|Home | About | Journals | Submit | Contact Us | Français|
New techniques to boost male and female fertility are being pioneered at a rapid pace in fertility clinics to increase the efficiency of assisted reproduction methods in couples in which natural conception has not been achieved. This study investigates the possible epigenetic effects of ooplasm manipulation methods on postnatal growth and development using a mouse genetic model, with particular emphasis on the possible effects of intergenotype manipulations. We performed interstrain and control intrastrain maternal pronuclear transfers, metaphase-II spindle transfers, and ooplasm transfer between C57BL/6 and DBA/2 mice, and found no major, long-term growth defects or epigenetic abnormalities, in either males or females, associated with intergenotype transfers. Ooplasm transfer itself was associated with reduced viability, and additional subtle effects of ooplasm strain of origin were observed. Both inter- and intrastrain ooplasm transfer were associated with subtle, transient effects on growth early in life. We also performed inter- and intrastrain germinal vesicle transfers (GVTs). Interstrain GVT females, but not males, had significantly lower body weights at birth and thereafter compared with the intrastrain GVT and non-GVT controls. No GVT-associated changes were observed in DNA methylation of the Mup1, Rasgrf1, H19, Snrpn, or Peg3 genes, nor any difference in expression of the imprinted Rasgrf1, Igf2r, or Mest genes. These results indicate that some ooplasm manipulation procedures may exert subtle effects on growth early in life, while intergenotype GVT can result in significant growth deficiencies after birth.
Human fertility clinics use many techniques to increase the efficiency of establishing pregnancy in couples when either parent is subfertile. Many of the techniques developed to treat human infertility have evolved from basic research on animal models, but some micromanipulative techniques, such as intracytoplasmic sperm injection (ICSI), were developed and applied clinically, without prior testing in animal models . Clinical data indicate that in vitro-fertilized (IVF) human eggs have decreased developmental success compared with embryos fertilized in vivo . Furthermore, some studies report increased rates of birth defects and low birth weights associated with babies born from either IVF or ICSI compared with babies resulting from natural conceptions, and an increased incidence of imprinting disorders in children conceived via ICSI [3–7]. While other modes of disease cannot be ruled out, genetic and epigenetic evaluations point to possible mosaic imprinting defects.
One of the common mechanisms of age-related female infertility is aneuploidy from nondisjunction of chromosomes at meiosis I [8–10] due to aberrant spindle alignment, presumably caused by ooplasmic insufficiencies [8, 11]. Also of concern is the mitochondrial health of oocytes [12–14]. Cytoplasm donation involves transferring a small amount of ooplasm (5%–15% egg volume) in conjunction with ICSI to boost mitochondrial number and quality, and increase the energy output necessary to fuel the embryo through preimplantation development . In some cases where a mother carries a mitochondrial defect, the entire germinal vesicle (GV) nucleus is transferred into an enucleated egg [16–19]. Previous animal models for GV transfer (GVT) focused on decreasing the incidence of aneuploidy, and understanding the changes in ooplasm to allow meiosis to progress after GV breakdown and extrusion of the first polar body, preparing the oocyte for fertilization [20, 21]. While some of these techniques have been used in clinics for nearly 10 yr, genetic studies evaluating the effects of “foreign” ooplasm on the maternal genome have not been performed. Reports following the growth and development to 1-yr-old children born after cytoplasm donation procedures revealed that mitochondrial heteroplasmy was still evident at 1 yr of age . Follow-up studies also revealed an XO genotype in 2 of 16 pregnancies. While the sample size is small and the significance of the latter outcomes can be debated, these observations raise concern about possible effects of ooplasm manipulation on meiosis and mitotic processes in the embryo .
Embryos containing only maternal (gynogenones/parthegenones) or paternal (androgenones) genomes demonstrate that both parental genetic contributions are necessary to ensure normal development [23–26]. Pronuclear transfer (PNT) studies demonstrated strain-specific differences in the epigenetic modification of the maternal and paternal genomes after fertilization. Differences in the abilities of mouse strains to confer epigenetic modifications postfertilization are primarily directed by components of the ooplasm [27, 28]. While the specific factors remain unknown, specific loci controlling the process have been mapped [28, 29]. While parental epigenetic marks are set at imprinted loci during gamete maturation, it is believed that the paternal epigenome must be modified after fertilization to make it compatible with maternal cytoplasmic components before zygotic transcription begins . Several studies revealed changes in DNA methylation after fertilization that may depend, in part, on prior methylation state [30–33]. The varying ability of different genotypes to direct the development of gynogenomes and androgenomes, and the strain-dependent differences observed for an egg to direct development after somatic cell nuclear transfer , indicate that the ooplasm modifies the gamete genomes after fertilization, and that this is subject to genetic variation [27, 34]. This genetic variability may thus affect the outcome of ooplasm manipulation procedures for treating infertility.
Of particular concern regarding ooplasm donation are potential epigenetic effects on postnatal growth and development. A previous study reported aberrant methylation patterns in nucleocytoplasmic (NC) hybrid mice after maternal PNT (mPNT) between C57BL/6 and DBA/2 mice. Adult male NC mice had decreased body weight, as well as decreased expression of a group of liver-specific proteins from the Mup1 (major urinary protein) family, coincident with increased methylation at the locus [35, 36]. Alterations were also seen in the brain olfactory marker protein .
While NC hybrids are not seen in nature, techniques in reproductive clinics that create NC hybrid babies led us to question the possible epigenetic consequences of such procedures. We examined the epigenetic effects of ooplasm donation techniques on the maternal genome. We performed cytoplasm donation and GVT, mPNT, and metaphase-II (MII) spindle transfer (ST). We examined the Mup1 locus for changes in methylation, as well as the imprinted loci, H19/Igf2, Snrpn, Peg3, Igf2r, and Rasgrf1. The results reveal no major consistent or long-term growth effects of interstrain maternal nuclear transfer, MII ST, or ooplasm transfer, although subtle effects of ooplasm transfer were noted. There was no change in methylation states of the Mup1 gene or imprinted loci. However, in female progeny mice created after interstrain GVT, a significant reduction in body weight was seen in neonates, which continued at least to 4 mo of age. These results indicate that ooplasm manipulation procedures can lead to long-term effects on growth and health of the progeny, and encourage caution for its clinical use in humans.
Adult C57BL/6 (Harlan Sprague-Dawley, Indianapolis, IN) and DBA/2 (Jackson Laboratories, Bar Harbor, ME) female mice at 8 to 12 wk of age were superovulated by sequential administration of 5 IU equine chorionic gonadotropin (eCG; Sigma-Aldrich, St. Louis, MO) and human chorionic gonadotropin (hCG; Sigma-Aldrich) 48 h apart. MII oocytes were isolated approximately 15 h post-hCG, and cumulus cells were removed by hyaluronidase treatment. To obtain fertilized embryos, females were mated to males of appropriate genotype at the time of hCG injection, and fertilized embryos were isolated approximately 20 h post-hCG. All mice in the study were fed Purina LabDiet 5001 (males) or 5015 (females). All procedures were approved by the Temple University Institutional Animal Care and Use Committee (IACUC) in accordance with all federal and state requirements for laboratory animal welfare. All embryos were cultured in Chatot-Ziomek-Bavister (CZB) medium or potassium simplex optimized medium (KSOM) [37, 38] under an atmosphere of 5% CO2 and 21% O2 in nitrogen at 37°C in a humidified modular incubator (Billups-Rothenberg, Del Mar, CA). After overnight culture, two-cell embryos were transferred into pseudopregnant CD1 females and allowed to develop to term.
PNT procedures were performed as described previously . Briefly, embryos were placed in CZB medium containing 5 μg/ml cytochalasin B (Sigma-Aldrich) and 0.2 μg/ml demecolcine (Sigma-Aldrich) for 0.5 h. The mPNTs were performed in drops of M2 containing 5 μg/ml cytochalasin B and 0.2 μg/ml demecolcine, using a PMM piezo micromanipulator (PMM; Primer Tech, Ltd., Ibaraki-ken, Japan), followed by electrofusion. Reconstructed embryos were cultured in CZB.
ST was performed with MII oocytes in drops of M2 containing 5 μg/ml cytochalasin B and 0.2 μg/ml demecolcine, using the PMM. Following electrofusion of “karyoplasts” containing donor spindles, oocytes were cultured in CZB for 2 h. Control oocytes were immediately put into CZB culture after isolation. To obtain fertilized embryos, ICSI was performed on both the ST and control oocytes, as previously described , and fertilized embryos cultured in CZB medium.
Cytoplasm transfers (CTs) were performed on MII oocytes in M2 medium containing 5 μg/ml cytochalasin B and 0.2 μg/ml demicolcine using the PMM. A volume of cytoplasm (approximately 10% of total oocyte volume) surrounded by plasma membrane was removed from a donor MII oocyte, transferred to the perivitelline space of the recipient oocyte, cultured in CZB for 2 h, subjected to electrofusion, and returned to CZB culture for another 2 h. Fertilized embryos were obtained by CT coupled ICSI (CT-ICSI) as described . Fertilized oocytes were cultured in CZB medium.
Fully grown GV-stage oocytes were collected from needle-punctured antral follicles 46–48 h after injection of eCG. Cumulus-enclosed GV oocytes were collected in modified Eagle medium (MEM)-α supplemented with 20% fetal bovine serum (FBS) and 0.2 mM 3-isobutyl-1-methyxanthine (IBMX) and incubated for 1 h. Cumulus cells were removed by pipetting, and oocytes were incubated in IBMX medium supplemented with 5 μg/ml cytochalasin B and 0.2 μg/ml demecolcine for 0.5 h, then transferred to M2 with 0.2 mM IBMX, 5 μg/ml cytochalasin B, and 0.2 μg/ml demecolcine for micromanipulation. GV transplantation was performed under an inverted Olympus microscope equipped with Narishige micromanipulator, as previously described [18, 42]. Briefly, slits were made in the zona pellucidae using glass needles, and plasma membrane-bound GV-containing karyoplasts with a minimal volume of surrounding ooplasm were removed by aspiration into a 25-μm pipet. The GV-containing karyoplasts were then transferred to the perivitelline space of enucleated GV-stage oocytes through a slit made by a sharp glass needle. Electrofusion was performed with a single 0.9 kV/cm DC fusion pulse for 10 μs in 0.27 M mannitol medium supplemented with 0.1 mM MgSO4, 0.05 mM CaCl2, and 0.3% bovine serum albumen. Reconstituted GV oocytes were placed in MEMα medium containing 20% FBS for 14 h for in vitro maturation (IVM). Mature oocytes were fertilized via ICSI, as previously described , and the embryos cultured in KSOM medium, according to previous studies producing mice after GVT.
Genomic DNA from adult liver was isolated as previously described  and digested with a combination of BamHI and HpaII, or BamHI and MspI. Digested DNAs were resolved on 0.8% agarose, Southern blotted, and hybridized with α-[32PO4]-dCTP-labeled probe. The probe, BS655, was generated by PCR .
Genomic liver or brain DNA was treated with sodium bisulphite, as previously described [44, 45]. PCR was performed using primers that specifically amplify bisulphate-converted genomic DNA at the imprinted gene loci, H19, Snrpn, Peg3, and Igf2r. Nested PCR was performed on mutagenized DNA using primers as previously described [46–48] under the following conditions: 94°C for 3 min, followed by 30 cycles of 94°C for 30 sec, 55°C for 30 sec, and 72°C for 1 min. For combined bisulfite restriction analysis (COBRA), PCR products were subjected to restriction endonuclease digestion with either HinfI (H19 and Snrpn) or TaqI (Peg3 and Igf2r) to assay the methylation state at select cytosine-phosphate-guanine (CpG) dinucleotides of each parental allele and products resolved on agarose. For DNA sequencing, PCR products were gel-purified with Wizard DNA Clean-up System (Promega) columns and cloned into the TOPO TA vector (Invitrogen, Carlsbad, CA). DNA sequencing was performed using a BigDye Terminator V3.1 sequencing kit (Applied Biosystems, Foster City, CA) at either the University of Pennsylvania Sequencing Facility (Philadelphia, PA) or Genewiz Inc. (South Plainfield, NJ).
Our overall objective was to employ a mouse genetic model to evaluate possible effects of intergenotype nuclear transfers or ooplasm transfers on the long-term growth and development of progeny. For this, we selected the DBA/2 and C57BL/6 mouse strains (hereafter, D2 and B6), for which earlier studies indicated that interstrain micromanipulations could lead to undesired epigenetic effects in progeny . Because such interstrain effects might arise at different points in the developmental pathway from the GV-stage oocyte to the fertilized zygote stage, and because clinical studies encompass experimental manipulations at all of these stages, we devised microsurgical schemes to test for effects in GV-stage oocytes, MII-stage oocytes, and fertilized embryos, using procedures that mirror the strategies employed clinically. Overall, we produced a total of 368 mice generated by mPNT, ST, CT, or GVT, and an additional 180 control mice, making this the largest such study of this type yet reported.
Perhaps the simplest and most long-established method of ooplasm manipulation is PNT using fertilized zygotes. We therefore sought first to test this method for possible epigenetic effects, like those reported previously . Because of differences in growth rates between inbred strains of mice, we devised a strategy to produce and compare embryos of F1 hybrid genotypes following mPNT and embryos not produced using mPNT (Fig. 1). To compare effects of inter- and intrastrain mPNT, it was necessary to compare progeny possessing reciprocal hybrid genotypes. Interstrain mPNTs were performed with homozygous D2 and B6 embryos to produce reconstituted zygotes containing: 1) B6 ooplasm, D2 maternal pronuclei (mPN), and B6 paternal pronuclei (pPN) (BDB), or 2) D2 ooplasm, B6 mPN, and D2 pPN (DBD); intrastrain mPNTs were performed with (D2B6) F1 and (B6D2) F1 hybrid embryos, designated, 3) DDB and 4) BBD, respectively. As a final control, we also exposed non-mPNT (D2B6) F1 (DB) and (B6D2) F1 (BD) embryos to culture and embryo transfer identically to the mPNT embryos.
The survival rates of BD and DB cultured embryos to parturition were similar (Table 1). The intrastrain BBDs had a similar birth rate to BD embryos, but DDB embryos had a significantly lower survival compared with DB embryos. Birth rates of the interstrain DBD embryos were significantly lower compared with both intrastrain BBD and BD controls. In contrast, while the interstrain BDB embryos had lower birth rates than DB embryos, they did not differ from intrastrain DDB embryos. These results are not surprising, as previous reports indicated that D2 eggs have decreased viability in culture compared with B6 eggs .
Comparisons of postnatal growth rates of hybrid embryos explanted to culture at the one-cell stage revealed that male and female DB mice were consistently larger than BD until at least 28 wk of age (P < 0.05 to 23 wk; P < 10−5 from 23 to 28 wk) (Fig. 1). If genotype alone controlled growth rate, then we predicted that growth rates would follow the pattern: BDB = DDB > DBD = BBD. This pattern would be altered if interstrain mPNT significantly affected growth rate. In general, the results of mPNTs followed the predicted pattern. For female mPNT mice (Fig. 1B), we found no effect of intrastrain mPNT for the DB genotype (DDB = DB). However, we observed a transient interstrain mPNT effect, transferring D2 mPN into B6 eggs (BDB), leading to a slight (9%) decrease in body weight at 4–6 wk of age (P < 0.03) compared with DB controls. The BDB weights appeared slightly lower at later stages; however, this difference was not statistically significant. Interstrain mPNT (BDB) females were significantly larger than intrastrain (BBD) females (P < 0.005) at 6–28 wk of age, but were not different from DDB females (i.e., intrastrain females of the same genotype). Intrastrain mPNT females of BD genotype (BBD, B6 mPN to B6 ooplasm) were not different from BD mice. Interstrain mPNT with the BD genotype (DBD, B6 mPN to D2 egg) displayed increased body weights compared with both BD (P < 0.05 from 6 to 8 wk; P < 0.008 from 23 to 28 wk) and intrastrain BBD (P < 0.05 from 6 to 28 wk) mPNT females.
For males (Fig. 1C), DB mice were again larger than BD animals (P < 0.05 to 20 wk). Intrastrain mPNT (DDB) resulted in increased growth rates compared with DB males (P < 0.05) to 20 wk. Interstrain transfer mPNT males (BDB) were smaller than intrastrain mPNT males (DDB; P < 0.013), but were not different from DB males. The interstrain DBD males were larger than the intrastrain BBD males only after 24 wk of age (P < 0.01), but were smaller than intrastrain DDB males (P < 0.016) from 6 wk of age. The transfer of either D2 or B6 mPN to D2 ooplasm resulted in increased body weights, but the transfer of either D2 or B6 mPN to B6 ooplasm did not alter the postnatal development of males.
Overall, with respect to postnatal growth rates through 28 wk of age, B6 ooplasm may mildly suppress growth, while D2 ooplasm may mildly enhance growth. However, no consistent, specific effect of interstrain versus intrastrain mPNT was observed. All of the differences observed are explicable on the basis of ooplasm strain of origin, or the effect of parental origin, which results in differential growth of DB versus BD mice, with the DB genotype yielding a greater rate of growth than the BD genotype.
Earlier studies of mPNTs between C57BL/6 and DBA/2 reported decreased expression of the Mup1a and Mup1b alleles, with a correlative increase in the methylation at the Mup1a/Mup1b genomic locus [35, 36]. We analyzed the methylation state of this locus in at least five females and males for each mPNT group by Southern blot and found no differences between experimental mPNT groups and controls (Supplemental Fig. S1 available online at www.biolreprod.org). Analysis of the methylation pattern at the imprinted genes H19, Snrpn, and Peg3 by bisulphite conversion followed by COBRA analysis of liver DNAs from five females and males from each group of mPNT mice revealed normal patterns of methylation for each of the imprinted genes analyzed (Supplemental Fig. S1). Thus, mPNT was not associated here with changes in DNA methylation for these loci.
The maternal genome could be more susceptible to effects of foreign ooplasm at an earlier stage (i.e., before fertilization). To test for effects of foreign ooplasm on the maternal genome at the MII stage, we performed inter- and intrastrain MII STs. As with the mPNTs, the inter- and intrastrain MII STs were constructed maintaining F1 hybrid genotype for the reasons outlined above (Fig. 2). Reconstructed oocytes and nonmanipulated control oocytes were fertilized by ICSI. These procedures resulted in embryos containing: 1) interstrain transfers with D2 ooplasm, a B6 maternal genome and D2 sperm (ST-DBD); 2) intrastrain MII STs with D2 ooplasm, a D2 maternal genome, and B6 sperm (ST-DDB); 3) unmanipulated B6 and D2 MII-stage oocytes fertilized with D2 and B6 sperm, respectively (ICSI-BD and ICSI-DB) (Fig. 2). Survival rates for ST and ICSI embryos during manipulation, culture, and development to parturition were not significantly different (Table 2).
Postnatal growth comparisons for female progeny (Fig. 2B) revealed that ICSI-BD mice were larger (16% on average) than ICSI-DB mice until 10 wk of age (P < 0.02), a trend that reversed after 16 wk, as ICSI-BD females reached a growth plateau and ICSI-DB females continued to increase in size (P < 0.015), resulting in a 36% difference in size at 28 wk. Interstrain ST-DBD females did not differ from intrastrain ST-DDB females; however, interstrain females had significantly lower body weights compared with ICSI-BD females until 10 wk of age (P < 0.003), presumably a result of the higher body weights of ICSI-BD females during early postnatal development. Among males (Fig. 2C), there were no significant differences between ICSI-BD and ICSI-DB mice during postnatal development. While neither intrastrain ST-DDB nor interstrain ST-DBD males showed changes in body weight compared to the same ICSI genotypes, the interstrain ST-DBD males were significantly smaller than ST-DDB males from 3 wk of age onward (P < 0.05), reflecting the effect of genotype noted above. Overall, there was no consistent, significant difference between interstrain and intrastrain ST mice; however, mild effects of the ST procedure on females were noted during early postnatal growth.
Analysis of the Mup1a/Mup1b genomic locus as well as imprinted loci in liver tissues revealed no changes in methylation between interstrain, ST-DBD, and intrastrain, ST-DDB controls (Supplemental Fig. S2). Thus, ST is not associated here with changes in DNA methylation for these loci.
CT has been performed clinically in an effort to enhance embryo development, and this has been performed via cytoplasm-coupled ICSI (CT-ICSI) . This creates a potential for localized effects of foreign ooplasm on the pPN as it forms. We wished to test whether CT-ICSI could lead to epigenetic effects on growth and DNA methylation. However, with mouse ICSI techniques, the transfer of a large volume of ooplasm in conjunction with ICSI proved highly inefficient, so we adopted a modified CT procedure, wherein a small amount of ooplasm is transferred with the sperm, and a separate, larger cytoplast is transferred to approximate the total volume of cytoplasm transferred clinically (~10%). Inter- and intrastrain donations were performed using unfertilized D2 MII-stage oocytes (Fig. 3A). Reciprocal transfers using B6 MII-stage ooplasts proved inefficient, hence only D2 MII-stage oocytes were employed. Manipulated and nonmanipulated oocytes were fertilized with B6 sperm via CT-ICSI and ICSI, respectively.
The survival of CT embryos after cytoplasm donation and ICSI, as well as birth rates, were significantly lower than other mouse models presented in this study; however, birth rates were equal between inter- and intrastrain CT mice (Table 3), with no major phenotypic abnormalities at birth. We performed postnatal growth comparisons between intra- and interstrain cytoplasm donation animals created with the same F1 hybrid genotype, DB (Fig. 3B). Intrastrain CT-DDDB females were larger (average = 10%) compared with ICSI-DB females until 13 wk of age (P < 0.01). Interstrain CT-DBDBs were slightly but significantly smaller (average = 11%) than intrastrain CT-DDDB at 8–16 wk of age (P < 0.05); however, interstrain CT-DBDB females did not differ statistically from ICSI-DB controls. A trend toward reduced body weight among the CT-DBDB females continued thereafter, but this difference did not reach statistical significance. Among males, body weights were greater in intrastrain, CT-DDDB males relative to ICSI-DB males (P < 0.05) from 4 wk continuing throughout adulthood, but no statistically significant difference was seen between interstrain CT-DBDB males and either intrastrain CT-DDDB or ICSI-DB males (Fig. 3C). Thus, there was no major effect of interstrain versus intrastrain CT on postnatal growth rates, but subtle effects were apparent, including transient effects on postnatal growth in females. Methylation-sensitive Southern blotting of the Mup1a/Mup1b locus and COBRA assays for imprinted loci revealed that the epigenetic state was unaltered at these loci (Supplemental Fig. S3).
The most extensive form of ooplasm manipulation is GVT, which effects a nearly complete cytoplasm replacement, and which has been employed in animal models and applied to human oocytes [16–21, 51]. GVT not only yields the greatest potential for intergenotype epigenetic effects, but also exposes the maternal genome to foreign ooplasm at an earlier stage than any of the other manipulations. To test whether interstrain GVT could exert an effect on embryo phenotype, intra- and interstrain GVTs were performed, followed by IVM and ICSI (GVT-BBD and GVT-DBD), and progeny compared to each other and to non-GVT control embryos, also produced following IVM and ICSI (IVM-BD, IVM-DB). GV karyoplasts from D2 GV-stage oocytes proved inefficient for GVT; hence, only B6 GVTs were performed. The inter- and intrastrain progeny were thus of identical genotypes. Fusion rates after GVT were lower for interstrain (DBD) compared with intrastrain (BBD) oocytes; however, rates of maturation to MII were higher for interstrain transfers compared with both intrastrain GVTs and IVM controls. No significant difference was observed among groups for birth rates following transfer to foster mothers, and GVT-derived embryos developed to term with no obvious phenotypic abnormalities (Table 4).
There was no significant difference in growth rates of GVT-BBD, IVM-BD, and IVM-DB mice of either sex (Fig. 4). However, interstrain GVT yielded females (GVT-DBD) that were significantly growth retarded compared with the intrastrain GVT-BBD females, and were, on average, 14% smaller at 3–8 wk (P < 0.03) and 22% smaller at 9–16 wk (P < 0.006). Interstrain GVT-DBD females were also significantly smaller than IVM-BD animals beginning at 1 wk of age, and were, on average, 19% smaller at 1–7 wk (P < 0.008) and 35% smaller from 8 to 16 wk P < 10−5), although no significant difference in growth was observed between intrastrain GVT-BBD and IVM-BD controls (Fig. 4B). These effects were not observed in males (Fig. 4C). It should be noted that, while the ooplasm in intra- and interstrain GVT progeny differs, the growth retardation in females cannot be attributed to this difference, because it was not observed in males, and because D2 ooplasm in the above studies tended to enhance, rather than suppress, postnatal growth.
Methylation-sensitive Southern blotting of the Mup1a/Mup1b locus and COBRA assays for H19, Igf2r, Snrpn, and Peg3 revealed no change in DNA methylation patterns of interstrain GVT females or males compared to intrastrain or IVM-ICSI controls (Supplemental Figs. S4 and S5). Further bisulphite sequencing of the Rasgrf1 locus in 10 IVM-BD, 10 GVT-BBD, and 12 GVT-DBD females revealed a possible hypermethylation for Rasgrf1 in the two smallest GVT-DBD females (Supplemental Fig. S6; nos. 704, 703, represents 2/4 “small” females); however, without specific markers to differentiate the parental alleles, hypermethylation cannot be confirmed. Altered expression of Rasgrf1, Igf2r, and Mest genes can affect growth [52–54], but none of these genes differed in expression between selected GVTs (n = 12) and IVM (n = 16) controls, nor between GVT-DBD females (n = 16) that varied in body weight (data not shown).
Techniques such as ooplasm donation and GVT were developed for clinical application to address potentially compromised ooplasm quality, and thereby to increase developmental success during assisted reproduction, while transmitting the maternal genotype. Studies in the mouse have revealed significant epigenetic changes after fertilization, such as changes in DNA methylation, nuclear reprogramming, changes in pronuclear function [27, 29, 34], genetic differences in some of these processes [23–26, 29, 34], and effects of exogenous factors, such as culture-induced loss of imprinting [55, 56], which may also be affected by genotype. Earlier studies indicated that foreign ooplasm could adversely affect pronuclear function and long-term development . These results raised concerns about possible adverse epigenetic effects of ooplasm manipulation procedures for clinical application.
We present here the most comprehensive study to date using a mouse model to test for adverse intergenotype effects of ooplasm manipulation on developmental potential and progeny health. We employed a combination of inbred strains that previously have been shown to display genetic differences in pronuclear modifications and in vitro developmental potentials . Moreover, these two strains are those that were employed previously in studies that reported a possible negative effect of interstrain nuclear transfer . Consequently, the combination of these two strains of mice was judged to be ideally suited to testing for possible negative epigenetic consequences of ooplasm manipulation procedures. We find that interstrain mPNT, ST, and ooplasm transfer produce little or no long-term adverse effects on growth rate or methylation states of the genes assayed. However, interstrain GVT can yield significant effects on progeny growth.
Our mPNT results differ from those previously reported [35, 36], with no apparent growth deficiency, developmental abnormalities, or changes in DNA methylation observed for interstrain constructs. It is noted, however, that the two previous studies employed different embryo culture systems and different sources of the inbred mouse lines, two factors that could account for the differences in outcome. Whatever the basis for the difference, our results indicate that growth deficiencies and malformations are not obligatory consequences of interstrain mPNT. This result mitigates this source of concern over the safety of ooplasm manipulation protocols.
To model clinical ooplasm transfers, a modified technique was used, which, like its clinical counterpart, exposed the emerging paternal pronucleus to a localized high concentration of foreign ooplasm, while also effecting a significant exchange of ooplasmic contents. The overall success of term development following ooplasm transfer was lower compared with other manipulations, indicating a possible cumulative negative effect of these combined manipulations on overall efficiency. Discovery of the underlying cause for this low efficiency will require further study, but may reflect a significant negative effect of the ooplasm transfer procedure on early developmental events.
Our data indicate that the interstrain ooplasm donation does not yield any disproportionate adverse effects on developmental potential, long-term growth phenotype, or overall visible health of progeny compared to intrastrain manipulations. There were no apparent adverse changes in DNA methylation state of the Mup1 locus or imprinted genes examined. Interestingly, ooplasm transfers between some genotypes failed to work efficiently (e.g., only ooplasm donations from B6 to D2 were successful, while B6 oocytes would not accept D2 ooplasts). Thus, the success of ooplasm transfer by cytoplast fusion could be sensitive to individual genotype, but ooplasm transfer by microinjection would not be so affected. Subtle effects of ooplasm strain of origin on postnatal growth were seen in some procedures, such as the mildly growth-suppressive effect of B6 ooplasm and the growth-stimulating effect of D2 ooplasm. We also note that earlier studies in mice revealed subtle effects of ooplasm on pronuclear function and embryo development [27, 29, 57]. Thus, although we observed no gross effects of intergenotype ooplasm transfer on long-term growth rate or epigenetic modification of the genes examined in these studies, the potential for more subtle or transient effects requires further study. Indeed, we note subtle effects of ooplasm transfer, including transiently increased body weights for intrastrain female CT mice compared with ICSI controls, and transiently reduced body weights for interstrain female CT progeny (B6 donor ooplasm) compared with intrastrain ooplasm transfer progeny, but not compared to ICSI controls. Such subtle, early effects on growth rate could reflect some undefined epigenetic effect of the procedure. Early growth deficiencies can be associated with long-term effects on health, particularly when overcome by compensatory weight gain during later life [58–61]. Concerns have also been raised about nuclear-mitochondrial incompatibilities, and potential effects of human germline modification that could arise following modification of the maternal mitochondrial legacy . Thus, our data do not yield evidence of a dramatic, negative effect of ooplasm transfer on long-term growth phenotype; however, more subtle effects may be possible. The actual efficacy of ooplasm donation (i.e., whether a 10% exchange of ooplasm is sufficient to yield a beneficial effect) has been questioned [1, 62]. Hence, the relative risk versus benefit of ooplasm donation clinically remains unresolved.
A second technique that has been applied in animals and to human oocytes [16–21, 51], GVT, is also intended to increase ooplasm quality. This procedure is intended to enable a woman to conceive when her own ooplasm is deficient for supporting embryonic development, or to prevent passing a heritable condition (e.g., mitochondrial mutation). We find a significant negative effect of GVT on the growth rates of female progeny; a body weight reduction not seen in males. Methylation analysis of the Mup1 locus and imprinted loci did not reveal any aberrant DNA methylation, nor did we find any changes in expression of several imprinted genes in GVT females compared to IVM controls. The basis for the growth retardation in the interstrain GVT females remains unexplained. The early onset of the defect (within 4 wk of birth) may reflect a growth hormone defect, as growth hormone-deficient mice display growth deficiencies as early as 10 days of age . The expression of several imprinted genes related to growth (Rasgrf1, Igf2r, and Mest) was unaltered, thus eliminating these genes as likely candidates for a direct cause of the deficiency. The fact that females are affected, but not males, indicates that androgens may ameliorate the growth retardation effect [64, 65], or that diet could play a role. Whatever the reason for the effect of GVT on growth in females, it is clear that the introduction of immature GV-stage nuclei into foreign ooplasm results in negative effects on the long-term growth rates of females, giving reason for caution in applying GVT clinically.
We thank Bela Patel for her technical assistance in the study.
1Supported by National Institutes of Health, National Center for Research Resources grants RO1 RR18907 and R24 RR15253, and by National Institute for Child Health grant RO1 HD41440.