitochondria, the killer organelles and their weapons.
J Cell Physiol
 
192
,
131
–137.
Reers M, Smith TW and Chen LB (
1991
) J‐aggregate formation of a carbocyanine as a quantitative fluorescent indicator of membrane potential.
Biochemistry
 
30
,
4480
–4486.
Reers M, Smiley ST, Mottola‐Hartshorn C, Chen A, Lin M and Chen LB (
1995
) Mitochondrial membrane potential monitored by JC‐1 dye.
Methods Enzymol
 
260
,
406
–417.
Reynier P, May‐Panloup P, Chretien MF, Morgan CJ, Jean M, Savagner F, Barriere P and Malthiery Y (
2001
) Mitochondrial DNA content affects the fertilizability of human oocytes.
Mol Hum Reprod
 
7
,
425
–429.
Richter C, Schweizer M, Cossarizza A and Franceschi C (
1996
) Control of apoptosis by the cellular ATP level.
FEBS Lett
 
378
,
107
–110.
Sathananthan AH and Trounson AO (
2000
) Mitochondrial morphology during preimplantational human embryogenesis.
Hum Reprod
 
15
(Suppl 2),
148
–159.
Smith LC and Alcivar AA (
1993
) Cytoplasmic inheritance and its effects on development and performance.
J Reprod Fertil
 
48
(Suppl),
31
–43.
St John JC (
2002
) Ooplasm donation in humans: the need to investigate the transmission of mitochondrial DNA following cytoplasmic transfer.
Hum Reprod
 
17
,
1954
–1958.
Steuerwald N, Barritt JA, Adler R, Malter H, Schimmel T, Cohen J and Brenner CA (
2000
) Quantification of mtDNA in single oocytes polar bodies and subcellular components by real‐time rapid cycle fluorescence monitored PCR.
Zygote
 
8
,
209
–215.
Sultan M, Wigle DA, Cumbaa CA, Maziarz M, Glasgow J, Tsao MS and Jurisica I (
2002
) Binary tree‐structured vector quantization approach to clustering and visualizing microarray data.
Bioinformatics
 
18(Suppl. 1)
,
S111
–S119.
Sutovsky P, Moreno RD, Ramalho‐Santos J, Dominko T, Simerly C and Schatten G (
2000
) Ubiquitinated sperm mitochondria selective proteolysis and the regulation of mitochondrial inheritance in mammalian embryos.
Biol Reprod
 
63
,
582
–590.
Thompson JG, Partridge RJ, Houghton FD, Cox CI and Leese HJ (
1996
) Oxygen uptake and carbohydrate metabolism by in vitro derived bovine embryos.
J Reprod Fertil
 
106
,
299
–306.
Thorburn DR and Dahl HH (
2001
) Mitochondrial disorders: genetics counseling prenatal diagnosis and reproductive options.
Am J Med Genet
 
106
,
102
–114.
Trimarchi JR, Liu L, Porterfield DM, Smith PJ and Keefe DL (
2000
) Oxidative phosphorylation‐dependent and ‐independent oxygen consumption by individual preimplantation mouse embryos.
Biol Reprod
 
62
,
1866
–1874.
Van Blerkom J, Davis PW and Lee J (
1995
) ATP content of human oocytes and developmental potential and outcome after in‐vitro fertilization and embryo transfer.
Hum Reprod
 
10
,
415
–424.
Van Blerkom J, Sinclair J and Davis P (
1998
) Mitochondrial transfer between oocytes: potential applications of mitochondrial donation and the issue of heteroplasmy.
Hum Reprod
 
13
,
2857
–2868.
Van Blerkom J, Davis P and Alexander S (
2000
) Differential mitochondrial distribution in human pronuclear embryos leads to disproportionate inheritance between blastomeres: relationship to microtubular organization, ATP content and competence.
Hum Reprod
 
15
,
2621
–2633.
Van Blerkom J, Davis P, Mathwig V and Alexander S (
2002
) Domains of high‐polarized and low‐polarized mitochondria may occur in mouse and human oocytes and early embryos.
Hum Reprod
 
17
,
393
–406.
Voisine C, Craig EA, Zufall N, von Ahsen O, Pfanner N and Voos W (
1999
) The protein import motor of mitochondria: unfolding and trapping of preproteins are distinct and separable functions of matrix Hsp70.
Cell
 
97
,
565
–574.
Wilding M, Dale B, Marino M, di Matteo L, Alviggi C, Pisaturo ML, Lombardi L and De Placido G (
2001
) Mitochondrial aggregation patterns and activity in human oocytes and preimplantation embryos.
Hum Reprod
 
16
,
909
–917.
Yang HW, Hwang KJ, Kwon HC, Kim HS, Choi KW and Oh KS (
1998
) Detection of reactive oxygen species (ROS) and apoptosis in human fragmented embryos.
Hum Reprod
 
13
,
998
–1002.
Zamzami N, Susin SA, Marchetti P, Hirsch T, Gomez‐Monterrey I, Castedo M and Kroemer G (
1996
) Mitochondrial control of nuclear apoptosis
J Exp Med
 
183
,
1533
–1544.