**Oocyte Selection for In Vitro Embryo Production in Bovine Species: Noninvasive Approaches for New Challenges of Oocyte Competence**

**Luis Aguila 1,\*, Favian Treulen 2, Jacinthe Therrien 1, Ricardo Felmer 3, Martha Valdivia <sup>4</sup> and Lawrence C Smith <sup>1</sup>**


Received: 20 October 2020; Accepted: 19 November 2020; Published: 24 November 2020

**Simple Summary:** The efficiency of producing embryos using in vitro technologies in cattle species remains lower when compared to mice, indicating that the proportion of female gametes that fail to develop after in vitro manipulation is considerably large. Considering that the intrinsic quality of the oocyte is one of the main factors affecting embryo production, the precise identification of noninvasive markers that predict oocyte competence is of major interest. The aim of this review was to explore the current literature on different noninvasive markers associated with oocyte quality in the bovine model. Apart from some controversial findings, the presence of cycle-related structures in ovaries, a follicle size between 6 and 10 mm, a large slightly expanded investment without dark areas, large oocyte diameter (>120 microns), dark cytoplasm, and the presence of a round and smooth first polar body have been associated with better embryonic development. In addition, the combination of oocyte and zygote selection, spindle imaging, and the anti-Stokes Raman scattering microscopy together with studies decoding molecular cues in oocyte maturation have the potential to further optimize the identification of oocytes with better developmental competence for in vitro technologies in livestock species.

**Abstract:** The efficiency of producing embryos using in vitro technologies in livestock species rarely exceeds the 30–40% threshold, indicating that the proportion of oocytes that fail to develop afterin vitro fertilization and culture is considerably large. Considering that the intrinsic quality of the oocyte is one of the main factors affecting blastocyst yield, the precise identification of noninvasive cellular or molecular markers that predict oocyte competence is of major interest to research and practical applications. The aim of this review was to explore the current literature on different noninvasive markers associated with oocyte quality in the bovine model. Apart from some controversial findings, the presence of cycle-related structures in ovaries, a follicle size between 6 and 10 mm, large number of surrounding cumulus cells, slightly expanded investment without dark areas, large oocyte diameter (>120 microns), dark cytoplasm, and the presence of a round and smooth first polar body have been associated with better competence. In addition, the combination of oocyte and zygote selection via brilliant cresyl blue (BCB) test, spindle imaging, and the anti-Stokes Raman scattering microscopy together with studies decoding molecular cues in oocyte maturation have the potential to further optimize the identification of oocytes with better developmental competence for in-vitro-derived technologies in livestock species.

**Keywords:** oocyte competence; livestock production; assisted reproductive technology; embryo development; micromanipulation; in vitro production

#### **1. Introduction**

In recent years, new knowledge in the field of assisted reproductive technologies (ART, has allowed researchers and practitioners to reach new hallmarks in oocyte and sperm in vitro competence. Gamete competence is the ability to undergo successful fertilization and develop a normal blastocyst that is capable of implanting in the uterus and generate viable offspring [1]. Many researchers are focused on identifying cellular and molecular markers to select the most competent oocyte and spermatozoon to produce embryos with higher implantation potential [2].

Although it is well known that the most common applications of ARTs in livestock species are for research purposes, some techniques, particularly in vitro embryo production (IVP), have become commercially viable and are extensively used for animal breeding [3]. Nonetheless, the efficiency of IVP technologies in livestock species, such as bovine, equine, and porcine, measured as the proportion of immature oocytes that reach the blastocyst stage, rarely exceeds the 30–40% threshold [4], which means that the proportion of oocytes that fail to develop following in vitro maturation, fertilization, and culture is considerably large. Contrary to humans, where eggs are mainly collected at the MII stage, in livestock species, the oocytes have to be matured in vitro due to the difficulty of obtaining a sufficient number of in vivo matured oocytes [5]. Additionally, given that the most frequent source of ovaries is slaughterhouse-derived animals, many important factors that influence oocyte quality, such as age of the donor, the stage of the estrous cycle, nutritional status, genetic potential, presence of a reproductive disorder, and others, are often unknown [6]. Therefore, it is almost impossible to avoid the retrieval of a heterogeneous population of oocytes that have a distinct ability to undergo maturation and support early embryonic development after fertilization, which is known as developmental competence or oocyte quality [7].

Considering that the intrinsic quality of the oocyte is one of the major factors affecting early embryonic development [8], and that embryo culture conditions have a crucial role in determining blastocyst quality [9], the precise selection of competent oocytes is vital for IVP technologies in livestock. Recently, the new arrival of bovine embryonic stem cells (ESCs) [10,11] emphasizes the already existing challenge in the selection of competent oocytes for the production of high-quality embryos through in vitro fertilization (IVF), intracytoplasmic sperm injection (ICSI) or somatic cell nuclear transfer (SCNT), and derivation of pluripotent stem cell lines, with promising applications in research or industry, such as in vitro breeding programs [12]. Usually, for IVP and micromanipulation procedures (ICSI and SCNT), the choice of the oocytes lie in morphological features that are easily assessed with light microscopy [13]. The major difference and/or advantage of conventional IVF compared to micromanipulation procedures is that fertilization can occur during gamete co-incubation when the oocyte has reached or is close to nuclear and cytoplasmic maturity [14]. Conversely, during micromanipulation procedures, the operator must accurately assess the maturity of the oocyte and, therefore, its competence [15]. Because the criteria used for grading and selecting oocytes vary among researchers, could be easily misinterpreted, and depend on the expert's evaluation and experience, the identification of noninvasive cellular or molecular markers that predict oocyte competence is a major research goal [16,17]. Despite efforts for finding molecular factors associated with oocyte quality, it is still challenging to find a visual marker that accurately predicts embryonic competence. Thus, this article reviews the current literature on different noninvasive markers that have been correlated with oocyte quality in cattle and explores the utility of each grading system.

#### **2. Morphological and Visual Markers for the Selection of the Best Oocytes**

#### *2.1. Ovarian Morphology*

During the retrieval of oocytes from slaughterhouse material, the collection of ovaries based on the presence or absence of estrus cycle structures, i.e., presence or absence of follicles and corpus luteum (CL), has been used as a straightforward noninvasive criterion to access developmentally competent oocytes. However, there are discrepancies among different studies in this regard. Early studies indicated that the presence of a dominant follicle (>10 mm) in one or both ovaries had a negative effect on in vitro developmental competence of oocytes derived from the subordinate follicles [18–20]. Manjunatha et al. [21] reported that embryonic development was higher in oocytes coming from ovaries with a CL and no dominant follicle, whereas gametes coming from ovaries that had a CL and a dominant follicle showed higher competence only when oocytes were derived from the dominant follicle. In agreement with this notion, Pirestani et al. [22] reported that oocytes derived from ovaries containing a large follicle (~20 mm) were less competent compared to those derived from ovaries containing a CL. Similarly, Penitente-Filho et al. [23] classified cumulus–oocyte complexes (COCs) under the stereomicroscope and indicated that ovaries with CL yielded a larger number of competent oocytes than ovaries without CL. However, the oocytes used in the latter study were not subjected to IVP to confirm their developmental competence. Overall, these studies indicate that the presence of a dominant follicle in the bovine ovary would negatively influence the subsequent embryo development, while the presence of a CL favors oocyte competence. In contrast, more recent studies indicated that the presence of a CL has negative effects on the developmental competence of ipsilateral oocytes [24,25]. However, this "negative" effect does not influence the competence of oocytes originated from large follicles (10–20 mm) as much as those derived from small and medium follicles (<9 mm) [25].

Ovaries without structures indicative of estrus cyclicity have less competent oocytes than others [21,26], as indicated by the presence of fewer than 10 follicles 2–5 mm in diameter and no large follicles [27]. In addition, other authors indicated that the developmental competence of bovine oocytes from antral follicles (2 to 8 mm) is not affected by either the presence of a dominant follicle or the phase of folliculogenesis [27–31]. Thus, despite the few discrepancies, it seems that the selection of ovaries based on the presence of cycle-related structures could help optimize access to oocytes with better developmental competence for in-vitro-derived technologies. Nevertheless, the positive or negative effects of ovarian structures on oocyte competence require further investigation to determine more precisely how these ovarian structures impact subsequent in vitro embryonic development.

#### *2.2. Follicle Size*

One of the most used criteria to obtain competent oocytes is the size of the follicle. Research over the past decades indicates that bovine oocytes gain competence at late stages of the follicular phase, when signs of atresia are observed for the first time, such as a slight expansion in the outer cumulus layers and some cytoplasmic granulations [7,32]. Therefore, the recommendation is that oocytes recovered from follicles between 6 and 10 mm develop more frequently to more advanced embryonic stages [7,33–36]. Although the acquisition of competence begins when the follicle reaches 3 mm and the effect of size becomes more important at 8 mm [19,37,38], success is not guaranteed even if the oocytes come from larger follicles [39].

The acquisition of oocyte competence seems to be due to the substrate support received and to the developmental phase at the time of removal from the follicle [7,32,34]. Recent reports indicate that the follicular fluid (FF) microenvironment of large follicles has higher levels of electrolytes, glucose, reactive oxygen species, glutathione, superoxide dismutase activity, lipids, cholesterol, pyruvate, and estradiol [33,40,41]. Moreover, oocytes derived from larger follicles also show a different transcriptional pattern for chromatin remodeling and metabolic pathways, such as lipid metabolism, cellular stress, and cell signaling, with respect to those coming from smaller sizes, which would favor their developmental

potential [41,42]. Therefore, these findings indicate that large follicles (>6 mm) provide an appropriate microenvironment for the oocyte leading to better embryonic development.

#### *2.3. Morphology of the Cumulus–Oocyte Complexes*

The quality of COCs can be influenced by multiple factors, both intrinsic and extrinsic. Intrinsic factors include breed, age, reproductive status, metabolic and nutritional status, hormonal levels, and stage of the estrous cycle [43], whereas key extrinsic factors include the timing between slaughter and oocyte withdrawal from the ovary, morphology and methods of collecting the COCs, storage temperature of the ovaries, collection media, and micromanipulation skills of the operator [44].

Since intrinsic factors are more difficult to control when using slaughterhouse ovaries from cows of unknown origin, the morphology of the COC is relatively easy to evaluate and is often the most common criterion used to select and classify a standard collection of bovine oocytes [45–47]. Morphological criteria include the number and appearance of cumulus layers and the cytoplasmic features of the oocyte, such as the texture or brightness of its cytoplasm. Basically, the healthiest COC quality (Class I) relates to a complete cumulus cover with several compact cell layers; medium quality (Class II) has only partial cumulus cover and/or slightly expanded cumulus containing fewer than five cell layers; lastly, the worst quality (Class III) has a darker cytoplasm and the presence of dark spots with expanded cumulus, all indicative of follicular atresia (Figure 1). However, such classification criteria vary among laboratories.

**Figure 1.** Representative images of bovine cumulus–oocyte complexes (COCs) after ovary withdrawal classified according to COC morphology. (**A**,**B**) Complete cumulus cover with several compacts (red arrows) and slightly loose (black arrows) cell layers; (**C**) partial cumulus cover and loose cell layers with signs of early atresia (red arrow); (**D**) COC showing clear signs of atresia (red arrow) and a black-punctate cytoplasm (black arrow).

The study by Wit et al. [30] classified COCs into three groups: (i) compact and bright, (ii) less compact and dark, and (iii) strongly expanded cumulus with dark spots, where developmental capacity, measured by in vitro embryo production, was correlated with COC appearance. Moreover, less compact and darker COCs showed faster meiotic resumption. Another study using similar categories reported that COCs with darker cumulus and ooplasm were the most competent in terms of cleavage and blastocyst yield after IVF and parthenogenetic activation [48]. In addition, this study showed that developmental competence was related to calcium currents in the plasma membrane and calcium stores in the cytoplasm of immature oocytes [48]. The report by Bilodeau-Goeseels et al. [49] divided COCs into six classes on the basis of their cumulus and ooplasm features. These authors found that, although oocytes with fewer than five layers of cumulus cells (CC) showed lower cleavage rates, their developmental potential to the blastocyst stage was similar to oocytes with more than five layers of CC. More recently, De Bem et al. [37] found that class III COCs, considered to be of poor morphological quality, were superior in terms of blastocyst development to the intermediate class II group, but similar to class I COCs, albeit without differences in blastocyst quality. Emanuelli et al. [50] indicated that COCs with partial (fewer than five cell layers) and expanded cumulus had higher levels of DNA fragmentation after in vitro maturation (IVM) and lower competence compared to healthier ones, in accordance with the report by Yuan et al. [51]. However, blastocysts derived from COCs with varied morphologies exhibited no variations in terms of quality assessed by the number of cells. In addition, Emanuelli et al. [50] further concluded that these differences were due to better nuclear maturation through enhanced maintenance of metaphase II (MII) block by COCs showing full cumulus coverage.

Thus, despite these contradictory results, most studies agree that COCs showing signs of early atresia yield high blastocyst rates compared to morphologically healthy COCs. Nonetheless, advanced atresia, with signs such as cytoplasmic granulations, fewer than five cumulus layers, and expanded cumulus with dark cellular masses or, strictly, its complete absence, show lower in vitro potential as measured by cleavage rates and blastocyst formation [30] (Figure 1C). Additionally, although morphological classification seems to influence the proportion of blastocysts formed, such criteria may not influence their quality. Therefore, when selecting COCs according to their cumulus investment and ooplasm texture, the ideal would be to target COCs with several cumulus cell layers (more than or at least five layers), compact and/or slightly expanded, with or without dark areas in the oocyte and cumulus.

#### *2.4. Lipid Content*

The morphological appearance of the ooplasm commonly assessed to select the oocytes [52,53] is influenced by lipid content in livestock species, such as cattle, pigs, and horses [54–56]. Lipids, in the form of lipid droplets (LDs), are signaling molecules with important roles in oocyte maturation and competence acquisition [57]. In the late stage of oocyte maturation and during preimplantation development, endogenous oocyte lipids work as an energy source [58,59] and as a lipid factory for energy reserve [60]. Failure to use lipids in oocytes has been shown to be related to inadequate nuclear maturation [61,62]. The number of LDs present in the cytoplasm increases as the oocyte grows [63] and, although the ooplasm organization does not undergo major changes during in vitro maturation to MII [56], the type and number of lipids in the LDs seem to be more dynamic and to undergo changes during meiotic progression to MII [59,64].

LDs aggregate in the form of dark clusters that can be seen in the ooplasm as a cytoplasmic darkness [55,65] (Figure 2). Cytoplasmic darkness can be homogeneous, affecting the entire cytoplasm or concentrated in the center, with a clear peripheral ring that gives the cytoplasm a darkened appearance (Figure 2B,D). This opaque appearance is more intense in pigs and domestic cats, followed by cows and finally sheep and goats, whose ooplasm is lighter. In the case of horses, lipid polarization is commonly observed, which facilitates the visualization of the spermatozoon within the oocyte [55,66].

**Figure 2.** Denuded MII bovine oocytes after 24 h of IVM. (**A**,**C**) oocytes showing a homogeneous dark cytoplasm. Black arrows depict the first polar body; (**B**,**D**) oocytes showing a heterogeneous pale and punctuated cytoplasm. Black arrows indicate the first polar body, while red arrows depict dark areas of intense lipid accumulation (cytoplasmic granulations).

Several studies investigated the relationship between oocyte lipid content and competence. For instance, cytoplasm color can be used as a marker of lipid content and as predictive of the embryonic potential [67], as oocytes with a uniform and brown or dark cytoplasm contain more intracellular lipids than oocytes with a granular or pale cytoplasm [65]. Most studies demonstrated that oocytes with rough granulations or very pale ooplasm yield a lower preimplantation development [49,53,67]. Jeong et al. [68] classified the ooplasm in three categories: dark, brown, and pale. In this study, the content of mitochondria and the proportion of oocytes that reached the blastocyst stage were higher in darker oocytes. Moreover, Nagano et al. [67] reported that sperm penetration, monospermic fertilization, cleavage, and blastocyst rates were higher in oocytes with a brown ooplasm compared to those with pale or very dark ones. Moreover, brown oocytes with a dark edge or with dark spots showed, under electron microscopy, an organelle arrangement similar to in vivo matured oocytes, and pale or black oocytes appeared to be degenerating and/or aging [67]. The authors concluded that a dark ooplasm is associated with a lipid accumulation and better developmental competence, while a pale ooplasm would indicate fewer organelles and poor developmental potential [69]. Interestingly, a study by Prates et al. [70] distinguished fat areas of different color shades using the Nomarski interference differential contrast (NIC) as the fat gray value of porcine oocytes, reflecting alterations in lipid content, and proposed this tool as an appropriate and noninvasive technique to evaluate the lipid content of a single oocyte before or after in vitro maturation. Recently, the study of Jasensky et al. [71] reported the use of anti-Stokes Raman scattering (CARS) microscopy as a new non-invasive tool for the quantification of lipid content in mammalian oocytes. This study showed that the ~2 min of laser exposure was enough for a quantitative comparison of lipid content in mice oocytes at different developmental stages, as well as in oocytes of others mammalian species, and, more importantly, without detrimental effects (without the need to attach fluorescence labels) for subsequent preimplantation development. Thus,

its application in live-cell imaging of oocytes is promising to provide alternative and/or additional information in order to improve the accuracy of subjective morphometric measurements.

Taken together, as stated by the review of Nagano and colleagues [69], a dark ooplasm indicates an accumulation of lipids and good developmental potential, a light-colored ooplasm indicates a deficiency of lipid stores and poor developmental potential, and a black ooplasm indicates aging and low developmental potential (Figure 2). Finally, the use of NIC and CARS should be further investigated as a potential noninvasive tool to evaluate the lipid content of single oocytes in livestock species.

#### *2.5. Cumulus Expansion and Oocyte Size*

Another parameter that is often used as an indirect indicator of oocyte quality is the degree of cumulus expansion following maturation, typically after 20 to 24 h of culture in an in vitro maturation environment. Grades 1 to 3 (sometimes 4) are attributed to increasing degrees of expansion (1: modest expansion, characterized by few morphologic changes compared to before maturation, 2: partial expansion, and 3: complete or almost complete expansion) [72–74].

Although the expansion of CCs has been described as the basis for oocyte maturation [75] and early reports supported the idea that quantity and quality of the expanded cumulus mass were correlated with developmental capacity [76], its usefulness as an indicator of developmental potential in bovine seems to be modest [77]. For instance, studies by Anchordoquy et al. [78], Dovolou et al. [79], and Rosa et al. [80] reported that, under different experimental conditions, the cumulus expansion index was not indicative of blastocyst yield or quality. Similarly, another study indicated that inhibition of cumulus expansion by enzymatic hyaluronidase degradation did not affect cleavage or blastocyst development [81]. Nonetheless, as shown by Fukui et al. [82], more than an indicator of developmental competence, CCs and their expansion play an important role in fertilization by inducing the acrosome reaction and, therefore, promoting higher fertilization rates.

In addition to follicle size, oocyte size has been used as a noninvasive quality parameter. Although it is difficult to measure the precise diameter of the oocyte during IVF, oocyte selection based on diameter can be used as a routine step during micromanipulation protocols. The study of Fair et al. [83] classified oocytes recovered from slaughterhouse ovaries into four groups (<100 microns, 100 to 110 microns, 110 to 120 microns, and >120 microns). Rates of resumption of meiosis to MII were higher for oocytes >110 microns. Moreover, oocytes <110 microns were transcriptionally active, suggesting that they were still in the growth phase of oogenesis [83,84]. Similarly, Anguita et al. [85] reported that cleavage and blastocyst rates were higher in oocytes >110 microns. Moreover, Otoi et al. [86] and Arlotto et al. [29] found that oocytes >115 microns had better rates of nuclear maturation and a lower incidence of polyspermy after IVF, but cleavage rates and development to the blastocyst stage were optimal in oocytes >120 microns. Huang et al. [87] and Yang et al. [88] compared oocytes collected from initial antral follicles (0.5–1 mm in diameter) cultured in vitro for 14–16 days with oocytes collected from antral follicles (2–8 mm in diameter), cultured, and submitted to IVM. The authors reported better maturation rate for oocytes >115 microns, optimal for oocytes >120 microns, but developmental competence was only high for oocytes collected from antral follicles and of size >120 microns.

These results suggest that bovine oocytes acquire meiotic competence with a diameter of 115 microns, but full developmental competence is acquired around 120 microns, possibly because smaller oocytes have not yet completed their growth phase [46]. Thus, the selection of follicles between 6 and 10 mm, with oocyte diameters >115 and <130 microns, has the potential to optimize developmental outcomes.

#### *2.6. First Polar Body Assessment*

At the end of IVM and after the removal of CCs, it is easy to perform a detailed observation of morphological features [13], including the assessment of oocyte shape, cytoplasm color and granulation, regularity and thickness of the zona pellucida, size of the perivitelline space, presence of vacuoles, and presence or absence of the first polar body (PB1) and its morphology. Extrusion of PB1 in mammalian

oocytes is a cellular landmark of meiotic maturation, and its assessment is frequently used as an indicator of nuclear maturation [89]. Thus, its absence indicates that the oocyte is immature or that it has degraded due to aging; however, its presence does not guarantee that the oocytes have completed their maturation process, and some of them remain incompetent despite exhibiting morphologic features of nuclear maturation [90].

In bovine species, extrusion of PB1 begins at 16–18 h after IVM [91–94]. Nonetheless, oocytes acquire the highest developmental competence at around 5–10 h after PB1 extrusion [14,95]. Dominko and First [95] indicated that oocytes that extruded their PB1 after 16 h of IVM were only capable of reaching higher developmental competence after 24 h of in vitro culture. Thus, cytoplasmic maturation in cattle occurs several hours after nuclear maturation, probably between 24 and 30 h after the beginning of IVM.

Unfortunately, there are no studies that analyzed the influence of the first PB morphology on oocyte competence in cattle. However, one study using porcine oocytes indicated that PB1 with a smooth or intact surface was indicative of a more advanced cytoplasmic maturation and better embryonic development in vitro than those with a fragmented or rough surface [96]. Despite lacking studies in domestic species, studies in humans investigated the association between PB1 morphology and oocyte competence [97,98]. Ebner et al. [99] conducted a retrospective study using 70 consecutive ICSI cases in which oocyte classification based on PB1 morphology revealed that oocytes with intact, well-shaped PB1 yield better fertilization and high embryonic quality. Later, Ebner et al. [97] confirmed the relationship among PB1 morphology, fertilization, and blastocyst quality, as well as a positive effect on implantation and pregnancy rates. Similarly, Rose et al. [100] reported that oocytes with an intact PB1 show better fertilization and embryonic development, whereas those displaying a PB1 with morphological abnormalities such as a larger size, irregularities, coarse surface, or fragmentation are less competent during an IVF protocol, having poor implantation capabilities after embryo transfer. In contrast, others did not report any correlation [101–103]. Thus, there is a lack of consensus on the impact of PB1 morphology on oocyte competence and embryonic development in humans. It is also important to note that some PB1 abnormalities may be an artefact of oocyte manipulation (mainly during the denudation process) or aging [104].

In summary, although the selection of oocytes with PB1 of a homogeneous, round shape with a smooth or intact surface may be indicative of a better oocyte, the usefulness of this selection criterion in livestock requires further research to establish its real predictive value for oocyte competence.

#### *2.7. Polarized Light Microscopy*

Polarized light microscopy (PLM) has been used in different mammalian oocytes since it allows the noninvasive assessment of subcellular features such as the meiotic spindle and zona pellucida birefringence (ZPB). To learn about the principles and equipment required for PLM in detail, readers are directed to excellent reviews on the subject [105,106].

#### 2.7.1. Evaluation of the Meiotic Spindle and Zona Pellucida Birefringence

Using PLM, it is possible to locate and evaluate the morphology of the meiotic spindle to confirm egg maturation, which has been positively correlated with developmental competence [90,107–109]. This method avoids damaging the spindle during the ICSI procedure, considering that the position of the PB1 can be altered when CCs are removed during preparation for ICSI [110]. Furthermore, PLM has been successfully used to remove the meiotic spindle and chromosomes (enucleation) in mice [111], bovines [112], and pigs [113], with an average efficiency of 90% and, more importantly, avoiding the exposure to ultraviolet (UV) rays and their detrimental effect on embryonic development.

In livestock species, the dark appearance of the ooplasm, attributed to high lipid contents, is known to interfere with spindle imaging [113] and, as in humans, precludes the detection of meiotic spindle abnormalities [102,113,114]. Therefore, spindle birefringence should be carefully considered as an index of gamete quality and chromosome alignment in some species. In pigs, a negative PLM signal was associated with to reduced maturation and poor development potential [113]. In the same study, when the PLM system was used for spindle removal, the overall enucleation efficiency was 92.6%, indicating that PLM is an effective tool for performing enucleation in pigs. A few years later, the same group evaluated the use of PLM to assess the meiotic spindle of in vitro matured bovine oocytes after vitrification and warming [115]. They were able to confirm the presence of the meiotic spindle in 99% of the analyzed eggs. Moreover, after vitrification and warming, meiotic spindles were detected in 79% of oocytes. Interestingly, thawed oocytes that displayed a positive PLM signal showed better competence in terms of cleavage and blastocyst rates after parthenogenetic activation, indicating that PLM can be a useful tool for assessing post-warming viability in vitrified bovine oocytes.

Overall, these studies demonstrate that PLM efficiently detects the meiotic spindle of livestock oocytes and does not affect early embryonic development. However, the selection of cattle oocytes on the basis of the presence of a PLM signal does not seem to offer improvement in IVP outcomes yet.

#### 2.7.2. Assessment of the Zona Pellucida Birefringence

In addition, PLM has been used for the evaluation of the ZPB, which in humans has been associated with oocyte quality [116–118], although this is still under debate [119,120]. The few studies in cattle showed that a lower ZPB is related to high-quality oocytes and improved blastocyst development [121,122], whereas two studies in horses reported conflicting results, indicating beneficial effects of both low ZPB [123] and high ZPB [124]. Because most of the studies with PLM were carried out in mice and humans with conflicting results, its potential application and practical use in cattle and other livestock species needs further assessment. Contrary to humans, where the number of highly valuable oocytes from donors is relatively low, livestock oocytes obtained from slaughterhouse ovaries allow a more stringent selection. Furthermore, assessment of the meiotic spindle can be a laborious procedure, which delays the overall process of in vitro manipulation and embryo production. Thus, its application will require showing a clear advantage over conventional approaches using the morphological criterion mentioned above for oocyte selection. However, PLM might be beneficial when individual oocytes are of high value, such as oocytes recovered from elite cows by ovum pick-up (OPU) [111,113].

#### *2.8. Brilliant Cresyl Blue (BCB) Staining*

Another approach that demonstrated predictive potential is the evaluation of glucose-6-phosphate dehydrogenase (G6PDH) activity via brilliant cresyl blue (BCB) staining. BCB is a dye that determines the intracellular activity of G6PDH. Activity of G6PDH is observed during the oocyte growth phase (BCB−: colorless cytoplasm, increased G6PDH) due to the demand of ribose-6-phosphate for nucleotide synthesis. This activity is low (BCB+: colored cytoplasm, low G6PDH) in oocytes that have completed their growth phase [125]. This technique has been successfully employed in various species, including cattle [125–127].

Although previous reports found that the developmental competence of oocytes with low G6PDH activity (BCB+) was higher than that of oocytes with a high G6PDH activity (BCB−), the absence of differences in terms of embryonic development between BCB<sup>+</sup> and the untreated control group decreases the utility of the BCB test in IVP technology [128]. However, it is unquestionable that BCB<sup>+</sup> oocytes have statistically higher developmental competence than BCB− oocytes, both in IVF and somatic cell nuclear transfer (SCNT) [128].

Later studies continued to show only a trend of BCB<sup>+</sup> oocytes toward greater developmental potential. Better blastocyst rates at day 7 were reported by Silva et al. [129], and a study by Fakruzzaman et al. [130] reported higher blastocyst quality on the basis of total apoptotic cells and mitochondria numbers. Similarly, Castaneda et al. [131] indicated that the higher lipid content of BCB<sup>+</sup> bovine oocytes might be associated with their better developmental competence. Interestingly, another article indicated that co-culture with BCB− oocytes during IVM affects negatively the capacity of BCB<sup>+</sup> oocytes to undergo embryonic development [132]. However, other authors suggested that

the BCB test is not sufficient for identification of the most competent gametes [133]. Nonetheless, the combination of oocyte and zygote selection using BCB staining would improve the efficiency of embryo selection [134]. Therefore, the BCB test can be a valuable tool when used together with classical morphological classification and could be useful for the selection of oocytes with a higher implantation potential. Nonetheless, an assessment of the effects of BCB staining on post-implantation development is necessary to elucidate its usefulness for IVP technologies, not only for research but also in the industry of animal production. A summary of the morphological and visual indicators associated with oocyte competence is shown in Table 1.


**Table 1.** Summary of the morphological and visual indicators of oocyte competence.

#### **3. Non-invasive Molecular Approaches**

Many studies are being performed in mammals in order to find molecular markers predictive of oocyte quality. So far, most of the data show considerable variations, perhaps due to different experimental conditions and/or the criterion of quality/competence, resulting in varied scientific views.

#### *3.1. Cell Death (Apoptosis) in Cumulus Cells*

Because morphological evaluation prior to maturation does not allow to discriminate the atretic oocytes from healthier ones [135], one of the earlier noninvasive markers of oocyte competence was the level of apoptosis in CC, seen as DNA fragmentation, externalization of phosphatidylserine (EP), and/or the expression ratio of anti-apoptotic (Bcl-2) and pro-apoptotic (Bax) genes (BCL-2/BAX). Early studies found that the CC of bovine COCs undergo progressive apoptosis during IVM [136], and this was negatively correlated with the oocyte developmental capacity [51]. However, results reported by Janowski et al. [137] supported the notion that follicular cells surrounding the more competent oocytes have a higher degree of apoptosis. Later, Warzych et al. [138] showed that the level of apoptosis in CC was not associated with morphology or the oocyte meiotic stage, suggesting that the extent of apoptosis

in CC is not a reliable quality marker for gamete competence. Similarly, the study of Anguita et al. [135] showed that embryonic developmental potential increased together with oocyte diameter, but this developmental competence was not related to the incidence of apoptosis. Recently, another study indicated that optimum control of the meiosis block, nuclear maturation, and developmental potential were associated with less DNA fragmentation in CC [50].

Similarly, in the human model, the majority of related studies have focused on granulosa cells (GC) isolated from FF during oocyte collection. Apoptosis, evaluated by EP, of GC was negatively associated with egg and embryo numbers in IVF/ICSI cycles, pregnancy rate, and live birth rate after IVF [139,140]. However, contrarily, it was also reported that the EP in GC is not related to follicular quality and oocyte competence during ICSI [141]. Thus, in the bovine and human models, it is still controversial whether apoptosis of GC and/or CC can impact the developmental potential of the oocyte.

#### *3.2. Transcriptomic and Proteomic of Cumulus Cells*

Many new genomic tools helped to deepen the understanding in the area of oocyte–cumulus communication, as well as molecular pathways required for the acquisition of competence in mammalian gametes and embryos. For instance, recent advances in RNA-Seq technology offer a global transcriptomic approach for identifying differentially expressed genes associated with competence and embryonic development.

Among the molecular approaches, study of the transcriptomic profile of the surrounding cumulus is one of the most popular attempts at finding molecular markers associated with gamete competence in mammals. The "noninvasive" strategy is based on profiling the gene expression of a small biopsy before IVM, maintaining COC integrity, and following the embryonic development of the respective oocyte. This is also called "oocyte fate" [142]. Although several studies in cattle already found several genes in CCs from germinal vesicle (GV) [16,35,143–151] and MII oocytes [144,152] to be associated with oocyte competence, only a few reports matched the oocyte fate with the transcriptomic profile obtained from the CCs or granulosa cells (Table 2). There is some consensus regarding pathways correlated positively with oocyte competence, including the cell cycle (CCND1, CCNB2, and CCNA2 genes) [143,145,153], cell growth and proliferation, (CD44, TGFB1, EGF, FGF11, PRL, and GH genes) [35,147–149,154], and steroidogenesis (HSD3B2 and CYP11A1 genes) [16,154]. On the contrary, genes related to cellular apoptosis would be associated with a low competence (ATRX, KRT8, ANGPT2, KCNJ8, and ANKRD1 genes) [142,147,152,155].


**2.**SummaryofstudiesperformingtranscriptomicandproteomicanalysisofCCand/or


**Table 2.** *Cont.* laser desorption/ionization-time of flight, ECM: extracellular matrix, MAPK: mitogen-activated protein kinases, PI3K: phosphatidylinositol 3-kinase, IVM: in vitro *maturation*. \* N.A not available.

=

#### *Animals* **2020**, *10*, 2196

On the other hand, studies analyzing the proteomic profile of the cumulus–oocyte complex (COC) are scarce. Moreover, most of them have done invasive analysis in a pool of oocytes; thus, oocyte fate could not be followed (Table 2). Nonetheless, the few studies described many proteins involved in cell signaling that may have a role in cumulus–oocyte communication and competence. Most of the proteins are involved in components of integrin, actin cytoskeleton, mitogen-activated protein kinases (MAPK) and phosphatidylinositol 3-kinase (PI3K) signaling pathways, extracellular matrix (ECM) receptor interactions, steroid biosynthesis, and glucose and carbohydrate metabolism, which may have implications in various reproductive processes such as oocyte development and maturation [156–158] (Table 2). A recent study reported a highly sensitive approach to characterize the CC proteome from a single COC after in vivo or in vitro maturation [156]. This method shows the potential to directly connect the cumulus proteome to the developmental potential of the corresponding oocyte, as already performed at the gene expression level.

#### *3.3. Follicular Fluid Analysis*

It is well known that the composition of FF has an impact on the developmental capacity of the oocyte and, thus, the resulting embryo. Excellent articles reviewed the importance of FF on oocyte physiology and fertility [159–161]. This fluid contains proteins, cytokines, growth factors, steroids, metabolites, and other indeterminate factors [159]. Therefore, by studying its composition, it should be possible to predict oocyte competence and fertilization outcomes [162–164]. Metabolites in the FF, such as glucose and potassium, have already been positively associated with oocyte quality in cattle [41,165]. However, studies linking the FF features with the respective oocyte fate in bovines have not been performed yet. Reports in humans have positively associated the presence of anti-Müllerian hormone (AMH) in FF with competence of the respective oocyte [166,167], although with some contradictory results [168,169]. Conversely, a recent study that used a large population of transferred embryos matching FF samples indicated that the AMH level in FF following withdrawal from the ovarian follicle is closely linked to the oocyte's competence, and it is a suitable predictor of a live birth after single embryo transfer [170]. In the cow, it was already reported that AMH concentrations can be predictive of the number of ovulations and embryos produced in response to ovarian stimulation by FSH [171–173], making it a suitable molecule to be related to the oocyte competence.

In addition, other molecules in FF of cattle that show promising results are microRNAs (miRNAs). The bovine FF contains free miRNAs, as well as some associated with exosomes [174,175]. Recently, the study of Pasquariello et al. [176] showed, for the first time, the miRNA content of different populations of oocytes categorized according to their competence. Interestingly, they discovered that the most differentially expressed miRNAs (miR-24, miR-10a, and miR-320a) in FF found in highly competent follicles were part of the regulation of the neurotrophin signaling pathway, which supports follicle formation and development, as well as the TGF-βsignaling pathway that controls the production of ovarian peptide hormones. Therefore, linking FF molecules such as AMH or miRNAs with gamete competence is an encouraging strategy in the field of oocyte selection. However, we have to consider that it will be applicable only when the fast collection and analysis of FF from individual follicles become practicable.

#### **4. Conclusions and Future Perspectives**

The classification and selection of oocytes in livestock species for in vitro embryo production and for micromanipulation techniques, such as ICSI and SCNT, can be one of the most important steps to reach superior embryonic development and quality. Although more sophisticated methods (qRT-PCR, global transcriptomic, and proteomic analysis) have been studied since a few decades ago, the lack of a quick enough method producing reliable results hinders the implementation of these technologies. Moreover, molecular analysis requires high-tech equipment and technical staff that would be cost-ineffective in most research laboratories. Thus, although oocyte selection based on morphologic criteria appears to be insufficient to distinguish more competent gametes, in real practice, when 100–300 oocytes are waiting to be processed during micromanipulation experiments, it seems to be the only available strategy so far. Furthermore, studies that perform embryo transfers are also important to effectively evaluate developmental potential, as successful embryo implantation is highly dependent on the quality of the embryo and the intricate relationship it establishes with the uterine endometrium. Ultimately, with the advent of bovine embryonic stem cells, greater scrutiny of oocytes with high developmental potential is necessary, for the production of stable pluripotent stem cell lines to be used in basic science, forward and reverse genetics, epigenetics, gene imprinting, and the production of animal models with applications in animal production. Thus, in addition to improving the conditions to support in vitro maturation, the implementation of new tools for the assessment of gamete competence, together with studies decoding molecular cues in oocyte maturation, will improve our understanding of this complex process and will more precisely identify the synchrony between nuclear and cytoplasmic maturation in livestock species.

**Author Contributions:** All authors contributed equally to reviewing the literature, as well as writing and editing the review. All authors read and agreed to the published version of the manuscript.

**Funding:** This article received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Cellular and Molecular Events that Occur in the Oocyte during Prolonged Ovarian Storage in Sheep**

#### **Alicia Martín-Maestro, Irene Sánchez-Ajofrín \*, Carolina Maside, Patricia Peris-Frau, Daniela-Alejandra Medina-Chávez, Beatriz Cardoso, José Carlos Navarro, María Rocío Fernández-Santos, José Julián Garde and Ana Josefa Soler \***

SaBio IREC (CSIC-UCLM-JCCM), ETSIAM, Campus Universitario, s/n, 02071 Albacete, Spain; alicia.martinmaestro@uclm.es (A.M.-M.); carolina.maside@uclm.es (C.M.); patricia.peris@uclm.es (P.P.-F.); daniela.medina@uclm.es (D.-A.M.-C.); beacardoso\_14@hotmail.com (B.C.); jnavarropedrosa@gmail.com (J.C.N.); mrocio.fernandez@uclm.es (M.R.F.-S.); julian.garde@uclm.es (J.J.G.)

**\*** Correspondence: irene.ssanchez@uclm.es (I.S.-A.); anajosefa.soler@uclm.es (A.J.S.)

Received: 1 November 2020; Accepted: 13 December 2020; Published: 17 December 2020

**Simple Summary:** Establishing efficient in vitro embryo production (IVP) protocols in sheep usually requires prolonged transportation of post-mortem ovaries since adult animals are often slaughtered in abattoirs far from laboratories. In this study, different analyses were carried out to investigate important cellular and molecular aspects of hypoxic injury on excised ovaries over time in order to understand the factors jeopardizing the development of competent oocytes during prolonged transport times. We observed that, when ovaries were stored for more than 7 h, the quality and developmental potential of oocytes and cumulus cells were greatly reduced. Moreover, the use of medium TCM199 over saline solution also had deleterious effects. Beyond transport time, strategies aimed at reducing these damages may improve oocyte quality and developmental competence.

**Abstract:** For the past two decades, there has been a growing interest in the application of in vitro embryo production (IVP) in small ruminants such as sheep. To improve efficiency, a large number abattoir-derived ovaries must be used, and long distances from the laboratory are usually inevitable when adult animals are used. In that scenario, prolonged sheep ovary transportation may negatively affect oocyte developmental competence. Here, we evaluated the effect of ovary storage time (3, 5, 7, 9, 11 and 13 h) and the medium in which they were transported (TCM199 and saline solution) on oocyte quality. Thus, live/dead status, early apoptosis, DNA fragmentation, reduced glutathione (GSH) and reactive oxygen species (ROS) content, caspase-3 activity, mitochondrial membrane potential and distribution, and relative abundance of mRNA transcript levels were assessed in oocytes. Afterin vitro maturation (IVM), cumulus cell viability and quality, meiotic and fertilization competence, embryo rates and blastocyst quality were also evaluated. The results revealed that, after 7 h of storage, oocyte quality and developmental potential were significantly impaired since higher rates of dead oocytes and DNA fragmentation and lower rates of viable, matured and fertilized oocytes were observed. The percentage of cleavage, blastocyst rates and cumulus cell parameters (viability, active mitochondria and GSH/ROS ratio) were also decreased. Moreover, the preservation of ovaries in medium TCM199 had a detrimental effect on cumulus cells and oocyte competence. In conclusion, ovary transport times up to 5 h in saline solution are the most adequate storage conditions to maintain oocyte quality as well as developmental capacity in sheep. A strategy to rescue the poor developmental potential of stored oocytes will be necessary for successful production of high-quality embryos when longer ovarian preservation times are necessary.

**Keywords:** sheep; ovary storage; transport; oocyte; in vitro embryo production

#### **1. Introduction**

Assisted reproduction technologies (ARTs) in small ruminants, such as sheep, have great potential for genetic improvement and dissemination programs, since they allow for a rapid and sustainable increase in animals of great genetic merit. Furthermore, ARTs are effective tools in the preservation of endangered species or breeds as well as in disease eradication programs [1].

Though the main lines of investigation in small ruminants have focused on germplasm banks and artificial insemination [2], over the past few decades, slight advances have been made toward the use of in vitro embryo production (IVP) [1–3]. Generally, improving the efficiency of IVP protocols in these species entails the use of ovaries of dead animals because a large number of samples should be collected. Unlike their in vivo counterparts, oocytes retrieved from dead animals exhibit reduced developmental potential [4]. Moreover, the slaughterhouses where adult animals are slaughtered are usually located in strategic places and often far from research laboratories. Storing ovaries for long periods of time due to long distances could compromise the viability of the oocyte. Considering that the quality of oocytes determines the developmental potential of embryos after fertilization [5], the preservation of oocyte integrity from the moment the animal dies until the ovaries are processed is of critical importance.

Immediately following death, the lack of blood flow prevents oxygen and energy supply and places the ovaries under ischemic conditions [6]. The main mechanism of injury in ischemia is hypoxia, and cells with high metabolic rates, including the ones that form the ovarian tissue, tend to be damaged very rapidly [7,8]. Acute hypoxia results in ATP depletion that triggers a switch to glycolysis, the major anaerobic pathway for ATP production [9]. ATP is broken down without being resynthesized, and eventually, decreased energy efficiency and accumulation of lactic acid produced by glycolysis also reduce intracellular pH, resulting in additional cellular dysfunction [6,9]. Notwithstanding the relationship between hypoxia/ischemia and organ damage is being established, the physiological mechanisms by which oocyte quality is affected with increased ovary storage time remain to be completely understood. It is therefore necessary to elucidate the events occurring in the oocyte at the cellular and molecular levels for the purpose of taking the appropriate measures to reduce this damage.

The goal of this study, therefore, was to evaluate the cellular and molecular events related to oocyte quality and developmental competence that occur throughout storage of sheep oocytes within ischemic ovaries after the death of the animal. This may help to create a better understanding of the mechanism of oocyte injury obtained from ischemic ovaries and to identify the temporal window for successful fertilization in IVP, particularly when ovary transport times are inevitably long. Ultimately, it will contribute to developing a strategy to reverse the poor developmental potential of stored oocytes within ovaries, which will have a great significance for ARTs.

#### **2. Materials and Methods**

The adult sheep ovaries were collected from an authorized slaughterhouse ("Ovinos Manchegos"), and sperm samples were obtained from the Germplasm Bank of the "Reproduction Biology Group", which is officially authorized for collecting and storing semen from sheep (ES07RS02OC). All chemicals were acquired from Merck Life Sciences (Madrid, Spain) unless otherwise stated.

#### *2.1. Oocyte Collection and In Vitro Maturation*

Adult sheep ovaries (*n*=1420) were obtained post-mortem and transported at 30 ◦C in physiological saline (8.9 gr/L NaCl) supplemented with penicillin (0.1 g/L) or at 38.5 ◦C in TCM199 medium supplemented with polyvinylpyrrolidone (PVA; 1 g/L), 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) (6.51 g/L), streptomycin (0.1 g/L), penicillin (0.1 g/L) and sodium bicarbonate (0.4 g/L) and were maintained for 13 h in the same media and at the same temperature. The mean age of animals was around 6 years old, and the sheep breeds were mainly Merino or mixed. Immature cumulus–oocyte complexes (COCs) were recovered by slicing the ovaries with a scalpel at 3, 5, 7, 9, 11 and 13 h post ovary collection. Then, a total of 4258 COCs from 8 replicates having a clear and homogeneous or moderate granular ooplasm and surrounded by at least three layers of tightly packed cumulus cells were selected and placed in TCM199 medium supplemented with HEPES (2.38 mg/mL), heparin (2 μL/mL) and gentamycin (4 μL/mL). In each replicate, the COCs of ovaries from the same treatments were mixed and homogeneously distributed. From those, 2327 COCs were mechanically denuded by vortex in phosphate-buffered saline (PBS) supplemented with 0.1% PVA (*w*/*v*; PBS-PVA) and oocytes were either directly analysed, fixed in 0.5% glutaraldehyde (*v*/*v*) and stored at 4 ◦C for terminal deoxynucleotidyl transferase mediated dUTP nick-end labelling (TUNEL) analysis or snap-frozen and stored at −80 ◦C for mRNA analysis. In addition, 1931 COCs, collected from the last 4 replicates, were matured, fertilized and cultured in vitro following the protocol by Sánchez-Ajofrín et al. [10]. Briefly, COCs were washed in TCM199-gentamycin (4 μL/mL) and randomly placed in four-well dishes containing 500 μL of TCM199 and 4 μL/mL gentamycin, 100 μM cysteamine, 10 ng/mL follicle stimulating hormone, 10 ng/mL luteinizing hormone and 10% fetal calf serum [11] under mineral oil (Nidacon, Gothenburg, Sweden) and an atmosphere of 5% CO2 at 38.5 ◦C with maximal humidity.

#### *2.2. In Vitro Fertilization (IVF)*

After approximately 22 h, COCs were partially denuded by gentle pipetting, divided into groups of 40–45 oocytes and placed in four-well plates containing 450 μL of synthetic oviductal fluid (SOF, Table S1), as described by Takahashi and First [12], with 10% oestrous sheep serum (ESS). Frozen-thawed spermatozoa were separated using a Percoll® density gradient (45%/90%) from two rams and capacitated for 15 min at 38.5 ◦C in 5% CO2 with SOF and 10% ESS. Spermatozoa were subsequently co-incubated with the oocytes at a final concentration of 106 spermatozoa/mL for 18 h at 38.5 ◦C in 5% CO2.

#### *2.3. In Vitro Culture (IVC)*

After 18 h post-insemination (hpi), presumptive zygotes were transferred to 25 μL IVC droplets (approximately one embryo per μL) containing SOF supplemented with 3 mg/mL of bovine serum albumin and cultured in a humidified atmosphere of 5% CO2, 5% O2 and 90% N2 in air until day 8 post-insemination (dpi). Cleavage rate and blastocyst yield were examined at 48 hpi and 6, 7 and 8 dpi, respectively. All expanded blastocysts were fixed in 0.5% glutaraldehyde (*v*/*v*) and stored for TUNEL analysis and cell-number evaluation.

#### *2.4. Early Apoptosis Assay*

To determine early apoptosis in oocytes, a total of 356 immature and denuded sheep oocytes were incubated in Annexin-V, fluorescein isothiocyanate (FITC) staining kit (Thermo Fisher Scientific, Barcelona, Spain) according to the manufacturer's instructions. Briefly, oocytes were stained for 15 min with Annexin-V/FITC and 100 μg/mL propidium iodide (PI) at 37 ◦C in the dark. After incubation, oocytes were washed thrice in PBS-PVA and mounted on slides. Samples were evaluated at ×20 augmentation by fluorescence microscopy (Eclipse 80i, Nikon Instruments Europe, Amsterdam, The Netherlands) with the Intensilight C-HGFI module. The filter for excitation and the emitted fluorescence were EX 450-490 nm (DM 505; BA 520). Oocytes were classified into the following groups: early apoptotic oocytes (Annexin-V positive signal and PI negative signal; Figure 1A-a), viable oocytes (Annexin-V and PI signals were both negative; Figure 1A-b) and dead oocytes (Annexin-V and PI positive signals; Figure 1A-c and negative Annexin-V signal in the membrane and positive PI signal; Figure 1A-d).

**Figure 1.** Effect of ovary storage time (3 to 13 h) on live/dead status and early apoptosis of immature sheep oocytes. (**A**) Representative images of sheep oocyte classification using Annexin-V staining: (a) viable oocyte, (b) early apoptotic oocyte and (c,d) dead oocytes. Scale bar = 50 μm. (**B**) Viability and early apoptosis rates (%): results are expressed as mean <sup>±</sup> SEM. a,b,c Different letters indicate differences (*p* ≤ 0.05) among storage times.

#### *2.5. Measurement of Glutathione (GSH) and Reactive Oxygen Species (ROS)*

A total of 350 immature oocytes were incubated in 50 μM Cell Tracker™ Blue (Thermo Fisher Scientific, Barcelona, Spain) and 10 μM of CM-H2DCFDA (Thermo Fisher Scientific, Barcelona, Spain) for 30 min at 37 ◦C in the dark to detect intracellular glutathione (GSH) and reactive oxygen species (ROS) levels, respectively. Oocytes were subsequently washed thrice in PBS-PVA and then placed on glass slides under cover slips. The fluorescence intensity was observed using ×20 augmentation by fluorescence microscopy (Eclipse 80i, Nikon Instruments Europe, Amsterdam, The Netherlands) and quantified using ImageJ 1.45s software (National Institutes of Health, Bethesda, MD, USA; Figure 2A-a GSH oocyte level and 2A-b ROS oocyte level).

**Figure 2.** Effect of ovary storage time (3 to 13 h) on intracellular glutathione (GSH) and reactive oxygen species (ROS) levels of immature sheep oocytes: (**A**) representative images of intracellular (**a**) GSH and (**b**) ROS oocyte levels, scale bar = 100 μm, and (**B**) fluorescence intensity of GSH and ROS levels. Results are expressed as mean <sup>±</sup> SEM. a,b Different letters indicate differences (*p* <sup>≤</sup> 0.05) among storage times.

#### *2.6. DNA Fragmentation Assay*

The TUNEL method was used to detect DNA fragmentation combined with PI staining (oocytes) or Hoechst 33342 staining (blastocyst). Fixed immature sheep oocytes (*n* = 370) and blastocysts (*n* = 120) were permeabilized in 0.5% Triton X-100 in PBS for 1 h at room temperature. Next, In Situ Cell Death Detection Kit (Merck Life Sciences, Madrid, Spain) was used for the detection of DNA strand breaks in oocytes and blastomeres. According to the manufacturer's instructions, samples were placed in 30 μL drops of TUNEL reagent with fluorescein isothiocyanate conjugated deoxyuridine 5-triphosphate (dUTP) and the enzyme terminal deoxy-nucleotidyl transferase and were incubated for 1 h at 37 ◦C. The positive control was incubated with DNAse (0.2 U/μL) at 37 ◦C in the dark for 1 h, while the negative control was incubated in the absence of enzyme terminal deoxynucleotidyl transferase. Immediately after, immature oocytes and blastocysts were washed three times in PBS-PVA and transferred onto slides in a drop of Slowfade™ with 6.25 μg/mL PI and 5 μg/mL Hoechst 33342 fluorescent dye, respectively. Samples were evaluated using a ×20 augmentation by fluorescence microscopy (Eclipse 80i, Nikon Instruments Europe, Amsterdam, The Netherlands). The DNA damage in oocytes was classified as TUNEL-positive (Figure 3A-a) and -negative (Figure 3A-b) according to fragmented cell nuclei. The DNA fragmentation in blastocysts was determined by the number of cells with fragmented nuclei (TUNEL-positive) in relation to the total cell number.

**Figure 3.** Effect of ovary storage time (3–13 h) on DNA fragmentation (positive terminal deoxynucleotidyl transferase mediated dUTP nick-end labelling (TUNEL) staining) of immature sheep oocytes: (**A**) representative images of (a) TUNEL-positive oocyte and (b) TUNEL-negative oocyte (scale bar = 50 μm) and (**B**) oocyte DNA fragmentation rates. Results are expressed as mean <sup>±</sup> SEM. a,b,c,d,e Different letters indicate differences (*p* ≤ 0.05) among storage times.

#### *2.7. Measurement of Caspase-3 Activity*

To monitor caspase-3 activity, 352 immature sheep oocytes were incubated for 30 min at 37 ◦C in 25 μL droplets of PBS-PVA containing 5 mM of PhiPhiLux-G1D2 (OncoImmunin Inc., Gaithersburg, MD, USA). After incubation, oocytes were washed twice in PBS-PVA and placed on slides under cover slips. Caspase activity was determined by fluorescence microscopy (Eclipse 80i, Nikon Instruments Europe, Amsterdam, The Netherlands), and intensity per unit area was quantified using ImageJ 1.45s software (National Institutes of Health, Bethesda, MD, USA; Figure S1).

#### *2.8. Mitochondrial Membrane Potential Analysis*

Membrane potential was determined by incubating 340 immature oocytes for 30 min at 37 ◦C in 0.5 μM of JC-1 dye (Thermo Fisher Scientific, Barcelona, Spain). After incubation, oocytes were washed twice for 5 min and then placed on glass slides. Oocytes were examined by ×20 augmentation by fluorescence microscopy (Eclipse 80i, Nikon Instruments Europe, Amsterdam, The Netherlands). Relative mitochondrial membrane potential was determined as the ratio of J-aggregate to J-monomer staining intensity with ImageJ 1.45s software (National Institutes of Health, Bethesda, MD, USA; Figure S2).

#### *2.9. Assessment of Mitochondrial Distribution*

Mitochondrial distribution patterns were examined by MitoTracker® Red CMXRos (Thermo Fisher Scientific, Barcelona, Spain). At least 363 immature oocytes were incubated in PBS-PVA supplemented with 100 nM dye at 37 ◦C for 20 min. Oocytes were washed and then placed on glass slides and examined under ×20 augmentation by fluorescence microscopy (Eclipse 80i, Nikon Instruments Europe, Amsterdam, The Netherlands). Mitochondrial distribution was classified into two categories: abnormal mitochondrial distribution (Figure S3-d) in the cytoplasm and normal distribution (Figure S3a–c).

#### *2.10. Quantification of Transcript Abundance*

A total of 196 sheep oocytes were subjected to RNA extraction, complementary DNA (cDNA) synthesis and quantitative real-time PCR (qPCR) analysis as previously reported by Sánchez-Ajofrín et al. with minor modifications [13]. The RNA from groups of approximately 10 oocytes (3 replicates) was extracted using Dynabeads® (Invitrogen, California, CA, USA) following the protocol by [14]. Briefly, oocytes were lysed at room temperature in 50 μL binding buffer for 5 min and hybridized with 10 μL magnetic beads for another 5 min. Then, samples were washed twice in 50 μL buffer A and twice in buffer B. Next, mRNA samples were eluted with 28 μL Tris-HCl. Following this, reverse transcription was carried out using the Fermentas™ First Strand cDNA Synthesis Kit (Thermo Scientific, Barcelona, Spain) in a total volume of 40 μL. After heating the samples at 65 ◦C for 5 min, cDNA was synthesized by adding 2 μL of reaction buffer (5×), 2 μL of dNTP Mix, 1 μL of RiboLock RNase Inhibitor, 1 μL of M-MuLV Reverse Transcriptase and 2.5 μL of nuclease-free water. Subsequently, reverse transcription reaction was performed by incubating for 5 min at 25 ◦C, followed by 60 min at 37 ◦C and 5 min at 70 ◦C.

After cDNA synthesis, PowerUp™SYBR® Green Master Mix (Thermo Fisher Scientific, Barcelona, Spain) and a LightCycler 480 II system (Roche, Barcelona, Spain) were employed to determine the relative abundance of mRNA transcripts by qPCR. A final volume of 20 μL was reached by adding 10 μL master mix, 400 nM each of forward and reverse primers, 2 μL of cDNA template and nuclease-free water. The following PCR amplification conditions were used: 50 ◦C for 2 min, 95 ◦C for 2 min, 40 cycles of 95 ◦C for 15 s and 60 ◦C for 1 min. Immediately after, a melting curve analysis was performed to eliminate contamination by heating the samples to 95 ◦C for 5 s in a ramp rate of 4.4 ◦C/s, followed by 65 ◦C for 1 min with a heating rate of 2.2 ◦C/s and continuous fluorescence measurement. Each sample was analysed in duplicate, and reactions without any cDNA template (2 μL nuclease-free water) were used as the negative control.

The comparative cycle threshold and 2−ΔΔCT methods [15,16] were used to calculate the relative transcript abundances of candidate genes: BCL2-associated X protein (*BAX*), BCL2 apoptosis regulator (*BCL2*), bone morphogenetic protein 15 (*BMP15*), caspase-3 (*CASP3*), fibroblast growth factor 16 (*FGF16*) and growth differentiation factor 9 (*GDF9*). Quantification was normalized against that of the endogenous control (Peptidylprolyl Isomerase A (*PPIA*)). Information on the qPCR primers is provided in Table S2.

#### *2.11. In Vitro Maturation and Fertilization Assessment*

After maturation, oocytes were stripped of the surrounding cumulus cells by gentle pipetting. To examine oocyte maturation and sperm penetration, cells were stained with Hoechst 33342 (1 μg/mL) for 10 min at room temperature, washed in PBS-PVA and then analysed with ×20 augmentation by fluorescence microscopy (Eclipse 80i, Nikon Instruments Europe, Amsterdam, The Netherlands). Maturation rate was defined as the number of oocytes with an evident polar body and metaphase II

(MII) plate relative to the total number of oocytes analysed. Oocytes containing both female and male pronuclei (regardless of stage of decondensation) relative to the total number of oocytes matured were considered fertilized and were classified as normal (2PN) according to the number of swollen sperm heads and pronuclei in the cytoplasm.

#### *2.12. Flow Cytometry Analysis of Cumulus Cells*

Cumulus cells were collected from in vitro-matured oocytes and centrifuged at 12,000 rpm during 5 min. Pellet was resuspended in 125 μL PBS-PVA preequilibrated at 37 ◦C, and samples were stained and analysed by flow cytometry. To quantify cell live/dead status and apoptosis, samples were incubated with 10 μM YO-PRO-1 and 0.5 μM PI; for mitochondrial activity, cells were incubated with 200 mM of MitoTracker™ Deep Red (Thermo Fisher Scientific, Barcelona, Spain) for 20 min at 38.5 ◦C in the dark and then stained with 10 μM YO-PRO-1 and 0.5 μM PI; and for oxidative status (GSH and ROS levels), cells were incubated with 10 μM of Cell Tracker™ Blue (Thermo Fisher Scientific, Barcelona, Spain) and 10 μM of CM-H2DCFDA (Thermo Fisher Scientific, Barcelona, Spain) for 30 min at 38.5 ◦C to assess GSH and ROS levels, respectively, and subsequently stained with 0.5 μM PI. The percentage of YO-PRO-1−/PI−showed the proportion of viable cells, while YO-PRO-1+/PI−showed that of apoptotic cells. Viable cells with active mitochondria were represented by the percentage of MitoTracker+/YO-PRO-1−. Finally, oxidative status was measured by GSH and ROS production only in viable cells (PI−).

Cumulus cells analyses were conducted using a FlowSight® imaging flow cytometer (Amnis, Merck-Millipore, Germany) equipped with violet, blue and red excitation lasers (405, 488 and 642 nm), 12 channels of detection and 10 available fluorescent channels. The system was controlled using INSPIRE® software (v.3). The flow cytometer was calibrated daily using calibration beads according to the manufacturer's instructions. A compensation overlap was performed before each experiment, and 1000 events were acquired per sample. In all cases, dot plots with aspect ratio and area were employed to exclude debris from cumulus cell populations and regions used to quantify cells subpopulation depended on the particular assay. The raw data were analysed using IDEAS® software (AMNIS) and out of focus cells, debris and cell clumps were excluded from the analysis.

#### *2.13. Statistical Analysis*

After determining that data were normally distributed and that variances were not heterogeneous, live/dead status, early apoptosis, DNA fragmentation, caspase-3 activity, GSH and ROS content, mitochondrial distribution and membrane potential, meiotic and fertilization competence, embryo production and total cell number were analysed by factorial ANOVA using the SPSS software (IBM, Armonk, NY, USA). For that, time of ovary storage (3, 5, 7, 9, 11 and 13 h) or type of ovary transport medium (TCM199 and saline solution) and the replicate (for IVM, IVF, embryo production, total number of cells and DNA fragmentation in blastocysts and cumulus cell analyses, four replicates were performed; for evaluation of oocyte viability and early apoptosis, DNA fragmentation, caspase-3 activity, GSH and ROS content, and mitochondrial membrane potential and distribution, eight replicates were performed) were considered fixed effects. Additionally, another factorial ANOVA was conducted to examine the relative abundances of mRNA transcripts, with time of ovary storage or type of medium and qPCR technical replicate (two replicates) as the fixed effects and the different target genes as the dependent variable. When a significant effect was observed, post hoc comparisons with Bonferroni correction were carried out. There was no evidence of statistically significant interactions between storage time and medium composition. Results are presented as mean ± S.E.M.

#### **3. Results**

#### *3.1. E*ff*ect of Ovarian Transport Time on Oocyte Viability and Quality*

As shown in Figure 1B, the percentage of viable immature sheep oocytes decreased (*p* < 0.05) from 7 h onwards (3 h = 65.62 ± 6.67% vs. 7 h = 20.00 ± 6.67%, 9 h = 3.12 ± 10.20%, 11 h = 3.12 ± 10.20% and 13 h = 1.25 ± 6.67%). Moreover, the lowest percentages (*p* < 0.05) of dead oocytes were observed at 3 and 5 h (25.62 ± 7.32% and 46.87 ± 11.18%, respectively) compared to 11 and 13 h (96.87 ± 11.18% and 96.25 ± 7.32%, respectively). Early apoptosis detected by phosphatidylserine localization using Annexin-V staining was not statistically different among groups (*p* > 0.05; Figure 1B).

A significantly higher level (*p* < 0.05) of ROS was recorded in immature sheep oocytes recovered from ovaries stored for 3 h (42.35 ± 3.69) compared to oocytes obtained from ovaries stored for 13 h (24.93 ± 3.69; Figure 2B). As also shown in Figure 2B, the different treatments did not exhibit different (*p* > 0.05) levels of GSH.

The number of immature sheep oocytes with TUNEL-positive fragmented DNA was lower (*p* < 0.05) at 3 h (1.94 ± 5.48%) of ovary storage, increased from 7 h (28.91 ± 5.48%) and recorded a maximum value at 13 h (75.88 ± 5.48%; Figure 3B).

The duration of ovary storage did not affect (*p* > 0.05) immature sheep oocyte caspase-3 intracellular activity (Figure S1), mitochondrial membrane potential (Figure S2) and mitochondrial distribution (Figure S3). Moreover, as shown in Figure S4, the relative abundance of mRNA transcripts of genes related to apoptosis (*BAX*, *BCL2* and *CASP3*) and oocyte quality (*BMP15*, *GDF9* and *FGF16*) did not show differences (*p* > 0.05) between 3, 7 and 13 h of ovary storage in immature sheep oocytes.

#### *3.2. E*ff*ect of Ovarian Transport Time on the In Vitro Maturation and Fertilization Potential of Oocytes*

The percentage of MII sheep oocytes recovered from ovaries stored for 3 h was higher (*p* < 0.05) compared to 7 and 13 h (Table 1. As expected, fertilization rate (2PN) was significantly increased after IVF in 3 h-derived oocytes compared to 7 and 13 h, with no significant differences between the latter groups (Table 1).


**Table 1.** In vitro maturation and fertilization of sheep oocytes retrieved from ovaries stored for 3, 7 and 13 h.

Data are expressed as mean ± SEM. The results represent four replicates. a,b,c Different letters indicate differences (*p* ≤ 0.05) among storage times.

#### *3.3. E*ff*ect of Ovarian Transport Time on Cumulus Cells from Cumulus–Oocyte Complexes (COCs)*

After maturation, COCs collected from ovaries stored for 3, 7 and 13 h were gently pipetted and detached cumulus cells were analysed by flow cytometry. Live/dead status, apoptosis, active mitochondria, and GSH and ROS levels were assessed. The results showed reduced (*p* < 0.05) cell viability as time increased (3 h = 82.06 ± 3.87% vs. 7 h = 59.16 ± 3.87% vs. 13 h = 26.53 ± 3.87%; Figure 4). Moreover, there was a lower (*p* < 0.05) percentage of dead cells at 3 h compared to 13 h (13.72 ± 6.08% and 58.50 ± 6.08%, respectively), although apoptosis did not show significant differences between storage time (*p* > 0.05; Figure 4).

**Figure 4.** Effect of ovary storage time on (**A**) live/dead status and apoptosis, (**B**) percentage of active mitochondria and (**C**) intracellular GSH and ROS levels and the GSH/ROS ratio of cumulus cells collected from mature oocytes retrieved from ovaries stored for 3, 7 and 13 h: the results are expressed as mean <sup>±</sup> SEM. a,b,c Different letters indicate differences (*<sup>p</sup>* <sup>≤</sup> 0.05) among treatments.

Our results also showed reduced (*p* < 0.05) mitochondrial activity with increased storage time (3 h = 51.08 ± 3.78%, 7 h = 31.45 ± 3.45% and 13 h = 14.14 ± 3.45%; Figure 4). Furthermore, GSH and ROS levels and the ratio of GSH/ROS were higher after 3 h (GSH =19890.59±3044.95, ROS = 3153.65 ± 418.98 and GSH/ROS = 6.89 ± 0.78) of ovary storage compared to 7 h (GSH = 4813.30 ± 3044.95, ROS = 1234.67 ± 418.98 and GSH/ROS = 3.98 ± 0.78) and 13 h (GSH = 2213.70 ± 3044.95, ROS = 1009.18 ± 418.98 and GSH/ROS = 2.31 ± 0.78; Figure 4).

#### *3.4. E*ff*ect of Ovarian Transport Time on In Vitro Embryo Development and Blastocyst Quality*

The proportion of sheep oocytes that progressed to the first cleavage stage after IVF was significantly lower (*p* < 0.05) with increasing ovary storage time (Table 2. The percentage of total expanded blastocysts and percentage of blastocysts relative to the number of cleaved embryos was drastically decreased (*p* < 0.05) after 7 h of storage compared 3 h. Sheep blastocysts were not produced after 13 h of ovary storage (Table 2). However, the total cell number (3 h = 134.30 ± 6.01 and 7 h = 150.16 ± 12.02) and proportion of TUNEL-positive blastomeres (3 h = 13.02 ± 0.68% and 7 h = 12.34 ± 1.36%) in sheep blastocysts were similar (*p* > 0.05) between 3 and 7 h.

**Table 2.** The effect of ovary storage time on rates of cleavage and blastocyst development in sheep.


Data are expressed as mean ± SEM. The results represent four replicates. a,b,c Different letters indicate differences (*p* ≤ 0.05) among storage times.

#### *3.5. E*ff*ect of Medium Type During Ovary Transport on Oocyte Developmental Competence and Quality*

The storage of sheep ovaries in TCM199 medium or saline solution did not show significant differences (*p* > 0.05) between the oocyte quality parameters studied, including oocyte live/dead status, apoptosis, caspase-3 intracellular activity, GSH and ROS levels, DNA fragmentation, mitochondrial distribution, mitochondrial membrane potential and relative mRNA transcript abundance (Figure S5).

The storage of ovaries with TCM199 resulted in lower (*p* < 0.05) rates of IVM, IVF, cleavage and blastocysts relative to the number of cleaved embryos (Table 3|). Nevertheless, the total cell number of blastocysts was decreased (*p* < 0.05) as a result of storing the ovaries in saline solution (125.46 ± 6.72%) compared to TCM199 (159.0 ± 10.83%). Moreover, the number of total expanded blastocysts and TUNEL-positive blastomeres was similar (*p* > 0.05) in both media (Table 3).

Cumulus cells were also affected by the type of ovary storage medium. Thus, cumulus cells from ovaries stored in saline solution showed a greater percentage (*p* < 0.05) of viable cells and active mitochondria, while the TCM199 medium exhibited higher rates (*p* < 0.05) of dead cells (Table 4). The proportion of apoptotic cells and the GSH and ROS levels did not show significant differences (*p* > 0.05) between media (Table 4).

#### **4. Discussion**

In the present study, storage of sheep ovaries beyond 7 h had a detrimental effect on oocyte quality and subsequent development to the blastocyst stage. Similar results were obtained in rat ovaries where apparent histological changes were observed after 3 h of ischemia [17]. Notably, an in-depth analysis revealed that, after 7 h, oocyte live/dead status was dramatically reduced along with increasing storage time. After organ removal, an immediate consequence of the cessation of blood supply is the deprivation of oxygen and nutrients as well as the accumulation of metabolic waste, which may lead to cellular damage [18]. The crucial event is ATP depletion, which occurs within the first few minutes of oxygen stoppage [19]. This early event results in a transition from aerobic to anaerobic metabolism, which in turn leads to a rise in lactate and H<sup>+</sup> levels that contribute in many mechanisms to cell injury related to ischemia [19]. In addition, the susceptibility of different types of cells to ischemic damage varies according to the degree of metabolic activity, and those with higher rates require a greater ongoing production of ATP [19]. For this reason, cells that form the ovarian tissue tend to be injured very rapidly by hypoxia [7,8].

Reduced viability of oocytes in post-mortem ovaries has been linked to degeneration of protein and DNA in horses [20] and domestic cats [21]. Remarkably, in the current study, we observed that the number of oocytes with fragmented DNA started to greatly increase after an ischemic time of 7 h, with no evidence of other oocyte apoptosis markers being significantly different (caspase-3 activity, phosphatidylserine binding by Annexin-V and mRNA transcripts). Moreover, there were no evident differences between storage times regarding mitochondrial membrane potential and distribution, which have been previously linked to many apoptotic stimuli [22–24]. Traditionally, apoptosis has been characterized by DNA damage as visualized by the TUNEL assay [25]. However, identification of terminal deoxy-nucleotidyl transferase (Tdt)–mediated deoxyuridine 5-triphosphate (dUTP) labelling in the nucleus of dying cells is not sufficient to demonstrate that cells are undergoing apoptosis, as the chromosomal DNA degradation and resulting DNA strand breaks also occur in necrotic cells [25,26]. Moreover, a pattern of TUNEL staining typical of necrotic cells was noticed in the present study with extensive staining of the cytosol, which may be due to the formation of large DNA fragments during karyorrhexis that are released into the cytosol upon nuclear disintegration [25,27].



 represent among



Data are expressed as mean ± SEM. The results represent four replicates. a,b Different letters indicate differences (*<sup>p</sup>* ≤ 0.05) among media.

Interestingly, intracellular ROS levels at 3 h were higher than at 13 h. It has been suggested that ROS quickly accumulate at the onset of ischemia despite limited O2 supply [28,29]. In the presence of xanthine oxidase (XOD) and O2, hypoxanthine can be converted to xanthine, which simultaneously produces superoxide anion (O2 •−) [30]. During ischemia, the accumulation of XOD and hypoxanthine results in increased O2 •− production [31]. Considering that molecular O2 is the limited substrate in this reaction and that its availability decreases over time, it was not surprising to find that, after 13 h of ovary storage, intracellular ROS levels were significantly lower.

Successful embryo development will largely depend on the optimal accumulation of organelles, metabolites and maternal RNAs during oocyte growth [32]. In our study, the duration of ovary storage did not affect mitochondrial membrane potential, mitochondrial distribution and mRNA transcript of genes related to oocyte quality, although the embryonic yield was lower beyond 7 h of storage. A possible explanation could be that the assessments were performed on immature oocytes and that, in this cellular state, fluorescent dyes displayed less sensitivity to intracellular changes. Thus, different studies have shown contradictory data in relation to mitochondrial activity between immature and mature oocytes [33,34].

Besides, our results revealed that the number of oocytes showing MII and 2 PN after fertilization were lower when ovaries were stored longer than 7 h. Gaulden [35] suggested that hypoxic conditions could reduce oocyte intracellular pH, influencing the organization and stability of the meiotic metaphase spindle. Moreover, other researchers found that normal oocytes collected from under-oxygenated follicles showed chromosomal defects such as a compact arrangement on the MII spindle [36]. In our case, rather than chromosomal alterations, it is likely that the low rates of MII and 2PN are due to the high oocyte mortality following acute deprivation of oxygen after ovary collection.

To gain further awareness of the biological consequences of prolonged transport time in stored oocytes, we examined cumulus cells and oocyte development in vitro. Cumulus cells play a critical role in oocyte maturation because they supply ions, metabolites and regulatory molecules that are necessary for meiotic progression, normal nuclear and cytoplasmic maturation of oocytes, and subsequent embryonic development after fertilization [37,38]. As expected, cumulus cells lost their supportive and protective functions when subjected to storage times longer than 7 h, since they showed decreased viability and impaired redox status and mitochondrial activity. Likewise, 7 and 13 h of ovarian storage resulted in drastic reduction in oocyte maturation, fertilization, cleavage and blastocyst rates. In fact, after 13 h of preservation, blastocysts were unable to develop. Therefore, reduced developmental potential of in vitro matured oocytes may also be related to impaired cumulus cell functions. Similar to our results, other studies have shown that the length of time that ovaries are held before oocyte recovery also affected developmental potential in several species. In sheep and pig, a delay of only 5–7 h reduced the maturation rate compared to that of oocytes placed immediately into maturation culture [39] or after 3 h of storage [6], respectively; in horses, 5–9 h had the same effects [40]. In addition, long-term storage (7–8 h) of ovaries reduced blastocyst formation rates after IVF in cattle [41] and intracytoplasmic sperm injection (ICSI) in horses [42].

Besides duration of ovarian storage, the type of medium where these organs are held plays an important role in determining appropriate transport conditions for oocyte survival and in vitro embryo development. Because TCM199 has more components (glucose, vitamins, amino acids and adenine sulphate) than saline solution and fully grown follicles are more metabolically active, we speculated that both follicles and oocytes may be better supported by the more complex medium. Although there were no differences between media for parameters used as indicators of oocyte quality, reduced oocyte developmental competence and cumulus cell quality were evidenced when ovaries were preserved in TCM199 medium. One possible explanation may be that ovaries from this group were kept at a higher temperature (38.5 ◦C) than the saline solution group (30 ◦C). In fact, preliminary results obtained by our group have indicated that the preservation of sheep ovaries in saline solution for 4 h at 38.5 ◦C negatively affects oocyte quality, IVM rates and cumulus cells compared to 30 ◦C (unpublished data). Moreover, it has been suggested that the use of low temperatures (4 ◦C) during

long ovary storage times preserves oocyte quality possibly due to a decrease in cellular metabolism [43]. However, in our laboratory, oocytes from ovaries stored in saline solution at 4 ◦C for 3 h showed lower rates of MII after maturation (unpublished data), suggesting that low transport temperatures may have a detrimental effect during the transport of sheep ovaries. Differences with other authors such as Goodarzi et al. [44] could be related to the use of different transport media or even to the size of follicles from which the oocytes are obtained. More studies using a combination of different transport solutions, temperatures and storage times to retrieve the largest number of competent oocytes are necessary.

#### **5. Conclusions**

Our study has provided new insight into the complex field of oocyte survival and in vitro development throughout ovary preservation. Moreover, it has contributed to understanding the effect of ovary storage in the physiological features of immature oocytes, which had not been evaluated up to date. We have demonstrated that transport ovary times up to 5 h in saline solution are the most adequate storage conditions to maintain oocyte quality as well as developmental capacity in sheep. After that, the quality and developmental potential of oocytes and cumulus cells dramatically decreases after storage of ovaries from 7 h. Based on these results, new strategies to evaluate the possibility of saving or rescuing the developmental potential of stored oocytes will be needed for successful production of high-quality embryos.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-2615/10/12/2414/s1, Table S1: Composition of solutions used in experiments, Table S2: List of primers used in qPCR of sheep immature oocytes, Table S3: List of abbreviations, Figure S1: Measurement of caspase-3 intracellular activity in sheep oocytes collected from ovaries stored for 3, 5, 7, 9, 11 and 13 h, Figure S2: Mitochondrial membrane potential in sheep oocytes collected from ovaries stored for 3, 5, 7, 9, 11 and 13 h, Figure S3: Mitochondrial distribution patterns in sheep oocytes collected from ovaries stored for 3, 5, 7, 9, 11 and 13 h, Figure S4: Relative mRNA transcript abundance pattern of genes of interest in sheep immature oocytes collected from ovaries stored for 3, 7 and 13 h, Figure S5: Live/dead status and apoptosis, GSH and ROS levels, DNA fragmentation, caspase-3 intracellular activity, mitochondrial membrane potential and distribution, and relative mRNA transcript abundance in sheep immature oocytes collected from ovaries stored with TCM or saline solution: the results are expressed as mean ± SEM.

**Author Contributions:** Conceptualization, I.S.-A. and A.J.S.; methodology: A.M.-M., I.S.-A., C.M., P.P.-F., D.-A.M.-C., B.C., J.C.N. and M.R.F.-S.; formal analysis: A.M.-M., I.S.-A. and C.M.; investigation: I.S.-A. and A.J.S.; resources: A.J.S. and J.J.G.; writing: A.M.-M. and I.S.-A.; original draft preparation: I.S.-A.; writing—review and editing: A.J.S. and J.J.G.; funding acquisition: A.J.S. and J.J.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Spanish Ministry of Economy and Competitiveness (AGL2017-89017-R). A.M.-M. was supported by a Ministry of Economy and Competitiveness scholarship. P.P.-F. and D.-A.M.-C. were supported by a University of Castilla-La Mancha scholarship. B.C. was supported by ERASMUS+−.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Reproductive Outcomes and Endocrine Profile in Artificially Inseminated versus Embryo Transferred Cows**

**Jordana S. Lopes 1,2, Estefanía Alcázar-Triviño 3, Cristina Soriano-Úbeda 1, Meriem Hamdi 4, Sebastian Cánovas 1,2, Dimitrios Rizos <sup>4</sup> and Pilar Coy 1,2,\***


Received: 1 July 2020; Accepted: 3 August 2020; Published: 6 August 2020

**Simple Summary:** Bovine embryos are nowadays produced in laboratories, frozen and transferred to other cows. However, the percentage of pregnancies obtained after these transfers as well as difficulties found during labor, especially due to increased size of calves, are a matter of great concern. One of the possible explanations for these problems relies on the embryo being produced in in vitro conditions (laboratory settings), more specifically the culture medium (liquid) used to develop these embryos. In an attempt to better mimic what happens naturally, female reproductive liquids (from oviducts and uterus) were used as a supplement to the culture of the embryos. As controls, embryos produced using the standard protocol in the laboratory were produced, as well as embryos derived from artificial insemination of cows (in vivo). An evaluation on the pregnancy rates, how the hormonal profile of the recipients changed during pregnancy, difficulties during parturitions, and phenotype of calves were recorded. Results showed that all the groups were very similar, but many differences were noted on the hormonal profiles during pregnancy. In conclusion, all systems provided safe production of calves, but long-term analysis of these calves is necessary to understand the future impact of the laboratory protocols.

**Abstract:** The increasing use of in vitro embryo production (IVP) followed by embryo transfer (ET), alongside with cryopreservation of embryos, has risen concerns regarding the possible altered pregnancy rates, calving or even neonatal mortality. One of the hypotheses for these alterations is the current culture conditions of the IVP. In an attempt to better mimic the physiological milieu, embryos were produced with female reproductive fluids (RF) as supplements to culture medium, and another group of embryos were supplemented with bovine serum albumin (BSA) as in vitro control. Embryos were cryopreserved and transferred while, in parallel, an in vivo control (artificial insemination, AI) with the same bull used for IVP was included. An overview on pregnancy rates, recipients' hormonal levels, parturition, and resulting calves were recorded. Results show much similarity between groups in terms of pregnancy rates, gestation length and calves' weight. Nonetheless, several differences on hormonal levels were noted between recipients carrying AI embryos especially when compared to BSA. Some calving issues and neonatal mortality were observed in both IVP groups. In conclusion, most of the parameters studied were similar between both types of IVP derived embryos and the in vivo-derived embryos, suggesting that the IVP technology used was efficient enough for the safe production of calves.

**Keywords:** embryo transfer; reproductive fluids; pregnancy; vitrification; calving

#### **1. Introduction**

In 2018, more than 1 million bovine embryos were produced worldwide from which the majority were in vitro produced (IVP) [1]. This represents a continuous growth of IVP-embryos globally despite the fact that IVP embryos have lower pregnancy rate than in vivo-derived embryos (IVD) [2,3]. Its success relies on the possibility of producing more embryos than in vivo, in a shorter amount of time and that the embryonic losses after implantation are no different from IVD, therefore, compensating the potential decrease in implantation rate [3,4]. In addition, the use of IVP-embryos is a better option during season of heat-stress [5] or in cases of repeat breeders [6].

Another technique on the rise is the cryopreservation of embryos, with more than 38% of 2018 s embryo transfers (ET) coming from frozen embryos [1]. The need for exchange of cattle genetics has popularized this embryo preservation technique, which has the added benefit to not have all the recipients synchronized at the time of fresh embryo production. Within these large numbers of ET of frozen embryos, almost half (>46%) were IVP-derived embryos, but only a very small percentage of these (<0.1%) were produced with oocytes from abattoir ovaries, the rest being obtained by ovum pick-up (OPU).

Although the increase in IVP is significant over the years, the quality of the embryos produced in vitro still remains inferior to IVD embryos [7–9]. This is a factor that is affected in all mammalian embryos, where culture conditions such as the type of medium, type of supplementations, gases concentration or even culture devices play a crucial role, determining the final yield and quality of the embryo [10]. Currently, research is leading us towards a more physiological approach to in vitro culture. Media supplementation is being updated, such as the addition of insulin-transferrin-selenium to culture medium [11] that increases the blastocyst yield, or the addition of reproductive fluids (RF) [12] that increases embryo quality by increasing their cryotolerance and protecting from oxydative stress. García-Martínez et al. [13] measured the levels of O2 within the different segments of the pig oviduct and uterus in vivo, and adapted the data obtained to the IVP, showing that decreasing the O2 concentration during IVF, though not affecting the IVF results directly, had a positive impact on blastocyst yield and quality. Also, a new device [14] named 3D oviduct-on-a-chip model was recently designed to promote a more accurate surface when culturing bovine embryos, allowing a proper fertilization of the oocytes and eliminating polyspermy and parthenogenic activation.

Several issues besides embryo quality have been pointed out regarding problems in IVP-pregnancies. Hasler [15] summarized them into four main points: (1) increased abortion rate; (2) reduced intensity of labour; (3) increased dystocia, birth weight, calf mortality and fetal abnormalities; and (4) higher percentage of male calves over female calves. However, in addition to these four issues, the fact that embryos are cryopreserved also affects their survival after warming [9,16,17], due to their high lipid content, as well as their implantation success, giving usually inferior pregnancy rates [15,18,19].

In order to further improve in vitro production and ultimately bovine-assisted reproductive technologies (ART), we designed the present study with the following specific objectives; (1) To find out whether culture media is improved by the addition of RF—acting closer to the physiological milieu—impacting the embryo implantation rate versus embryos produced in vitro with conventional culture media (BSA) or in vivo (AI); (2) To find out if hormonal levels in recipients, including P4, E2, AMH and cortisol, at day of ET and across pregnancy could help to predict ART-associated problems such as decrease implantation rate; and (3) To find out if RF-derived calves have more similarities with AI-derived calves, relative to conceptus size, gestation length, calving difficulty, and calf birth weight, than BSA-derived calves.

To fulfil our objectives, we used cow recipients allocated in three experimental groups, according to the origin of the embryo transferred: the first group received an in vitro produced embryo grown in culture media improved by supplementation with reproductive fluids (RF group); the second group received an embryo produced at the same time as those at the previous group but using conventional culture media, lacking reproductive fluids (BSA group); and the third group consisted of animals artificially inseminated with frozen-thawed semen from the same bull used in the two IVP groups (AI group).

#### **2. Materials and Methods**

All chemicals were purchased from Sigma-Aldrich Chemical Company (Madrid, Spain), unless otherwise indicated.

#### *2.1. Ethics*

The experimental work was submitted to evaluation by the CEEA (Comité Ético de Experimentación Animal) from University of Murcia. After approval, authorization from "Dirección General de Agricultura, Ganadería, Pesca y Acuicultura", Región de Murcia- nr A13170706 was given to perform the animal experiments.

#### *2.2. Oocyte Collection and In Vitro Maturation*

Ovaries from crossbred beef cycling heifers and cows were collected at the slaughterhouse. The protocol has already been described elsewhere [12]. Briefly, follicles between 2–8 mm were aspirated and intact cumulus-oocyte complexes (COC) were selected for in vitro maturation (IVM). Groups of 50 COCs were cultured in maturation medium that consisted of TCM-199 supplemented with 10% fetal calf serum and 10 ng/mL epidermal growth factor for a period of 24 h at 38.5 ◦C, under 5%CO2 and high humidity.

#### *2.3. In Vitro Fertilization*

Commercially bought semen doses from one bull (Asturian Valley breed, ASEAVA, Asturias, Spain) were used for all the cycles of embryo production. Frozen semen straws were thawed in a water bath at 38 ◦C, and following the manufacturer's instructions from Bovipure® (Nidacon, Sweeden), the gradient with semen was centrifuged at 300× *g* for 15 min and then washed for 5 min at 300× *g*. Matured oocytes were washed in Fert-TALP [20] medium, transferred to a new dish and inseminated with 1 <sup>×</sup> 106 spermatozoa/mL. Fertilization was left to occur during 18–20 h, at 38.5 ◦C, under 5%CO2 and high humidity.

#### *2.4. Embryo Culture*

Presumptive zygotes were denuded from cumulus cells by vortex for 3 min. In each replicate, putative zygotes were divided in two groups according to embryo culture medium (SOF) supplementation: bovine serum albumin (BSA group) or reproductive fluids (RF group). BSA group received supplementation of 3 mg/mL of bovine serum albumin from day 1 to day 8. RF group received supplementation of 1.25% (*v*/*v*) of oviductal fluid (from early luteal phase of the estrous cycle, NaturARTs BOF-EL—Embryocloud, Spain) from day 1 to day 4, and 1.25% (*v*/*v*) of uterine fluid (from mid-luteal phase of the estrous cycle, NaturARTs BUF-ML—Embryocloud, Spain) from day 4 to day 8, as previously described [20]. Putative zygotes were washed twice in the corresponding medium and put into culture in groups of 25 per 25 μL microdrop, covered with parafin oil (Nidoil, Nidacon, Sweden). Incubation conditions were 38.5 ◦C, 5% CO2 and 5% O2. On day 4 of culture, all embryos were washed twice in the new corresponding medium and put in a new culture dish.

#### *2.5. Embryo Vitrification and Warming*

Commercial vitrification media (Kitazato-Dibimed, Spain) and an open-system Cryotop were used following manufacturer's instructions and as previously described [21]. Embryos on day 7 or 8 of culture and on stage 6 or 7 of development [22] were submitted to vitrification and stored in liquid nitrogen until use. Commercial thawing media (Kitazato-Dibimed, Spain) was used to warm vitrified embryos. Warming was performed according to manufacturer's instructions and performed less than 4 h before embryo transfer. Embryos were loaded in 0.25 mL straws with commercial medium (BO-Transfer, IVF-Bioscience, Denmark).

#### *2.6. Recipient Synchronization and Embryo Transfer*

Holstein multiparous dairy cows from a commercial farm (El Barranquillo SL, Spain) were synchronized using double-Ovsynch protocol. Reproductive ultrasound evaluation on the previous day of the transfer was made in order to discard recipients that failed or had delayed ovulation. Recipients were on their day 6, 7 or 8 of the estrus cycle and embryo transfer (ET) was made non-surgically to the ipsilateral uterine horn from ovulation (one embryo per recipient).

#### *2.7. Artificial Insemination*

Cows were artificially inseminated with frozen-thawed semen from the same Asturian-Valley bull used for IVP. Synchronization was made the same way described earlier (double-Ovsynch protocol) and cows were inseminated on day 0 (presumptive day of estrus).

#### *2.8. Pregnancy Detection and Follow-up until Parturition*

Pregnancy was detected at day 30 ± 3 of gestation by rectal ultrasonography (Easi-Scan™, BCF Technology, Scotland, UK). Measurement of the crown-rump length (mm) was performed to the conceptuses. Confirmation of pregnancy was repeated at days 60, 90, 150, and 210 of pregnancy. If parturition had not occurred by day 283 ± 2 of gestation, labour was induced with 0,150 mg d-cloprostenol q24h. No C-sections were performed but human intervention was available when calving was difficult. Calving was considered "easy" if little or no help was necessary and "difficult" if heavy assistance was needed. Calves' weight was assessed between 0 to 4 h after birth, using a weight scale.

#### *2.9. Blood Collection and Analysis*

Blood from recipient cows was collected via puncture of the median caudal vein with plain tubes (Vacutainer, BD Spain) on the day of ET or 7 days after-estrus for AI group, from here on referred to as "day 7" in both cases. Blood from pregnant recipients was also collected on day 30, 90, 150, and 210. Samples were centrifuged at 1000× *g* for 20 min at room temperature and plasma collected and stored (−20 ◦C) until analysis. Hormone levels—anti-Müllerian (AMH, ng/mL), estradiol (E2, pg/mL), progesterone (P4, ng/mL) and cortisol (nmol/L)—were measured with ECLIA assay (electrochemiluminescence immunoassay) using a Cobas® e801 system (Roche Diagnostics GmbH, Mannheim, Germany). Five samples of estradiol were not correctly measured and gave values below the detection (<5 ng/mL; a mean value of the previous/following measurement of the individual cow was made and used as a replacement).

#### *2.10. Statistical Analysis*

For statistical analysis, we followed two different approaches: In the first one, we evaluated the three groups (AI, BSA and RF) independently; in the second one, we compared the data from AI group versus the pooled data of BSA and RF groups, with the intention of getting information about the effect of the ART procedure used (i.e., in vivo fertilized embryos (AI) versus in vitro produced embryos, where IVP = RF + BSA data).

Data were tested for normality using Shapiro–Wilk test (*p* > 0.05) followed by either one-way parametric ANOVA/t-test when approved for normality, or non-parametric Kruskal–Wallis/Mann Whitney U when not normally distributed. Multiple comparisons tests (Tukey/Dunn) were used when significant difference was found (*p* < 0.05). Hormonal parameters were analysed using ANOVA for

repeated measures, with Geisser–Greenchouse's correction applied when data did not follow sphericity. Multiple comparisons tests were performed using Tukey for group analysis and Sidak for type of ART analysis, and *p* < 0.05 was considered significant.

Data presented are mean ± SEM, unless otherwise indicated. The software used was GraphPad Prism version 8.4.0 for Windows (GraphPad Software, San Diego, CA, USA).

#### **3. Results**

*3.1. Rate of In Vitro Produced Embryos, Pregnancy Maintenance and Conceptus Size after Embryo Transfer*/*Artificial Insemination, Were Similar between Groups*

Cleavage and blastocyst rate with BSA or RF as supplements to the embryo culture medium was not significantly different (Table 1).

**Table 1.** Total presumptive zygotes, cleavage and blastocyst rate of in vitro produced embryos bovine serum albumin (BSA) or reproductive fluids (RF).


Percentages are shown as mean ± SEM.

Pregnancy rates were also similar between groups. Table 2 describes the confirmation of pregnancy at different timelines and no statistical difference was found between groups, both in terms of pregnancy rates as well as pregnancy loss.

**Table 2.** Pregnancy confirmation during gestation and parturitions of artificial insemination (AI), bovine serum albumin (BSA), reproductive fluids (RF) and IVP (BSA and RF groups data combined) groups.


Percentages are means and n are number of animals.

Conceptus size was also not significantly different between groups (Figure 1).

**Figure 1.** Distribution of conceptus size at day 30 of gestation of embryos produced by artificial insemination (AI), bovine serum albumin (BSA), reproductive fluids (RF), or in vitro-produced (IVP, corresponding of BSA + RF). The boxplot is represented by the first quartile, the median and the third quartile, with minimum and maximum as whiskers.

#### *3.2. Recipient's Hormonal Levels at Day 7, 30, 90, and 210 of Gestation Showed Differences at Specific Time-Points*

Figure 2 represents the evolution of hormonal levels through day 7, 30, 90, and 210 of pregnancy; their statistical differences and influence of the day of collection; the group; and day x group interaction. Cortisol levels were significantly affected by the day of collection, and particularly at day 7, showed a tendency to be higher in BSA recipients than RF recipients (Figure 2A, *p* < 0.09). AMH levels of AI recipients showed a significant decrease on day 30 when compared to IVP (Figure 2B, *p* < 0.05). P4 concentrations were influenced by the day of collection, being significantly lower on AI recipients when facing RF or IVP (*p* < 0.01) but just a tendency when compared to BSA (Figure 2C, *p* < 0.09). E2 had a tendency to be influenced by the group or day x group interaction, but no other significant difference was shown (Figure 2D). E2/P4 ratio was highly variable due to the day, group and day x group interaction, having AI recipients higher mean levels on day 7 vs. BSA/RF/IVP (*p* < 0.01) and a tendency on day 210 vs. BSA (Figure 2E, *p* < 0.09).

**Figure 2.** Distribution of mean ± SEM values, of (**A**). Cortisol, (**B**). Anti-Müllerian hormone (AMH), (**C**). Progesterone (P4), (**D**). Estrogen (E2) and (**E**). E2:P4 ratio, in pregnant recipients from artificial insemination (AI) group, bovine serum albumin group (BSA), reproductive fluids group (RF) or in vitro produced (IVP, corresponding of BSA + RF) across gestation time. \* *p* < 0.05, \*\* *p* < 0.01 and # *p* < 0.09.

#### *3.3. Gestation Length, Parturitions and Neonatal Period*

Mean gestation length, minimum and maximum length for the non-induced parturitions is shown in Table 3.

**Table 3.** Gestation length of non-induced parturitions from pregnant cows of artificial insemination (AI), bovine serum albumin (BSA), reproductive fluids (RF) or in vitro-produced (IVP, corresponding of BSA + RF).


Gestation length is represented as mean ± SEM; \* this parturition was not induced by decision of the farmers.

One premature calf was delivered naturally with 240 days and another recipient, due to productivity reasons, was not induced and the gestation lasted 298 days. Induced parturitions were necessary in some cases for BSA and RF groups (Table 4), mostly due to the fear of increased birth weight and consequently calving difficulties, and resulted in a higher tendency of IVP vs. AI of induced parturitions (*p* = 0.0573). Calving ease showed no statistical differences between groups. Percentage of male calves was high in absolute values, but without statistical differences between groups.

**Table 4.** Induced parturitions, calving ease score, male calves, and neonatal mortality within groups of artificial insemination (AI), bovine serum albumin (BSA), reproductive fluids (RF) or in vitro produced (IVP, corresponding of BSA + RF).


Calving ease was scored as easy if it required little to no assistance or difficult if it needed moderate to heavy assistance (i.e., surgery/veterinarian intervention). † AI vs. IVP *p* = 0.0573; \* AI vs. RF *p* = 0.0752.

Weight at birth was also not different between groups (Figure 3).

**Figure 3.** Weight at birth of calves born through artificial Insemination (AI) group, bovine serum albumin group (BSA), reproductive fluids group (RF) or in vitro-produced (IVP, corresponding of BSA + RF). The line represents the median and each symbol represents one animal.

Neonatal mortality happened in one calf for BSA group and in four calves for RF group and statistically did not show any difference between groups (Table 4), but a tendency of higher mortality in RF vs. AI happened. Details on age at death, sex, birth weight, and cause of death are shown in Table 5.

**Table 5.** Cases of neonatal mortality from calves born by embryo transfer of vitrified-warmed embryos produced with reproductive fluids (RF) or bovine serum albumin (BSA) as supplements to culture medium.


ND Stands for not determined.

#### **4. Discussion**

The in vitro embryo production in cattle industry has been drastically increasing over the past few years. Today, it is necessary not only to optimize its final yield, but also to assure that we are producing high quality and healthy animals and that calving problems are minimum. To this end, culture conditions and recipient selection and monitoring appear as main points to keep under control. Culture conditions of IVP are still in need of improvement, not only in cattle but in mammals in general [10,23,24]. The pre-implantation embryo represents a critical stage where all environmental conditions might reach higher relevance than at any other stage. As reviewed by Vajta et al. [10], current culture conditions, ranging from media composition to temperature of incubation, are not a unanimity, suggesting that there is still a lot of unknown factors in the biological environment that need to be studied and addressed. The addition of RF, such as oviductal and uterine fluids, has been proposed before [12,25] as a potential tool to improve the embryo quality. Hamdi et al. [12], using both fluids as supplements to embryo culture media, obtained similar blastocyst rates to those reached with fetal calf serum or even BSA supplementation. Despite not showing any improvement of blastocyst yield, embryos produced with RF supplementation showed higher survivability after vitrification-warming than serum-derived embryos, a downregulation of genes related with oxidative stress in comparison with BSA-derived embryos as well as significantly lower reactive oxygen species when compared to both serum and BSA groups. However, no data related to the ability of these embryos to implant or to develop to term after being transferred into recipient cows had been published until now, thus our work is the first producing live birth calves from IVP embryos cultured with reproductive fluids as supplements.

Pregnancy rates from vitrified-warmed IVP embryos are known to be inferior when compared to fresh IVP embryo transfers [5,26,27], with very few studies showing equal values [28]. Our results showed similar values at the first diagnosis for both BSA and RF groups, and these rates were slightly lower when compared to some studies that used abattoir-derived oocytes in IVP plus vitrification [5,26,29], but within the range or even higher than others that used slow freezing method [6,30]. Our AI pregnancy rates were lower than expected, probably explained by the high temperatures during the season in southern Spain, which induces heat-stress to the cows. Pregnancy loss was not significantly different between groups, and the percentage of lost pregnancies is in accordance to another study [5] that used vitrified embryos and had 25.9% pregnancy loss between diagnoses and calving.

Progesterone, being a key hormone associated with uterus preparation to receive the embryo, is one of the most studied hormones in cattle. According to Stronge et al. [31], low levels of P4 during the first 5 days after estrus are associated with low fertility. In our data, we found the normal physiological response that is the increase of P4 between day 7 and day 30. However, between the pregnant animals of our experimental groups, lower levels of P4 in AI recipients were found when compared to RF group or even IVP. This difference might be related to the fact that IVP recipients

were chosen carefully, with evaluation of corpus luteum on the day prior to ET, and non-conforming recipients were discarded. On the contrary, AI recipients were inseminated after synchronization, without later confirmation of ovulation. It could be that AI recipients, even though pregnancy was achieved, may have ovulated later than expected in comparison to IVP recipients and consequently, the P4 levels were lower. Nevertheless, both our AI and IVP groups had P4 values that are within the range of normality for day 7 [32–36]. After this difference on day 7, by day 30 and onwards, P4 concentrations in all groups were similar. This could also help to explain why we did not find any differences regarding the conceptus size. Higher levels of circulating P4 have been associated with conceptus elongation as early as day 14 [36,37]. Supplementation of progesterone in early days post-estrus has also been responsible for higher conceptus length [38,39]. Although day 7 P4 concentrations from our groups differ significantly between AI recipients and RF/IVP recipients, the fact that by day 30 P4 levels were similar, might have led to non-existing differences in size of conceptus. Whether the rise of P4 concentrations on AI group between day 7 and day 30 was sufficient to catch up with IVP groups or if the possible higher embryo quality (and interferon-τ secretion) from AI group were responsible for this lack of difference, remains to be explained.

The role of cortisol during pregnancy is still not totally understood. It is known that glucocorticoids have roles in many physiological processes (reviewed by [40,41]), but its importance on modulating the uterine environment and consequently, influencing embryo implantation, has been the subject of attention. In cattle, interferon-τ is part of the maternal-recognition system and it is secreted by the trophoblast into the uterus especially between day 13–21 of pregnancy, and as so, Majewska et al. [40] correlated the relationship between cortisol and interferon-τ. This study, after using epithelial and stromal cells from pregnant and non-pregnant cows, concluded that interferon-τ modulates the conversion of cortisone to cortisol just after 12 h of incubation, but that response is also dose-dependent, meaning that there is a necessary dose of interferon-τ to achieve this modulation. This is an important step, as by increasing cortisol levels in vivo, the prostaglandins will be down-regulated and the cow will be able to maintain the corpus luteum active. In our study, the day of collection showed a high influence over the results. Mean levels of cortisol by day 30 were very similar between groups, and then, in general, showed a decrease during the rest of the pregnancy. The only non-conforming group was BSA recipients that showed really high levels at day 7, then a decrease on day 30 continuing until day 90, and finishing with a slight increase at day 210. However, this response did not influence, as far as we understand, the outcomes of the pregnancies. Whether the vitrified-warmed IVP embryos differ from those produced in vivo in the expression of interferon-τ is not yet clear (reviewed by [42]).

Anti-Müllerian hormone is a growth factor produced by granulosa cells that has been positively correlated with pregnancy rate and maintenance, among other traits [43,44]. This hormone might be used as a biomarker for fertility, and even though its value varies much between individuals and breeds [43], it is quite stable within one individual and does not vary much within the estrus cycle. In general, our AI recipients showed lower levels of AMH when compared to the other groups, but this difference was only significant when comparing to IVP data on day 30. All the obtained values are in the mean average range of AMH concentration in Holstein cows [43], which is around 0.250 ng/mL, and could be considered an intermediate value (not too high nor too low). This is important because the same study [43] unveiled that cows with low AMH had greater risk for early pregnancy loss, which was not the case of our AI recipients. Nevertheless, these studies need to be taken into careful consideration, since they only use AI as a pregnancy method, and we are not fully confident that the embryo does not itself have an influence over their mother's hormonal values, as it happens with foetal sex [45].

Oestrogen levels were quite similar within groups of pregnant cows. However, when evaluating levels of E2, it is important to compare with P4 levels, thus the necessity of E2:P4 ratio. This ratio showed a high influence of the day, group and day x group interaction, and much higher levels for AI recipients vs. BSA/RF/IVP on day 7, as expected after the differences found on P4 levels. Later on, at day 210 there was again a tendency of difference between AI and BSA recipients. With the exception

of day 7 differences, where the reason for the higher E2:P4 was the low P4 concentrations, the day 210 tendency was due to high levels of E2. Fuchs et al. [46] described that concentrations of E2 tend to decrease after day 20 of estrus until day 250 where they start to rise again to prepare for parturition. Nonetheless, they also describe E2 concentrations at day 20 around 24 pg/mL, which is within the range for our AI group, meaning that it is not that AI recipients have high levels of E2, but rather that BSA cows expressed lower levels than expected. However, all these eight BSA recipients gave birth to live calves.

Gestation lengths are reported to be longer for IVP pregnancies in some studies [15,47–52], but similar to AI in others [5,53]. It should be noted that the percentage of induced parturitions was higher in IVP (regardless of the RF or BSA supplementation) pregnancies than AI pregnancies (with a tendency to be significantly different), and those pregnancies lengths (*n* = 7) were not accounted for mean gestation length, which if it did, would probably contribute to an increase in the mean gestation length. Moreover, one of the RF-pregnancies was not induced and was led to term without induction, lasting 298 days. Although gestation length is considered a highly heritable trait (reviewed by [54]), it is important to refer to the fact that the same bull was used in both AI and IVP groups. The mean value for gestation length for calves from Asturian Valley breed is 286.6 days for female and 287.5 days for male calves [55], but since AI cows were Holstein and there is no traceability on the female donors for IVP embryos (crossbred beef), it is not possible to attribute these differences to any of the estimated breed values.

Parturitions in IVP-derived pregnancies are known to have a tendency to be more difficult than AI pregnancies [29,51,52,56]. In our study, three IVP pregnancies were considered difficult while all AI calving were evaluated as easy, not being statistically relevant. The proportion of male calves being higher in IVP vs. AI has also been pointed as a distinct characteristic [15]. In fact, some studies reported higher percentage of male calves [3,50,56,57], while on the contrary others reported a similar proportion [58–60]. Nevertheless, our results were that AI group had the highest proportion of males vs. female, followed by BSA group, and the lowest proportion was given by the RF group.

Weight at birth was not statistically different between groups. This is in disagreement with previous studies that reported higher birth weight for IVP calves [30,47–50,52,53,57,59,61,62], but in agreement with others [63]. Interestingly, our heaviest calf was IVP-derived (56.2 kg) as well as our lightest (21.5 kg). The AI *Asturian x Holstein* combination brought a high proportion of calves with weights within 40–50 kg.

Mortality during the neonatal period happened only with IVP calves. Numabe et al. [64], Behboodi et al. [57] and Van Wagtendonk-De Leeuw et al. [53] reported higher incidence of perinatal mortality in IVP-derived calves than AI-derived calves. Both calves that died during parturition were heavy calves (>50 kg), which is an attribute given to the IVP-origin. The calf that died 2 days after birth was in agreement with Jenkins et al. [54] that associated longer gestation to higher perinatal mortality. Another calf died after surviving 12 days of premature deliver (240 days). Finally, the last calf that died at 13 days old was due to one of the most common causes of death in neonatal calves, diarrhoea.

#### **5. Conclusions**

In this study we applied new strategies in ART towards the more physiological approach of IVP, by including natural reproductive fluids in the culture media. With our limited sample size, we were not able to find worthy differences at the first stages of embryo development, implantation and parturitions, derived from the addition of such fluids. By contrast, we found that most of the parameters studied were similar between both types of IVP derived embryos and the in vivo-derived embryos, suggesting the IVP technology used was efficient enough for the safe production of calves. Since there are changes that may only be detected phenotypically later in life, as one may understand from other works [53], it is imperative to study the development and growth of these animals in order to reach more consistent conclusions. Moreover, it is important to assess if the changes that have been previously found in pre-implantation stages are in any form present in the live animals,

particularly regarding large offspring syndrome. The fact that these ART are being used globally to improve genetics of the herd or even to overcome heat stress in affected areas, make them essential tools that need to be studied thoroughly.

**Author Contributions:** Conceptualization, P.C.; methodology, J.S.L., D.R. and P.C.; formal analysis, J.S.L.; investigation, J.S.L., E.A.-T., C.S.-Ú., M.H. and S.C.; writing—original draft preparation, J.S.L. and P.C.; writing—review and editing, J.S.L., E.A.-T., C.S.-Ú., M.H., S.C., D.R. and P.C.; funding acquisition, D.R. and P.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by European Union, Horizon 2020 Marie Sklodowska-Curie Action, grant number REPBIOTECH675526 and as well as by the Ministry of Economy and Competitiveness (Spain), grants number AGL2015-66341-R & AGL2015-70140-R MINECO-FEDER and Fundación Séneca, grant number 20040/GERM/16.

**Acknowledgments:** The authors are grateful to Transformación Ganadera De Leganés SA; Matadero Madrid Norte, San Agustin de Guadalix; and Carnica Colmenar SC, in Madrid, Spain for providing access to biological material (ovaries); also to Roberth and rest of the staff from El Barranquillo SL for their precious assistance with the animals. Extended gratitude towards Francisco Javier Ibáñez López for statistical support and Mikhael Poirier for English editing.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Study of the Metabolomics of Equine Preovulatory Follicular Fluid: A Way to Improve Current In Vitro Maturation Media**

**Pablo Fernández-Hernández 1,2, María Jesús Sánchez-Calabuig 3,4, Luis Jesús García-Marín 1,5, María J. Bragado 1,6, Alfonso Gutiérrez-Adán 3, Óscar Millet 7, Chiara Bruzzone 7, Lauro González-Fernández 1,6,**† **and Beatriz Macías-García 1,2,\*,**†


Received: 20 April 2020; Accepted: 14 May 2020; Published: 19 May 2020

**Simple Summary:** Commercial in vitro embryo production in horses by ICSI (intracytoplasmic sperm injection) is currently used to produce embryos clinically. However, the successful pregnancy and foaling rates obtained after ICSI are only 10% of the oocytes matured in vitro. Conditions used for oocyte in vitro maturation are not optimized for equine oocytes. Hence, in the present work, we aimed to elucidate the major metabolites present in equine preovulatory follicular fluid obtained from postmortem mares using proton nuclear magnetic resonance spectroscopy (1H-NMR). Twenty-two metabolites were identified; among these, nine of them are not included in the composition of in vitro maturation media. Hence, our data suggest that the currently used media for equine oocyte maturation can be further improved.

**Abstract:** Production of equine embryos in vitro is currently a commercial technique and a reliable way of obtaining offspring. In order to produce those embryos, immature oocytes are retrieved from postmortem ovaries or live mares by ovum pick-up (OPU), matured in vitro (IVM), fertilized by intracytoplasmic sperm injection (ICSI), and cultured until day 8–10 of development. However, at best, roughly 10% of the oocytes matured in vitro and followed by ICSI end up in successful pregnancy and foaling, and this could be due to suboptimal IVM conditions. Hence, in the present work, we aimed to elucidate the major metabolites present in equine preovulatory follicular fluid (FF) obtained from postmortem mares using proton nuclear magnetic resonance spectroscopy (1H-NMR). The results were contrasted against the composition of the most commonly used media for equine oocyte IVM: tissue culture medium 199 (TCM-199) and Dulbecco's modified eagle medium/nutrient mixture F-12 Ham (DMEM/F-12). Twenty-two metabolites were identified in equine FF; among these, nine of them are not included in the composition of DMEM/F-12 or TCM-199 media, including (mean ± SEM): acetylcarnitine (0.37 ± 0.2 mM), carnitine (0.09 ± 0.01 mM), citrate (0.4 ± 0.04

mM), creatine (0.36 ± 0.14 mM), creatine phosphate (0.36 ± 0.05 mM), fumarate (0.05 ± 0.007 mM), glucose-1-phosphate (6.9 ± 0.4 mM), histamine (0.25 ± 0.01 mM), or lactate (27.3 ± 2.2 mM). Besides, the mean concentration of core metabolites such as glucose varied (4.3 mM in FF vs. 5.55 mM in TCM-199 vs. 17.5 mM in DMEM/F-12). Hence, our data suggest that the currently used media for equine oocyte IVM can be further improved.

**Keywords:** IVM; oocytes; equine; metabolomic

#### **1. Introduction**

Production of equine embryos in vitro is currently a commercial technique and a reliable way of producing embryos for vitrification or uterine/oviductal transfer [1]. In order to produce those embryos, immature oocytes are retrieved from postmortem ovaries or live mares by ovum pick-up (OPU) [2], matured in vitro (IVM), fertilized by intracytoplasmic sperm injection (ICSI), and cultured until day 8–10 of development [3]. However, among all the oocytes used for ICSI, in the best of the scenarios, roughly 10% of them end up in successful pregnancy and foaling [1,4]. Surprisingly, when in vivo matured equine oocytes are transferred to oviducts of live mares to produce equine offspring, the likelihood of pregnancy rises to 75%, contrasting with the 40% obtained when the oocytes transferred are matured in vitro [1]. These results highlight the fact that the media and conditions used for oocytes matured in vitro (IVM) largely differ from the physiological conditions required for correct nuclear and cytoplasmic maturation in horses, therefore decreasing the oocyte's developmental competence.

It has to be noted that the base media more commonly used for equine IVM are tissue culture medium 199 (TCM-199) or Dulbecco's modified eagle medium/nutrient mixture F-12 Ham (DMEM/F-12), which are generally chosen depending on the preferences of the laboratory where IVM is performed, and core differences exist among them [5]. Furthermore, none of these media have been developed specifically for equine IVM; instead, they were developed for cell culture, albeit equine cumulus–oocyte complexes (COCs) are capable of maturing with similar efficiency in either media [3,6,7].

To try to better understand the physiological conditions that equine COCs require and improve current IVM conditions, several reports have tried to address the metabolic requirements of equine COCs in vitro [5,6,8], the differences between the proteomic profiles of equine COCs maturated in vivo or in vitro [8], or the differential expression and localization of glycosidic residues in equine COCs matured in vitro vs. in vivo [9], among other approaches. All these reports have revealed a specific metabolic profile of equine COCs matured in vitro and important differences between equine COCs matured in vitro vs. in vivo. However, no research has been conducted to develop a defined oocyte equine-specific maturation medium. Hence, in the present work, we aimed to elucidate the metabolomic composition of equine preovulatory follicular fluid (FF). To do this, the major metabolites present in equine preovulatory follicular fluid were analyzed by high-field proton nuclear magnetic resonance spectroscopy (1H-NMR) and the results were contrasted against the composition of the formerly mentioned media according to the manufacturer's specifications.

#### **2. Materials and Methods**

#### *2.1. Collection of Equine Follicular Fluid*

Follicular fluid was obtained immediately postmortem at a commercial slaughterhouse, on four separate days. At the time of evisceration, the entire mare reproductive tract was extracted and carefully inspected. The ovaries were examined and those tracts having a preovulatory follicle ≥35 mm in diameter, associated with uterine edema on examination of the opened endometrial surface (vivid endometrial folds with a gelatinous appearance), were sampled as preovulatory. The fluid was collected using a 10 mL plastic syringe attached to a 20 g hypodermic needle. The fluid obtained was separated into 1.5 mL Eppendorf tubes and centrifuged for 2 min in a microcentrifuge at room temperature (RT) to remove large cellular masses. The supernatant was retrieved, transferred to a clean tube, and placed in dry ice until its arrival at the laboratory (4–5 h). Once at the laboratory, the fluid was thawed and centrifuged at 16,000× *g* at 4 ◦C for 20 min, and the supernatant was transferred to a clean tube. The samples were then kept at −80 ◦C until analysis.

#### *2.2. Sample Preparation*

Samples of follicular fluid from six different mares (one sample per mare) at the preovulatory stage (PRE) were used (*n* = 6). For the preparation of the nuclear magnetic resonance samples (NMR), the follicular fluids were pretreated. A methanol extraction was performed with the following protocol: samples were defrosted at room temperature for 30 min slowly on ice and 170 μL of follicular fluid of each sample were placed in a 1.5 mL Eppendorf tube and 1.3 mL of a mixture of methanol:deuterated water in a ratio 2:1 was added to the follicular fluid. The Eppendorf tube with the extraction was placed at 4 ◦C with agitation for 4 h. The mixture was centrifuged at 4 ◦C at 25,000× *g* for 30 min. The supernatant was transferred to a new 2 mL Eppendorf tube and the samples were plunged in liquid N2. Once the mixtures were frozen, samples were subjected to lyophilization. For sample analysis, the lyophilized product was resuspended with 500 μL of 0.2 M potassium phosphate buffer in deuterium oxide (D2O) with a pH of 7.4 ± 0.5 and 1.11 μL of TSP (3-(Trimethylsilyl) propanoic acid), to reach a final volume of 500 μL. Samples were briefly vortexed and 500 μL of the follicular fluid/buffer mixture were finally pipetted into a 5 mm NMR tube. In all cases, sample preparation was manually done at RT.

#### *2.3. NMR Measurements*

Samples were measured at 298 K in an 800 MHz Bruker spectrometer (AVANCE III, Bruker Biospin GmbH, Reinsthetten, Germany) equipped with a 1H detected cryoprobe with z-gradient and automatic tuning and matching unit. Optimization of experimental conditions included automated tuning and matching, automated locking, and automated shimming using TopShim. The 90◦ hard pulse was optimized to be sample-specific and the presaturation field strength was adjusted to 25 Hz. To minimize the interference of the water content in the NMR spectrum, solvent suppression techniques were applied.

For each sample, one-dimensional (1D) 1H-NMR spectra were collected using a Carr–Purcell– Meiboom–Gill (CPMG) pulse sequence; 2D *J*-resolved included 800 × 40 points. Data analysis was done using the TopSpin 3.5 software (Bruker Biospin GmbH, Reinsthetten, Germany). Free induction decays were multiplied by an exponential function equivalent to 0.3 Hz line-broadening before applying a Fourier transformation. All transformed spectra were automatically corrected for phase and baseline distortions and referenced to the DSS singlet at 0 ppm for further analysis.

The 2D-*J*res experiment was also routinely included in the acquisition package, along with the 1D 1H-NOESY. This experiment separates J-couplings and chemical shifts in the 2D plane and provides a useful and simplified proton-decoupled projection spectrum. A standard pulse sequence with a water peak suppression was used. After 16 dummy scans, 2 free induction decays (FIDs) were accumulated into 8 k × 40 data points at a spectral width of 16 ppm.

For assignment purposes, a battery of experiments including 2D-1H, 13C-HSQC, 2D-1H, 1H-TOCSY, and 2D 1H-1H-NOESY were recorded in a Bruker Avance III 800 MHz spectrometer (Figure S1). The chemical shift, multiplet type, and number of contributing nuclei to each metabolite are provided in Table 1 and were determined following previously validated methods [10–12].


**Table 1.** Chemical shift assignment, multiplicity, and number of contributing protons for the identified metabolites.

Multyplet type: s—singlet; d—doublet; t—triplet; dd—doublet of doublets; q—quadruplet; m—multiplet.

#### *2.4. Commercial Media Composition*

The composition of commercial media routinely used in our laboratory: TCM-199 with Earle´s salts (Ref. 31150022; Thermo Fisher Scientific (Waltham, MA, USA)) and DMEM/F-12 (Ref. 11320033; Thermo Fisher Scientific (Waltham, MA, USA)) were directly extracted from the manufacturer´s website.

#### *2.5. Statistical Analysis*

Data were analyzed using descriptive statistics to establish the mean, standard error of the mean, minimum, and maximum for each metabolite using the software Sigma Plot (ver. 12.0) for windows (Systat Software, Chicago, IL, USA).

#### **3. Results**

#### *3.1. Metabolite Identification*

The chemical shift assignment for each metabolite was performed using a random follicular fluid sample; the spectra of the FF (Figure S1) was contrasted against the identified metabolites that were chosen based on our previous study in horses [13]. The list of the chemical shift for each proton nucleus of each metabolite is provided in Table 1. The measured concentrations of 22 metabolites, which were selected based on the report of González-Fernández et al. (2020) [13], and the known identification capabilities of the NMR facility for preovulatory follicular fluid samples are presented in Table 2. Pyruvate and succinate are characterized by one single peak with the same chemical shift and cannot be discriminated; therefore, they are presented as the sum of both metabolites. All metabolites were detected in all the samples except for acetylcarnitine, which could not be detected in one sample submitted to NMR analysis.


**Table 2.** Concentrations of metabolites detected in equine preovulatory follicular fluid.


**Table 2.** *Cont*.

The results are presented as mean ± SEM (minimum value−maximum value); the values correspond to 6 different mares (*n* = 6).

#### *3.2. Comparison of the Metabolites Present in Commercial Media and Equine Preovulatory Follicular Fluid*

Among the 22 metabolites identified in native preovulatory FF, nine of them are not present in TCM-199 or DMEM/F-12 according to the manufacturer's specifications (Table 3). These metabolites were: acetylcarnitine, carnitine, citrate, creatine, creatine phosphate, fumarate, glucose-1-phosphate, histamine, and lactate. Other metabolites such as acetate is present in FF and TCM-199 but not in DMEM/F-12, while pyruvate is included in the composition of DMEM/F-12 and possibly is present in FF (as it cannot be discriminated from succinate) but not in TCM-199. Vivid differences exist in the concentration of core metabolites such as lactate (27.3 mM in FF vs. 0 mM in TCM-199 and DMEM/F-12), glucose (4.3 mM in FF vs. 5.55 mM in TCM-199 vs. 17.5 mM in DMEM/F-12), alanine (1.1 mM in FF vs. 0.28 mM in TCM-199 vs. 0.05 mM in DMEM/F-12), aspartate (2.7 mM in FF vs. 0.22 mM in TCM-199 vs. 0.05 mM in DMEM/F-12), or glycine (3.2 mM in FF vs. 0.67 mM in TCM-199 vs. 0.25 mM in DMEM/F-12).


**Table 3.** Presence and concentration of the metabolites found in equine preovulatory follicular fluid, and in TCM-199 and DMEM/F-12 media (manufacturer's specifications).

#### **4. Discussion**

In the present work, the metabolome of equine preovulatory FF was investigated using 1H-NMR. Our work revealed the presence of at least 22 metabolites including carbohydrates, amino acids, and intermediate metabolites (Table 2). The metabolome of equine FF at different dominant follicular stages (early dominant, late dominant, and healthy preovulatory stage) has previously been described by Gérard et al. in 2002 [14]. In their work, they did not detect apparent differences in the pattern or concentration of the metabolites detected among the studied stages. These authors described eight peaks corresponding to chemical groups of sugar chains and N-acetyl groups of glycoconjugates, CH3 groups of lipoproteins, trimethylamines, acetate, alanine, creatine/creatinine, and polyamines, plus a non-identified peak at 3.1 ppm, but quantitative identification of the peaks is not provided [14]. In our work, we detected 21 peaks (as succinate and pyruvate were overlapped; Table 1); an explanation for the differences found in the number of peaks (8 vs. 21) can be easily explained as Gérard et al. (2002) used a 200 MHz Bruker spectrometer (in our work we used an 800 MHz spectrometer and a cryoprobe) and the samples were directly diluted in deuterated water (instead of being previously subjected to a methanol extraction and lyophilization as in the present work), likely resulting in higher water interferences and lower spectra resolution [14]. Hence, in the formerly mentioned work, some peaks as citrate are suspected, while in our work, it was detected in all samples due to a better resolution of current NMR spectrometers (Tables 1 and 2). It must be mentioned that citrate and fumarate are not routinely added to base equine IVM media (Table 3) as both are intermediate metabolites produced by the metabolism of glucose in the tricarboxylic acid cycle (TCA); specifically, citrate comes from acetyl-CoA and oxaloacetate in the tricarboxylic acid cycle (TCA). Citrate acts as a key substrate for epigenetic modifiers in the oocyte [15,16] and is a link between TCA, β-hydroxybutyrate, and lipid metabolism in FF [17,18], so adequate supplementation could be important during equine IVM and should be added to TCM-199 (Table 3). Regarding pyruvate, this molecule has been previously reported to range between 0.03 and 0.13 mM in FF from early and late dominant equine follicles, respectively [19]. This metabolite has been demonstrated to be crucial for adequate oocyte metabolism [5] and is also involved in active reactive oxygen species scavenging [20]. In equine oocytes, it has been demonstrated

that when DMEM/F-12 is supplemented with pyruvate at 0.15 mM, this induces an increase in glycolytic activity, without affecting mitochondrial oxidative phosphorylation [5]. The concentration of pyruvate above reported for equine FF [19] coincides with our work in which 0.16 mM ± 0.03 was observed. However, in our experiments, succinate could not be discriminated from pyruvate, and thus, exact values cannot be provided; nevertheless, as per previous reports, pyruvate addition to equine IVM media and the concentration at which it is supplemented needs to be seriously considered.

The research group of Gérard et al. also published in 2015 another report in which they performed a comparative metabolomic study of porcine, equine, and bovine FF [21]. In this report, even when the extraction method remained the same as in their previous work, a 500 mHz 1H-NMR spectrometer was used and the resolution of the analysis was improved. In this work, they described the presence and concentration of some metabolites that coincide with the ones reported in the present work such as acetate, alanine, citrate, and glucose (alpha and beta, while in our work, glucose and glucose-1-phosphate were detected) [21]. However, the concentration of other metabolites reported in their work such as histidine (0.05 ± 0.009 mM in our work vs. 0.27 mM [21]) and valine (0.13 ± 0.02 mM in our work vs. 0.37 mM [21]) do not agree with our findings. These differences can be attributed to the different equipment used, the sample extraction method, or the fact that Gérard et al. (2015) recovered the FF by transvaginal aspiration of follicles around 33 mm and in our work postmortem FF was obtained from bigger follicles (35–45 mm). Interestingly, lactate concentration largely differs in our report compared with that obtained by Gerard et al. [21] (27.3 mM ± 2.2 vs. 0.70 mM ± 0.18, respectively), the time between mare decease and sample obtention being around 30 min. This time lapse could lead to lactate postmortem accumulation [22] as previously observed for equine oviductal fluid [13]. Nevertheless, the high lactate concentration observed in the present work (19.3–35.02 mM) cannot be solely explained as the metabolism of the typical amount of available glucose (2 molecules of lactate per 1 of glucose consumed [17]), as the concentration of glucose observed in our work matches previous reports [14,19,21,23], and thus appears to reflect a high level of lactate in the equine FF. However, it has to be noticed that the equine FF analyzed by Gérard et al. (2015) [21] was retrieved when the dominant follicle reached 33 mm, while it has been demonstrated that in equine follicles reaching 39 mm 24 h after equine crude gonadotropin administration, the lactate values reach 4 mM, this concentration of lactate being closely related to adequate meiosis resumption in equine oocytes [19]. Similar findings among lactate production during IVM and oocyte competence in the horse have been recently reported in vitro [5], highlighting the important relationship existing between anaerobic glycolysis and oocyte meiotic competency in the horse, strongly suggesting that FF may be providing energy to the oocyte in the form of lactate as postulated in humans [17]. Furthermore, high lactate concentrations in the FF have been linked to improved pregnancy rates in humans [24], and this metabolite could be an important additive contributing to osmolarity adjustment, as reported in human oviductal fluid [25]; hence, lactate would need to be supplemented to equine IVM media (Table 3).

Notably in our NMR spectra, different intracellular metabolites such as glucose-1-phosphate, creatine, creatine phosphate, and carnitine were found. These metabolites cannot be incorporated into the oocytes in their molecular form and may have been released from the granulosa cells due to cellular damage. However, recent investigations have demonstrated that a wide variety of intracellular metabolites are also present in the oviductal fluid of cows and horses [13,26], embedded in extracellular vesicles [27]. The compounds included in these vesicles could enter the oocytes/embryos via fusion as previously demonstrated in mouse embryos [28], contributing to the oocyte's developmental competence and metabolism as observed in cows [29]. It is known that carnitine plays a major role in the catabolism of lipids, allowing the transport of fatty acids from the cytosol to the mitochondria, where they are metabolized through beta-oxidation, which has also been identified in human FF [30]. Surprisingly, this work demonstrated the presence of carnitine in the FF but did not find expression of the enzymes involved in the carnitine synthesis either in oocytes or in cumulus cells. In contrast, the enzymes related to beta-oxidation were highly expressed in the oocyte and cumulus cells as also demonstrated in horses [8]. Therefore, this compound may also need to be somehow included in

equine IVM media to ensure adequate oocyte lipid metabolism [30], as low concentration of carnitine in the FF has been associated with low reproductive performance outcome in sows [31]. The presence of creatine in the equine FF has been previously reported at similar concentrations to the ones reported in the present work [14]. Creatine and creatine phosphate are produced as a result of arginine and glycine metabolism. Even when arginine was not found in equine FF (Tables 1 and 2), arginine was present in TCM-199 (0.33 mM) and DMEM/F-12 (0.7 mM). Interestingly, arginine depletion during the final 6 h of IVM of human oocytes was associated with a higher maturation potential [32,33]. Hence, even when arginine was not detected in our experiments (Table 1), our data suggest that an active metabolism of arginine could be occurring during in vivo maturation of equine oocytes explaining the depletion of this amino acid in the FF and thus, the arginine present in the equine IVM media (TCM-199 and DMEM/F-12) could be needed. Glycine is known to be an organic osmolyte that regulates osmolarity in cells and embryos [34] and is one of the most abundant amino acids in follicular and uterine fluids [35]. Interestingly, its concentration in the FF has been demonstrated to predict the cleavage rate of oocytes after insemination (being a stronger marker of cleavage capacity in lower grade oocytes) as well as to be a good marker of the blastocyst rate in bovine [36]. Hence, considering the vivid difference existing in glycine concentration among TCM-199, DMEM/F-12, and native equine preovulatory FF (Table 3), the concentration at which this amino acid is added to equine IVM media should be carefully evaluated.

Interestingly, glucose-1-phosphate was detected in equine preovulatory FF (Table 1). This molecule is derived from glycogen, which is generally metabolized in the liver but is also a metabolic source used by granulosa cells in humans, pigs, bulls, and mice [15,37,38]. Thus, in view of our data, the equine oocyte also relies on glycogen metabolism as is also proposed in previous reports [5]. This is an interesting finding as, in our experiments, the amount of glucose-1-phosphate found (6.9 mM ± 0.4) surpassed the quantity of glucose (4.3 ± 0.4 mM), indicating that glycogen metabolism could support equine oocyte maturation in vivo, and more research is needed in this field, as this energy source is not generally considered in equine oocytes.

Other metabolites such as histamine, which has recently been found in equine oviductal fluid [13], are also present in horse FF and could be involved in ovulation induction, as observed in rabbits and rats [39,40] and as also postulated in horses [41].

Our results demonstrate that the base media used for equine IVM and the composition of preovulatory equine FF greatly differ. One limitation of the present study is that the composition of fetal bovine serum (FBS) that is usually added at concentrations ranging from 10% to 20% to equine IVM media has not been considered [4,42]. It is well known that FBS composition greatly varies among batches, and thus, a significant error could be introduced if the composition of a single batch was considered; this is why FBS composition is not considered in the present report.

#### **5. Conclusions**

In conclusion, our data provide new insights into equine preovulatory follicular fluid composition comparing it with the most widely used commercial media available for equine IVM (TCM-199 and DMEM/F-12). Our data provide new metabolite information that should be considered to design specific equine IVM media and help to improve current in vitro fertilization outcomes in the equine species. More research is warranted to better understand the metabolic requirements of equine oocytes and the relationship among metabolism, oocyte meiotic competence, and developmental competence.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-2615/10/5/883/s1, Figure S1. Metabolite assignment on 1D 1H NMR spectra of an FF sample. (**A**) Chemical shift region from 6.4 to 8.0 ppm, (**B**) chemical shift region from 3 to 4.2 ppm, (**C**) chemical shift region from 1.8 to 2.5 ppm, (**D**) chemical shift region from 0.8 to 1.5 ppm.

**Author Contributions:** Conceptualization, L.G.-F. and B.M.-G.; Data curation, P.F.-H. and Ó.M.; Formal analysis, P.F.-H., Ó.M., C.B., and B.M.-G.; Funding acquisition, M.J.S.-C., L.J.G.-M., M.J.B., A.G.-A., and B.M.-G.; Investigation, M.J.S.-C., A.G.-A., C.B., and L.G.-F.; Methodology, M.J.S.-C. and Chiara Bruzzone; Project administration, B.M.-G.; Resources, Ó.M.; Software, C.B.; Supervision, Ó.M.; Validation, C.B.; Writing—original draft, P.F.-H., M.J.B., A.G.-A., and B.M.-G.; Writing—review & editing, M.J.S.-C., L.J.G.-M., Ó.M., and L.G.-F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by Spanish Ministry of Economy, Industry and Competitiveness and Fondo Europeo de Desarrollo Regional (FEDER) (AEI/FEDER/UE); References: RTI2018-093548-B-I00, AGL2017-84681-R and RYC-2017-21545 (this last awarded to B. Macías-García). L. González-Fernández was supported by regional grant "Atracción y retorno de talento investigador a Centros de I+D+i pertenecientes al Sistema Extremeño de Ciencia, Tecnología e Innovación" from "Junta de Extremadura" (Spain); Reference: TA18008. P. Fernández-Hernández was supported by a grant "Acción II" from the University of Extremadura (Ref. Beca RC4).

**Acknowledgments:** The help of the veterinary team (Gerardo, Carmen, and Jesús) at the slaughterhouse (Incarsa, Burgos, Spain) is warmly appreciated. This work is dedicated to all the victims of Covid-19 who will not directly benefit from our results but would possibly be relieved to know that science never stops, not even in difficult scenarios. To Daniela and Bruno González-Macías and all the kids who are valiantly coping with confinement.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Animals* Editorial Office E-mail: animals@mdpi.com www.mdpi.com/journal/animals

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18