| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Center for Conservation and Research of Endangered Wildlife,3 Cincinnati Zoo and Botanical Garden, Cincinnati, Ohio 45220
Department of Animal Sciences,4 Purdue University, West Lafayette, Indiana 47907
Department of Biological Sciences,5 University of Cincinnati, Cincinnati, Ohio 45221
ABSTRACT
The objective of this study was to define the physiologic needs of domestic cat embryos to facilitate development of a feline-specific culture medium. In a series of factorial experiments, in vivo-matured oocytes (n = 2040) from gonadotropin-treated domestic cats were inseminated in vitro to generate embryos (n = 1464) for culture. In the initial study, concentrations of NaCl (100.0 vs. 120.0 mM), KCl (4.0 vs. 8.0 mM), KH2PO4 (0.25 vs. 1.0 mM), and the ratio of CaCl2 to MgSO4-7H2O (1.0:2.0 mM vs. 2.0:1.0 mM) in the medium were evaluated during Days 1–6 (Day 0: oocyte recovery and in vitro fertilization [IVF]) of culture. Subsequent experiments assessed the effects of varying concentrations of carbohydrate (glucose, 1.5, 3.0, or 6.0 mM; L-lactate, 3.0, 6.0, or 12.0 mM; and pyruvate, 0.1 or 1.0 mM) and essential amino acids (EAAs; 0, 0.5, or 1.0x) in the medium during Days 1–3 and Days 3–6 of culture. Inclusion of vitamins (0 vs. 1.0x) and fetal calf serum (FCS; 0 vs. 5% [v/v]) in the medium also was evaluated during Days 3–6. Development and metabolism of IVF embryos on Day 3 or Day 6 were compared to age-matched in vivo embryos recovered from naturally mated queens. A feline-optimized culture medium (FOCM) was formulated based on these results (100.0 mM NaCl, 8.0 mM KCl, 1.0 mM KH2PO4, 2.0 mM CaCl2, 1.0 mM MgSO4, 1.5 mM glucose, 6.0 mM L-lactate, 0.1 mM pyruvate, and 0x EAAs with 25.0 mM NaHCO3, 1.0 mM alanyl-glutamine, 0.1 mM taurine, and 1.0x nonessential amino acids) with 0.4% (w/v) BSA from Days 0–3 and 5% FCS from Days 3–6. Using this medium, ~70% of cleaved embryos developed into blastocysts with profiles of carbohydrate metabolism similar to in vivo embryos. Our results suggest that feline embryos have stage-specific responses to carbohydrates and are sensitive to EAAs but are still reliant on one or more unidentified components of FCS for optimal blastocyst development.
assisted reproductive technology, early development, embryo, in vitro fertilization
The composition of embryo culture media can affect embryonic morphology, metabolism, and gene expression [1–6]. Following embryo transfer these alterations influence implantation, fetal growth, the incidence of birth defects, gestation length and, ultimately, offspring health [7–11]. Even as little as 6 h of culture in inappropriate conditions can affect embryo transfer success [2]. As a result, embryo physiology in laboratory rodents, domestic livestock, and humans has received considerable attention to formulate culture media that support growth of embryos with the capacity to develop into healthy offspring.
In contrast, very little is known about the physiology of the feline embryo, even though successful in vitro fertilization (IVF) [12] and blastocyst development [13] were first reported in the domestic cat more than 25 years ago. The effects of gas atmosphere [14], protein source [15, 16], culture temperature [14], oviductal co-culture [17], and energy source [18] on embryonic development have been examined, but all of these studies used basal media intended for somatic cells or the embryos of other species. Cat embryos will develop to the blastocyst stage in a wide range of media [19–23], and kittens have been produced following culture in several of these media [24]. However, in vitro development to the blastocyst stage is reduced and delayed relative to in vivo embryos, and implantation and pregnancy rates are low compared with natural mating [25, 26]. Similarly, embryo transfer success is considerably lower in cats than in other species [24, 27–29]. The lack of knowledge concerning feline embryo physiology and reduced viability of cultured embryos limit the usefulness of assisted reproductive technologies for the genetic management of domestic and endangered, nondomestic cat populations.
In other species, tailoring the concentrations of ions, carbohydrates, and amino acids present in the culture medium to the needs of the embryo greatly improves embryo viability [30–32]. Early embryos appear to have a reduced ability to maintain ionic homeostasis, especially in the first few hours after fertilization [3]. As a result, inappropriate ionic conditions can affect gene expression, metabolic activity, and embryonic development in vitro and in vivo [3, 31]. Similarly, concentrations of carbohydrates in the medium influence the metabolic activity of the embryo, which impacts ATP production as well as the intracellular pH and redox state [33, 34]. Concentrations of carbohydrates also vary in different regions of the reproductive tract, suggesting that the carbohydrate requirements of embryos are stage specific [35, 36]. Whereas the amino acids designated as nonessential (NEAAs) for somatic cells appear to be beneficial throughout preimplantation development, essential amino acids (EAAs) exhibit concentration- and stage-specific effects on embryo development [27, 37, 38]. Given the effects these basic medium components have on embryonic physiology, it is not surprising that blastocyst development is improved in other species by culture in relatively simple media containing optimized concentrations of ions, carbohydrates, and amino acids [29, 39, 40].
A variety of approaches to culture medium formulation have been used in other species, each with their own set of advantages and disadvantages [41, 42]. The challenge common to all of these approaches is how to determine which treatment is optimal. Embryo transfer provides the best measure of embryo viability, but the large numbers of treatments involved in studies of culture medium composition make embryo transfers impractical in most species. In addition, embryo transfer success depends on the quality of the uterine environment and the quality of the embryo. Without protocols for consistent synchronization of an appropriate uterine environment, which are not available for the cat, it is difficult to evaluate embryonic viability with embryo transfers. Embryo morphology typically is used as an alternative, but viability can vary greatly among morphologically similar embryos [8]. In contrast, nutrient consumption and developmental kinetic data provide useful indicators of embryo viability when comparable data are available from in vivo-grown embryos [8, 43].
The overall goal of this study was to define the physiologic needs of preimplantation cat embryos using a novel, sequential approach to allow development of a feline-optimized culture medium (FOCM). Our specific objectives were to: 1) develop a basal medium containing optimized ion (NaCl, KCl, KH2PO4, CaCl2, and MgSO4-7H2O) concentrations for all subsequent experiments (experiment 1); 2) establish physiologic norms of embryo development and metabolism for oviductal and uterine-stage embryos grown in vivo (experiment 2); 3) assess the effects of carbohydrates (glucose, L-lactate, and pyruvate) and EAAs on embryo physiology during Days 1– 3 (experiments 3 and 4) and Days 3–6 (experiments 5 and 6) following IVF (Day 0); and 4) evaluate the impact of vitamins [2, 44] and serum during Days 3–6 of culture (experiment 7).
Chemicals and Media Preparation
All chemicals were purchased from Sigma Chemical Co. (St. Louis, MO) unless specified otherwise. Stock solutions were prepared in 18.2 M
water and stored at 4°C for either 1 wk (pyruvate, glutamine, bicarbonate) or 1 mo (basal salt solution, glucose, L-lactate [ICN Biomedicals, Costa Mesa, CA], taurine). Aliquots of an alanyl-glutamine stock solution were stored at –80°C and used within 1 mo of thawing [45]. Stock solutions of gentamicin, amino acids (minimal essential medium concentrations; ICN Biomedicals), vitamins (minimal essential medium concentrations; ICN Biomedicals), and fetal calf serum (FCS; Hyclone, Logan, UT) were stored according to supplier instructions and were used directly for medium preparation. All media contained 4.0 mg/ml BSA (Crystalized, true Cohn BSA 81-001-3; Serologicals Proteins, Inc., Kankakee, IL) unless otherwise specified. Working solutions of all media were prepared for each replicate, filtered (0.22 µm; MillexGV; Millipore, Billerica, MA), and equilibrated in the appropriate gas atmosphere. Cumulus-oocyte complexes or embryos were cultured in 50-µl drops of medium (~8–12 per drop) covered with 10 ml embryo-tested mineral oil in 10 x 60 mm plastic dishes (Falcon 1007; Becton Dickinson Labware, Franklin Lakes, NJ) and were maintained at 38.7°C with 6% CO2.
All animal procedures were approved by the Cincinnati Zoo and Botanical Garden's Institutional Animal Care and Use Committee. Group-housed, anestrual queens (ages 1–8 yr; Liberty Research Inc., Waverly, NY) with basal serum progesterone concentrations (
1 to 2 ng/ml; Target Rapid Feline Progesterone Kit; Biometallics, Princeton, NJ) were treated i.m. with 150 IU eCG (Sioux Biochemical Inc., Sioux Center, IA; or Sigma Chemical Co.) followed 84–86 h later with 100 IU hCG (Sioux Biochemical Inc. or Sigma Chemical Co.). At 24–27 h after hCG, females were anesthetized and subjected to laparoscopic follicle aspiration [26]. Mature follicles (
2 mm) were aspirated with a 22-gauge needle attached to a foot pedal-operated vacuum pump (~1.5 mm Hg) into HEPES-buffered FOCM (FOCMH, Fraction V BSA A-3311 [Sigma Chemical Co.]; Table 1) containing 40 U/ml heparin (Elkins-Sinn Inc., Cherry Hill, NJ). Mature cumulus-oocyte complexes were washed twice in FOCMH without heparin, three times in the appropriate fertilization medium for each experiment (Table 1), placed into 45-µl drops of fertilization medium, and maintained at 38.7°C in 6% CO2 in air until insemination.
|
Semen from one of the two males used for natural breeding was collected for each replicate using an artificial vagina. Males were alternated from replicate to replicate. Recovered semen was diluted 1:4 with FOCMH and centrifuged for 10 min at 600 x g, and the pellet was resuspended in 50–100 µl FOCMH. After determining sperm concentration and motility, aliquots were diluted in fertilization medium (Table 1) to 2 x 106 motile spermatozoa per milliliter. Aliquots (5 µl) then were added to insemination drops containing cumulus-oocyte complexes for a final volume of 50 µl containing 2 x 105 motile spermatozoa per milliliter. Gametes were coincubated for 18–22 h. In experiment 1, IVF was performed in 2.0 ml medium in a 2.0-ml cryovial (Nalge Nunc International, Rochester, NY) containing the same sperm concentration. After re-equilibration of the medium after insemination, vials were sealed and transported (~3 h) in a portable incubator to another laboratory. Vials then were incubated at 38.7°C in 6% CO2 in air without lids throughout the fertilization period.
At 18–22 h after insemination, presumptive zygotes were placed into 100 µl FOCMH containing 80–160 U/ml hylauronidase in a 1.5-ml microcentrifuge tube. Loosely bound spermatozoa and remaining cumulus cells were removed by vortexing for 2–3 min. Denuded zygotes were washed twice in FOCMH and either randomly allocated to treatments (experiments 1, 3, and 4) or placed in FOCM IVC1 (experiments 5–7). All subsequent cultures (Days 1–6) were performed in an atmosphere of 6% CO2, 5% O2, and 89% N2. On Day 3, cleavage was evaluated and embryos were either transferred into fresh medium (experiment 1) and cultured until Day 6, evaluated for metabolic activity and stained for total cell number (experiments 3 and 4), or allocated to treatments and cultured until Day 6 (experiments 5–7). On Day 6, blastocyst development and metabolism (experiments 5–7) were evaluated, and all embryos were stained to determine total cell number (experiments 1 and 5–7) [46, 47]. Only embryos containing 30 cells with a visible blastocoel cavity were considered blastocysts.
In Vivo Embryo Recovery (Experiment 2)
Group-housed female domestic cats (n = 20, ages 1–7 yr) were naturally bred with one of two male domestic cats three times per day for 2 days beginning on the second to fifth day of behavioral estrus [25]. On Day 4 or 7 after the first mating, or approximately 3 or 6 days after presumed ovulation and fertilization [25], mated queens were ovariohysterectomized. Oviducts and uterine horns were separated and flushed with FOCMH to recover cleavage-stage embryos (Day 3) or blastocysts (Day 6). Within 1 h of collection, metabolic activity and total cell number were determined.
Embryos were washed once in FOCMH, twice in 50-µl drops of metabolism medium, and held in a third drop of metabolism medium until assayed (<1 h after removal from culture). The medium used in experiment 3, including 1.5 mM unlabeled glucose, 3.0 mM L-lactate, and 0.1 mM unlabelled pyruvate, was used for all metabolic assessments. Metabolism of glucose via glycolysis and pyruvate oxidation through the tricarboxylic acid cycle was simultaneously measured using 5-3H-glucose (0.0125 mM, 0.25 µCi/µl; Perkin-Elmer NEN Life Sciences Inc., Boston, MA) and 2-14C-pyruvate (0.276 mM, 0.0014 µCi/µl; American Radiolabeled Chemicals Inc., St. Louis, MO). Assays were performed at 38.7°C in 6% CO2, 5%O2, and 89% N2 as described by Herrick et al. [47].
Zygotes (experiments 1, 3, and 4) or embryos (experiments 5–7) from each replicate were divided among as many treatments as possible (incomplete block design) so that each treatment contained an equivalent number of embryos (7–12 per treatment per replicate). All treatments were used once before any were replicated. All treatments were replicated two (experiment 1 only) or three times so that each treatment group had a total of approximately 20–30 embryos.
In experiment 1, the effects of NaCl (100.0 or 120.0 mM), KCl (4.0 or 8.0 mM), and KH2PO4 (0.25 or 1.0 mM) concentrations and two CaCl2:MgSO4-7H2O ratios (1.0:2.0 mM or 2.0:1.0 mM [3, 31]; Table 1) in the medium were evaluated. The osmolarity of the media was not kept constant, so the effects of NaCl on embryonic development cannot be differentiated from osmotic effects (100 mM NaCl, 255.3 ± 1.1 mOsm; 120 mM NaCl, 292.3 ± 1.4 mOsm). Zygotes were allocated to treatments on Day 1, cleavage was evaluated on Day 3 after transfer to fresh medium, and all embryos were cultured until Day 6, when blastocyst development and total cell numbers were evaluated.
In experiment 3, the effects of glucose (1.5, 3.0, or 6.0 mM), L-lactate (3.0, 6.0, or 12.0 mM) [48, 49], and pyruvate (0.1 or 1.0 mM) were evaluated during Days 1–3 in the optimal medium from experiment 1 (Table 1). The final, equilibrated pH of media with different lactate concentrations was maintained at 7.3 ± 0.1 by adding 1.0 M NaOH to the medium before equilibration. In experiment 4, embryos were cultured from Days 1–3 in the optimal medium from experiment 3 containing 0, 0.5, or 1.0x the concentrations of EAAs (Table 1). Because the inhibitory effects of EAAs are attributed to their spontaneous breakdown to NH4, alanyl-glutamine was used instead of glutamine in experiment 3 to reduce NH4 produced from sources other than the EAAs [45, 50, 51]. In experiments 3 and 4, embryo cleavage, total cell numbers, and metabolic activity (picomoles of substrate per embryo or cell per 3 h) were assessed on Day 3 and compared among treatments and to in vivo embryos recovered 3 days after ovulation. The final medium resulting from experiments 1, 3, and 4 (Table 1) was designated FOCM IVF (includes 50 µg/ml gentamicin) and FOCM IVC1 (in vitro culture Days 1–3) and was used throughout the remainder of the study.
For experiment 5, the same carbohydrate treatments used in experiment 3 were tested during Days 3–6 (Table 1). Due to the beneficial effects of culturing with higher-quality embryos [52], only cleaved embryos were selected and randomly allocated to treatments on Day 3. In experiment 6, the effects of 0, 0.5, and 1.0x EAAs were evaluated in the best two treatments from experiment 5 (Table 1). Again, alanyl-glutamine was used in place of glutamine to reduce background concentrations of NH4. In experiment 7, Day 3 embryos were randomly allocated to one of three media: 1) the optimal medium from experiment 6; 2) the optimal medium from experiment 6 with 1x vitamins [2, 44]; or 3) the optimal medium from experiment 6 with 5% (v/v) FCS instead of BSA (Table 1). In experiments 5–7, total cell numbers of all embryos and blastocyst metabolism (picomoles of substrate per embryo or cell per 3 h) were evaluated on Day 6 and compared between treatments and to in vivo embryos recovered 6 days after ovulation. The ratio of the amounts (picomoles) of pyruvate oxidized to glucose metabolized through glycolysis was also calculated for each embryo. This ratio provides an overall measure of metabolic activity that is independent of both cell number and cell volume, providing a better basis of comparison between embryos with greatly different cell counts. The optimal medium at the conclusion of experiment 7 is designated FOCM IVC2 (In Vitro Culture Days 3–6).
All comparisons were made by analysis of variance in the PROC MIXED procedure of the SAS System [53]. For all endpoints, each treatment (e.g., NaCl, glucose, vitamins, etc.) and any interactions of those treatments (e.g., NaCl*KCl*KH2PO4*CaCl2:MgSO4, glucose*lactate, etc.) were considered fixed factors. Replicate (all endpoints) and the interaction between replicate and the treatments (e.g., replicate* NaCl*KCl*KH2PO4*CaCl2:MgSO4; cell number and metabolic activity) were considered random factors. For comparison of in vitro embryos to age-matched embryos grown in vivo, treatment was the only fixed factor, and there were no random factors, since replicates for the in vitro experiments were different from replicates for the in vivo experiments.
In experiment 1, the proportion of cultured zygotes that cleaved, the proportions of cleaved embryos containing at least 30, 50, or 70 cells, and the proportion of cleaved embryos developing to the blastocyst stage on Day 6 were calculated for each treatment in each replicate. In experiments 5, 6, and 7, development to each stage was based on the proportion of embryos with nine or more cells. If the embryos contained fewer than nine cells on Day 6 (1 standard deviation below the mean of Day 3), development most likely stopped prior to application of the treatments. Proportional data were transformed (arcsine of the square root of the proportion) prior to analysis.
Metabolic data (picomoles of substrate per embryo or cell per 3 h or picomoles pyruvate oxidized/picomoles glucose metabolized) were analyzed in a similar manner as developmental data, except that transformation was unnecessary. On Day 3, embryos containing fewer than nine cells were considered to be developing abnormally, so metabolic data from these embryos were excluded. Similarly, only data from in vivo embryos with nine or more cells were used for in vivo versus in vitro comparisons.
Cell count data were analyzed using the generalized mixed model (GLIMMIX) macro in PROC MIXED of SAS. A log link function was used, and the error was designated as having a Poissson distribution [53]. For Day 6 embryos, cell numbers of all embryos containing nine or more cells, including blastocysts, were included in the analysis.
For all analyses, an F-test using the ratio of the largest and smallest variances was used to determine homogeneity of variance [54]. When this test was significant (P < 0.2), a separate variance for each treatment was used in the ANOVA. Fisher least significant difference test was used for pairwise comparisons of treatments when the ANOVA indicated a significant (P 0.05) effect or statistical trend (0.05 < P 0.10) for each factor (e.g., glucose) or interaction of factors (glucose*lactate). To avoid error associated with multiple comparisons, individual treatments were not compared in experiments 3 or 5 when the glucose*lactate*pyruvate effect was significant. For the analyses of in vivo and in vitro-produced embryos, only comparisons between in vivo embryos and individual treatments were used. Finally, in experiments 4–7, some treatments were pooled when the initial analysis revealed that they were not significantly different. Probability values 0.05 were considered significant, and values 0.05 < P 0.1 were considered to indicate a statistical trend. All values are reported as mean ± SEM.
A total of 348 zygotes were cultured in the 16 treatments, resulting in 174 embryos (two or more cells). Cleavage rates were higher (P = 0.03) with 100 mM than with 120 mM NaCl (Table 2). In addition, there was a trend (P = 0.08) toward higher cleavage rates when zygotes were cultured in 1.0 mM instead of 0.25 mM KH2PO4 (Table 2). Total cell numbers on Day 6 and the proportion of embryos with 30 cells were not affected (P > 0.10) by treatments (Table 2). However, the proportion of embryos with >50 cells on Day 6 tended (P = 0.10) to be higher in 8.0 mM than in 4.0 mM KCl (Table 2). The proportions of embryos with 70 cells also was higher (P = 0.01) when zygotes were cultured in 100 mM NaCl and 1.0 mM KH2PO4 (Table 2). In addition, there was a trend (P = 0.06) for the NaCl*KCl*KH2PO4*CaCl2:MgSO4 interaction to affect the proportion of embryos with 70 cells, and 100.0 mM NaCl, 8.0 mM KCl, 1.0 mM KH2PO4, and 2.0:1.0 mM CaCl2:MgSO4 was numerically (11.3%) the best treatment (Table 2; Supplemental Table 1 available online at www.biolreprod.org). The proportion of blastocysts on Day 6 was low (3.4% overall) and was not affected (P > 0.10) by treatment (Supplemental Table 1). No embryos in any treatment contained 100 cells on Day 6.
|
Experiment 2: In Vivo Embryo Recovery on Day 3 or 6 Postovulation
A total of 14 (70%) of 20 mated females ovulated using the described mating regimen (Table 3). Ovulating females had 4.6 ± 0.2 (range: 3–6) corpora lutea at the time of ovariohysterectomy (Table 3). Overall embryo recovery rate (based on the number of corpora lutea) was 83.6% ± 7.4%, but this was influenced by embryo quality. Six ovulating females (42.9%) produced only degenerate embryos, and seven ovulating females (50%) produced only good-quality embryos (uniformly pigmented, symmetrical blastomeres). The recovery rate for females producing at least one good-quality embryo was 95.0% ± 3.3% (Table 3). On Day 3, embryos (n = 18 from four females) contained 15.5 ± 2.0 cells, with a range of 2–32 cells per embryo (Table 3). Among the embryos from a single female, the average difference between the largest and smallest embryo was 15.1 ± 1.4 cells. Day 3 embryos metabolized 0.75 ± 0.09 pmol glucose and 0.37 ± 0.11 pmol pyurvate per embryo per 3 h or 0.049 ± 0.008 pmol glucose and 0.025 ± 0.007 pmol pyruvate per cell per 3 h. On Day 6, all recovered embryos (n = 17 from four females) were at the blastocyst stage and contained 330.0 ± 31.3 cells (range: 102–576 cells per embryo; Table 3). From a single female, cell numbers between the largest and smallest embryos differed by 167.5 ± 24.3 cells. In vivo blastocysts metabolized 9.46 ± 1.21 pmol glucose and 1.80 ± 0.24 pmol pyruvate per embryo per 3 h or 0.029 ± 0.004 pmol glucose and 0.006 ± 0.001 pmol pyruvate per cell per 3 h.
|
Experiment 3: Carbohydrates (Days 1–3)
From 547 zygotes placed into culture, a total of 403 embryos (2 or more cells) developed in the 18 treatment groups. Cleavage rate was not affected (P > 0.10) by treatment (data not shown). The mean final cell number of all embryos on Day 3 (18.8 ± 0.5 overall) was affected (P = 0.02) by an interaction between glucose, lactate, and pyruvate, and there was a trend (P = 0.07) toward an interaction between lactate and pyruvate. Cell numbers were not different (P > 0.10) from in vivo embryos (15.5 ± 2.0 cells) in 12 treatments (Supplemental Table 2, available online at www.biolreprod.org). In the remaining six treatments, cell numbers were greater (P 0.05, five treatments; 0.05 < P 0.1, one treatment) than those of in vivo embryos (Supplemental Table 2).
Of the cultured embryos, 333 (82.6%) had nine or more cells on Day 3, and 326 of those were used for metabolic comparisons (Supplemental Table 2). Glycolytic activity per embryo was affected (P = 0.04) by lactate, and rates were higher after culture in 3.0 or 6.0 mM than they were in 12.0 mM. The interaction between glucose and lactate also affected (P = 0.02) glycolytic rate per embryo, with 3.0 mM glucose and 3.0 mM lactate and 6.0 mM glucose and 6.0 mM lactate producing the highest rates. The glucose, lactate, and pyruvate interaction tended (P = 0.09) to affect glycolytic activity expressed per cell. None of the other tested factors affected (P > 0.10) glycolytic activity per cell. Similarly, pyruvate metabolism per embryo and per cell were not affected (P > 0.10) by any of the tested factors. When compared to in vivo-grown embryos, glycolytic activity per embryo was not different (P > 0.10) in 13 treatments (Supplemental Table 2). In 16 treatments, glycolytic activity per cell also was not different (P > 0.10) from in vivo embryos (Supplemental Table 2). In contrast, only six and eight treatments produced embryos with rates of pyruvate activity per embryo or per cell, respectively, that were not different (P > 0.10) from in vivo embryos (Supplemental Table 2).
Overall, only six treatments did not differ (P > 0.10) from in vivo embryos for all four metabolic endpoints (Supplemental Table 2). Of these six, only two (1.5 mM glucose, 6.0 mM lactate, and 0.1 mM pyruvate; and 1.5 mM glucose, 12.0 mM lactate, and 1.0 mM pyruvate) also contained more (P 0.05) cells than in vivo embryos (Supplemental Table 2). Since cell numbers in vivo were affected by different times of ovulation relative to embryo recovery, in vitro treatments resulting in embryos with higher cell numbers were considered optimal and not indicative of abnormally accelerated growth rates. Metabolic activity of embryos cultured in 1.5 mM glucose, 6.0 mM lactate, and 0.1 mM pyruvate were numerically closer to in vivo embryos, so this combination was used in all subsequent experiments for IVF and culture from Days 1–3 (Supplemental Table 2).
Experiment 4: Essential Amino Acids (Days 1–3)
From 185 zygotes, 140 embryos (2 or more cells) were cultured in this experiment. There was no difference (P > 0.10) between 0.5 and 1.0x EAAs for any of the endpoints (data not shown), so these treatments were pooled for comparison to embryos cultured in the absence of EAAs and to embryos grown in vivo (Table 4). The addition of EAAs decreased (P 0.05) the proportion of embryos that cleaved (69.6% ± 5.6%) compared with embryos cultured without EAAs (89.0% ± 3.6%; Table 4). However, the total cell number of embryos on Day 3 was not affected (P > 0.10) by the presence of EAAs, and both in vitro-produced treatment groups (with and without EAAs) were not different (P > 0.10) from in vivo-grown embryos (Table 4).
|
For metabolic analysis, 116 embryos (82.9%) containing nine or more cells were used. The presence of EAAs did not affect (P > 0.10) the rates of glycolysis or pyruvate oxidation on a per embryo or per cell basis (Table 4). Although glycolytic activity per embryo was higher (P 0.05) for in vitro embryos than in vivo-grown embryos, glycolytic activity per cell was not different (P > 0.10; Table 4). Pyruvate oxidations per embryo and per cell were higher (P 0.05) in the in vitro-produced embryos than in those produced in vivo (Table 4). Due to the inhibitory effects of EAAs on cleavage rates, IVF and culture from Days 1–3 were performed without EAAs in subsequent experiments.
Experiment 5: Carbohydrates (Days 3–6)
A total of 624 zygotes were cultured from Days 1–3, resulting in 478 embryos. Of these embryos, 450 (94.1%) contained 9 or more cells and were used for analysis. Overall, embryos contained 53.6 ± 1.1 cells on Day 6, with 85.1% having 30 cells, 54.9% having 50 cells, 22.0% having 70 cells, and 4.2% having 100 cells (Supplemental Table 3, available online at www.biolreprod.org). There was a trend (P = 0.10) toward higher cell numbers following culture in 0.1 mM pyruvate than in 1.0 mM, but no other factors affected final cell number. The proportion of embryos with 70 cells also was higher (P = 0.03) after culture in 0.1 mM pyruvate than in 1.0 mM. In addition, more (P = 0.01) embryos contained 100 cells after culture in 1.5 mM glucose and 6.0 mM lactate than embryos cultured in all other glucose and lactate combinations except 3.0 mM glucose and 3.0 mM lactate (P > 0.10) or 6.0 mM glucose and 3.0 mM lactate (P = 0.06). Overall, 24.9% of embryos developed to the blastocyst stage, and these blastocysts contained 68.1 ± 2.0 cells (Supplemental Table 3). The proportion of embryos reaching the blastocyst stage was not affected (P > 0.05) by any of the carbohydrates, but blastocyst cell number tended (P = 0.10) to be higher after culture with 0.1 mM pyruvate than it did with 1.0 mM. Blastocyst cell number also was affected (P = 0.01) by an interaction between glucose and pyruvate, with 1.5 mM glucose and 0.1 mM pyruvate being the best treatment numerically.
Although the proportion of in vitro embryos containing 30 cells (85.1%) on Day 6 was not different (P > 0.10) from in vivo-grown embryos (100.0%), the proportions of in vitro embryos with 50 cells (54.9%; 0.05 < P 0.10), 70 cells (22.0%), 100 cells (4.2%), or at the blastocyst stage (24.9%) were all reduced (P 0.05) relative to in vivo-grown embryos (100.0%, 100.0%, 100.0%, and 100.0%, respectively; Supplemental Table 3). Blastocyst cell numbers (68.1 ± 2.0) also were lower (P 0.05) in all treatments compared with in vivo-derived blastocysts (330.0 ± 31.3; Supplemental Table 3). Numerically, the most blastocysts (43.7% ± 15.4%) were produced in 3.0 mM glucose, 3.0 mM lactate, and 0.1 mM pyruvate, but blastocyst cell numbers were highest (114.3 ± 4.8) after culture in 1.5 mM glucose, 6.0 mM lactate, and 0.1 mM pyruvate (Supplemental Table 3).
A total of 80 blastocysts were used for metabolic analysis (Supplemental Table 4, available online at www.biolreprod.org). Glucose tended to affect glycolytic activity on both a per embryo (P = 0.07) and a per cell (P = 0.06) basis. In both cases, glycolysis was more active after culture in 1.5 mM glucose than in 6.0 mM glucose. Lactate also affected rates of glycolysis on both a per embryo (P = 0.03) and per cell (P = 0.08) basis, with higher concentrations of lactate stimulating glycolytic activity. Glycolysis per embryo also was affected (P = 0.04) by an interaction between glucose, lactate, and pyruvate. In addition, pyruvate affected (P = 0.04) glycolytic activity per cell, with more glucose metabolized following exposure to 1.0 mM pyruvate compared with 0.1 mM. The rate of pyruvate oxidation per cell tended (P = 0.06) to be affected by an interaction between glucose and lactate, with 3.0 mM glucose and 6.0 mM lactate producing higher metabolic rates than all other treatments except 6.0 mM glucose and 12.0 mM lactate. Pyruvate oxidation per embryo and the ratio of the amount of pyruvate oxidized to the amount of glucose metabolized through glycolysis were not altered by the carbohydrate composition of the medium.
None of the treatments resulted in blastocysts with glycolytic activity per embryo similar to blastocysts grown in vivo (Supplemental Table 4). Similarly, only two treatments resulted in levels of pyruvate oxidation per embryo that were not different (P > 0.10) from in vivo blastocysts (Supplemental Table 4). However, 10 and 16 treatments produced blastocysts with rates of glycolysis and pyruvate oxidation per cell, respectively, that were not different (P > 0.10) from in vivo blastocysts (Supplemental Table 4). All treatments resulted in blastocysts with ratios of pyruvate oxidized to glucose metabolized through glycolysis (overall: 0.19 ± 0.03, n = 80) that were not different (P > 0.10) from in vivo embryos (0.21 ± 0.02, n = 17; Supplemental Table 4). In addition, when the ratios of oxidized pyruvate to glucose metabolized through glycolysis were pooled across in vitro treatments (no significant effect of any individual carbohydrate or interaction between carbohydrates on this parameter; Supplemental Table 4), there was no difference (P > 0.10) between in vitro- and in vivo-grown embryos. There were 10 treatments whose metabolic activity per cell and the ratio of pyruvate to glucose metabolized were all similar (P > 0.10) to in vivo blastocysts (Supplemental Table 4). One of these 10 treatments resulted in the most blastocysts (3.0 mM glucose, 3.0 mM lactate, and 0.1 mM pyruvate), and another resulted in blastocysts with the most cells (1.5 mM glucose, 6.0 mM lactate, and 0.1 mM pyruvate; Supplemental Table 4). Since both blastocyst quantity (frequency) and quality (cell number and metabolism) are important, both of these treatments were used in experiment 3b.
Experiment 6: Essential Amino Acids (Days 3–6)
From 226 cultured zygotes, 180 embryos were allocated to the six treatments, resulting in 161 embryos (89.4%) with 9 or more cells on Day 6 for comparisons. Although carbohydrate treatments did not affect (P > 0.10) any of the developmental endpoints, there was a significant (P = 0.004) interaction between EAA concentration and carbohydrate combination in the proportion of embryos with 50 cells (data not shown). Therefore, data could not be pooled to examine the effects of culturing embryos with (0.5x and 1.0x EAAs combined) and without EAAs. Both 0.5x and 1.0x EAAs significantly (P 0.05) inhibited final cell number and the proportions of embryos with 30 cells, 50 cells, or 70 cells (Table 5). The proportion of embryos containing 100 cells, the proportion of embryos at the blastocyst stage, and the number of cells per blastocyst were not affected (P > 0.10) by the presence of EAAs (Table 5). All treatments were significantly different (P 0.05) from in vivo embryos for all endpoints, except (P = 0.052) the proportion of embryos (88.1% ± 8.6%) with 30 cells after culture in 1.5 mM glucose, 6.0 mM lactate, and 0.1 mM pyruvate and 0x EAAs (data not shown).
|
Unlike developmental data, blastocyst metabolism data could be pooled to examine the effects of culturing embryos with (0.5x and 1.0x EAAs; data not shown) and without EAAs (Table 6). The EAAs did not affect (P > 0.10) any of the metabolic endpoints evaluated, but there was trend (P = 0.06) toward a difference between pyruvate oxidation per cell (0.009 ± 0.003 without EAAs vs. 0.002 ± 0.001 with EAAs; Table 6). When compared to in vivo blastocysts, in vitro blastocysts from both groups had different (P 0.05) rates of metabolism per embryo (Table 6). In contrast, glycolytic activity per cell, pyruvate oxidation per cell, and the ratio of pyruvate to glucose metabolized were not different (P > 0.10) between in vivo blastocysts and blastocysts cultured in the absence of EAAs (Table 6). Culturing embryos with EAAs altered (P 0.05) both glycolysis and pyruvate oxidation per cell, as well as the ratio of pyruvate to glucose metabolized, in resulting blastocysts relative to in vivo blastocysts (Table 6). For subsequent experiments, no EAAs were used in FOCM during Days 3–6. Although there were no significant differences between the two tested carbohydrate treatments in either experiments 5 or 6, 1.5 mM glucose, 6.0 mM lactate, and 0.1 mM pyruvate was selected based on blastocyst cell numbers in experiment 5.
|
Experiment 7: Vitamins and FCS (Days 3–6)
In the final experiment, 110 zygotes resulted in 89 embryos, and 86 (96.6%) of these had nine or more cells and were used for analysis. Supplementation of the medium with vitamins for Days 3–6 did not affect (P > 0.10) any developmental or metabolic endpoint (Table 7). Therefore, data from embryos cultured in the presence of BSA (with or without vitamins; data not shown) were pooled for comparison to embryos cultured in medium with FCS instead of BSA. Final cell numbers on Day 6 were increased (P 0.05) following culture with FCS (106.6 ± 9.0) compared with embryos cultured with BSA (54.5 ± 3.3; Table 7). Although the proportion of embryos with 30 cells was not different (P > 0.10) between groups, the proportions of embryos with 50 cells, 70 cells, 100 cells, and at the blastocyst stage were all higher (P 0.05) after culture with FCS (Table 7; Fig. 1). Similarly, blastocyst cell numbers tended to be higher (P = 0.053) after culture with FCS (121.0 ± 8.8) than culture with BSA (73.3 ± 11.2; Table 7; Fig. 1). Despite the improved development when serum was added to the culture medium, both in vitro treatments differed (P 0.05) from in vivo-grown embryos for all developmental endpoints (Table 7).
|
|
There were no differences (P > 0.10) between the metabolism of blastocysts cultured in BSA and those cultured in FCS for any metabolic endpoint (Table 8). Metabolic activity per embryo of in vitro-produced blastocysts was different (P 0.05) from in vivo blastocysts for both glycolysis and pyruvate oxidation (Table 8). However, when the metabolic activities of each pathway were expressed per cell, there were no differences (P > 0.05) between in vitro- and in vivo-grown blastocysts (Table 8). In addition, the ratio of pyruvate to glucose metabolized also was similar (P > 0.10) between in vitro and in vivo blastocysts (Table 8).
|
The primary objective of this series of experiments was to critically investigate the physiology of domestic cat embryos in order to develop an embryo culture medium that would better support the viability of feline embryos during culture. Using a systematic, multifactorial approach to assess the effects of varying concentrations of several medium components on embryo development and metabolism, we have shown that in vivo- and in vitro-generated feline embryos exhibit unique physiologic properties relative to embryos from other mammalian species. Our findings also facilitated formulation of a culture medium that supports development of 70% of cultured embryos to the blastocyst stage within a normal, or "in vivo-like," timeframe. In addition, these embryos exhibited similar profiles of carbohydrate metabolism to that of in vivo embryos on both Day 3 and Day 6 of culture. Although the ability of these embryos to produce healthy offspring was not evaluated in this study, improved development in vitro and an enhanced understanding of the physiology of feline embryos provides important information for further development of in vitro fertilization and embryo transfer technologies for felids.
Since ovarian tissue from domestic cats is readily available from most veterinary clinics following the surgical sterilization of pets, most studies of feline embryo culture have used in vitro-matured oocytes. However, the developmental competence of in vivo-matured oocytes is better than oocytes matured in vitro, and in vivo-matured oocytes can be repeatedly collected from the same female rather than a single oocyte harvest after ovariohysterectomy or postmortem [22]. Therefore, the use of in vivo-matured oocytes is more applicable to the genetic management of domestic and nondomestic cat populations, and in vivo-matured oocytes were used throughout this study. In vitro development following IVF of feline oocytes matured in vivo differs greatly between laboratories using a variety of culture media [19, 20, 26]. No blastocysts were produced following culture for up to 10 days in Ham F-10 medium (contains 126.7 mM NaCl, 3.8 mM KCl, 6.1 mM glucose, 0.0 mM lactate, 1.1 mM pyruvate, and 2.0 mM glutamine) supplemented with 5% FCS [26]. However, if the volume of the culture drops is reduced (~10 embryos in 20 vs. 100 µl) and embryos are moved to fresh medium at 48-h intervals, Ham F-10 medium with 5% FCS can support up to 50% of embryos derived from high-quality in vitro-matured oocytes in development to the blastocyst stage [52]. This discrepancy suggests embryo-derived growth factors and the concentration of ammonium in the medium are important determinants of embryonic development in this medium [40, 52]. A sequential culture system using Tyrode salt solution as the basal medium (contains 137.0 mM NaCl, 2.7 mM KCl, 5.6 mM glucose, 2.2 mM lactate, 0.36 pyruvate, and 1.0 mM glutamine) also has been used to culture feline embryos [20]. Supplementation of this medium with 10% FCS from Days 4–7 after IVF allows ~50% of embryos to develop into blastocysts with ~75–450 cells on Day 7 [20]. Blastocyst development in either Ham F-10 or the Tyrode-based medium appears to be reduced and delayed (~50% of cleaved embryos on Day 7 after IVF) relative to blastocyst development achieved in FOCM (~70% of cleaved embryos on Day 6 after IVF). To our knowledge there has only been one other report of feline blastocyst development within 6 days of IVF using in vivo-matured oocytes [19]. Using a modified version of Earle balanced salt solution, designated MK-1 (contains 116.4 mM NaCl, 5.4 mM KCl, 1.0 mM glucose, 3.6 mM lactate, 0.36 pyruvate, and 2.0 mM [assumed from MEM, but not stated] glutamine) supplemented with 10% human serum, Kanda et al. [19] reported 71.7% of cleaved embryos developing into blastocysts with ~170 cells on Day 6 after IVF. However, Kanda et al. [19] included human serum throughout the culture period, and cumulus cells remained attached to the embryos until at least the morula stage. The presence of cumulus cells at this time is abnormal, since embryos recovered at the 16-cell stage following in vivo development did not have any cumulus cells attached (data not shown). In addition, cumulus cells exhibit high metabolic activity [47], which can alter the concentrations of nutrients available to the embryo. Although Kanda et al. [19] demonstrated excellent viability of these embryos following embryo transfer, it remains unclear how embryonic development and viability were affected by cumulus cell coculture.
The ionic composition of culture media has dramatic effects on early embryos, including altered gene expression, reduced ability to maintain homeostasis, abnormal metabolic activity, and impaired embryonic development in vitro and in vivo [31, 55–57]. As in other species [58–61], development of feline embryos was improved in a reduced concentration of NaCl. Although the inhibitory effects of elevated NaCl can often be alleviated by including glutamine, taurine, and the NEAAs in the medium, these amino acids failed to protect feline embryos [58, 61, 62]. Possibly the feline embryo has a reduced ability to use these amino acids as osmolytes and/or an increased sensitivity to NaCl or medium osmolarity. Another interesting finding was the beneficial effects of an elevated concentration of phosphate. In some species, phosphate is omitted from the medium used for early cleavage-stage embryos to prevent inhibitory effects on embryo metabolism [31, 63]. Hamster embryos are especially sensitive to phosphate, with concentrations as low as 2.5 µM affecting mitochondrial organization, metabolism, and embryonic development [57]. Early hamster embryos also have a reduced ability to maintain intracellular Ca2+ concentrations during culture, making them particularly sensitive to the ratio of Ca2+ to Mg2+ in the medium [3, 31]. Feline embryo development was not affected by the two Ca2+:Mg2+ ratios tested, again illustrating the difference between feline embryos and those of other species. The results of this study indicate that the feline embryo manages ionic homeostasis differently than embryos from other species, especially rodent embryos. Determination of the specific mechanisms controlling regulation of intracellular osmolarity and the role of extracellular phosphate await further research.
Although our primary objective for collecting in vivo-grown embryos was to establish normative values for cell numbers and metabolic activity, this experiment also provided valuable information on feline embryonic development in vivo. In the first 3 days of development after fertilization, feline embryos underwent approximately four cell divisions resulting in ~16 cells, similar to what has been reported based on morphology [25]. Feline embryos underwent another four to five cell divisions from Days 3–6, indicating a slight acceleration in growth rate. In contrast, previous reports based on gross morphology suggested that growth rates after the five- to eight-cell stage slowed to one division per day [25, 26]. Our data also revealed a large range in development among embryos of the same litter. On Day 3, cell number per embryo varied by ~15 cells within individual females. On Day 6, this range increased to ~168 cells. Swanson et al. [25] reported that ~30% of ovulated oocytes fail to implant. In pigs, early embryonic loss is correlated with the relative time of ovulation and fertilization for each oocyte within the litter, resulting in two cohorts of embryos at different stages of growth with variable developmental potentials [64]. A similar phenomenon may be occurring in cats, but inherent differences in embryo quality cannot be ruled out.
In vivo-derived embryos also were subjected to metabolic assessments, representing the first report of metabolism by carnivore embryos. Based on the amount of pyruvate oxidized (14CO2 from 2-14C-pyruvate) relative to glucose metabolized through glycolysis (3H2O from 5-3H-glucose; picomoles pyruvate:picomoles glucose), feline embryos metabolized approximately twice as much glucose as pyruvate on Day 3 and approximately five times as much glucose on Day 6. The increased utilization of glucose through glycolysis during compaction and blastocyst formation is common among species [30, 32]. However, the ratio of pyruvate oxidation and glycolytic activity appears to be both stage specific and species specific. In 16-cell cattle embryos, the amounts of glucose and pyruvate metabolized through these pathways are approximately equal [65]. By the blastocyst stage, glycolytic activity is two to three times greater than pyruvate oxidation [65–67] in bovine embryos, suggesting an increased reliance on pyruvate and/or a decreased reliance on glucose at both stages relative to feline embryos. In contrast, porcine embryos appear to depend even more heavily on glucose, metabolizing approximately 10 to 18 times more glucose than pyruvate at the eight-cell and morula stages, and approximately 20 to 27 times more glucose at the blastocyst stage [68]. Although comparisons between species evaluated in different studies must be made cautiously, these data suggest that carbohydrate metabolism in feline embryos differs from that in bovine and porcine embryos, reinforcing the need to customize culture media for each species.
During the first 3 days following fertilization, the combination of 1.5 mM glucose, 6.0 mM lactate, and 0.1 mM pyruvate produced embryos with higher cell numbers than in vivo embryos and metabolic profiles that were not statistically different from in vivo embryos. In contrast, late-stage feline embryos exhibited more sophisticated culture requirements. None of the tested carbohydrate combinations supported in vitro development beyond the 30-cell stage at rates equivalent to that of in vivo embryos. However, when metabolic activity was expressed as a ratio of the amount of pyruvate oxidized to glucose metabolized through glycolysis, in vitro embryos (0.19 pmol pyruvate per pmol glucose) were nearly identical to in vivo embryos (0.21 pmol pyruvate per pmol glucose). Together, these results suggest that carbohydrate composition of the culture medium, at least within the tested range, is not a major determining factor of feline embryo development, and carbohydrate metabolism does not seem to be compromised in cultured cat embryos. Similarly, embryos from rabbits, pigs, and sheep can develop to the blastocyst stage in the absence of carbohydrates [69–71]. Possibly, amino acids or lipids are more critical for blastocyst development in cats, and monitoring activity of these metabolic pathways would provide a more accurate indicator of blastocyst viability.
It is well established that in vitro- and in vivo-derived embryos have different metabolic profiles [2, 68, 72, 73], but there has been relatively little research [34, 49, 71, 73, 74] exploring the effect of media composition on metabolic activity. In this study, the use of a factorial design allowed us to assess the impact of carbohydrate concentrations on feline embryo metabolism. Increasing the glucose concentration lowered the glycolytic activity of the feline blastocyst. In contrast, increasing extracellular glucose increases glycolytic activity in both cleavage- and blastocyst-stage mouse embryos [30, 33]. While it is certainly possible that species-specific responses do occur, differences in experimental conditions (e.g., the presence of NEAAs and glutamine) could explain the different responses of cat and mouse embryos to glucose [30]. Glycolysis in feline embryos also was affected by lactate concentrations at both early and late developmental stages. Elevated lactate was inhibitory to glycolysis in early cleavage-stage embryos, but stimulatory to glycolysis in blastocysts. The responses of murine embryos to extracellular lactate also are dose specific and stage specific. In mouse zygotes, increasing lactate concentrations from 0 to 10 mM increases glycolytic activity, whereas concentrations >10 mM decrease glycolysis [49]. However, lactate concentrations between 0 and 30 mM do not affect glycolytic activity in mouse morulae [49]. In the mouse embryo, the ability to maintain intracellular pH, lactate dehydrogenase activity, the change in redox state in response to lactate, and lactate metabolism also change during development from the zygote to the blastocyst stage [34, 49]. Similar mechanisms may operate in feline embryos in response to extracellular lactate. These results help to illustrate the complicated, species-specific nature of metabolic control that occurs within the developing embryo.
In several species, EAAs are inhibitory to the development of early embryos, although responses do vary in different media [37–39]. In contrast, EAAs are stimulatory to blastocyst cell number, inner cell mass formation, and hatching when present at later developmental stages [27, 37]. Many of the EAAs' inhibitory effects have been attributed to NH4 produced when these amino acids are metabolized by the embryo or spontaneously break down at culture temperatures [27, 51, 75]. By reducing the EAA concentration and only including them in media for later-stage embryos, many of these inhibitory effects can be avoided [27, 37, 38]. In this study, EAAs were found to be inhibitory to early-stage feline embryos, as in other species. However, the EAAs also proved to be inhibitory to later stage feline embryos, even at reduced (i.e., 0.5x MEM) concentrations.
Species-specific responses to EAAs in culture could result from differences in systemic or embryonic physiology. In mice, increasing the proportion of dietary protein leads to elevated NH4 concentrations in reproductive tract fluids, and to altered gene expression and reduced viability among embryos grown in vivo [1]. Increasing circulating urea and/or ammonia/ammonium concentrations in ruminants also increases concentrations of these substances in reproductive tract fluids, affecting oocyte competence, embryo metabolism, and developmental kinetics [76–79]. Because rodents and ruminants typically do not ingest high-protein diets, the sensitivity of oocytes and embryos from these species to EAAs and NH4 in vitro is not surprising. In contrast, the cat is an obligate carnivore whose diet always consists of a very high proportion of protein [80, 81]. However, unlike other species that recycle urea through the gut to conserve nitrogen, cats constitutively excrete a large amount of nitrogen; having little need to conserve a common nutrient in their diet and an increased need to avoid hyperammonemia [81–84]. Due to the cat's efficient ability to clear nitrogenous waste from circulation, feline embryos may lack adaptive mechanisms to counter elevated NH4 concentrations. Analysis of feline oviductal and uterine fluids would be useful in determining the extent of NH4 exposure experienced by feline embryos in vivo.
Alternatively, the sensitivity of feline embryos to EAAs may reflect differences in embryonic physiology. In mice, the amount of metabolic NH4 in the medium remains relatively stable (~20–30 µM) regardless of the EAA concentration, suggesting that spontaneously produced NH4 is responsible for the inhibitory effects of elevated EAAs [27]. Since basic medium composition and culture conditions in the present study are similar to those used in other studies, cat embryos must be more sensitive to spontaneously produced NH4 or metabolize more EAAs than murine embryos [27, 28]. Feline embryos also could be better able to utilize extracellular protein to meet their amino acid needs, resulting in elevated metabolic NH4 even in the absence of exogenous EAAs. If the inhibitory effects of EAAs on feline embryos are indeed due to elevated amino acid metabolism, embryos grown in vivo must be exposed to and/or metabolize less EAA or be better able to cope with the resulting NH4. Studies concerning the control of amino acid and protein metabolism in feline embryos will be necessary to answer these questions.
The inclusion of FCS in the culture medium from Days 3 to 6 greatly improved embryonic development, allowing ~70% of cleaved embryos to reach the blastocyst stage in the same timeframe (i.e., by Day 6) as in vivo blastocyst development. Beneficial effects of serum on embryo development have been reported in other species [85, 86]. However, when ions, carbohydrates, and amino acid concentrations are optimized, high rates of blastocyst development can be achieved in rodents [39, 51, 57], ruminants [28, 40, 67, 86], and primates [29] without serum supplementation. Serum can be detrimental to embryonic ultrastructure [4], gene expression [10, 87], metabolism [40], and morphology [7, 40, 88], and can alter fetal growth rate and gestation length [7]. Serum also contains a wide variety of substances, including energy substrates, growth factors, cytokines, and hormones, that can vary from batch to batch [89]. Therefore, although FCS improved embryonic development, it should not be considered a final solution for feline embryo culture. Even with serum supplementation, blastocyst cell numbers in the present study were only one third of those for in vivo-produced embryos. Studies investigating the roles of growth factors, lipids, and other substances found in FCS would be useful in moving toward a defined medium that supports embryonic viability as effectively as, and preferably better than, FCS-supplemented media.
The beneficial effects of FCS on embryonic development also are interesting relative to our findings concerning EAAs. Even though FCS contains low concentrations of EAAs [89, 90], development was still greatly stimulated. The use of lower EAA concentrations would reduce direct negative effects of one or more inhibitory amino acids, as well as indirect negative effects caused by elevated concentrations of NH4, on embryonic physiology. Alternatively, serum contains a variety of substances that may allow the embryo to cope better with NH4. In fact, feline blastocysts are routinely produced in media containing EAAs, but most of these media also contain serum [19–22]. Bovine embryos cope with NH4 in the medium by using it to convert pyruvate to alanine, which is secreted into the medium without affecting embryonic viability [91]. The pyruvate concentration used for feline embryos in FOCM was determined in the absence of EAA, but the optimal concentration of pyruvate may be dependent on the presence of EAAs. In support of this, the only study of cat embryos in which a relatively high (39.0%) proportion of embryos reached the blastocyst stage in the presence of EAAs and the absence of serum used a culture medium with three times more pyruvate (0.36 mM) than FOCM (0.1 mM) [92]. Although bovine embryos do not convert excess NH4 to urea in vitro [91], this pathway could be active in feline embryos. Cats have a reduced capacity to synthesize citrulline and ornithine, making them dependent on adequate dietary arginine or supplementation of these intermediates for proper urea cycle function and prevention of hyperammonemia [80, 81]. If the urea cycle is active in feline embryos, the presence of citrulline and ornithine in serum may explain the beneficial effects of FCS supplementation on development [89, 90].
In summary, this study represents the first attempt to intensively assess the physiologic requirements of feline embryos and tailor culture conditions to meet these specific needs. Our findings suggest that feline embryos exhibit unique developmental and metabolic requirements compared with the embryos of other species. Using this new information, a culture medium was formulated that produces embryos on Day 3 with cell numbers and metabolic profiles that are similar to age-matched in vivo-grown embryos. If serum is included in this medium after Day 3, a high proportion (~70%) of these embryos can develop into blastocysts on Day 6 with "in vivo-like" profiles of carbohydrate metabolism. However, total cell numbers in resulting blastocysts and overall development are still compromised relative to in vivo-produced embryos, and the in vivo viability of these cultured embryos has yet to be evaluated. Although additional research will be needed before truly "normal" feline embryos can be routinely produced in vitro, this study has helped to improve our understanding of feline embryo physiology and furthered our capacity to culture cat embryos.
ACKNOWLEDGMENTS
The authors are extremely grateful to Jackie Newsom, Suzanne Booker, and Carla Mascari for excellent care of all of the cats used in this study. Numerous volunteers are also thanked for their assistance maintaining the cats and monitoring anesthesia during surgical procedures. We are grateful to Dr. Ann Manharth for veterinary assistance, Ted Hunt for help with radioactive materials at University of Cincinnati, Roger Ruff for help with laboratory space at University of Cincinnati, Dr. Rebecca Spindler for comments on an earlier version of this manuscript, and Dr. Normand St.-Pierre for his assistance with statistical analysis. Finally, special thanks to Desperado and Cassanova for their cooperation and dedication throughout this study.
FOOTNOTES
1Supported by 5R24RR15388-5 from the National Institute of Health to W.F.S. 
Correspondence: 2Jason R. Herrick, C.R.E.W./Cincinnati Zoo and Botanical Garden, 3400 Vine St., Cincinnati, OH 45220. FAX: 513 569 8213; e-mail: jason.herrick{at}cincinnatizoo.org
Received: 11 October 2006.
First decision: 15 November 2006.
Accepted: 22 January 2007.
REFERENCES
