Blastocyst Formation Rates In Vivo and In Vitro of In Vitro-Matured Equine Oocytes Fertilized by Intracytoplasmic Sperm Injection 1

Transcription

1 BIOLOGY OF REPRODUCTION 7, (24) Published online before print 26 December 23. DOI 1.195/biolreprod Blastocyst Formation Rates In Vivo and In Vitro of In Vitro-Matured Equine Oocytes Fertilized by Intracytoplasmic Sperm Injection 1 Y.H. Choi, 3 L.M. Roasa, 3 C.C. Love, 3 D.D. Varner, 4 S.P. Brinsko, 4 and K. Hinrichs 2,3,4 Departments of Veterinary Physiology and Pharmacology 3 and Large Animal Medicine and Surgery, 4 College of Veterinary Medicine, Texas A&M University, College Station, Texas ABSTRACT This study was conducted to evaluate in vivo and in vitro development of in vitro-matured equine oocytes fertilized by intracytoplasmic sperm injection. Oocytes were collected from slaughterhouse-derived ovaries, matured in vitro, and injected with frozen-thawed stallion sperm. In vivo development was assessed after transfer of injected oocytes to the oviducts of recipient mares. Mares were killed days after transfer and the uterus and oviducts flushed for embryo recovery. Of 132 injected oocytes transferred, 69 (52%) were recovered; of these, 25 (36%) were blastocysts with a blastocoele and capsule. In vitro development was assessed in three culture systems. Culture of zygotes in modified Chatot, Ziomek, Bavister medium with BSA containing either 5.5 mm glucose for 7.5 days or.55 mm glucose for 3 days, followed by 3 mm glucose for 2 days, then 4.3 mm glucose for 2.5 days, did not result in blastocyst formation. Culture of zygotes in Dulbecco modified Eagle medium (DMEM)/F-12 with 1% fetal bovine serum with and without coculture with equine oviductal epithelial explants yielded 16% and 15% blastocyst development, respectively. Development to blastocyst was significantly lower in G1.3/2.3/BSA than in DMEM/F-12/BSA or in either medium with 1% added serum (2% vs. 18%, 18% or 2%; P.5), suggesting that requirements for equine embryo development differ from those for other species. These results indicate that in vitro-matured equine oocytes are sufficiently competent to form 36% blastocysts in an optimal environment (in vivo). While we identified an in vitro culture system that provided repeatable blastocyst development without coculture, this yielded only half the rate of development achieved in vivo. assisted reproductive technology, embryo, fertilization, in vitro fertilization, ovum INTRODUCTION In vitro culture (IVC) of equine embryos has had low success compared with that for embryos of other domestic animals. A major reason for the delay in development of equine IVC systems has been lack of embryos with which to work, due to the failure of in vitro fertilization (IVF) to be repeatably successful. Recently, techniques for intracytoplasmic sperm injection (ICSI) using a piezo drill have been developed in the horse. These techniques have result- 1 Supported by the Link Equine Research Endowment Fund (Texas A&M University). A portion of these results have been presented at the meeting of the International Embryo Transfer Society, January Correspondence: Katrin Hinrichs, TAMU 4466, College Station, TX FAX: ; khinrichs@cvm.tamu.edu Received: 1 October 23. First decision: 1 November 23. Accepted: 1 December by the Society for the Study of Reproduction, Inc. ISSN: ed in high fertilization and cleavage rates (69 89%) after injection of in vitro-matured (IVM) equine oocytes [1, 2]. However, the reported rates of in vitro blastocyst formation after piezo-driven ICSI remain low (9 14% (combined morula and blastocyst); [2, 3]), and are similar to those typically reported for conventional equine ICSI (3 4% [4 6]). Only three laboratories have succeeded in producing foals [7; unpublished results] or pregnancies [2] by transfer of equine blastocysts produced using an IVM/ICSI/IVC system. Because ICSI of in vitro-matured oocytes can now provide embryos for study, work has begun on the requirements of the equine embryo for development to the blastocyst stage in vitro. Several laboratories have investigated the use of synthetic oviductal fluid (SOF)-based media, including commercial G1/G2 media, for equine embryo culture. We found that a modified Chatot, Ziomek, Bavister (CZB) medium (CZB-C) was capable of supporting equine embryo development in vitro for 4 days [8] and that glucose concentration in this medium (.55 or 5.5 mm) did not affect development at 4 days [9], but embryos were not cultured further. Culture of injected oocytes for 7 days in G1.2/G2.2 yielded up to 9% blastocyst formation [3]; no difference was seen between change of medium at 3 or 4 days. Maclellan et al. [6] produced 4% (2/57) blastocysts by culture of equine zygotes in modified SOF. Galli et al. [2] obtained 14% development to the compact morula/blastocyst stage using SOF with BSA and amino acids. These rates of blastocyst formation in G1.2/G2.2 or SOF are similar to that reported in the pig (7.8%; [1]), but lower than that reported in cattle (3%; [11]). Other laboratories have used complex culture systems including coculture with somatic cells for equine embryo culture. Li et al. [12] produced the first two IVC equine blastocysts, after IVF of zona-drilled oocytes, by coculturing presumptive zygotes in TCM199 1% fetal bovine serum (FBS) on a bovine oviductal cell monolayer with mouse embryos. Choi et al. [13] produced 1 (3 8%) equine blastocysts from partially zonaremoved, fertilized oocytes by culturing presumptive zygotes in trophoblast-conditioned medium or in TCM/Dulbecco modified Eagle medium (DMEM) with 1% FBS. Blastocyst formation from ICSI equine zygotes was also reported by laboratories using coculture systems with Vero cell monolayers [4, 5] or cumulus cell monolayers [7]. It is unclear whether the poor development of equine zygotes to the blastocyst stage in vitro is a result of suboptimal embryo culture systems or is related to low developmental competence of IVM equine oocytes. To circumvent in vitro culture, Grondahl et al. [14] transferred equine ICSI oocytes to the oviducts of mice or mares for 3 days; however, none of five oocytes transferred to mouse oviducts were cleaved and only one of the five (2%) transferred to mare oviducts was at the two-cell stage. Galli et al. [2] used

2 1232 CHOI ET AL. the sheep oviduct as an in vivo culture system for 5 days after ICSI of equine oocytes. They reported that 5% of injected oocytes showed advanced development (no differentiation was made between compact morula and blastocyst). This rate was significantly higher than that of the in vitro-cultured embryos in that study (14%). We studied the development of ICSI-derived equine zygotes 96 h after transfer to the oviducts of mares; this group showed similar cleavage to that for in vitro-cultured embryos (85% and 8% for in vivo and in vitro, respectively) but the average nucleus (blastomere) number was 16, twice that of in vitrocultured embryos [1]. These latter two studies are the first to suggest that IVM/ICSI equine zygotes, if exposed to the proper environment, may have the competence to develop at rates similar to that found for bovine IVM/IVF zygotes. There is no information available on the rate or pattern of development to blastocyst, ability to descend from the oviducts into the uterus, or ability to form an embryonic capsule, of equine ICSI zygotes transferred to the mare s oviduct. This study was conducted to examine blastocyst development in equine IVM/ICSI zygotes either after transfer to the oviduct in vivo or in different culture systems in vitro. In vivo development was assessed to determine the intrinsic developmental competence of IVM/ICSI zygotes when cultured in the optimum environment. In vitro development was evaluated in three culture systems to evaluate the effects of glucose, oviductal cell coculture, and medium (G1.3/2.3 vs. DMEM/F-12) on equine blastocyst development. MATERIALS AND METHODS Oocyte Collection Equine ovaries were transported from two local slaughterhouses to the laboratory at room temperature (3 4 h transport time) during the months of January April (experiment 1), August and September (experiment 2), September and October (experiment 3), and September December (experiment 4). Temperature of ovaries on arrival at the laboratory was C. They were trimmed of connective tissue and adnexa with scissors and cleaned with sterilized gauze. All visible follicles were opened with a scalpel blade and the granulosa layer of each follicle was scraped using a.5-cm bone curette. The contents of the curette were washed into individual Petri dishes with Hepes-buffered TCM199 with Hanks salts (Gibco Life Technologies, Inc., Grand Island, NY) plus ticarcillin (.1 mg/ml; SmithKline Beecham Pharmaceuticals, Philadelphia, PA). The contents of the Petri dishes were examined using a dissection microscope at 1 2 magnification. Oocyte-cumulus complexes were classified as compact, expanded, or degenerating depending on the expansion of both mural granulosa and cumulus as described previously [15, 16]. Oocytes with any sign of expansion of either the cumulus or the mural granulosa (from having individual cells visible protruding from the surface to having full expansion with copious matrix visible between cells) led to the classification of expanded. Oocytes having both compact cumulus and compact mural granulosa were classified as compact. Only oocytes with expanded cumuli were used for this study. In Vitro Maturation Selected oocytes were washed twice in maturation medium (TCM199 with Earle salts [Gibco], 5 U/ml FSH [Sioux Biochemicals Inc., Sioux Center, IA], 1% FBS [Gibco], and 25 g/ml gentamycin [Gibco]). Oocytes were cultured in droplets at a ratio of 1 l medium per oocyte under light white mineral oil (Sigma Chemical Co., St Louis, MO) at 38.2 C in5%co 2 in air for 24 h. After culture, oocytes were denuded by pipetting in.5% hyaluronidase and those with a polar body were used for ICSI. Oocytes not having a polar body were fixed in buffered formol saline, mounted on a slide with 6.5 l of 9:1 glycerol:pbs containing 2.5 g/ml Hoechst 33258, and examined using fluorescence microscopy to determine the chromatin configuration. FIG. 1. A morula having more than 32 nuclei (central mass) recovered from the oviduct of a mare 7.5 days after transfer of ICSI oocytes. Presumptive maternal cell nuclei are present within the perivitelline space. Bar 5 m Experiment 1: In Vivo Development of Equine Oocytes After ICSI Sperm preparation. Frozen semen from one fertile stallion was used for all experiments. Semen straws were thawed at 37 C for 3 sec and 2 l of semen was placed on the bottom of a 5-ml tube containing 1 ml of modified TALP (Sp-TALP, [17]) for swim-up. After a 2-min incubation in an atmosphere of 5% CO 2 in air,.6 ml of medium was collected from the top of the tube and centrifuged at 327 g for 3 min in a 1.7-ml polypropylene tube. The supernatant was discarded and the sperm pellet was resuspended and washed once with the same medium. The percentage of motile spermatozoa after washing was 6 75%. Intracytoplasmic sperm injection. The ICSI procedure was conducted as previously described [1]. Briefly, the outside diameter of the pipette used for ICSI was 7 8 m. For holding oocytes, a m (outside diameter) pipette was used. Immediately before injection, 1 l of sperm suspension was placed in a 3- l droplet of Sp-TALP containing 1% polyvinylpyrrolidone (Sigma) under oil. ICSI was carried out in a separate 5- l drop of Hepes-buffered TCM199 containing 1% FBS. Each spermatozoon was immobilized by applying a few pulses with a piezo drill (Prime Tech Ltd., Ibaraki, Japan) to the tail immediately before injection [18]. All manipulations were performed at room temperature. Injected oocytes were held for 2 min at room temperature in Hepes-buffered TCM199 containing 1% FBS to heal the broken membrane slowly. Oocyte transfer and recovery. Three mares scheduled for euthanasia due to nonreproductive abnormalities were used as recipients. All animal work was performed according to the U.S. government s Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research and Training and was approved by the Laboratory Animal Care Committee at Texas A&M University. Injected oocytes were transferred to the mares oviducts within 2 h of ICSI. For transfer, mares were administered detomidine (3 5 mg i.v.; Pfizer Inc., Lees Summit, MO) and butorphanol (5 mg i.v.; Fort Dodge Animal Health Co., Fort Dodge, IA). Transfer of injected oocytes to the oviduct was performed via standing flank laparotomy, as previously described for oocyte transfer [19]. Oocytes were transferred to both oviducts of each mare. At 7.5 or 8.5 days after transfer, the mares uteri were flushed for embryo recovery with four flushes of 75 1 ml each of Dulbecco PBS with 1% FBS. Then the mares were killed and the uterus, ovaries, and oviducts were removed. The uteri were rinsed again with 1 ml Dulbecco PBS with 1% FBS to recover any embryos that may have remained after the initial flushing. Within 2 h of the mare s death, the oviducts were trimmed and straightened, then flushed by cannulating the oviductal papilla with a blunt 25-gauge needle and infusing 5 ml Dulbecco PBS through the oviduct into six-well multidishes. Recovered embryos were evaluated under a dissection microscope and then fixed and stained with Hoechst as described above for oocytes. Embryos were classified as morulae if they contained more than 32 embryonic nuclei (nucleated blastomeres) but did not have a blastocoele (Fig. 1), and as blastocysts if they had a blastocoele apparent on examination with the dissection microscope (Fig. 2) and on staining contained more

3 DEVELOPMENT OF EQUINE ICSI EMBRYOS 1233 FIG. 2. Thirteen blastocysts recovered from the uterus of one mare 8.5 days after transfer of ICSI oocytes to the oviducts. Eleven blastocysts were recovered on uterine flush ex vivo, and two additional blastocysts were recovered from the uterus postmortem (left corner). Bar 5 m. than 64 nuclei. Embryos were examined for presence of atypical nuclei, presumed to represent cells of maternal origin, within the zona pellucida. Oocytes that were flattened and ovoid with minimal cytoplasm visible were classified as oviductal in origin (originating from the recipient mare) and were not included in the recovery rate. Experiment 2: In Vitro Development of Equine Zygotes Cultured in Semidefined Media Sperm preparation and injection were performed as described in experiment 1, except that the media used for oocyte holding, swim-up of sperm, and manipulation during ICSI were those based on CZB [2] as modified by Choi et al. ([8], CZB-H, Sp-CZB, and CZB-M, respectively, for the different procedures). After ICSI, injected oocytes were cultured in CZB-C [8] containing nonessential amino acids and.5% BSA supplemented with either low (.55 mm) or high (5.5 mm) concentrations of glucose, at a ratio of 5 l medium per injected oocyte at 38.2 C inan atmosphere of 5% CO 2,5%O 2, and 9% N 2. At 72 h culture, half the volume of medium in both treatments was exchanged with CZB-C containing essential amino acids and 5.5 mm glucose, resulting in a glucose concentration of 3.3 mm in the low-glucose group and 5.5 mm in the high-glucose group. Noncleaved oocytes and those having only two cells were removed at this time. We repeated the same medium-refreshing process at 12 h, resulting in a glucose concentration of 4.26 mm in the lowglucose group and 5.5 mm in the high-glucose group. After 7.5 days of culture, embryos were fixed and stained as described above. Embryos were classified as morulae if they contained more than 32 cells but did not have an organized outer rim of cells (Fig. 3a) and as blastocysts if they contained more than 64 cells and had started organization of outer presumptive trophoblast cells (Fig. 3b). Experiment 3: In Vitro Development of Equine Zygotes Cocultured with Oviductal Epithelial Explants Preparation of oviductal epithelial explants. Oviductal epithelial explants were harvested the day before ICSI was performed, from oviducts that were transported from the slaughterhouse to the laboratory (4 h) either on ice or at room temperature. The connective tissues surrounding the oviduct were trimmed with scissors. The ampulla of the oviduct was opened using sterile scissors and the interior was scraped using a.5-cm bone curette. The recovered tissue was collected in a Petri dish and washed twice in Hepes-buffered TCM199. Cells were transferred to 5 l DMEM/F-12 (Sigma) with 1% FBS and cultured in 5% CO 2 in air at 38.2 C. After 24 h, 3 5 of the resulting vesicle-like structures formed from the oviductal tissue were recovered from the culture dish, washed twice in the same medium, and placed in a 5- l well of the same medium for embryo culture. Sperm preparation and injection were conducted as for experiment 2. Injected oocytes were cultured in DMEM/F-12 1% FBS either alone or in coculture with oviductal vesicles under mineral oil at 38.2 C in5% CO 2 in air. Each 5- l well contained 1 24 zygotes. Every 48 h, half of the medium in each well was removed and replaced with fresh medium. Embryos were cultured for 7.5 days, then fixed and stained to examine the number and status of nuclei, as described above. Experiment 4: In Vitro Development of Equine Zygotes Cultured in Two Different Media Sperm preparation and ICSI were conducted as for experiment 2. Injected oocytes were cultured in microdroplets at the medium ratio and culture conditions described for experiment 2. Zygotes were placed into one of four media: 1) G1.3/G2.3/BSA (which contains 8 mg/ml BSA; IVF Science, Denver, CO), 2) G1.3/G2.3/BSA 1% FBS, 3) DMEM/F-12 with 5 mg/ml BSA, or 4) DMEM/F-12/BSA with 1% FBS. The G1.3 media treatments were completely refreshed with new medium at 48 h. At 96 h, embryos in G1.3 media were transferred to G2.3/BSA with or without FBS (corresponding to early culture conditions) and were completely refreshed with the same media at 144 h. For embryos in DMEM/F-12 treatments, media were completely refreshed every other day. Embryos in all treatments were cultured for 7.5 days, then fixed, stained, and evaluated as described above. Statistical Analysis Three to five replicates were performed for each treatment. Differences in proportions of cleaved embryos and proportions of embryos forming morulae and blastocysts were compared among groups using Chi square analysis, with Fisher exact test used when the expected value for any parameter was less than five. RESULTS For these experiments, 623 ovaries were processed and 341 follicles were scraped (5.5 follcles per ovary). A portion of one replicate of oocytes having compact cumuli was accidentally lost after classification; it was estimated to contain 4 compact-cumulus oocytes. Using this estimate, a total of 1867 oocytes was recovered (55% recovery rate). FIG. 3. Images of an in vitro-produced equine morula (a) and blastocyst (b) stained with Hoechst Bar 5 m.

4 1234 CHOI ET AL. Total no. (%) blastocysts* 6 (26) 6 (38) 13 (43) 25 (36) TABLE 2. In vitro development of equine ICSI zygotes cultured in CZB- C in varying levels of glucose for 7.5 days.* Glucose (mm) Injected oocytes cultured n Oocytes cleaved Oocytes developed to Morula Blastocyst Total no. (%) morulae* 11 (48) () 2 (7) 13 (19) (74) 47 (85) 3 (5) 3 (5) () () * All percentages are calculated from the number of injected oocytes cultured. Concentration of glucose:.55 mm glucose for 3 days, then 3.3 mm for 2 days, then 4.26 mm. Concentrations of glucose: 5.5 mm glucose throughout culture. TABLE 1. In vivo development of equine oocytes fertilized by ICSI and transferred to the oviduct (recovered at days after ICSI). Recovered from uterus Recovered from oviduct Degenerated Morulae Blastocysts Degenerated Morulae Blastocysts No. (%) of transferred oocytes recovered No. of injected oocytes transferred Recipient (58) 16 (38) 3 (6) 69 (52) Total * Percentages of morulae and blastocysts were calculated from the total number of injected oocytes recovered. Of these, 1181 (63%) had expanded cumuli, 52 (27%) had compact cumuli, and 184 (1%) were degenerating. For these experiments, 114 oocytes with expanded cumuli were used. The remaining oocytes were used in separate studies. When the oocytes were evaluated after 24 h of culture in maturation medium, 25 were broken during denuding and 989 were observed for the presence of a polar body. A polar body was visualized in 619 oocytes (63%). Of 37 oocytes without visible polar bodies, 34 were found to be in MI and 15 in MII on fixation and staining, and the remainder were degenerating. The 619 oocytes with polar bodies were subjected to ICSI and 611 (99%) were successfully injected with sperm. Of these, 55 oocytes were used for this study and 61 oocytes were used on a different project. Experiment 1 A total of 132 injected oocytes were transferred to the oviducts of recipient mares (Table 1). Three mares were used; oocytes were transferred to both oviducts in each mare. After being killed, 69 (52%) of the embryos were recovered, 48 from the uterus and 21 from the oviducts. An additional six oocytes (one from the uterus and five from the oviducts) were classified as oviductal, i.e., originating from the recipient mares, and were not included in the recovery rate. Twenty-five (36%) of the recovered embryos were blastocysts; all of these were recovered from the uterus. Of these, one was an early blastocyst and the remainder were expanded blastocysts with apparently normal morphology, including a capsule (Fig. 2). Thirteen (19%) of the recovered embryos were morulae; 3 were recovered from the uterus and 1 from the oviducts. All morulae had atypical nuclei within the perivitelline space, suggesting invasion by maternal cells (Fig. 1). Blastocysts that still had remnants of the zona pellucida also contained apparent maternal cells visible inside the zona pellucida, but external to the capsule; those blastocysts having shed the zona appeared to have only embryonic nuclei present. Experiment 2 The degree of embryo development achieved in injected oocytes cultured in CZB-C with.55 or 5.5 mm glucose in early culture is shown in Table 2. There were no significant differences in cleavage (74 85%) or development to morula (5%) between the two concentrations of glucose for early culture. None of zygotes from either group developed to the blastocyst stage.

5 DEVELOPMENT OF EQUINE ICSI EMBRYOS 1235 TABLE 3. In vitro development of equine oocytes fertilized by ICSI and cultured in either DMEM/F-12 1% FBS or DMEM/F-12 1% FBS with equine oviductal epithelial explants for 7.5 days.* Media Injected oocytes cultured n Oocytes cleaved DMEM/F-12 DMEM/F-12 oviductal epithelial explants (61) 41 (66) * All percentages are calculated from the number of injected oocytes cultured. Oocytes developed to Morula 5 (11) 7 (11) Blastocyst 7 (15) 1 (16) Experiment 3 The degree of embryo development achieved in injected oocytes cultured in DMEM/F12 1% FBS, with and without oviductal coculture, is presented in Table 3. The rates of cleavage (61 66%) and development to the blastocyst stage (15 16%) were not significantly different between treatments (P.1). Blastocoele expansion at 7.5 days was limited (Fig. 4) and capsules were not formed. Experiment 4 The degree of development of equine zygotes cultured in either G1.3/G2.3/BSA or DMEM/F-12/BSA, both with or without 1% FBS, is shown in Table 4. There were no significant differences in cleavage (59 64%) among treatments. Development to blastocyst in G1.3/2.3/BSA (2% blastocysts) was significantly (P.5) lower than that for the other three treatments. Development in DMEM/F- 12/BSA (18% blastocysts) was not significantly different from that for media with serum (18 2%). Blastocoele expansion at 7.5 days was limited and capsules were not formed. DISCUSSION This is the first report to examine the rate of blastocyst formation of equine ICSI zygotes after transfer to the mare s oviduct in vivo. Oviductal transfer has been used previously for production of pregnancies and foals after assisted reproduction techniques, including ICSI, oocyte transfer, and gamete intrafallopian tube transfer in the horse [19, 21 26]. Scott et al. [25] found a significantly lower pregnancy rate at Day 15 after oviductal transfer of IVM oocytes than after transfer of ex vivo-collected preovulatory oocytes (1% vs. 82%, respectively). Hinrichs et al. [26] reported that the fertilization rate of IVM oocytes transferred to the oviduct of inseminated mares and recovered 48 h after transfer was 77%; this was significantly higher than the rate of IVF ( 22%). In the present study, development of equine ICSI zygotes transferred to the oviduct was superior to that for in vitro culture (36% vs. 2% blastocysts). This is consistent with our previous finding that the average nucleus number at 96 h for ICSI embryos transferred to the oviduct was twice that of in vitro-cultured embryos [1]. The high percentage of blastocyst formation found after transfer of ICSI oocytes to the oviduct suggests that about 35% of equine oocytes matured in vitro and fertilized by ICSI have the capability to develop to the blastocyst stage. The difference between the relatively high fertilization/blastocyst development reported after oviductal transfer of IVM oocytes ([26], this study) and the low 15-day pregnancy rate achieved after similar transfers [25] suggests that some blastocysts originating from IVM oocytes may fail to develop further within the uterus. Alternatively, there may have been differences in oocyte selection for culture, methods of IVM, and selection of oocytes for transfer that affected the viability of transferred oocytes differentially between laboratories. Further study is needed to determine the proportion of blastocysts resulting from equine IVM oocytes that can go on to establish normal pregnancy. Blastocysts represented only 19% of injected oocytes initially transferred to the oviduct; however, we calculated the rate of blastocyst formation based on the number of embryos recovered from the tract rather than the number transferred. We did this under the hypothesis that embryos were lost soon after transfer, without regard for developmental stage, based on the following findings: 1) the mare normally retains oocytes from previous ovulations within the oviduct for a period of months [27]; 2) when in vitromatured oocytes were transferred to the oviducts of inseminated mares and recovered 2 days later, the proportion of recovered oocytes that were nondegenerating (72%) was not significantly different from that of oocytes similarly selected and fertilized in vitro (68 8%; [26]); 3) when sperm-injected oocytes were transferred to the oviduct and recovered 4 days later, the cleavage rate of recovered oocytes (85%) was not significantly different from that of injected oocytes cultured in vitro (8%; [1]; and 4) the proportion of ICSI embryos developing to morula or blastocyst in vivo in the present study (38/69, 55%) agrees well with the proportion of injected equine oocytes developing to morula/blastocyst after transfer to the oviducts of sheep (5%) when embedded in agar allowing 1% recovery [2]. During ICSI, the piezo drill generates a small hole (7 8 m in diameter) in the zona pellucida. We previously found that ICSI embryos recovered from oviducts 96 h after trans- FIG. 4. Five blastocysts, including two that have started hatching through the hole incurred during ICSI, cultured in DMEM/F-12 1% FBS for 7.5 days in vitro. Bar 5 m.

6 1236 CHOI ET AL. TABLE 4. In vitro development of equine oocytes fertilized by ICSI and cultured in two different media with BSA alone or with BSA FBS for 7.5 days.* Media G1.3/G2.3/BSA G1.3/G2.3/BSA/FBS DMEM/F-12/BSA DMEM/F-12/BSA/FBS Injected oocytes cultured n Oocytes cleaved 29 (59) 29 (59) 32 (64) 32 (64) * All percentages are calculated from the number of injected oocytes cultured. a,b Within columns, values with different superscripts differ significantly (P.5). Oocytes developed to Morula 1 (2) () 1 (2) 2 (4) Blastocyst 1 (2) a 9 (18) b 9 (18) b 1 (2) b fer had numerous presumptive maternal cells within the perivitelline space [1]. Invasion of maternal cells after transfer of micromanipulated embryos to the oviduct has been discussed by Willadsen [28] in regard to sheep, mice, and cattle embryos that had large rents in the zona or had only partial zonae. Presence of maternal cells was associated with embryonic death, and imbedding of embryos in agar was effective in preventing cellular invasion. In the present study, the hole in the zona was small, but maternal cells apparently were still able to pass through to the perivitelline space. Because such cells were likely to have been present by 96 h (as seen in our previous study), yet embryos in the present study were able to develop to morulae and blastocysts, the importance of such maternal cellular infiltration in embryos subjected to ICSI is unclear. In those blastocysts that still had remnants of the zona pellucida, atypical nuclei were visible between the zona and the capsule. Atypical nuclei were not seen in blastocysts having shed the zona, suggesting that the embryo was capable of excluding the presumptive maternal cells during capsule formation. Use of agar embedding or encapsulation with sodium alginate [29] may help to determine if maternal cells interfere with embryo development in oocytes fertilized using a piezo drill. The equine oviduct is known to retain nonfertilized oocytes at the ampullary-isthmic junction, while allowing viable embryos to pass through into the uterus [27]. The 5- day equine embryo, at the late morula stage, produces the signal (PgE) that induces its passage [3]. However, because unfertilized oocytes may sometimes be found in mares when the uterus is flushed for embryo recovery, it has been hypothesized that, when the oviduct opens to allow passage of an embryo, adjacent oocytes may accompany that embryo to the uterus. It was notable in this study that blastocysts descended into the uterus without apparent difficulty even when large numbers of embryos were transferred. While the data appear to indicate that morulae were maintained within the oviduct, 9 of the 1 oviductal morulae in this study were located within one oviduct of one mare. This oviduct was noted to have an atypical morphology (a sharp bend in the ampulla) at the time it was flushed. The high number of morulae in the oviduct was associated with a lower blastocyst recovery (26%) from this mare than was obtained from the other two mares (38 43%). If this was a result of a blockage, it would suggest that equine embryos cannot develop normally into blastocysts within the equine oviduct. In the remaining oviducts, the passage of degenerating oocytes or embryos into the uterus (more were found in the uterus than in the oviducts) supports the hypothesis that adjacent structures may pass into the uterus at the time of descent of a viable embryo. The in vitro-produced equine blastocysts in this study did not expand normally between Day 6 and Day 7, the embryonic capsule did not form, and the zona pellucida did not thin and break off the embryos as it would in vivo. By Day 7.5 in vitro, embryos started to hatch through the hole made in the zona during the ICSI procedure (Fig. 4). Hatching, associated with failure of both zona dissolution and capsule formation, has been noted previously in equine embryos recovered on Day 5 ex vivo and cultured to the blastocyst stage in vitro [31]. Freeman et al. [32] also reported hatching of in vitro-cultured embryos collected ex vivo between Day 5 and Day 7, but noted that almost half of the cultured embryos experienced near-normal thinning of approximately half of the zona, and expanded within this area. These authors did not report whether the thinning was seen more commonly in embryos collected on the later days (that had been longer in the uterus). It may be hypothesized that the uterine environment is necessary not only for normal capsule formation but also for promoting thinning and elasticity of the zona. During in vitro culture, the rigid zona may prevent the embryo from expanding normally; thus, continued embryo growth follows the path of least resistance (the ICSI-induced breach in the zona) or, in intact embryos, causes the zona to rupture open. Once free of the zona, equine blastocysts can expand in vitro up to at least 2.4-mm diameter [31]. We previously found high cleavage rates (69 94%) in equine zygotes cultured in the semidefined medium CZB- C BSA, with either low or high concentrations of glucose [9]. The average nucleus number at 96 h in that medium was also good, at However, when in the present study we extended the culture period to 7.5 days, no embryos developed to the blastocyst stage in CZB-C, in either low- or high-glucose treatments. These results are similar to those of Azuma et al. [33], who failed to produce blastocysts using a completely defined medium. Although semidefined media such as G1.2/G2.2, G1.3/2.3, or SOF have been used to produce equine blastocysts in vitro, the rates have been low (2 9%, [3, 6]), similar to the 2% blastocyst formation seen with G13./2.3 in the present study. A 14% advanced development rate per injected oocyte after culture in modified SOF was reported by Galli et al. [2]; however, this data represented only five embryos and included both blastocysts and morulae. These authors later reported a blastocyst development rate of 6 9% in the same medium [34]. Thus, the 18% blastocyst development achieved with semidefined medium (DMEM/F-12/BSA) in this study represents a notable advance in equine embryo culture. Blastocyst development rates in all treatments of DMEM/F-12 (with BSA or FBS or both; Tables 3 and 4) were higher than previously reported for equine in vitro-produced embryos in the absence of coculture, and were repeatable between experiments 3 and 4. DMEM/F-12 may thus serve as a base medium for detailing the requirements of the equine embryo for development to the blastocyst stage.

7 DEVELOPMENT OF EQUINE ICSI EMBRYOS 1237 We were surprised to find that culture in DMEM/F-12 alone yielded high rates of equine blastocyst development. This medium was originally chosen for use in this study to support growth of oviductal cells when coculture was performed. In contrast with G media, which have been optimized to reflect the requirements of early and late bovine and human embryos [35, 36], DMEM/F-12 is formulated to support proliferation of somatic cells. As such, it has components that have been reported to be detrimental to embryo development in other species, such as high concentrations of glucose, Hepes, and phosphate [37, 38]. The fact that, in the absence of serum, this medium supports higher equine embryo development than does G1.3/2.3 suggests that equine embryos differ from bovine and human embryos in their requirements for early development. This hypothesis is supported by our findings on culture of bovine ICSI oocytes in DMEM/F-12/FBS: none of 39 bovine zygotes developed past 4 days, in contrast with 8/3 (27%) bovine zygotes that developed to morula or blastocyst in G1.3/2.3 (data not shown). While the blastocyst development in vitro in this study was relatively high, it was only half that seen after in vivo transfer (36%). These data indicate that the historically poor development of in vitro-produced equine embryos is likely to lie with the culture system rather than with an inherent developmental defect in IVM equine oocytes. In other species, coculture of embryos with oviductal epithelial cells served to improve development when culture systems were not well established [39, 4], but subsequent development of optimized media has made coculture essentially obsolete [36]. However, in the horse, even when complex culture systems are used, the in vitro production of blastocysts from IVM oocytes has been disappointing [4, 5, 7]. In the current study, we did not see a positive effect of oviductal cell coculture on equine embryonic development over use of DMEM/F-12 alone. ACKNOWLEDGMENT The authors thank Ms. L.B. Love for excellent technical support. REFERENCES 1. Choi YH, Love CC, Love LB, Varner DD, Brinsko S, Hinrichs K. Developmental competence in vivo and in vitro of in vitro-matured equine oocytes fertilized by intracytoplasmic sperm injection with fresh or frozen-thawed spermatozoa. Reproduction 22; 123: Galli C, Crotti G, Turini P, Duchi R, Mari G, Zavaglia G, Duchamp G, Daels P, Lazzari G. Frozen-thawed embryos produced by ovum pickup of immature oocytes and ICSI are capable to establish pregnancies in the horse. Theriogenology 22; 58:75 78(abstract). 3. Choi YH, Love CC, Varner DD, Love LB, Hinrichs K. Effects of gas conditions, time of medium change, and ratio of medium to embryo on in vitro development of horse oocytes fertilized by intracytoplasmic sperm injection. Theriogenology 23; 59: Dell Aquila ME, Cho YS, Minoia P, Traina V, Lacalandra GM, Maritato F. Effects of follicular fluid supplementation of in vitro maturation medium on the fertilization and development of equine oocytes after in vitro fertilization or intracytoplasmic sperm injection. Hum Reprod 1997; 12: Guignot F, Ottogalli M, Yvon JM, Magistrini M. Preliminary observations in in vitro development of equine embryo after ICSI. Reprod Nutr Dev 1998; 38: Maclellan LJ, Sims MM, Squires EL. Effect of invasive adenylate cyclase during oocyte maturation on development of equine embryos following ICSI. Proceedings of the 5th International Symposium on Equine Embryo Transfer, Havemeyer foundation Mono Series 21; 3:35 36(abstract). 7. Li X, Morris LHA, Allen WR. Influence of co-culture during maturation on the developmental potential of equine oocytes fertilized by intracytoplasmic sperm injection (ICSI). Reproduction 21; 121: Choi YH, Chung YG, Walker SC, Westhusin ME, Hinrichs K. In vitro development of equine nuclear transfer embryos: effects of oocyte maturation media and amino acid composition during embryo culture. Zygote 23; 11: Choi YH, Love LB, Varner DD, Hinrichs K. Factors affecting developmental competence of equine oocytes after intracytoplasmic sperm injection. Reproduction 24;(in press). 1. Swain JE, Bormann CL, Krisher RL. Development and viability of in vitro derived porcine blastocysts cultured in NCSU23 and G1.2/G2.2 sequential medium. Theriogenology 21; 56: Krisher RL, Lane M, Bavister BD. Developmental competence and metabolism of bovine embryos cultured in semi-defined and defined culture media. Biol Reprod 1999; 6: Li LY, Meintjes M, Graff KJ, Paul JB, Denniston RS, Godke RA. In vitro fertilization and development of in vitro-matured oocytes aspirated from pregnant mares. Biol Reprod Mono 1995; 1: Choi YH, Chung YG, Seidel GE Jr, Squires EL. Developmental capacity of equine oocytes matured and cultured in equine trophoblastconditioned media. Theriogenology 21; 56: Grondahl C, Hansen TH, Hossaini A, Heinze I, Greve T, Hyttel P. Intracytoplasmic sperm injection of in vitro-matured equine oocytes. Biol Reprod 1997; 57: Hinrichs K, Williams KA. Relationships among oocyte-cumulus morphology, follicular atresia, initial chromatin configuration, and oocyte meiotic competence in the horse. Biol Reprod 1997; 57: Hinrichs K, Schmidt AL. Meiotic competence in horse oocytes: interactions among chromatin configuration, follicle size, cumulus morphology, and season. Biol Reprod 2; 62: Parrish JJ, Susko-Parrish J, Winer MA, First NL. Capacitation of bovine sperm by heparin. Biol Reprod 1988; 38: Kimura Y, Yanagimachi R. Intracytoplasmic sperm injection in the mouse. Biol Reprod 1995; 52: Hinrichs K, Matthews GL, Freeman DA, Torello EM. Oocyte transfer in mares. J Am Vet Med Assoc 1998; 212: Chatot CL, Ziomek CA, Bavister BD, Lewis JL, Torres I. An improved culture medium supports development of random-bred 1-cell mouse embryos in vitro. J Reprod Fertil 1989; 86: Carnevale EM, Ginther OJ. Defective oocytes as a cause of subfertility in old mares. Biol Reprod Mono 1995; 1: Squires EL, Wilson JM, Kato H, Blaszczyk A. A pregnancy after intracytoplasmic sperm injection into equine oocytes matured in vitro. Theriogenology 1996; 45:36(abstract). 23. Cochran R, Meintjes M, Reggio B, Hylan D, Carter J, Pinto C, Paccamonti D, Godke RA. Live foals produced from sperm-injected oocytes derived from pregnant mares. J Equine Vet Sci 1998; 18: Carnevale EM, Maclellan LJ, Coutinho da Silva MA, Scott TJ, Squires EL. Comparison of culture and insemination techniques for equine oocyte transfer. Theriogenology 2; 54: Scott TJ, Carnevale EM, Maclellan LJ, Scoggin CF, Squires EL. Embryo development rates after transfer of oocytes matured in vivo, in vitro, or within oviducts of mares. Theriogenology 21; 55: Hinrichs K, Love CC, Brinsko SP, Choi YH, Varner DD. In vitro fertilization of in vitro-matured equine oocytes: effect of maturation medium, duration of maturation, and sperm calcium ionophore treatment, and comparison with rates of fertilization in vivo after oviductal transfer. Biol Reprod 22; 67: van Niekerk CH, Gerneke WH. Persistence and parthenogentic cleavage of tubal ova in the mare. Onderstepoort J Vet Res 1966; 33: Willadsen SM. Micromanipulation of embryos of the large domestic species. In: Adams CE (ed.), Mammalian Egg Transfer. Boca Raton, FL: CRC Press; 1982: Adaniya GK, Rawlins RG, Quigg JM, Roblero L, Miller IF, Zaneveld LJ. First pregnancies and livebirths from transfer of sodium alginate encapsulated embryos in a rodent model. Fertil Steril 1993; 59: Weber JA, Freeman DA, Vanderwall DK, Woods GL. Prostaglandin E 2 hastens oviductal transport of equine embryos. Biol Reprod 1991; 45: Hinrichs K, Schmidt AL, Memon MA, Selgrath JP, Ebert KM. Culture of 5-day horse embryos in microdroplets for 1 to 2 days. Theriogenology 199; 34: Freeman DA, Butler JE, Weber JA, Geary RT, Woods GL. Co-culture

8 1238 CHOI ET AL. of Day-5 to Day-7 equine embryos in medium with oviductal tissue. Theriogenology 1991; 36: Azuma T, Choi YH, Hochi S, Oguri N. Effect of glucose in the culture medium on development of horse oocytes matured and microfertilized in vitro. Reprod Fertil Dev 1995; 7: Tremoleda JL, Stout TAE, Lagutina I, Lazzari G, Bevers MM, Colenbrander B, Galli C. Effects of in vitro production on horse embryo morphology, cytoskeletal characteristics, and blastocyst capsule formation. Biol Reprod 23; 69: Gardner DK. Mammalian embryo culture in the absence of serum or somatic cell support. Cell Biol Int 1994; 18: Lane M, Gardner DK, Hasler MJ, Hasler JF. Use of G1.2/G2.2 media for commercial bovine embryo culture: equivalent development and pregnancy rates compared to co-culture. Theriogenology 23; 6: Keskintepe L, Brackett BG. In vitro developmental competence of in vitro-matured bovine oocytes fertilized and cultured in completely defined media. Biol Reprod 1996; 55: Ludwig TE, Squirrell JM, Palmenberg AC, Bavister BD. Relationship between development, metabolism, and mitochondrial organization in 2-cell hamster embryos in the presence of low levels of phosphate. Biol Reprod 21; 65: Gandolfi F, Moor RM. Stimulation of early embryonic development in the sheep by co-culture with oviduct epithelial cells. J Reprod Fertil 1987; 81: Thibodeaux JK, Menezo Y, Roussel JD, Hansel W, Goodeaux LL, Thompson DL Jr, Godke RA. Coculture of in vitro fertilized bovine embryos with oviductal epithelial cells originating from different stages of the estrous cycle. J Dairy Sci 1992; 75: