Fertility in seasonal-calving pasture-based lactating dairy cows following timed artificial insemination or timed embryo transfer with fresh or frozen in vitro produced embryos

The objective was to compare pregnancy per service event (P/S) in lactating dairy cows following timed artificial insemination ( AI ) or timed embryo transfer ( ET ) using either fresh or frozen in vitro produced ( IVP ) embryos. Oocytes were collected once per week for up to 9 weeks using transvaginal ovum pick-up from elite dairy donors (ET-DAIRY; n = 40; Holstein-Friesian and Jersey) and elite beef donors (ET-ELITE-BEEF; n = 21; Angus). Both ET-DAIRY and ET-ELITE-BEEF donors were comprised of heifers and cows. In addition, oocytes were collected from the ovaries of beef heifers of known pedigree following slaughter at a commercial ab-attoir (ET-COMM-BEEF; n = 119). Following in vitro maturation and fertilization, presumptive zygotes were cultured in vitro to the blastocyst stage. Grade 1 blastocysts were either transferred fresh or frozen for on-farm thawing and direct transfer. 1106 recipient cows (all lactating, predominantly Holstein-Friesian) located on 16 herdlets were blocked based on parity, calving date and economic breeding index, and randomly assigned to receive AI (n = 243) or ET (n = 863) after estrous synchronization with a 10-d Progesterone-synch protocol. Cows assigned to ET were further randomized to receive fresh (n = 187) or frozen (n = 178) ET-ELITE-BEEF embryos, fresh (n = 169) or frozen (n = 162) ET-DAIRY embryos, or fresh (n = 80) or frozen (n = 87) ET-COMM-BEEF embryos. Pregnancy was diagnosed using trans-rectal ultrasound on d 32 to 35 after synchronized ovulation and confirmed on d 62 to 65, at which time fetal sex was determined. Pregnancy per service event at d 32 was not different between AI (48.8%) and ET (48.9%) and did not differ between dairy and beef embryos (50.3% vs 48.1%, respectively). However, P/S was less on d 32 following transfer of frozen embryos (41.6%) compared with fresh embryos (56.1%). Pregnancy loss between d 32 and 62 was greater for ET (15.1%) compared with AI (4.7%), with greater losses observed for frozen beef (18.5%), fresh beef (17.3%) and frozen dairy (19.2%) compared with fresh dairy (6.0%) embryos. Serum P4 concentration on d 7 was associated with P/S at d 32 and d 62. Cows in the quartile with the least serum P4 concentrations (Q1) had less probability of being pregnant on d 32 (33.4%) compared with cows in the 3 upper quartiles for serum P4 (45.7%, 55.6% and 61.2 for Q2, Q3 and Q4, respectively). Sex ratio (M:F) at d 62 was skewed toward more male fetuses following ET (61.1:38.9) compared with AI (43.2:56.8) and was consistent with the sex ratio among in vitro blastocysts (61.2:38.8). In conclusion, P/S was similar for AI and ET, although pregnancy loss between d 32 and d 62 was greater for ET than for AI.


INTRODUCTION
Until relatively recently, artificial insemination (AI) with conventional (non-sorted) semen from high genetic merit dairy bulls was the only option available to commercial dairy farmers to impregnate cows and to generate replacement heifers.While the use of conventional semen has provided long-term genetic gain, its use results in approximately 52:48 male-to-female sex ratio at birth (Xu et al., 2000, Roche et al., 2006, Berry and Cromie, 2007).While a small proportion of these male calves (around 0.1%) are genetically elite and of value as potential future AI bulls (De Vries et al., 2008), the majority of male dairy calves have low economic value because of their poor future beef value.These surplus male calves present welfare, social and environmental concerns (Ritter et al., 2019, Shivley et al., 2019).As a consequence, the use of both sex-sorted dairy semen and conventional beef semen in the dairy herd is increasing (reviewed by Crowe et al., 2021).Use of sex-sorted dairy semen on the dams with the best genetic merit to generate replacement females and beef semen to generate all remaining pregnancies (non-replacement calf crop) is growing in popularity (Bittante et al., 2020, Pahmeyer and Britz, 2020, Cabrera, 2022), facilitating genetic gain in replacement stock while enhancing the beef value of surplus calves.For example, in the United States, the number of beef breed AI straws used on dairy herds more than doubled in the period from 2015 to 2019 (McWhorter et al., 2020).Similarly, in 2018, 45% of calves derived from Irish Holstein-Friesian dams were sired by beef bulls (Department of Agriculture, 2019) an increase from 32% 5 years previously (Department of Agriculture, 2014).Although beef-cross calves (derived from AI or natural service) have greater economic value than male dairy calves, further gains are potentially feasible if the percentage of beef-breed genetics in non-replacement stock could be increased to ≥ 75%.
Widespread uptake and usage of female sex-sorted semen on the best genetic merit dams may inadvertently decelerate genetic gain as a result of fewer elite male dairy calves from which to choose potential AI sires.One solution to this undesirable side-effect of sex-sorted semen usage would be to generate future AI bulls by design, using ovum pick-up (OPU) to harvest oocytes from elite dairy dams, and in vitro embryo production (IVP) to generate blastocysts suitable for transfer to recipients.In seasonal, pasture-based, systems of production, however, excellent fertility in the short (12-week) breeding season is essential.The economic consequences of poor fertility, manifested as poor pregnancy per service event (P/S) and/or excessive embryo loss after conception, are amplified in such systems compared with year-round calving systems (Shalloo et al., 2014).
The overall objective was to compare pregnancy outcomes in lactating dairy cows following timed AI or timed ET using either fresh or frozen IVP embryos from dairy or beef breeds.Specifically, we tested the hypotheses that cows that received a fresh embryo would have similar P/S as cows artificially inseminated with conventional semen, and that both AI and ET with fresh embryos would achieve better P/S than that achieved with frozen embryos.

MATERIALS AND METHODS
All experimental procedures involving animals were approved by the Teagasc Animal Ethics Committee and authorized by the Health Products Regulatory Authority in Ireland, in accordance with Statutory Instrument No. 543 of 2012 under European Union legislation (Directive 2010/63/EU) for the Protection of Animals used for Scientific Purposes.The experimental design is illustrated in Figure 1.Experimental procedures with animals were conducted between March and August 2021.

Live donors used for oocyte collection
Oocytes were collected weekly from the ovaries of elite dairy donors (n = 40 Holstein Friesian (HF), and Jersey (JE); ET-DAIRY) and elite beef donors (n = 21 Angus (AA); ET-ELITE-BEEF) using transvaginal OPU.The ET-DAIRY donors were a mixture of maiden heifers (n = 7 HF, n = 1 JE; mean age (±SD) 13.3 ± 0.4 mo), and cows (n = 22 HF, n = 10 JE; mean parity (±SD) 3.7 ± 1.9; mean DIM (±SD) on the day of the first OPU session 42.1 ± 13.6).The ET-ELITE-BEEF donors were a mixture of heifers (n = 15; mean age (±SD) 14.1 ± 1.2 mo) and nonlactating cows (n = 6; mean parity (±SD) 2.3 ± 1.5; mean number of days since last parturition (±SD) on the day of first OPU session 627 ± 148.5).The ET-DAIRY donors were sourced from the Next Generation Herd, which was established at Teagasc Moorepark as a sentinel research herd (top 5% of females nationally for Economic Breeding Index, EBI) to investigate the anticipated phenotypic performance of future animals selected using the EBI (O'Sullivan et al., 2020).The ET-ELITE-BEEF donors were sourced from 2 herds at the upper end of the Dairy Beef Index (DBI) for the AA breed (Berry et al., 2019).
Ovum pick-up was conducted once per week on each donor for up to 9 weeks, with a total of 21 OPU days during the study.Immediately before OPU, donors received caudal epidural anesthesia (5 mL Procaine Hydrochloride; Adrenacaine, Norbrook Laboratories, Monaghan, Ireland).The rectum was emptied, and the vulva and perineal area were cleaned.Subsequently, the ultrasound transducer (ExaPad, C614P Microconvex probe 128 element 5.0 -7.5 MHz; IMV Imaging, Bellshill, United Kingdom) was placed in the vagina, while ovaries were manipulated and positioned per rectum for follicle puncture.Follicles were punctured using an 18-guage 3-inch needle, the follicular fluid was aspirated (Bovine Follicular Aspiration Pump; WTA, College Station, Texas, USA) at a vacuum pressure of approximately 100 mmHg and cumulus oocyte com- All media used for oocyte and embryo handling were proprietary media from Vytelle LLC (Hermiston, OR, USA).Throughout the entire process, oocytes/embryos from individual donors were processed separately.Following OPU, COCs from each donor were identified using a stereomicroscope in a mobile laboratory at the farm.They were then transferred to 2-mL plastic tubes (one tube per donor; Becton Dickenson (BD), Franklin Lakes, New Jersey, USA) containing 1 mL maturation medium (Vytelle LLC), covered by 500 μL of mineral oil in a portable incubator (iQ2, MicroQ Technologies, Scottsdale, AZ, USA) and transported to the IVF laboratory.In the laboratory, the tubes were transferred to an incubator maintained at 38.5°C with 5% CO 2 in air to allow completion of maturation, approximately 24 h from the time of the OPU.

Heifers used for oocyte collection post-slaughter
On one day per week for 6 weeks, COCs were collected from the ovaries of beef heifers of known pedigree following slaughter at a commercial abattoir (ET-COMM-BEEF; n = 119 heifers).These heifers were comprised of AA x HF crossbreds (n = 90), Limousin (LM) x HF crossbreds (n = 18) and other beef or beef x dairy crossbreds (n = 11).The mean age (±SD) at slaughter was 24.6 ± 1.6 mo, and the mean carcass weight (±SD) was 291.9 ± 28.7 kg.Ovaries were removed from the reproductive tract in the abattoir and stored in flasks of PBS at 35°C until arrival at the laboratory (approximately 3 h post slaughter).Once in the laboratory, all surface visible follicles were aspirated using a 10-mL syringe and 18-guage needle to recover the COCs.In vitro maturation of oocytes was conducted as described for those collected from the live donors.ET-DAIRY) and elite beef donors (n = 21 Angus; ET-ELITE-BEEF) using transvaginal ovum pick-up, OPU).Following OPU, COCs from each donor were transported to the IVF laboratory where they completed maturation, approximately 24 h after the time of OPU.On one day per week for 6 weeks, COCs were collected from the ovaries of beef heifers of known pedigree following slaughter at a commercial abattoir (ET-COMM-BEEF; n = 119 heifers).Ovaries were removed from the reproductive tract in the abattoir and stored in flasks of PBS at 35°C until arrival at the laboratory where all surface visible follicles were aspirated to recover the COCs.In vitro maturation (IVM) was conducted as described for those collected from the live donors.Following IVF, presumptive zygotes were cultured in vitro (IVC).The resulting grade 1 blastocysts were either transferred fresh (all d 7) or frozen (d 6, d 7 or d 8) for on-farm thawing and direct transfer (ET).Timed AI was carried out for 243 control cows and ET occurred on d 7 after synchronized estrus for 863 cows.

In vitro fertilization and embryo culture
Frozen thawed semen straws from 13 dairy-breed bulls (n = 7 HF and n = 6 JE), 6 beef-breed bulls (n = 5 AA and n = 1 LM) were used for the fertilization of oocytes harvested from ET-DAIRY donors, ET-ELITE-BEEF donors and ET-COMM-BEEF donors, respectively.A motile sample of sperm was obtained by density gradient separation (Vytelle LLC).Matured COCs were washed in one drop of washing medium and one drop of fertilization medium (Vytelle LLC), transferred in groups of up to 30 oocytes to droplets of fertilization medium (Vytelle LLC) under mineral oil and inseminated with a concentration of approximately 1 million sperm/ml.Gametes were co-incubated for approximately 18 to 24 h in an atmosphere of 5% CO 2 in air at 38.5°C.Presumptive zygotes were cultured in vitro in first-step culture medium (Vytelle LLC) in an atmosphere of 5% CO 2 , 6% O 2 , 89% N 2 at 38.5°C.On d 4, cleaved embryos were transferred to secondstep culture medium (Vytelle LLC).Resulting grade 1 blastocysts, classified according to guidelines of the International Embryo Technology Society (Barfield and Demetrio, 2022), were either transferred fresh (all d 7) or frozen (on d 6, d 7 or d 8) for on-farm thawing and direct transfer.

Embryo cryopreservation
Grade 1 blastocysts and expanded blastocysts were removed from culture on d 6 p.m. (n = 87), d 7 a.m.(n = 220), d 7 p.m. (n = 42) or d 8 a.m.(n = 49) and exposed to a freezing first-step and freezing secondstep medium (Vytelle LLC).Embryos were loaded into 0.25-mL yellow direct transfer straws, placed in a central column surrounded by 4 columns of the freezing second-step medium separated by air bubbles.After loading, the straws were placed in a freezing machine (EFT-3002, Beltron Instruments, CO, USA) that had been previously stabilized at −6°C.Two min after being placed in the machine, crystallization ("seeding") of the columns immediately below and above the embryo column was conducted.The freezing curve was then initiated, lowering the temperature 0.5°C per min to −35°C, after which straws were immersed directly into liquid nitrogen, where they were stored until on-farm thawing for direct transfer into the recipients.

In vitro assessment of embryo survival post freezethawing
To obtain an estimate of embryo survival post-thawing, representative samples of frozen blastocysts (n = 107, 3 replicates) were thawed (10 s in air followed by 30 s in water at 35°C) and cultured in vitro, as described above, for 72 h.Survival (re-expansion) and hatching were recorded at 24, 48 and 72 h post thawing.

Recipient synchronization for timed AI and timed ET
Recipient cows (all lactating, predominantly HF but including a small number of JE or HF x JE crossbreds; mean parity = 2.9 +/− 1.5) were located in 10 herds, broken into 16 herdlets, with some herds (n = 6) having 2 separate cohorts of cows enrolled on the synchronization protocol followed by AI and ET 2 weeks apart, and others (n = 4) having a single cohort of cows enrolled on the synchronization protocol followed by AI and ET.Recipients were blocked based on parity, calving date and EBI and randomly assigned to receive AI or ET with an ET-DAIRY, ET-ELITE-BEEF or ET-COMM-BEEF embryo.Recipients were synchronized using a modified 10-d Progesterone-synch protocol (Figure 2) as described by Drake et al. (2020).On d −10 relative to the farm mating start date, a 2 mL i.m. injection of GnRH analog (Ovarelin®, 100 μg of gonadorelin diacetate tetrahydrate; Ceva Santé Animale, Libourne, France) was administered, and a progesterone-releasing intravaginal device containing 1.55 g progesterone (PRID® Delta; Ceva Santé Animale) was inserted.On d −3, a 5 mL i.m. injection of PGF2α (Enzaprost®, 25 mg of dinoprost trometamol; Ceva Santé Animale) was administered.On d −2, a second 5 mL i.m. injection of PGF2α was administered and the PRID was removed.On d −1 (32 h after PRID removal) a second i.m. injection of GnRH was administered.AI was carried out 16 h later and ET occurred on d 7.
The industry target for pregnancy to first service in seasonal-calving systems in 60%.Based on previous publications, the P/S for cows following transfer of a frozen IVP embryo has been reported to be 15 to 20 percentage points less than P/S for cows following either AI or transfer of a fresh IVP embryo (Carrenho-Sala et al., 2016, Pereira et al., 2016, Hansen, 2020).For the current study, a power test indicated that 150 cows per treatment were required to have an 80% chance of detecting, as significant at the 5% level, a decrease in the pregnancy per service event from 60% (target for AI and fresh ET treatments) to 45% in the frozen ET treatments.To mitigate against cows being dropped from the study for various reasons, a greater number of cows was initially enrolled.
A total of 1197 recipients were synchronized as described above, of which 243 (20.3%) were assigned to receive AI (16 h after second GnRH) and 954 (79.7%) were assigned to receive ET on d 7 after presumptive estrus.In 12 of the herdlets, cows assigned to AI (n = 183, 20%) were inseminated using frozen-thawed semen from dairy sires and cows assigned to ET (n = 648) were further randomized to receive ET-ELITE-BEEF (20% fresh, 20% frozen) or ET-DAIRY (20% fresh, 20% frozen) blastocysts.On the remaining 4 herdlets, cows assigned to AI (n = 60) were inseminated with semen from a beef sire.On one of these 4 herdlets, cows assigned to ET (n = 48) all received ET-ELITE-BEEF (40% fresh, 40% frozen) blastocysts and on 3 herdlets, cows assigned to ET (n = 167) all received ET-COMM-BEEF (40% fresh, 40% frozen) blastocysts.On the day of scheduled ET, recipient reproductive tracts were examined by trans-rectal ultrasound to assess corpus luteum (CL) status, presence or absence of cystic structures and uterine health status.Following this examination, 91 cows (9.6%) were deemed unsuitable for ET on the basis of having no dominant follicle or CL present (n = 41), abnormal ovarian structures (presumed follicular cyst or luteal cyst; n = 26) or small CL (<15 mm diameter; n = 22), and 2 cows were removed for non-compliance with the synchronization protocol.Thus, the final number of recipients that were suitable for embryo transfer was 863.The uterine horn ipsilateral to the ovary bearing the CL was identified at the time of the ultrasound scan.For cows with corpora lutea on both ovaries (n = 2), the uterine horn ipsilateral to the ovary bearing largest CL was identified.

Embryo Transfer
Embryos for fresh transfer were removed from culture on the morning of d 7 (d 0 = day of IVF) and loaded into transportation medium (Vytelle LLC) in a clear straw.Fresh embryos were maintained at 38.5°C in a portable incubator (MicroQ Technologies) during transport from the laboratory to the farm where the recipients were located.At the farm, fresh straws were loaded into ET guns, which were then placed into a temperature-controlled gun warmer (IFT Instruments, Montevideo, Uruguay) at 35°C until immediately before transfer.Straws containing frozen embryos were removed from liquid nitrogen, held in air for 10 s and immersed in water at 35°C for 30 s.The straw was dried and loaded into an ET gun, which was then placed into a gun warmer until transfer.All embryos were transferred to the uterine horn of the recipient ipsilateral to the ovary bearing the CL, or to uterine horn ipsilateral to the ovary bearing the largest CL if both ovaries had a CL present.All transfers were conducted by one of 2 experienced technicians.

Pregnancy diagnosis and fetal sexing
Returns to estrus were recorded by visual observation of estrous activity and/or tail paint removal.Pregnancy status was diagnosed in cows that had not returned to estrus by trans-rectal ultrasound scanning on d 32-35.For cows that were diagnosed pregnant, a second ultrasound examination was conducted at d 62-65 after synchronized ovulation to determine whether the pregnancy had been maintained or lost.Fetal sex was determined at the ultrasound exam on d 62-65 in 436 pregnant cows based on the detection and location of the genital tubercle.Embryonic loss was calculated as the percentage of those cows pregnant on d 32 that were no longer pregnant on d 62.

Determination of progesterone concentration on d 7
On the day of ET, blood samples were collected into serum tubes (BD Vacutainer, BD, Plymouth, UK) by coccygeal venipuncture from all cows enrolled in the study (n = 1197), including those that received AI, ET and those rejected for ET, to measure serum progesterone (P4) concentration.Blood samples were stored at 4°C for 24 h before centrifugation at 1922 x g at 4°C for 15 min.Using a Pasteur pipette, serum was separated and stored at −20°C until analysis of P4 concentrations by solid-phase radioimmunoassay using a PROG-RIA-CT Kit (DIAsource ImmunoAssays S.A., Louvainla-Neuve, Belgium) according to the manufacturer's instructions.The sensitivity of the assay was 0.05 ng/ ml.The inter-assay coefficients of variation for quality control samples were 12.1% (low), 7.4% (medium), and 5.5% (high), respectively.The intra-assay coefficients of variation were 16.3% (low), 13.0% (medium), and 8.7% (high).

Sex determination of IVP embryos
Day 7 (n = 63) and d 8 (n = 40) IVP blastocysts produced in 3 replicates (i.e., 3 independent days of ovary collection) using abattoir-derived ovaries as the source of oocytes were snap-frozen, and crude DNA lysates were prepared by adding 20.0 μL of PCR Buffer supplemented with Proteinase K (1 mg/ml) to each sample.The samples were incubated at 56°C for 30 min with gentle shaking and then heat inactivated at 94°C for 10 min.A 20.0 μL PCR assay targeting the Amelogenin gene was prepared using 5.0 μL of the crude lysate, 0.5 μM of each primer, AML-X and AML-Y, 1 X DreamTaq PCR buffer and 0.5 Units of DreamTaq DNA polymerase (ThermoFisher, Waltham, Massachusetts, USA).The PCR amplification consisted of an initial 5-min activation at 95°C, followed by 25 cycles of amplification; denaturation (60 s at 95°C), annealing (60 s at 60°C) and extension (90 s at 72°C), followed by a final incubation at 72°C for 10 min and a hold at 4°C.The resulting PCR products were analyzed using gel electrophoresis on a 2.5% agarose gel and the sex of each blastocyst was determined.Male blastocysts produced 2 PCR products at 241 bp and 178 bp and female blastocysts produced a single PCR product at 241 bp (Gokulakrishnan et al., 2012).

Statistical Analysis
All statistical analyses were conducted using SAS v. 9.4 (SAS Institute, Cary, NC).Before analysis, data were assessed for normality and transformed using optimum Box-Cox transformations where necessary (TRANSREG).Data relating to oocyte collection data, blastocyst percentage and blastocyst number were analyzed using generalized linear mixed models (GLIMMIX) with repeated measures.Several models were constructed to compare the oocyte collection and embryo production between donors and embryo production between sires used.Due to parity structure differences between elite dairy donors and elite beef donors, data were analyzed separately for each type of donor.In each model assessing oocyte collection, the model fixed effects included donor, parity of donor (nulliparous vs parous) and the collection number.
In each model assessing blastocyst rate (%) and total blastocyst yield (n) per OPU session, the model fixed effects included donor, parity of donor and sire.
A total of 1197 cows were enrolled in the study and synchronized for either AI or ET.Before the d 32 pregnancy diagnosis, 97 cows were removed from the study: 2 cows had no recorded service event, 91 cows were deemed unsuitable on the day of ET, 3 cows were culled for reasons unrelated to the study and one cow died due to reasons unrelated to the study; therefore, 1100 cows were included in the d 32 analysis.Before the d 62 pregnancy diagnosis, a further 4 cows were removed: 3 cows were culled for reasons unrelated to the study and one cow was sold; therefore, 1096 cows were included in the d 62 analysis.Pregnancy data were analyzed using generalized linear mixed models, with a binary distribution specified.P/S did not differ between AI services that used dairy or beef semen (P = 0.969), and these services were combined and reported as AI.P/S did not differ between the ET-ELITE-BEEF and ET-COMM-BEEF (P = 0.395), and these services were combined and are reported as ET-BEEF.Multiple variables and interactions were included as fixed effects and retained in the final GLIMMIX models when P ≤ 0.25.Several models were constructed to compare the type of service event: (1) AI vs. ET (all cows); (2) AI vs. ET-BEEF vs. ET-DAIRY; (3) AI vs. Fresh ET vs. Frozen ET; and (4) AI vs. ET-BEEF-Fresh vs. ET-BEEF-Frozen vs. ET-DAIRY-Fresh vs. ET-DAIRY-FROZEN.In each model, treatment, parity and serum P4 concentration (categorized into quartiles; Q1: < 5.79 ng/ml, n = 274; Q2: 5.79 -7.36, n = 274; Q3: 7.37 -9.42, n = 274 and Q4: > 9.43, n = 275) were included as fixed effects and herd was included as a random effect.Box-and-whisker plots were used initially to visualize the variation in serum P4 concentrations on d 7 after synchronized estrus in cows assigned to AI vs. ET, including a separate plot for cows that were rejected for ET.Cows that were deemed unsuitable for ET on the scheduled day of transfer were then added back into the same models to determine the effect of treatment on pregnancy/cow synchronized (P/Sync).For assessing the effect of each type of service event on the incidence of embryonic loss, 538 cows that were diagnosed pregnant on d 32 were available for inclusion in the analysis.The GLIM-MIX model included treatment and parity, with herd included as a random effect.For all binary outcome variables analyzed, the GLIMMIX model output values for treatment means are reported, which equate to the predicted probability of that event.Finally, an additional analysis was undertaken on the data related to frozen ET only.For the 397 cows that received frozen ET, a Chi-Square test of independence was completed to determine the association between the day of culture (age) of the embryo at the time it reached blastocyst stage and was cryopreserved and the associations with

Factors associated with oocyte recovery and in vitro embryo production
The oocyte recovery rates and in vitro embryo development results are summarized in Table 1 and Figure 3A.The donor used for OPU was associated with the mean number of oocytes collected per OPU session (mean: 15.6, range: 4.3 to 36.3), the mean number of viable embryos per OPU/IVF session (mean: 3.8, range: 0 to 15) and with the percentage of oocytes that developed to transferable embryos (mean: 25.0%, range: 0% to 75%).The yield of transferrable embryos per donor per OPU/IVF session remained relatively constant over time (Figure 4).Blastocyst developmental rates per sire are presented in Table 2 and Figure 3B.The sire used in IVF was associated with blastocyst yield (mean: 23.1%, range: 5% to 46%) and the number of transferable embryos per IVF session (mean: 3.6, range: 0.5 to 7.1).

Blastocyst survival post-thawing
Blastocyst survival and hatching results following freezing/thawing are summarized in Table 3 and illustrated in Figure 5.By 24 h after thawing, 94.4% of blastocysts had re-expanded and 19.6% had hatched.By 48 h, 96.3% had re-expanded and 51.4% had hatched.At the final assessment, 72 h after thawing, 96.3% had re-expanded and 72.0% had hatched.
To compare the efficiency of AI (i.e., all synchronized cows inseminated) versus ET (9.6% of synchronized cows deemed unsuitable for ET), data were analyzed to determine P/Sync for each treatment.The predicted probability of P/Sync at d 32 was not different between AI and ET (47.9% vs 43.8%; P = 0.279).P/Sync was not different between ET-DAIRY and ET-BEEF at d 32 (45.4% vs. 42.8%;P = 0.746) and was greater following transfer of a fresh embryo than following transfer of a frozen embryo (50.5% vs 37.1%; P = 0.0002).The magnitude of the difference between fresh embryos and frozen embryos was larger for ET-DAIRY (53.5% vs. 37.2%; P = 0.024) than ET-BEEF (48.6% vs. 37.0%; P = 0.0498) (Figure 7C), primarily driven by the greater percentage of cows that became pregnant in the ET-DAIRY treatment after transfer of a fresh embryo.By d 62, P/Sync was less (P = 0.01) for cows assigned to ET (35.9%) compared with cows assigned to AI (45.4%).P/Sync was not different between ET-DAIRY and ET-BEEF at d 62 (38.5% vs. 34.3%;P = 0.438) and was greater for fresh embryos than frozen embryos (42.8% vs 29.1%; P < 0.0001).The magnitude of the difference between fresh embryos and frozen embryos was larger for ET-DAIRY (48.9% vs. 28.2%;P = 0.0009) than for ET-BEEF (39.0% vs. 29.7%;P = 0.141; Figure 7D).

Factors affecting the probability of embryonic loss between d 32 and d 62
Across all cows that were diagnosed pregnant on d 32, the predicted probability of pregnancy loss between d 32 and 62 was greater (P = 0.004) for ET (15.1%; 11.0, 20.5) compared with AI (4.7%; 2.0, 10.4) (Figure 7E).There was no overall effect of embryo type (fresh, 12.2% vs. frozen, 18.8%; P = 0.141 or dairy vs beef; P = 0.347) on probability of embryo loss; frozen beef (18.5%), fresh beef (17.3%), frozen dairy (19.2%) and fresh dairy (6.0%).Embryo loss following the transfer of fresh dairy embryos tended to be less than the transfer of frozen dairy embryos (P = 0.06).There was no association between P4 concentration (P = 0.934) or days in milk (P = 0.431) on embryonic loss between d 32 and 62, but parity tended to be associated with embryo loss across all treatments (P = 0.071).Embryonic loss between d 32 and d 62 occurred in 14.9% of parity 1 cows compared with 6.3% of parity 2, 11.2% of parity 3 and 15.6% of parity 4 cows.
Day of culture when blastocysts were cryopreserved, based on when the embryo reached the blastocyst stage, tended to be associated with P/S on d 32 (P = 0.078) and on d 62 (P = 0.052), but was not associated with embryo loss between d 32 and d 62 (P = 0.346).These data are summarized in Table 5.

DISCUSSION
In vitro embryo production is now an established technology in the toolbox of assisted reproductive technologies available to farmers and breeding companies.Despite greater fixed costs, it offers significant advantages over traditional superovulation and embryo transfer (MOET) including increased numbers of embryos produced per donor per unit of time and greater flexibility in sire usage.In addition, IVF facilitates more efficient use of rare or high-cost semen straws, and thanks to the predictability of a donor's performance (once a donor has been collected previously) in terms of oocyte and embryo yields, IVF can simplify the logistics of recipient synchronization and management.All of these factors have contributed to the marked increase in the use of IVF compared with MOET as a method of generating embryos for transfer worldwide (Viana, 2022).
The mean number of oocytes recovered (Dairy: 16.9, Beef: 15.8) and mean yield of transferable embryos (Dairy: 4.0, Beef: 4.5) per OPU session was highly variable between donors, with means ranging from 4.3 to 36.3 oocytes per donor per OPU session and zero to 15 transferrable embryos per donor per IVF session.This observation is consistent with published commercial data (Demetrio et al., 2020), where dairy donors produced an average of 15.6 oocytes and 3.6 viable embryos per OPU session and beef donors produced an average of 19.2 oocytes and 5.2 viable embryos per OPU session.Similar to superovulation, yield per session was repeatable within donor.In agreement with other studies (e.g., Ortega et al., 2018), significant variation was noted between sires used in IVF, with blastocyst yield ranging from 14.3% to 45.5%.Such sire variation is not unusual; for example, significant variation in field fertility exists among bulls used in AI, despite rigorous assessments of sperm quality before semen is released (Fair and Lonergan, 2018).Furthermore, sire field fertility does not correlate well with IVF success (Al Naib et al., 2011).
Over the past decade, the success of commercial IVP has been reflected in the number of IVP embryos transferred annually now superseding that of in vivo-derived embryos (Crowe et al., 2021, Viana, 2022).Nevertheless, embryos generated in vitro still differ from their in vivo-produced counterparts, particularly in terms of cryotolerance, P/S and embryo loss after freeze-thawing (Pontes et al., 2009, Sartori et al., 2018).Given that ET bypasses potential issues in the recipient cow related to oocyte quality, sperm transport and fertilization, one would expect P/S to be greater than that achieved with AI or natural service; however, this potential is typically not observed in practice (Hansen, 2020).It is well accepted that the quality of the oocyte at the start of the process is the key factor determining the proportion  of oocytes developing to the blastocyst stage (Rizos et al., 2002, Lonergan andFair, 2016).Culture conditions throughout IVP, particularly during post-fertilization culture, influence the quality, including cryotolerance, and developmental potential of the early embryo (Rizos et al., 2002, Gad et al., 2012).In the current study, P/ ET was similar to P/AI when embryos were transferred fresh, demonstrating the potential for fresh ET to be used effectively in a seasonal dairy production system without compromising subsequent calving pattern for recipient dams.It is important to note, however, that across all cows assigned to be recipients, 9.6% were deemed unsuitable on the day of scheduled ET.Hence, the proportion of synchronized cows that could become pregnant following ET would be less than reported for cows assigned to be artificially inseminated.One benefit of removing unsuitable cows was that they could be re-bred, receive veterinary treatment for any problems identified, or be assigned to be culled without further expenditure on reproductive interventions or insemination costs.The financial cost and loss of days in milk that arises following synchronization of cows that are subsequently unsuitable for transfer also need to be considered.Nonetheless, the results of this study illustrate the feasibility of using IVF technology in a seasonal system.Pregnancies per ET have generally been reported to be less for IVP embryos compared with in vivo-derived embryos (Pontes et al., 2009, Carrenho-Sala et al., 2016, Pereira et al., 2016, Sartori et al., 2018).In addition, pregnancy loss has been reported to be greater for IVP embryos than for either in vivo-derived embryos or pregnancies from AI, further reducing reproductive efficiency achieved with IVP embryos (Carrenho-Sala et al., 2016, Pereira et al., 2016, Sartori et al., 2018).Thus, strategies to improve P/ET and reduce pregnancy loss are needed to maximize the efficiency of IVP-ET programs in cattle.
The majority of cryopreserved embryos that were later transferred to recipient cows were frozen on the morning of d 7. Day of culture when blastocysts were cryopreserved, based on when the embryo reached the blastocyst stage, tended to affect P/S on d 32 or d 62. Failure to find a significant association in this study was likely due to the relatively small number of embryos in each age category.It is interesting to note that P/  3. ET appeared to diminish and embryo loss appeared to increase as embryo age at cryopreservation increased.This is consistent with previous observations that the timing of blastocyst formation affects blastocyst quality (Dinnyes et al., 1999).This association between embryo age at cryopreservation and P/S merits further investigation in future studies that are adequately powered to specifically address this issue.
In the current study, pregnancy loss between d 32 and d 62 was significantly greater following ET, particularly following transfer of frozen-thawed embryos (12.2% vs 18.8% loss in fresh vs frozen embryos respectively).Seasonal production systems require excellent herd reproductive performance, and hence any management strategy that reduces P/S and increases embryonic losses will reduce profitability (Shalloo et al., 2014).Therefore, uptake of IVP and ET in seasonal dairy production systems will largely focus on use of fresh ET.Further investigation is necessary to improve the cryopreservation of IVP embryos.
The relationship between circulating P4 and uterine receptivity has been well described (Lonergan and Sanchez, 2020).Elevated P4 concentrations in the first week after conception have been associated with accelerated post-hatching conceptus elongation, mediated through advancement in the regular temporal changes in the uterine endometrial transcriptome (Forde et al., 2009) and alterations in the uterine lumen fluid (histotroph) composition (Simintiras et al., 2019).Consistent with other studies that have shown positive linear (Herlihy et al., 2013) and quadratic relationships (Diskin et al., 2006), there was a positive, quadratic relationship between serum P4 on d 7 and likelihood of pregnancy establishment, irrespective of method of breeding (AI or ET); cows with P4 concentrations in the quartile with the least P4 concentrations were almost half as likely to become pregnant than those in the upper quartile.Furthermore, Wallace et al. (2011) reported that administration of hCG at the same time as ET increased incidence of accessory CL formation, serum P4 in pregnant recipients, and P/S and reduced early embryonic losses after transfer.Treatment of cows with GnRH on d 5 after estrus was reported to also reduce embryo loss between d 33 and d 60 in heifers that received fresh IVP embryos (Garcia-Guerra et al., 2020).In the current study, P/S increased quadratically as P4 concentration on d 7 increased, plateauing at around 5-8 ng/ml.
To determine whether the deviation in sex ratio was due to inadvertent preferential selection of male embryos for transfer or was due to preferential survival of male embryos post transfer, a representative number of embryos (n = 103) was sexed post-thawing.Our observations indicate that the bias toward male fetuses on d 62-65 in recipient cows was mirrored by a similar sex bias in IVP blastocysts on d 7 and d 8, indicating similar survival of male and female embryos after transfer to recipients on d 7. A deviation in sex ratio toward more males following IVF in cattle was first described over 30 years ago (Avery et al., 1991(Avery et al., , 1992)).Subsequently, several studies reported that this bias was due to impaired imprinted X chromosome inactivation (Gutierrez-Adan et al., 2001, Wrenzycki et al., 2002, Tan et al., 2016).A similar phenomenon of greater male birth rate following IVF has been reported in humans (Dean et al., 2010, Maalouf et al., 2014).
The difference in developmental rates between male and female embryos in vitro (Gutierrez-Adan et al., 1996) has been attributed to differences in metabolic activity of X-linked enzymes involved in energy metabolism (Tiffin et al., 1991).The X-linked gene glucose-6-phosphate dehydrogenase (G6PD) has been considered a likely candidate for involvement in sex differences due to impaired imprinted X chromosome  inactivation in IVF female bovine embryos (Wrenzycki et al., 2002).Furthermore, glucose concentration >2.5 mM during embryo culture impairs bovine embryo de-velopment and increases the sex ratio toward males, most likely as a result of increased pentose-phosphate pathway activity in female embryos (Kimura et al., 2008).Irrespective of the cause of the deviation in sex ratio in IVP embryos, this can be overcome through the use of sexed semen in IVF (Bermejo-Alvarez et al., 2010).
In conclusion, P/S was similar for AI and ET on d 32 post-estrus, but subsequent pregnancy loss was greater for ET than for AI.Transfer of frozen embryos resulted in fewer pregnancies and tended to be associated with greater embryo loss between d 32 and d 62 compared with fresh ET or AI.Nonetheless, our results demonstrate the potential for using OPU/IVF to produce embryos from high genetic merit donors within the calendar constraints of the seasonal-calving system.In the present era where beef on dairy is very topical, IVP and ET provides a clear route accelerating genetic gain in both dairy breeds and beef breeds suitable for crossing with dairy dams, and also producing beef breed embryos using abattoir-derived ovaries to allow premium quality beef to be produced from dairy dams.Additional research is required to improve fertility performance and reduce embryo loss of IVP embryos.

Figure 1 .
Figure 1.Experimental design.Oocytes were collected weekly from the ovaries of elite dairy donors (n = 40 Holstein Friesian and Jersey;ET-DAIRY) and elite beef donors (n = 21 Angus; ET-ELITE-BEEF) using transvaginal ovum pick-up, OPU).Following OPU, COCs from each donor were transported to the IVF laboratory where they completed maturation, approximately 24 h after the time of OPU.On one day per week for 6 weeks, COCs were collected from the ovaries of beef heifers of known pedigree following slaughter at a commercial abattoir (ET-COMM-BEEF; n = 119 heifers).Ovaries were removed from the reproductive tract in the abattoir and stored in flasks of PBS at 35°C until arrival at the laboratory where all surface visible follicles were aspirated to recover the COCs.In vitro maturation (IVM) was conducted as described for those collected from the live donors.Following IVF, presumptive zygotes were cultured in vitro (IVC).The resulting grade 1 blastocysts were either transferred fresh (all d 7) or frozen (d 6, d 7 or d 8) for on-farm thawing and direct transfer (ET).Timed AI was carried out for 243 control cows and ET occurred on d 7 after synchronized estrus for 863 cows.
Crowe et al.: FERTILITY FOLLOWING TIMED AI OR TIMED ET Figure 2. Protocol for synchronization of recipients for timed artificial insemination (AI) or timed embryo transfer (ET).On d −10 relative to the planned breeding date, a 2-mL i.m. injection of GnRH analog was administered, and a progesterone-releasing intravaginal device (PRID) was inserted.On d −3, a 5-mL i.m. injection of PGF2α was administered.On d −2, a second 5-mL i.m. injection of PGF2α was administered and the PRID was removed.On d −1 (32 h after PRID removal) a second i.m. injection of GnRH was administered.Timed AI and ET were carried out 16 h and 7 d later, respectively.Pregnancy diagnosis (PD) was carried out by ultrasound scanning on d 32-35 and d 62-65.
Crowe et al.: FERTILITY FOLLOWING TIMED AI OR TIMED ET P/S on d 32 and 62, and with embryo loss between d 32 and d 62.

4B3-
B5 were used for ET-ELITE-BEEF oocytes only.*Blastocyst rate (P = 0.025) and the mean number of transferable embryos per IVF session differed by sire (P = 0.011).
Crowe et al.: FERTILITY FOLLOWING TIMED AI OR TIMED ET

Figure 5 .
Figure 5. Representative images of blastocyst survival and hatching post-freezing/thawing.Frozen blastocysts (n = 107, 3 replicates) were thawed and cultured in vitro for 72 h.Survival (re-expansion) and hatching were recorded at 24, 48 and 72 h post thawing.Representative images of blastocysts were captured at 10 min, 24 h, 48 h and 72 h post thawing.For more details, see text and Table3.

Figure 7 .
Figure 7. A. Predicted probability of pregnancy/service event (P/S) (%) depending on service type at d 32 and d 62, and predicted probability of embryo loss (%) between d 32 and d 62 depending on service type.B. Predicted probability of pregnancy/cows synchronized (P/Sync) (%) depending on service type at d 32 and d 62. P/Sync includes cows that were synchronized for ET but were deemed unsuitable to receive ET on the day of timed-ET.Values not sharing a common letter (a, b, c) differ (P < 0.05).Numbers under each bar are actual numbers of cows that were pregnant, received AI or ET and were synchronized.
Crowe et al.: FERTILITY FOLLOWING TIMED AI OR TIMED ET

Figure 6 .
Figure 6.Serum progesterone (P4) concentrations on d 7 after synchronized estrus for all cows that were assigned to be receive AI or ET.(A) Box and whisker plot indicating variation in serum progesterone concentration on d 7 for cows that received AI or ET, and for cows that were rejected for ET following transrectal ultrasound examination on d 7. Progesterone concentration differed between cows that were rejected and cows that received AI or ET (P < 0.0001).The ends of the whiskers represent the maximum (Quartile 3 + 1.5*Inter-quartile range) and minimum (Quartile 1 -1.5*Inter-quartile range) P4 concentration for each group.The upper and lower quartiles make up the boundaries of the box.The height of the box represents the interquartile range, and the median is indicated by the horizontal line within the box.The arithmetic mean serum P4 concentration for each group is indicated by the circle within the box and outliers are identified by the circles outside of the ends of the whiskers.(B) Interquartile ranges in d 7 serum P4 concentrations in cows that received AI or ET that were non-pregnant (blue) or pregnant (red) 32 d after estrus.Mean serum P4 concentration did not differ between cows that were diagnosed pregnant or non-pregnant on d 32 after the service with AI or ET (P = 0.113).(C) Interquartile ranges in d 7 serum P4 concentration in cows serviced using AI or ET that were non-pregnant (blue) or pregnant (red) 62 d after estrus.Mean serum P4 concentration did not differ between pregnant and non-pregnant (d 62) cows serviced using AI or ET (P = 0.262).(D) Distribution of d 7 serum P4 concentrations in lactating dairy cows that were pregnant (red) or non-pregnant (blue) on d 32 after AI or ET service events.
Crowe et al.: FERTILITY FOLLOWING TIMED AI OR TIMED ET

Table 1 .
Crowe et al.: FERTILITY FOLLOWING TIMED AI OR TIMED ET Oocyte recovery and embryo production from ET-DAIRY, ET-ELITE-BEEF and ET-COMM-BEEF donors

Table 2 .
Effect of sire used for IVF on outcomes related to embryo development.Table includes sires used for ≥10 IVF sessions 2B1 was used for IVF on oocytes from both ET-COMM-BEEF (n = 63 IVF sessions) and ET-ELITE-BEEF (n = 22 IVF sessions) donors; data are included here together (abattoir and ovum pick-up from live donors).3B2was used for ET-COMM-BEEF only.

Table 3 .
Survival and hatching at 24, 48 and 72 h post freezing/thawing of in vitro produced d 7 or d 8 bovine blastocysts 2as a percentage of surviving blastocysts.

Table 5 .
Effect of blastocyst age at freezing on pregnancy/embryo transfer at d 32 and d 62. Grade 1 blastocysts and expanded blastocysts were removed from culture on d 6 p.m., d 7 a.m., d 7 p.m. or d 8 a.m.Day of blastocyst cryopreservation tended to be associated with P/S on d 32 (P = 0.078) and on d 62 (P = 0.052), but was not associated with embryo loss between d 32 and d 62 (P = 0.346).Embryo loss was calculated as n pregnant d 32 minus n pregnant d 62 divided by n pregnant d 32.

Table 4 .
Predicted probability values for both pregnancy/service event (P/S) and embryo loss in lactating dairy cows separated into quartiles based on serum progesterone (P4) concentration on d 7. Values in parentheses represent 95% CI -c Mean values in the same column with different superscripts differ (P < 0.025). a