Final Report Summary - FECUND (Optimisation of early reproductive success in dairy cattle through the definition of new traits and improved reproductive biotechnology)
Executive Summary:
The FP7 European research project Fecund ended in January 2017. For 48 months, the research partners and SMEs addressed the needs of the cattle industry: the problems of reproductive success in the cow using a systems biology approach to deliver targets for genomic selection of fertility, the improvement of biotechnologies for reproduction and the identification of molecular indicators to support early stages of reproduction were investigated. The project structure was defined by 9 scientific work-packages (WP1-9), one transfer and dissemination WP (WP10) and a management WP (WP11).
WP1 was focussed on the establishment of two animal models to independently examine the effects of (i) the metabolic status of the cow and (ii) genetics on the factors involved in establishing a pregnancy. Animals in both models were used to generate follicular fluid, oocytes, embryos, oviduct tissue and fluid, conceptuses, endometrial tissue and uterine lumen fluid which were analysed in subsequent WPs. In WP2 transcriptomic and proteomic analysis were carried out to investigate biological systems directly involved in the formation of oocytes. In WP3 advanced techniques were developed for sampling. Analysis on oviduct cells, sperm transport and fertilization and embryos were performed.In WP4 the question whether metabolic condition or genetic merit for fertility affects the conceptus-endometrial crosstalk in dairy cattle was addressed. WP5 addressed novel approaches to identify candidate recessive lethal mutations, identify variations within key genes involved in early reproductive biology identified in previous WPs.
WP6 carried out activities to identify robust biomarkers predictive of the reproductive status of the cow in biological fluids easy to obtain as plasma and vaginal swabs. WP7 managed to improve the techniques for assisted reproduction taking advantage of the novel findings derived from the animal models developed in previous WPs and testing novel hypothesis reported in recent literature in order to improve oocyte and embryo developmental competence. WP8 provided the bioinformatic structure to manage the data produced and the tools for primary data analysis.
WP9 evaluated the data and technological advances achieved in earlier WPs and developed or tested direct applications that can be used by the industry in the short term to start reversing the negative reproductive trend. WP10 covered the main dissemination tasks through a public website, periodic newsletters, leaflets and workshops.
Main objectives achieved by FECUND were to:
carry out appropriate scientific studies focused on early embryonic development to create new knowledge; develop improved ways of defining, measuring and recording fertility traits including the identification of direct and surrogate measurements of physiological variables;identify phenotypes associated with metabolic profiles and to correlate them with low or high fertility; identify biomarkers associated with early aspects of reproduction that can be used to quantify the new phenotypes;identify genetic loci associated with reproductive performance in the breeding populations; add functional validation to genetic information using system biology studies; contribute to the improvement of reproductive biotechnologies for efficient livestock production; disseminate the information to the wider cattle breeding industry and to end users.
The collaboration and interaction among partners of the consortium led to the achievement of the objectives of the project.
Project Context and Objectives:
FECUND is focussed on fertility of dairy cattle, where the greatest reproductive problems exist and where the financial benefits of improved reproduction efficiency are greatest. The project addressed the reproductive biology of ovulation through implantation which is the period which is associated with greatest reproductive losses. FECUND worked together with the PROLIFIC project, which was funded under the same KBBE call for proposals, to address the most important reproductive problem facing cattle production at a European level.
Concomitant with intensive selection for increased milk yield, reproductive performance of dairy cows has declined in recent decades, in part due to an unfavourable genetic relationship between these traits. Today the decline in fertility has become a major concern of farmers and the dairy industry. The causes of this decline in reproductive success are multi-factorial (Pretheeban et al., 2009) with higher milk yield being shown to have an inverse relationship with the pregnancy rates observed in dairy cows (Faust et al., 1988). Embryonic loss is a major cause of this reduced pregnancy rate (Hansen, 2002). Successful implantation and establishment of pregnancy is dependent on good quality embryos, an appropriate endometrial environment, and the successful interaction between the embryo and the endometrium (Chen et al., 1999). While selection for productivity has had a negative impact on fertility and some genetic loci have been shown to have opposing effects on production and fertility, other loci act on only one or other of these traits. Understanding the relationship between metabolism and fertility and the interaction with genetics will allow producers to optimise productivity and reproductive success.
Good reproductive performance is crucial for economic sustainability as well as animal welfare. However, reproductive success is a highly complex trait and fertility problems in dairy cattle include an increase in postpartum failure to return to cyclicity, poor expression of oestrus, defective oogenesis, poor embryo survival and susceptibility to uterine infections. Both genetic and management changes can be envisaged that may improve reproductive success but these could compromise productivity. Therefore, it is important to understand the interaction between the various facets of reproduction and fertility to devise appropriate strategies to manage and improve reproductive success.
The challenges facing improvement in dairy cattle fertility can be addressed through two interacting factors: biological changes in dairy cattle, resulting from genetic selection, and changes in dairy herd management. The economic environment under which the dairy industry operates leaves little margin for error before dairy farming becomes unsustainable. Human and economic factors differ among countries, regions and individual herds, and so local management decisions may impact directly on fertility and profitability. Improvements in the definition and measurement of fertility in relation to the changing biological characteristics of dairy cattle would help to optimise these management decisions, and have direct economic benefits for the farmer. After the successful first calving each cow is expected to calve at an interval that maximizes the economic output from milk production. However, in the face of declining reproductive efficiency it is difficult to maintain the desired reproductive period, which results in early culling of low fertility cows, with the associated need to breed an increased number of replacements. Indeed the herd life of a modern Holstein cow is little more than 3 lactations (Oltenacu and Broom, 2010), in part caused by reproductive problems. Reduced age at the first calving has, to some extent, decreased the cost of breeding replacements, however, the short herd life is a major concern and large economic burden.
Metabolic demands placed on the cow to sustain a high level of milk production impact negatively on the reproductive function of postpartum cows. Factors other than fertility per se contribute to the decreasing reproductive efficiency of dairy herds. Among these are changes in expression and detection of oestrous behaviour, which results in mistimed insemination. Inappropriate cow management post partum during a period of high metabolic stress amplifies the problem. Nevertheless, the underlying fertility of dairy cows has declined and new knowledge is required to develop successful management systems to meet the metabolic and nutritional needs of highly productive cattle. In addition, housing and social needs of increasingly productive animals impact on their inherent capacity to achieve and maintain pregnancy.
Fertility is a complex trait affected by the environment, management and genetics. Efficiency of selecting for improved fertility using both traditional genetic selection and now genomic selection relies on having accurate fertility data; however, traditional measures of fertility have low heritability making selection of sires based on genetic merit for fertility traits unreliable. This is partly because indicator traits have been used that do not accurately reflect the true underlying biological variation. Furthermore, reproductive success is a trait that is expressed late in life, and so good information on fertility is not available when breeding decisions have to be made. Advanced, technically complex and expensive recording technologies that could be used to improve the measurement of fertility in experimental situations cannot be applied in a commercial setting. Novel approaches and indicator traits are therefore needed to measure fertility. These will be used to identify genetic loci and gene variants to improve selection, to assist cow management and to improve assisted reproductive systems.
The FECUND project addressed the problems of reproductive success in the cow using a systems biology approach to deliver targets for genomic selection of fertility, improve biotechnologies for reproduction and identify molecular indicators of the ability of the cow to support early stages of reproduction, when most embryonic loss occurs.
The overarching objectives of FECUND were therefore:
• to carry out appropriate scientific studies focused on early development to create new knowledge;
• to develop improved ways of defining, measuring and recording fertility traits including direct and surrogate measurements of physiological variables;
• to identify phenotypes associated with metabolic profiles and to correlate them with low or high fertility;
• to identify biomarkers associated with early aspects of reproduction that can be used to quantify the new phenotypes;
• to identify genetic loci associated with good reproductive performance in the breeding populations;
• to add functional validation to genetic information using system biology studies;
• to contribute to the improvement of reproductive biotechnologies for efficient livestock production;
• to disseminate the information to the wider cattle breeding industry and to end users.
Realising these objectives was essential to underpin the sustainability and competitiveness of the European cattle industry.
Project Results:
The Fecund consortium achieved all the planned objectives and deliverables for the project. The consortium has been very active and collaborative to perform the research, and then transfer and disseminate scientific outcomes to researchers and scientific community, policy makers and stakeholders.
WP1: Animal Models
Pregnancy establishment in mammals, including dairy cows, involves several key checkpoints along the developmental axis including the follicle/oocyte, oviduct/early embryo and uterus/conceptus. WP1 was focussed on the establishment of two animal models to independently examine the effects of (i) the metabolic status of the cow and (ii) genetics including genetic merit for milk production on the factors involved in establishing a pregnancy in order to address whether dairy cow infertility is primarily a genetic or metabolic problem. As such, WP1underpinned WPs 2, 3 and 4 as it was responsible for the generation of tissues and samples which were then distributed amongst the partners and used in the subsequent three WPs.
To breed and manage cows under well defined and controlled experimental conditions in order to produce information, biological tissues and fluids from animals of divergent metabolic and genetic status were collected for analysis in subsequent WPs.
The specific objectives are to:
• establish optimised experimental protocols and write SOPs for the experimental work and sample collection
• identify, source and breed suitable animals for the studies
• collect data and experimental materials following the standard operating procedures
• to provide the samples and data to partners for the execution of work in subsequent WPs
The Energy Balance model was based on a comparison of lactating and nonlactating postpartum cows and maiden heifers. Forty in-calf primiparous Holstein-Friesian heifers and 20 non-pregnant Holstein-Friesian heifers with a similar Economic Breeding Index were purchased and enrolled into the study. All cows calved between February and May 2013. At calving, cows were randomly assigned to one of two groups 1) lactating (n = 20) or 2) non-lactating (n = 20). From calving, animals in the lactating group were milked twice per day (0700 and 1600 h), while those in the non-lactating group were dried off immediately after calving (i.e. never milked. Pre-calving, all animals were fed 30 kg grass silage/head/day with pregnant heifers getting an additional 3 kg concentrates/head/day. Post calving, dry cows received ad lib grass silage plus 4 kg concentrates/day, while lactating cows received 24 kg maize silage/16 kg grass silage plus 7 kg concentrates. Beginning two weeks prior to expected calving date, all animals were weighed, body condition score (BCS) recorded and blood sampled twice weekly until Day 85 postpartum. Similar measurements were recorded for the non-pregnant heifers over the same period. A thorough metabolic analyses of all animals was carried out by measuring concentrations of NEFA, BHB, IGF-1, insulin, and glucose.
The Genetic model comprised three groups of heifers (n=45) to assess the impact of genetics on variations in fertility which were used to generate samples to be be used in subsequent WPs. The groups were:
- 15 Prim Holstein heifers with high genetic merit for fertility (FERTI+)
- 15 Prim Holstein heifers with low genetic merit for fertility (FERTI-)
- 15 Montbéliarde heifers, a French local breed with a good fertility (positive control group)
Animals were similarly characterised to those in the Energy Balance model.
Animals in both models were used to generate follicular fluid, oocytes, embryos, oviduct tissue and fluid, conceptuses, endometrial tissue and uterine lumen fluid which were analysed in subsequent WPs.
SOPs were generated for the collection and distribution of these tissues and biological fluids to optimize downstream data collection.
WP2: Follicular fluid protein profiles
WP 2 investigated the biological systems that are involved in the formation and development of oocytes. Therefore cumulus cells, oocytes and follicular fluid generated in Work package 1 were analysed. Proteome analysis in follicular fluid is generally biased by the extremely high abundance of a few individual proteins (e.g. serum albumin, complement c3). Nevertheless, more than 200 protein analytes could be addressed in the samples of both models. Concerning follicular fluid of the metabolic model, 251 proteins were identified, and quantitative differences in the follicular fluid proteome were detected using LC-MS/MS followed by statistical and bioinformatic analyses: Compared to heifers, 4 proteins showed reduced levels in dry cows and 6 proteins in lactating cows. Another six proteins differed between lactating and dry animals. In follicular fluid samples of the genetic model, 247 proteins were identified. Compared to high fertility heifers (LFH), the levels of 3 proteins were significantly enhanced in low fertility heifers. The comparison of low fertility heifers to Montbeliard revealed three proteins enhanced and three decreased in LFH. In the comparison high fertility heifers (HFH) to Montbeliard, 1 protein was less abundant and four were more abundant in HFH.
Thus, the data revealed clear differences in the follicular fluid protein composition between the relevant groups of both animal models. The affected proteins represent mainly the biological processes, metabolic process, response to stimulus and immune system process. According to GO analysis the most prevalent molecular functions of these proteins are catalytic activity, e.g. glutathione peroxidase 3 and serotransferrin, binding and enzyme regulator activity with the involvement of lactadherin, alpha-2-macroglobulin and C8G protein.
RNA expression profiles in cumulus cells: Cumulus cells (CC) are involved in oocyte nutrition and protection and essential for oocyte metabolism, thus influencing oocyte quality. Using NGS technology (RNAseq), we compared RNA expression profiles in cumulus cells of Holstein-Friesian and Monbeliarde animals in metabolic and genetic models. In the CC from Heifers and Lactating cows, 187 DEG enriched were evidenced DEGs were enriched in GO related to lipoprotein binding and steroidogenesis, transcription factor (TF) and protein kinase activity. Thus, NGS analysis of bovine cumulus cells in metabolic model evidenced a strong impact of lactation on transcriptional activity in cumulus cells by the regulation of a large number of genes in particular those involved in intracellular lipid metabolism. Within the genetic model, the highest number of DEGs were found in the comparison of Montbéliard and Fertiplus (205 DEG : 37 higher in Montbéliard, 168 higher in Fertiplus), while the comparison of Montbéliard and Fertiminus revealed only 102 DEG. Interestingly, these 102 DEGs were able to discriminate Montbeliarde and FertiMin groups using PCA, whereas no clear discrimination was observed between Monbéliarde and FertiPlus. The data indicate a strong impact of lactation on transcriptional activity in cumulus cells by the regulation of a large number of genes in particular those involved in intracellular lipid metabolism.
RNA expression profiles in oocytes: Oocytes accumulate the RNA during their growth period for using maternal factors during transcription silencing through oocyte maturation and first embryo cleavages. RNA expression profiles in the immature oocytes of Holstein-Friesian and Montbeliarde animals in metabolic and genetic models was compared using NGS technology (RNAseq). In the metabolic model, oocytes of dry and lactating cows showed 8 DEG, and Oocytes from Dry cows and Heifers showed two DEGs, whereas Heifers and Lactating cows showed 51 DEG: 19 genes were up-regulated and 32 down-regulated in lactating cows. These genes are involved in cellular biosynthesis and transcription regulation. GO terms were enriched in genes of apoptosis and TGF-beta signalling. In the genetic model, 36 DEGs were found between the oocytes from FertiPlus and FertiMinus heifers: 26 genes were up-regulated and 10 down-regulated in FertiPlus. A comparison of Montbéliarde and FertiPlus Holstein oocytes revealed 202 DEGs, 83 up-and 119 down-regulated in Montbéliarde. 68 DEGs were detected by comparison of Montbéliarde and FertiMinus oocytes, 28 were up-regulated and 40 down-regulated in Montbéliarde. DEGs were enriched (p<0.05) in molecular function GO ‘”nuclear acid binding”.
In summary, NGS analysis of bovine oocytes in metabolic model showed that oocyte transcriptomes were moderately affected by the different energy balance due to lactation. In the genetic model, oocytes from the Monbéliarde heifers showed the most discriminative transcriptome profile in comparison to FertiMinus Heifers, which may at least partly explain the difference in their reproductive performance.
Factors influencing oocyte quality: Cumulus cells transcriptome analysis in different models revealed that differently expressed genes (DEG) were involved in particular in pathways related to lipid metabolism with a high expression level of genes involved in fatty acid (FA) synthesis, transport, elongation, degradation and oxidation. Therefore we analysed whether the factors related to FA metabolism may influence oocyte quality: by RT PCR of thirty genes at different times of IVM culture, by expression at the protein level by Western blotting and immunofluorescence, and by mass spectrometry to establish lipid fingerprints. Functional studies using pharmacological modulators of lipid metabolism demonstrated that during IVM (i) C75 neither affected oocyte maturation rate nor viability of CC and oocytes (DO and CEO); (ii) Mildronate had no impact on oocyte/CC viability but significantly diminished oocyte maturation rate in dose-dependent manner (CEO); (iii) Etomoxir dose-dependently blocked meiosis progression in both CEO and DO and affected CC viability, and in consequence, oocyte survival. Mass spectrometry analysis revealed significant differences between lipid fingerprints in CC before and after IVM, and numerous differentially abundant lipid species were determined (p<0.05). Moreover, CC showed significant accumulation of lipid droplets throughout IVM as shown by microscopy analysis.
In addition, it could be shown that signalling pathways were affected by lipid metabolism regulators in oocytes and cumulus/granulosa cells: in oocytes, etomoxir significantly decreased MAPK 3/1 phosphorylation after 6h IVM, and in cumulus/granulosa cells, this agent decreased phosphorylation of MAPK14 (alias P38) and SMAD1/5/8 (p<0.05) but not pMAPK3/1. In summary, functional studies based on in vitro maturation of bovine oocyte-cumulus complexes demonstrated a crucial role of fatty acid metabolism in CC on oocyte maturation and survival. Numerous genes known to be expressed only in lipogenic tissues (adipose, liver, muscles) were investigated in cumulus cells for the first time.
The data suggest that an active regulation of lipid metabolism promotes the successful completion of oocyte maturation and affects oocyte developmental competence.
Effects of selected factors on oocyte quality: Data derived in FECUND describe significant differences in the composition of the follicular fluid of postpartum cows induced by lactation, which could potentially influence the quality of the oocyte ovulated and subsequent fertility. Using these data generated from our unique model of postpartum dry and lactating cows and maiden heifers we have examined the direct effect of certain follicular fluid components on oocyte quality in vitro. Expression level of several key genes involved in lipid metabolism was measured by real-time PCR. Analysis of gene expression in cumulus cells after 22h IVM showed that 100 µM docosahexaenoic acid (DHA) induced a three-fold increase in the FA transporter CD36 and a two-fold decrease in FA synthase (FASN) gene expression in the cumulus cells of corresponding oocytes, whereas lower DHA concentrations had no effect. We could show that administration of a low physiological dose of DHA (1 µM) during IVM may have beneficial effects on oocyte developmental competence in vitro with no effect on nuclear maturation or lipid metabolism in surrounding cumulus cells. In contrast, high DHA concentration (100 µM) had negative effect on embryo development and FA synthesis in cumulus cells. Based on data from Task 2.1 highlighting differences in components of follicular fluid in maiden heifers and postpartum dry and lactating cows, we tested the effect of Stearic acid, Linoleic acid, Linolenic acid, Myristic acid, Docosahexaenoic Acid, and Eicosapentaenoic acid during IVM on oocyte cleavage and blastocyst yield following IVF. At the concentrations tested, exposure of COCs to stearic acid during in vitro maturation significantly decreased cleavage and development of bovine oocytes which is consistent with the published literature on the effect of this fatty acid on oocyte quality. There was no apparent effect of the other fatty acids tested.
The results highlight that some of the differences described between lactating and nonlactating animals may be implicated in the subfertility observed in postpartum lactating dairy cows.
Epigenetic competence of oocytes: the epigenetic modification of the oocyte is important for correct programming of subsequent development. Our data showed that chromosomal compartment of oocytes collected from prepubertal sheep, although genomically stable, is hypomethylated and that impaired DNA methylation machinery persists also in blastocysts developed from those oocytes.
In the bovine metabolic model, serum analysis revealed reduced concentration of homocysteine in lactating cows compared to both not lactating ones and maiden heifers and increased concentration of folate and vitamin B12 in lactating cows compared to not lactating ones. These results confirm the adequacy of three groups as metabolic models. When oocytes collected from the three groups were subjected to qPCR for DNMT1, DNMT3A and DNMT3B genes, no significant differences were observed between maiden heifer and not lactating cows. However, deregulated expression of DNMT1 and DNMT3A was observed in oocytes from lactating cows. These results suggest dysfunction of DNA methylation machinery in a critical developmental window, which will affect the development of embryos and consequently of the ongoing pregnancy.
WP3: Oviduct-embryo interactions
The oviduct, narrow duct between ovary and uterus is hosting several events that are determinant for the success of reproduction. The oviduct is capturing the cumulus oocyte complex at ovulation and drives the female gamete to the site of fertilization. Sperm cells are stored at the uterotubal junction, transported toward the oocyte and activated to allow fertilization. Oviduct is also the site of early development during the first five days, during which the activation of embryonic genome occurs and transports the growing embryo to the uterine cavity for implantation. More than 50% of reproductive failures in cattle, occur during the first days after insemination, indicating that they may take place in the oviduct, highlighting the importance of this organ. However, the oviduct being poorly accessible, it represents a kind of black box in which it is very difficult to know what happened (failure of fertilization, low oocyte quality, early embryo death...). In this context, the FECUND project allowed to gain a high amount of valuable information that will help understanding how the oviduct is regulating early reproductive events and how the genetic and metabolic status of the female may affect this finely tuned environment.
To study oviduct physiology and functions, oviduct samples have been collected after slaughter at Day 3 of a synchronized estrous cycle of heifers (n=5), lactating (n=4) and dry (n=5) cows in the energy balance model in Ireland and from Montbéliarde (n=3), high (n=5) and low (n=4) fertility Holstein cows of the genetic models in France. Oviduct have been dissected and processed to flush the fluid for proteomic and sperm survival function studies and scrap the epithelial cells for transcriptomic studies. Oviduct ipsi- and contralateral side from ovulation, as well as isthmus and ampulla segments of each oviduct were processed separately. In parallel, oviduct have been collected in France at slaughterhouse and classified as preovulatory, postovulatory, early luteal, mid luteal and late luteal phase according to ovarian morphology (follicles and corpus luteum). In addition, to evaluate the impact of variable maternal environment on the set up of gene expression of early developing embryos, in vitro fertilized 2-4 Cell stage embryos were cultured in vitro for 5 days or transferred in the oviduct of females from the different groups for the same period of time, using an innovative laparoscopic approach. The transcriptomic and proteomic analysis have been done at LMU and INRA (proteomic of estrous cycle fluid). The analyze of sperm survival supporting activity of the fluid was done in PTP. The Fecund project clearly showed that oviduct provides a finely tuned environment suitable for sperm survival, fertilization and early development. This environment is under the control of genetic parameters and seems poorly sensitive to variations of the metabolic status of the female, providing steady conditions for the success of early reproductive events. The big data set generated by the project is still under analysis and will be opened to the scientific community.
Advanced techniques developed for sampling: progress in performing tissue, fluid and oocyte sampling including in vivo culture of in vitro derived embryos, embryo collection and embryo transfer have been made. These procedures have been standardized and used in both animal models, metabolic and genetic model. Using this design it was possible to generate extensive molecular genetic and metabolic profiles of pathophysiological events during early reproduction including follicles, oviducts and uterine horns.
Study focusing on the in vivo collection of oviductal fluid: another study has been designed in order to establish a method for in vivo collection of oviductal fluid. This fluid accurately provides the physiological environment timely adjusted to all reproductive processes and requirements, which occur in the Fallopian tube. Since the oviduct provides a very low amount of fluid in which gametes enter the oviduct and accomplish fertilization, and in which embryos develop and migrate many efforts have been performed to access the oviduct for fluid collection. However, these attempts were very invasive and it cannot be ruled out a biased effect on the analytical outcome.
A novel method for collection of tubal proteins and fluid in heifers was established. For this purpose beads were pre-tested in vitro for its binding capacity, normally used for affinity chromatography, selected the ones fulfilling the best criteria, transferred them into the oviduct of heifers and performed re-collection one hour later. A second group of animals served for the collection of tubal fluid. For this purpose a flushing cannula was introduced into the oviduct and flushed the ampulla three times using 0.5 to 0.8 ml medium. Oviducts obtained from slaughtered animals have also been flushed and served as controls. Both techniques have been used in animals at Day 1 and Day 3 of the estrous cycle which accords with two different activity status of the oviduct. Re-collected beads and flushing samples were either used to measure the total protein content or were prepared for mass spectrometer analysis (LC-MS/MS). The experiment showed that both, transferring beads and flushing of the oviduct can be successfully used to collect proteins and the total fluid, respectively. It has been shown that both methods provided enough total protein for analysis. In total, 3165 proteins could be identified together in all groups. According to the physical characteristic of each collection procedure, group-specific protein clusters have been found. This method can be frequently used in cows, it is minimally invasive and does not need much time to accomplish. This is the first study aiming at the collection of tubal protein and fluid in heifers using endoscopy.
As technically developed and demonstrated, this method can now be used in cows for a detailed screening of proteins as well as total oviductal fluid containing many components such as lipids, epithelial cells and immune cells, exosomes etc.
It also suggested to combine these collection methods with in vivo embryo development using single or superovulated animals or by in vivo culture techniques in order to perform tubal sampling according to a precise embryo cleavage stage.
Once, more knowledge has been gained about the functional groups of proteins which might have a pivotal role in embryo development, the bead technology can be improved and refined using beads with specific physical surface characteristics to selectively bind and identify these proteins.
In total it is concluded that within the FECUND project a new method has been described and developed. This method provides new experimental designs which might be very helpful in finding ovidcutal components necessary for a better understanding of processes in early reproduction including in vitro production of embryos.
Oviduct transcriptomic: a very high number of DEGs was observed in the oviduct for Montbéliarde compared to Holstein (both F+ and F-), but very few DEGs were recorded between F+ and F- genotypes. Inverted situation was recorded for the endometrial tissue with few DEGs between breeds and higher number of DEG between F+ and F-. Considering the metabolic model, the number of oviduct DEGs (ipsilateral isthmus) was lower between lactating and dry cows (n=15) than between heifers and lactating or dry cows (n=127 and 223, respectively). This indicates that the breed (Holstein vs. Montbéliarde) and the physiological status (Heifers vs. Cows) are more important than fertility genotype and energy balance in determining the oviduct gene expression profile. The complete lists of DEGs are available in deliverable 3.1. They highlight the importance of several pathways in the oviduct regulation, such as steroids biosynthesis, complement and coagulation cascades.
Oviduct proteomic: the high throughput protein contents of bovine oviduct fluid along the estrous cycle has been published for the first time within Fecund project, showing the fine tuning of protein secretion and pointing out protein candidates specifically expressed by the time of ovulation, fertilization and early development, representing interesting candidates for further studies of oviduct function (Lamy et al., 2016). In addition, it has been shown that many of the oviduct secreted proteins are packed in extracellular vesicles and that these vesicles may have important functions in the regulation of early development (see WP7). In total, 1,740 and 1,902 proteins were identified by nanoLC-MSMS in the NEB and genetic models, respectively. The comparative analysis revealed differentially expressed proteins (DEP) between isthmus and ampulla regions of the oviduct, highlighting the functional specialization of these two segments. These DEP will be interesting candidates to follow up for a better understanding of oviduct functions. A relatively low number of DEP was observed between the groups of the NEB model, while this number was slightly higher between the groups of the genetic model (Figure 4). The full lists of DEP are available in deliverable 3.2. Promising biomarker candidates have been identified by gene ontology analysis of the abundance altered proteins, which revealed that most proteins are assigned to the term metabolic process, such as Peroxiredoxin 2 (PRDX2), SERPIND1 or Glutathione S-transferase theta-1 (GSTT1), which were found to differ significantly among the groups in both models.
Sperm transport and survival: the first approach was to test the effects of the oviductal fluids collected from both models on sperm motility and kinetics parameters, using Computer Assisted Semen Analysis (CASA) system, to reveal if effects observed in vivo can be mimicked in vitro. The motility parameters of sperms decreased significantly during co-incubation time in all the groups. However, this decrease was significantly faster for semen co-incubated in SOF medium alone (P<0.001) than for those incubated in oviduct supplemented medium, whatever the group in both models. Overall the fluids collected in the ampulla seem to have a more positive effect on sperm motility and kinetics parameters conservation, compared to the fluids collected in isthmus. The second approach consisted in the use of a microendoscopic imaging system (Cellvizio) that INRA has developed in sheep to follow the progression of fluorescent spermatozoa in genital tract in vivo, after cervical insemination. The objective was to transfer this technology in cow and compare the migration of fluorescent labeled spermatozoa from the same bull in the genital tracts of cows for both models. A fluorochrome providing stable bright fluorescence staining of the whole spermatozoa without affecting their viability was selected. Fluorescent sperms were inseminated through the cervix and followed in the uterus at 30 min and 3h after AI by insertion of the Cellvizio 1.5mm probe by the same way. The spermatozoa, motile and immotile, could be clearly detected and tracked 30 min after insemination. In contrast, very few spermatozoa were found in the uterine horns 3 hours after AI. The collection of the female genital tract after the slaughter of one cow after imaging revealed that the spermatozoa were absent from the body of the uterus but stored in the highest part of the uterine horns, the utero-tubal junction, which cannot be reached by the probe. Interestingly, inseminating fluorescent beads with sperms allowed to show that beads are rapidly drawn back to the vagina after insemination, evidencing an adverse flow that may also evacuate dead sperms. In conclusion, the method was adapted to observation of uterine cavity in cattle, however, the quick migration of sperms to the uterotubal junction impaired the use of the technique to compare the fate of sperms in the different models.
Embryo transcriptomic: given that oviducts of lactating cows are less able to support early development than oviducts of heifers or dry cows, this task was designed to see if variable maternal environment can modify embryo genome activation expression pattern. Therefore, in vitro fertilized embryos were transferred at day 2 after fertilization into the oviducts of either lactating and nonlactating (dry) Holstein cows and heifers (Metabolic Model) and Holstein heifers with high (F+, n=4) or low (F-, n=4) EBV for fertility compared to Montbéliarde heifers (n=4) as an additional control (Genetic Model). The embryos were recovered at day 4 and investigated by RNA-sequencing (LMU) and compared to in vitro developed ones. Using a unique model of endoscopic embryo transfer coupled with next generation sequencing, this study has demonstrated the influence of metabolic status and estimated breeding value on the evolution of embryonic genome activation in bovine embryos. The effect of metabolic status (heifer vs postpartum lactating and nonlacting cow) was more extreme than that observed for the genetic model where high vs. low estimated breeding values for fertility have only a small effect on the transcriptome embryos that were cultured for two days in their oviducts.
Embryo proteomic: to study the effect of the oviduct on the early embryo, in vitro-produced 4- to 8-cell embryos (Day 2) were transferred into the oviducts of females from the metabolic and genetic models. Two days later, on Day 4 post estrus, the embryos were recovered from the oviducts and were analyzed using a multiplexed “Selected Reaction Monitoring” (SRM) assay, together with D2 embryos and D4 in vitro cultured ones. This multiplexed SRM assay was applied for the relative and absolute quantification of 27 proteins (see deliverable 3.5) in each of the pools of embryos, deriving from the different animals. The proteins were chosen based on previous results obtained in our group, demonstrating that they play an important role and can serve as markers for the different stages in early embryonic development. To absolutely quantify the 27 proteins, two stable isotope-labeled proteotypic peptides (isotopically labeled at incorporated C-terminal Arg or Lys) per protein, i.e. 54 peptides in total, were used as internal standards. Absolute protein concentrations could be determined for 19 proteins in embryos from the metabolic model and 20 proteins in embryos from the genetic model. Proteins could be quantified down to the attomole range per embryo. We observed a clear influence of the embryo environment on individual proteins (in vivo vs. in vitro). We also observed statistically significant differences in protein concentrations between groups of females in both models (GSTM3, HSPB1 and CLU in the metabolic model, and ASTL, KPNA7 and TACC3 in the genetic model).
WP4: Uterus and conceptus
In this work package we addressed the question whether metabolic condition or genetic merit for fertility affects the conceptus-endometrial crosstalk in dairy cattle. This was analyzed by molecular profiling of endometrium and conceptus. In addition, uterine fluid was analysed for proteome changes in a holistic manner.
Previous studies have reported differences in the endometrial response of lactating cows compared to non-lactating cows to the major embryonic embryonic pregnancy recognition signal, interferon tau (IFNT) . Moreover, the endometrium is capable of distinguishing between conceptuses with different developmental trajectories i.e. conceptuses produced by AI compared to in vitro and cloned embryos, and can modify its transcriptomic response accordingly.
Thus, we tested hypothesis that metabolic changes in circulation that are associated with lactation would impede the ability of the uterus to support adequate conceptus elongation and thus the ability the conceptus to provide the appropriate pregnancy recognition signal to endometrium to establish an environment suitable for implantation to begin. In addition, we tested if differences in the estimated breeding index (EBI) for fertility are associated with significant transcriptome changes in the endometrium at the time of implantation. Therefore, endometrium samples from Holstein heifers with high (Fertiplus) or low (Fertiminus) EBI for fertility were investigated by RNA-sequencing, including corresponding samples from Montbéliard heifers as an additional control.
Effects on the endometrium: in total, 256 differentially expressed genes (DEGs) were revealed by comparing endometrial transcriptome profiles of heifers vs. non-lactating cows (FDR 5%, 131 increased and 125 decreased in heifers compared to non-lactating cows, respectively). The comparison of heifers with lactating cows revealed 238 DEGs, 141 of which were increased and 97 were decreased in expression. At a FDR of 10%, 28 DEGs were identified 8 of which increased and 20 decreased in endometria from lactating compared to non-lactating cows on Day 19 of pregnancy. Overall, there was a substantial overlap (92 genes) between the DEGs of the comparisons heifer vs. lactating and heifer vs. non-lactating cows, while there was little overlap found for the DEGs between endometria from lactating and non-lactating cows with the comparisons to the heifer group. Functional annotation clustering showed that among the DEG with lower expression in heifers compared to non-lactating cows functional categories such as ‘vesicle’, ‘signal transduction’, ‘steroid metabolic process’, and ‘carbohydrate binding’ were overrepresented. Categories ‘regulation of cell migration’, ‘cell death’, and ‘oxidoreductase’ were enriched among DEG with higher expression in heifers. Corresponding to the overlap of the DEGs for heifers vs. lactating cows and heifers vs. non-lactating cows, respectively, similar functional categories were found as overrepresented. In addition, DEG with lower expression in heifers included genes with potential transcription factor binding sites for HFH3, FREAC7, and RSRFC4 and ‘cell adhesion’, while DEG with higher expression in heifers were enriched in the functional categories ‘fatty acid metabolic process’, ‘defence response’, and ‘lymphocyte activation’. For the DEGs in endometrium of lactating vs. non-lactating cows, only three overrepresented functional annotation clusters were obtained: ‘response to hormone stimulus’, ‘cell migration’, and genes with potential transcription factor binding sites for IRF7, MYOD, and OCT.
In addition to the holistic approach by RNA sequencing, we analysed the expression of a selection of endometrial genes known to be regulated by the presence of the conceptus (PLET1, SOCS6) and by interferon tau (STAT1, RSAD2, SOCS1, SOCS3) or involved in progesterone (FOXL2, SCARA5) prostaglandin (PTGS2) and oxidative stress (CAT, SOD1, SOD2) molecular pathways by RT-PCR. Variance analyses revealed a significant impact of the maternal metabolic status on FOXL2 (P<0.002; DRY/LACT fold-ratio=2.1) and SOD2 (P<0.04; DRY/LACT fold-ratio=0.8) mRNA expression. Based on these findings, the contribution of FOXL2 to endometrial physiology has been investigated using primary cultures of bovine endometrial cells.
Furthermore, we tested if differences in the estimated breeding index (EBI) for fertility are associated with significant transcriptome changes in the endometrium at the time of implantation. Therefore, endometrium samples from Holstein heifers with high (Fertiplus) or low (Fertiminus) EBI for fertility were investigated by RNA-sequencing, including corresponding samples from Montbéliard heifers as an additional control. Interestingly, the highest number of differentially expressed genes were found in the comparison of Fertiplus and Fertiminus (276 higher in Fertiplus, 418 higher in Fertiminus), while the comparison of these two groups with Montbéliard revealed fewer differentially expressed genes.
In conclusion, this study demonstrated the influence of parity and metabolic stress on the transcriptome of the endometrium independently of the quality of the transferred embryos. Compared to the high fertility uterine environment of heifers, cows exhibited significant transcriptome changes of the endometrium transcriptome at day 19. In addition, a number of transcriptome differences were found between endometrium samples from non-lactating vs. lactating cows suggesting that metabolic stress in the latter group affects gene expression in the endometrium and eventually hampers its ability to respond physiologically to pregnancy recognition signals released by the conceptus. In addition to effect of metabolic condition on gene expression in the endometrium this study showed that high vs. low estimated breeding values for fertility have a clear effect on the endometrium transcriptome at day 19 of pregnancy, which may – at least in part – explain the difference in reproductive performance.
Effects on the uterine fluid: in addition to candidate transcripts, differentially abundant proteins in the uterine fluid of the different recipient groups may serve as biomarkers for fertility. Therefore, this work package included a differential proteome analysis of uterine fluid recovered from the different recipient groups. For the metabolic model (lactating cows, dry cows, heifers) the proteins with the most pronounced abundance differences were:
1. Dry cows vs. heifers
a. More abundant in dry cows: Lysozyme C, milk isozyme; Histone H3.3; Histone H2B; Histone H4 (Fragment); Histone H2A.J
b. Less abundant in dry cows: Ran-specific GTPase-activating protein; CBX3 protein; Prothymosin alpha; Trophoblast Kunitz domain protein 4; Density-regulated protein
2. Dry cows vs. lactating cows
a. More abundant in dry cows: Aminopeptidase N; Superoxide dismutase [Cu-Zn]; Phosphatidylethanolamine-binding protein 4; Ganglioside GM2 activator 2; Phospholipase A(2)
b. Less abundant in dry cows: Apolipoprotein A-IV; Density-regulated protein; Glycerol-3-phosphate dehydrogenase [NAD(+)]; Protein HP-25 homolog 1; Prothymosin alpha
Compared to metabolic condition, high vs. low estimated breeding value for fertility had an even more pronounced effect on the proteome of uterine fluid. Out of 358 proteins showing differences in abundance between Fertiplus and Fertiminus, the ten most prominent were Growth/differentiation factor 8; Tubulointerstitial nephritis antigen-like; Lactoferrin (Fragment); Platelet-activating factor acetylhydrolase; Polymeric immunoglobulin receptor (less abundant in Fertiplus); and NID2 protein; GFPT1 protein; Perilipin; Thymosin beta-10; CALML5 protein (more abundant in Fertiplus).
Consequences for the conceptus: lactation in high yielding dairy cows results in elevated non-esterified fatty acids (NEFA) and beta hydroxy butyrate (BHBA) in response to a deficit between energy intake and energy output as well as decreased circulating concentrations of insulin, insulin-like growth factor 1 (IGF1) and glucose, creating a compromised environment for embryo development.
This study tested the hypothesis that the uterine environment of lactating cows would affect conceptus gene expression. In order to isolate effects of lactation on the conceptus from those on the oocyte and early embryo we used embryo transfer of good quality embryos recovered from nulliparous Holstein heifers. The specific aim of the study was to characterize the transcriptome of the bovine conceptus at the initiation of implantation in a cohort of age-matched postpartum primiparous dairy cows that were either milked postcalving (i.e. lactating) or were dried off immediately at calving (i.e. never milked, non-lactating) as well as a group of contemporaneous heifers.
Principal Component Analysis (PCA) of transcriptomic data revealed moderate separation between the overall expression patterns of conceptuses recovered from lactating cows compared to maiden heifers. PCA of gene expression patterns of conceptuses from non-lactating cows vs. heifers or from non-lactating cows vs. lactating cows revealed no clear separation of these groups. Comparison of conceptus transcriptome profiles from lactating cows vs. heifers revealed 269 DEG, 100 with higher and 169 with lower expression in conceptuses recovered from lactating cows. Gene ontology analysis of these DEGs revealed 37 biological processes falling into six broad categories: (i) the regulation of biological processes; (ii) cellular and metabolic processes; (iii) cell component/cell cycle regulation; (iv) cell death and apoptosis; (v) developmental processes and (vi) mitochondria.
In conclusion, this part of the work package demonstrated the influence of metabolic stress on the transcriptome of grade one embryos. Compared to the high fertility uterine environment of heifers, lactation significantly modifies the transcriptome of a conceptus at day 19. The lack of differences between non-lactating cows and heifers indicated that it is not just an artefact of the source of conceptus i.e. generated by AI compared to superovulation. Moreover, even with no differences in the transcriptome of the conceptus e.g. non-lactating and heifer, here does seem to be differences in the utilization of amino acids (detected by a parallel study of amino acid concentrations in the uterine luminal fluid) i.e. the metabolic activity of the conceptus. It is unclear if these differences in conceptus transcriptome and amino acid utilization are a result of conceptus ‘plasticity’ in response to the lactating environment of have a definite effect or consequence for pregnancy outcome.
Potential practical use of the results:candidate genes identified by expression profiling of endometrium and conceptus provide an interesting source for the identification of genetic polymorphisms (e.g single nucleotide polymorphisms, SNPs) which are associated with fertility and metabolic stability. The candidate genes discovered within this work package can be used by breeding associations for SNP discovery and validation studies as outlined in Ref. [6]. In addition, proteins in uterine fluid of animals with different fertility may serve as putative biomarkers for reproductive status.
WP5: Population genetics
This work package addressed novel approaches to identify candidate recessive lethal mutations, identify variations within key genes involved in early reproductive biology identified in previous WPs. Preliminary results have been validated and tested at the population level: the activity brought to the identification and localisation of putative recessive lethal mutations, the identification of variations in candidate genes identified in earlier WPs and the validation of the effects of the recessive lethal and candidate genes variations on fertility.
To create a list of candidate recessive lethal loci from gene sequence analysis, two complementary approaches were used: 800K SNP genotypes of 1009 progeny tested Italian Holstein sires were analysed to identify putative recessive loci and exome sequences of 20 Italian Holstein sires having extreme values for male and female fertility EBVs. A total of 191 variants in 163 genes were identified as strong candidate deleterious variants for lethal recessive mutations.
Genetic variants, falling within sequences of candidate genes/loci and identified in previous WPs as having an effect on dairy cattle fertility, have been identified. A pipeline of analysis has been realised, to ensure reproducibility, modularity and scalability of the experiment.
First, from an input list of 745 genes 604 have been annotated and positioned on the UMD3.1 genome. Next, polymorphisms falling within candidate regions have been identified in the whole-genome sequence data coming from the 1000 bull genomes project (http://www.1000bullgenomes.com/).
In all the 604 regions, a total count of 758,684 polymorphisms have been identified. The overall distribution of the polymorphisms per gene indicates a very skewed right distribution, as expected.
The Minor Allele Frequency (MAF) for all polymorphisms was calculated. 23,415 variant sites out of 758,684 showed MAF = 0 and were therefore monomorphic. The MAF showed a mean value of 0.059732 (std=0.112234) and a median value of 0.004156.
A genotyping assay for candidate recessive lethal/deleterious polymorphisms coming from the results obtained in deliverable 5.1 and polymorphisms within fertility candidate genes/QTLs has been designed, resulting in a total of 71 potentially tested combinations of variations/haplotypes. Such 71 potential deleterious variants (including variations and haplotypes) were fully assessed to have a complete picture of the genomic effects, their frequencies, the flanking sequences (useful to design the probes in the array), and other variations present in the flanking regions which could difficult the hybridization of the probes. In addition, a further analysis on these 71 variants considering all the information collected was performed to target a reduced subset of the most promising candidates to run the tests on the population. Two variants (DelVar_6 and DelVar_9) were identified as the strongest candidates, considering the complete co-segregation in the population of 17 individuals with both HD and exome data. Candidate variants from D5.2 obtained from whole-genome sequence data were also processed: results are analogous to those for D5.1.
Most promising candidates (DelVar_6 and DelVar_9) were selected for large scale testing in the population in about 5200 Italian Holstein bulls and cows. The frequency of the deleterious alleles in the dataset agreed with the one found in the WES and haplotype data analyses (0.08%). In the WES and SNPchip data we didn’t find homozygotes for the deleterious allele (variant or haplotype). The population data didn’t confirm this behavior: 32 (0.62%) and 37 (0.72%) homozygotes for the deleterious were found in DelVar_6 and DelVar_9, respectively. Also the Hardy-Weinberg equilibrium test on the population data didn’t confirm the haplotype results. The two DelVars resulted in fact in Hardy-Weinberg equilibrium. The p-value of the test was 0.9987 and 0.5435 for DelVar_6 and DelVar_9 respectively. the access to the “1000 bull genome” dataset permitted to use the 450 Holstein animals with WGS as test population for all the 67 variants. Out of the 42 variants with no observed homozygotes in WES, only 18 were confirmed in the 1000 bull genome dataset.
The chi-square test of Hardy-Weinberg equilibrium identified 11 variants with p-value below 0.05. Two of them (DelVar_15 and DelVar_28/29) had a p-value under the threshold of 0.05 corrected for multi-testing (0.00075).
These two variants were further assessed mining the 1000 bull data of other breeds, in particular Brown Swiss (BRO: 105 animals), Simmental type (SIM, 274), Jersey (JER, 66), Angus (ANG, 141) and Hereford (HER, 41). DelVar_28/29 was in Hardy-Weinberg equilibrium in all breeds investigated. DelVar_15 was significantly out of equilibrium in all other breeds investigated and had no homozygotes with the exception of a single one found in Simmental (0.36% of the total SIM dataset) but we can’t exclude a sequencing error or allelic dropout at this nucleotide.
WP 6: Surrogate markers and indicative traits
The WP6 has carried out activities to identify robust biomarkers predictive of the reproductive status of the cow in biological fluids easy to obtain as plasma and vaginal swabs. The activities have been organized in four tasks. The first 3 tasks have been focused on the identification of biomarkers related to fertility status in blood samples collected from animals belonging to WP1;in vaginal swabs; in blood samples and vaginal swabs collected during the negative energy balance (NEB) of cows in their transition phase until the first insemination.
In the last task, the potential markers found in plasma and vaginal fluid as measure of fertility have been validated in a greater population.
The main results obtained in the WP can be summarized in the following points:
Correlated markers for fertility in blood: the purpose of this activity was to discover less invasive biomarkers in blood serum to avoid complex procedures, often invasive, and expensive assays. Blood samples were collected from animals after the insemination up to day 19 for microRNA, proteins and metabolites analysis. Overall, 51 animals of WP1 belonging to tasks 1.2 (Impact of energy balance; n=27) and 1.3 (impact of genetic merit; n=24) have been analysed. Blood samples obtained from task 1.2 belonged to three groups: LACT (n=10), cows that calved normally and had a standard lactation; DRY (n=11), cows that calved normally but dried off immediately after calving; MH (n=6), heifers of breeding age. Blood samples obtained from task 1.3 belonged to three groups: HOLST- (n=8), Holstein heifers with low reproductive indices; HOLST+ (n=8), Holstein heifers with high reproductive indices; MONT (n=8), Montbeliarde heifers breed with a higher level of reproductive success. All animals of both experiments were synchronized with the same procedure, and LACT and DRY show a estrous around 70 days in milk.
The miRNAs in plasma at d19 showed important results in the metabolic model, but not in the genetic model, where only heifers have been used. Some of the miRNAs with different concentrations between high and low fertility groups in the metabolic model are linked with the immune tolerance and the bta-mir140 (up-regulated in pregnant cows) seems a good candidate as an early biomarker of fertility.
The metabolic indices measured at the insemination time in the two models of the project gave the possibility to clarify that: (i) cows dried off immediately after calving or heifers in comparison with lactating cows differed in some functions of the liver (lower concentrations of cholesterol, index of the synthesis of the lipoproteins, GOT, total bilirubin), which reflects a lower metabolic activity of this organ. (ii) pregnant vs non-pregnant cows that follow a standard lactation showed a better liver functionality (higher concentrations of cholesterol and albumin) and lower concentrations of inflammatory markers (positive acute phase proteins: as haptoglobin and ceruloplasmin; globulin); (iii) pregnant vs non-pregnant heifers showed (in both models, but not in HOLST+) lower concentrations of cholesterol and albumin, likely in relationship to the different energy status.
The candidates of fertility emerged from this task include parameters of the liver functionality and inflammation for lactating dairy cows; parameters related to the energy status and markers of long duration of inflammation for heifers.
Biomarkers for fertility in vaginal samples: this task aimed at establishing a sampling procedure of vaginal swabs for the analysis of biomarkers (i.e. metabolites, miRNA) in the vaginal fluid; (ii) setting up methods to measure biomarkers (i.e. metabolites, proteins and miRNA) in samples; (iii) collecting samples from dairy cows in different physiological conditions to correlate the biomarkers measured in the vaginal fluid with measures of fertility and to identify some putative biomarkers to be monitored in a new dedicated experiment (Task 6.4).
The procedure to collect the vaginal fluid has been defined and dedicated methods for the determination of the potential candidates of fertility status in this fluid have been developed. After setting up the analysis for miRNA, possible biochemical markers have been investigated, demonstrating the possibility to detect the following parameters: glucose, NEFA, BHB, total cholesterol, triglycerides, urea, creatinine, AST/GOT, GGT, ALP; lactate dehydrogenase, total bilirubin, Ca, Mg, P, haptoglobin, ceruloplasmin, Zn, globulin, total protein, albumin, paraoxonase, advanced oxidation protein products (AOPP), total antioxidant (ferric ion reducing antioxidant power or FRAP) and thiol groups. Biochemical and miRNA profiles demonstrated a very light presence of molecules in vaginal samples collected from dry cows, while showed a good variability of many molecules in the samples collected at 30 days in milk. These results confirmed the possibility to use these biomarkers as possible candidate of the reproductive traits. To verify this we realized samples of vaginal fluids at 30 days after calving and 19 days after the first insemination in 20 cows submitted to the procedure of estrus synchronization (the same used in the WP1). Within this population, 10 subjects have been selected in accordance to the success (5 cows) or failure (5 cows) of the insemination. The pregnant in comparison to the non pregnant cows showed significant lower concentration of glucose, triglycerides, lactate dehydrogenase, urea, phosphorus and thiol groups and also statistical differences in several miRNA (i.e. bta-mir-2411; bta-mir-150; bta-mir-342; bta-mir-574; bta-mir-98; bta-mir-204; bta-mir-34c; bta-mir-221). Some miRNAs are specific for the high fertile group (bta-mir-451; bta-mir-92a-2; bta-mir-92a-1), others are specific for the hypo fertile group (novel:18_15941; Novel:1_1781; Novel:has-miR-4505; Novel:has-miR-4467; bta-mir-200b; Novel:4_36799) while others are common to both groups (bta-mir-486; Novel:17_14769; bta-mir-215; Novel:9_45060; Novel:13_7606; bta-mir-192; bta-mir-143; bta-mir-126). Nevertheless, none of the significant miRNAs detected in vaginal swabs resulted significant at plasma level. This demonstrates that identified miRNAs are tissue-specific and depending also on the time point analysed.
Markers of NEB in transition cows: for this activity blood samples from the dry period to the first insemination have been collected to assess a wide inflammo-metabolic profile and to select some potential reliable biomarkers of fertility status for a further confirmation in field conditions (WP9). The experiment has been done in a well-controlled herd on 52 subjects (22 primiparous and 30 multiparous). Cows have been classified in accordance to the success (PR, 15 cows, 40% primiparous) or failure (37 cows) of the insemination (AI), done after the estrus synchronization (the same of WP1). Non pregnant cows after the first AI were divided in two further groups: pregnant within 250 days in milk (LatePR, 24 Subjects; 46% primiparous), non pregnant or repeated breeders (e.g. pregnant with more than 3 AI and over 200 days in milk; NoPR, 13 Subjects, 38% primiparous). In this experiment, we have found some biomarkers that statistically differ among fertile (PR) and less fertile (LatePR) or infertile (NoPR) cows.
The most promising plasma biomarkers resulted: (i) indexes of the prolonged effects of the inflammation (high concentrations of ceruloplasmin and globulin, lower level of the ratio albumin/globulin) during the transition period and also around the first AI; (ii) some negative acute phase proteins, during the periparturient period and in some cases also around the first AI. The more fertile cows showed higher concentration of cholesterol (index of lipoproteins), retinol (index of Retinol Binding Protein), total bilirubin (index of enzymes necessary to their clearance); (iii) some molecules with anti-inflammatory effects, as tocopherol and beta-carotenes, which remained higher in more fertile cows after calving and around the first AI. In opposite, the concentration of the total antioxidants and the indexes of the oxidative stress (both ROM and AOPP) do not show significant differences among the groups; (iv) concentrations of energy balance indicators (NEFA and BOHB, respectively at two or four weeks after calving) which showed the lowest concentration in more fertile cows in comparison to less fertile or infertile (but not at the time of the first AI); (v) concentrations of phosphatase alkaline and some minerals (mainly calcium and phosphorous), both around the calving and around the time of first AI. In particular, the lower concentrations of plasma phosphatase alkaline and phosphorus and the higher one of calcium are correlated with conditions of worsened fertility. All these biomarkers appear good candidates for the detection of cows with poor fertility in field conditions and seem able to anticipate the identification of cows that need specific treatments to restore the normal physiology of the reproductive apparatus.
Validation: to validate previous results we collected vaginal swabs after the calving, from the same cows involved in task 6.3 experiment in order to select the most reliable biomarkers of fertility status for a further confirmation in field conditions (WP9). On the same 52 subjects enrolled in task 6.3 two vaginal swabs were taken at 28 days in milk and 19th days after the AI for the measurements of parameters identified in previous activities. Among the wide biochemical profile investigated, the most promising biomarkers of fertility status in the vaginal swabs collected at 30 days in milk include indexes of negative energy (cholesterol, NEFA and BOHB) and protein (urea and creatinine) balance, indices of mineral (Ca, P, Mg, Zn) metabolism, and indices of the inflammation (globulin, ceruloplasmin, total bilirubin), all lower in fertile vs less fertile cows. The indices of oxidative stress do not seem so important in this context. Otherwise, only two biomarkers in vaginal swabs collected at 19 days after first AI (around 90 days in milk) have the potential to identify subjects with low fertility, and in particular zinc and phosphatase alkaline, with lower or higher concentrations respectively in fertile vs less fertile cows. In the whole, these information appear to have a practical meaning when the checks occur at 30 days in milk. For this reason the investigation to confirm the results obtained in this WP and planned in the WP9 on an higher number of cows belonging to more herds, should be realized only around 30 days in milk.
WP7 Reproductive Biotechnologies
With the recent development of animal genomics cattle breeding companies are in the process of designing new breeding schemes by combining reproductive biotechnologies with genomic information. Their objective is to make selection decisions based on genomic estimated breeding values (GEBVs), as they are becoming rapidly available through widespread genotyping of cattle herds, rather than having to wait for phenotypic data from progeny and performance testing. GEBVs integrate genomic and pedigree information and can be calculated very early in the life of the animals, well before they reach the reproductive age. For this reason the potential of taking full advantage of genomic selection relies on reproductive biotechnologies capable of deriving progeny from young heifers and bulls. Therefore the objective of this workpackage has been the improvement of techniques for assisted reproduction taking advantage of the novel findings derived from the animal models developed in previous workpackages and testing novel hypothesis reported in recent literature in order to improve oocyte and embryo developmental competence.
Improving in vitro culture
To improve in vitro bovine embryo culture conditions the specific effect of oviduct factors on early embryo gene expression were tested. Therefore we compared the transcriptomes of embryos developed in vitro either in co-culture with Bovine Oviduct Epithelial Cells (BOEC) (Cordova, Perreau et al. 2014; Schmaltz-Panneau, Cordova et al. 2014) or with another feeder cell type also known to improve bovine embryo early development the kidney epithelial VERO cells (Menck, Guyader-Joly et al. 1997). Because co-culture with BOEC implies to use a high oxygen concentration (20%), we also compared embryo gene expression in 20% or 5% oxygen. We used the new custom 60K bovine microarray developed during the first part of FECUND project. This array displays 26 055 transcripts (corresponding to 23 926 genes) which represents 97.4% of the Ensembl annotated transcripts. Embryonic gene expression was analysed at two different stages (16-cell and Day 8 blastocyst stages). While high level of oxygen induced a significant decrease of blastocyst proportion (but did not change cleavage rates), no impact of oxygen level was observed on embryo transcriptome at both developmental stages. No significant difference of cleavage, 16-cell and blastocyst rates was observed in presence of feeder cells when compared to 5% of oxygen. No impact of feeder cell type was observed on bovine early development kinetic and rates. Regarding to embryonic gene expression, very few genes were directly impacted by the feeder cell types. However, comparing each co-culture system to culture in absence of any feeder revealed greater proportions of differentially expressed genes. Nevertheless, no significant difference in impacted biological pathways identified by Ingenuity Pathway Analysis Software was observed. These regulated pathways are mainly related to cell cycle, free radical scavenging and glucose and lipid metabolisms. Thus, it seems that there is no major impact of feeder cell types (BOEC or VERO) on bovine early development in terms of kinetic, developmental rate and transcriptome.
After investigation the dialogue between the developing embryo and bovine oviduct epithelial cells (BOEC) in an in vitro co culture model showing a beneficial effect of the cells on development rate and quality (see previous report), and considering that many genes involved in vesicles transport were influenced by energy balance in COC and oviduct, we searched for the presence of exosomes in oviduct fluid and in BOEC culture medium. Extracellular sesicles were found in both cases and were analysed for proteic contents, showing that these vesicles are able to transport protein known to be present in oviduct fluid (ex OVGP1). We showed that exosomes are able to cross the zona pellucida and to be internalized by embryo cells and the addition of in vivo exosomes to the embryo culture medium (SOF) had a positive effect on embryo development rate and quality, showing that the effect of BOEC on development may be at least partially mediated by extracellular vesicles.
Another aspect addressed in the project was to investigate the different metabolic requirement of oocytes derived from young heifers versus adult cows followed by the field trial corresponding to deliverable 7.5. The study was carried out in a period of approximately 1 year and during this period the animals were allocated to the 2 groups considering their age at the time of oocyte collection. For this purpose a database for data recording was constructed that included the date of birth of each animal in order allocate the embryo production data accordingly, meaning that the same animals were at first in group 1 (OPU performed when age was below 12 months) and later they were moved in group 2 (OPU performed when age was above 12 months). In total 392 OPU session were performed on over 50 donor animals, divided in the 2 age groups. The number of embryos obtained was 480 of which 359 were transferred in recipient females giving rise to 177 pregnancies. A significant difference was observed between the low pyruvate and high pyruvate medium only in the group of donors older than 12 months and only for the percentage of embryo calculated on the fertilized oocytes (14.94 vs 23.21 chi square test p<0.05). A general tendency for lower pregnancy rate was observed in the young age group (heifers aged less than 12 months) versus the older age groups. Therefore the indication from this experiment is that most likely the culture requirements for donors aged more than 12 months are different from those of younger heifers and that increasing the concentration of pyruvate in the culture media is beneficial only for older donors.
Improvement of oocyte competence in vitro: in this activity we compared two different in vitro maturation systems testing the working hypothesis that the developmental competence of the oocytes can be increased by a period of prematuration during which the oocytes are blocked in meiotic arrest by exposure to cAMP. Resumption of meiosis occurs after a drop in oocyte cAMP levels, which also regulate gap junctional communication between the oocyte an the cumulus cells. It is thought that cAMP levels are regulated with two different mechanisms: the diffusion through gap junctions and the inhibition of its degradation. The diffusion of cAMP from the granulosa and cumulus cells to the oocyte and the consequent increase of cAMP levels prevent oocyte maturation. In the oocyte the degradation of cAMP is catalysed by the specific type 3 phosphodiesterase (PD3A) which is inhibited by cyclic Guanosine Monophosphate (cGMP). Several studies demonstrated that somatic cell derived cGMP enters the oocyte through the gap junctions, thus inhibiting PDE3A and maintaining meiotic arrest. On the background of this knowledge the modulation of intracellular cAMP amount was obtained in this study acting on three different molecular targets through the use of three different meiotic inhibitors: 3-isobutyl-1-methylxanthine (IBMX), forskolin (FSK) and cilostamide. FSK raises the levels of cAMP by stimulating adenylate cyclase activity.IBMX is a methylated xanthine derivate, which acts both as a competitive nonselective PDE inhibitor and as a nonselective adenosine receptor antagonist. The effect of the exposure of COCs to IBMX is the raise of intracellular cAMP and the subsequent activation of PKAs in the cumulus cells and in the oocyte. Cilostamide is a specific inhibitor of PDE3 and is used to selectively modulate cAMP levels in the oocyte. The aim of the present study was to develop a defined in vitro maturation system and to compare two different protocols for pre-maturation of oocytes both modulating cAMP levels. The assessment of the experimental conditions was done first evaluating kinetics and maturation rate and secondly following fertilization evaluating embryo development to blastocyst. Two different prematuration treatments were compared with two control groups. Maturation kinetics was affected by the treatment indicating a delay in the experimental groups up to 24h of maturation while at 30h of maturation no differences were observed. A similar trend was observed during embryo development where the prematuration groups were developing at slightly lower pace in the fist days after fertilization but at the blastocyst stage, 7 days after IVF, no difference was aobserved between the groups. This finding indicate that a prematuration strategy could be used successfully in a practical commercial context to provide more flexibility in the field operation and allowing to plan the fertilization timing according to the field and laboratory logistics.
Validation of culture supplements that influence methylation:Oocyte maturation process includes the establishment of proper epigenetics marks, fundamental to ensure successful pregnancy. Epigenetic maturation of the oocyte depends on one-carbon-metabolism (OCM), whose key elements (cobalamin, folate) are crucial cofactors for the transfer of methyl groups onto chromatin. Commercially available IVM-media only partially supports oocyte metabolic requirements, and thus may negatively affect epigenetic maturation and further development. Of relevance, cobalamin, one of the OCM cofactors normally present in follicular fluid, is missing in IVM-media. Aim of this work was to investigate if cobalamin supplementation of IVM media affects sheep oocyte developmental competence in term of subsequent embryo development and epigenetic pattern. Data demonstrated that Cobalamin supplementation improves oocytes’ competence and partially corrects epigenetic defects (DNMTs expression, genome-wide methylation profile) in pre-implantation embryos and placental tissues at 20-days of pregnancy. Moreover, early placental vascularization, in terms of vessels maturity and vasculogenetic factors’ expression, was improved by the treatment. Data have provided useful information for the improvement of in vitro maturation condition and epigenetic competence of IVP embryos.
WP8: Bioinformatics and statistical analyses
This work package provided the bioinformatic structure to manage the data produced and the tools for primary data analysis. Analyses were carried out by combining and mining data to identify candidate genetic loci, putative biomarkers and novel traits that can be used to predict fertility and to examine the contributions of genetic and metabolism to fertility problems. Shortly, a project database to manage and store project data has been created, bioinformatics tools and pipelines were used to analyse and mine data across WPs and thus investigate genetic and metabolic impact on fertility.
First, a Fecund database and genome browser was put in place to store and allow user-friendly retrieval of information from the other WPs. Bioinformatics pipelines and tools were made available via a bioinformatics platform. In particular software tools for mass spectrometry based protein identification (i.e MASCOT, Andromeda), for quantitative proteomics (Scaffold, MaxQuant), data interpretation (Cytoscape, ClueGO/CluePedia), RNA-Seq and Methyl-Seq data analysis (Trimmomatic, TopHat2, Bismark, ...) Metabolomics (Agilent Chemstation MSD) enabled cutting-edge approaches to process the genomics and proteomics data and ensured high quality analyses and outputs.
Using phenotypic and genotypic data of Montbéliarde and Holstein cows, a genome-wide association study for milk production and cow fertility has been performed. On the same line of thoughts, from the GWAS results we have identified SNP with a significant effect on both fertility and milk production. These SNP were then tested for association in another cattle populations comprising Holstein cows from dairy herds in Northern Italy. In total, 16 associations from the GWAS on French Holsteins and Montbeliarde cattle were validated in an independent Italian Holstein population. These indicate SNP and related genomic regions with a signifcant effect on both milk production and cow fertility. The next step was to pursue these signals further in a functional analysis (gene ontologies and metabolic pathways) to uncover genes and biological functions behind fertility and milk production in dairy cattle.
Differentially expressed genes from transcriptomics experiments in WPs 2, 3 and 4 have been used to explore the gene ontologies and molecular pathways implicated in milk production and fertility. Gene expression was studied in five tissues related to the reproductive function: the cumulus cells, the endometrium, the oocytes, the oviduct and the embryo. A bioinformatics pipeline to automate and standardise gene functional analysis (analysis of gene ontologies and metabolic pathways) was developed for the purpose. The pipeline comprises several steps related to two major groups:
•gene ontololgies (GO terms) + gene set enrichment (wherever possible)
•pathway analysis (metabolic pathways)
Additionally, there were steps for conversion of gene IDs, names and symbols (e.g. to go from SwissProt proteins codes to gene symbols) and for plotting/summarizing results. The pathway comprises scripts in R, Bash and Python, plus using external software when needed. The following biological databases are queried to retrieve ontologies and pathways: Ensembl www.ensembl.org/, NCBI (www.ncbi.nlm.nih.gov) KEGG www.genome.jp/kegg/, Reactome (www.reactome.org/). A list of 4,416 differentially expressed genes was obtained from experiments in WPs 2, 3 and 4. The Gene Ontologies that were retrieved from biological databases are categorized in three groups: biological processes (BP), cell compartments (CC) and molecular functions (MF). The most represented GO terms, for each category, associated to the genes detected in Fecund's experiments were summarized in D8.4. Metabolic pathways that were found to be associated to fertility have been also investigated and reported.
WP9: Field application of validated biomarkers
FECUND has produced a vast amount of information that have added considerably to knowledge of early reproduction. The objectives of the WP 9 were to evaluate the data and technological advances achieved in earlier WPs and to develop or test direct applications that can be used by the industry in the short term to start reversing the negative reproductive trend. The activities performed, results obtained and perspectives open in this WP are summarized below.
Validation of putative biomarkers of Negative Energy Balance and fertility outcomes: candidate biomarkers have been identified in WP6 to predict, in early post-partum, the severity of the NEB (D6.3) and reproductive outcomes at the 1st AI (D6.4). The objective of the tasks 9.1 and 9.2 was to validate the NEB and fertility biomarkers on a higher number of dairy cows in early post-partum. This kind of biomarkers could be useful for farmers for the optimisation of their animal’s management and also to define potential new phenotypes for breeding companies in order to refine actual traits.
Plasma samples have been collected on 155 cows of different breeds, in early post-partum, at 3±1 and 28±2 days in milk (DIM). Vaginal swabs have been collected at 28±2 DIM for 52 out of previously cited females. Body Condition Score (BCS) was registered at each time point (note from 0 to 5). The biochemical determination in plasma samples included the list established in the D6.4 and according to the procedures reported in the D6.3. The biochemical determination in vaginal swabs included the list of the putative biomarkers identified in D6.4 according to the procedures reported in the same document. Complementary data have been extracted from the French national database containing all the milking and reproduction information for each animal.
Cows have been grouped in subjects with low, moderate or severe NEB in accordance to the value of the BCS variation between 3±1 and 28±2 DIM: “low NEB”, “moderate NEB” and “severe NEB” respectively for a variation of 0 (no variation of BCS during this 1st month of lactation); [0;-0.5] and [-0.5;-1.0]. Moreover, cows have been classified in “Pregnant AI1” or “Not pregnant AI1” at the 1st AI based on the non-return in heat 90 days after the 1st AI post-partum. A cow present in the herd 90 days after the 1st AI and not re-inseminated was classified in the “Pregnant AI1” group whereas a cows inseminated twice or more was classified in “not Pregnant AI1”. All the cows not inseminated or inseminated once within a delay shorter than 90 days post 1st AI have been classified in “Not determinable” group and excluded from the analysis.
It was observed that the variation of BCS during the 1st month of lactation can reflect the severity of NEB which significantly impact the fertility success of the 1st AI with 70.5% of pregnancy rate for “Low NEB” females; 52.2% for “Moderate NEB” females and 16.7% for “Severe NEB” females respectively. Thus, this parameter could be considered as a good predictive marker of fertility outcomes but can be difficult to measure in routine on the field without specific monitoring devices.
Blood and vaginal biomarkers were assessed in this trial to have an indirect measure of the severity of NEB of each female. Even if it could be difficult to imagine a massive and systematic collection of vaginal swabs on post-partum females, the measurements made on that fluid highlighted some markers (Urea, Total protein, Globulins, AST/GOT, GGT, Haptoglobin and LDH) significantly different according to the NEB severity groups of females. Easier to obtain, measurements in blood highlighted some others biomarkers (Sodium, Zinc, Cholesterol, Ceruloplasmin, Albumins, total Bilirubin). Finally, the synthetic criteria Liver Functionality Index (LFI) developed by Bertoni and Trevisi is significantly different between females with low, moderate or severe NEB and therefore represents a good biomarker of NEB severity. As biomarkers the determination of a threshold to identify post-partum females with severe NEB with the maximum accuracy is still needed. In this trial, the low number of females presenting a severe NEB doesn’t allow to perform this kind of analysis. Thus, further experiments are still needed to go towards on field application allowing farmers to identify critical females and adapt their management.
Concerning the validation of fertility biomarkers, few of the metabolic parameters suggested in the D6.3 (plasma) and D6.4 (vaginal swabs) have been confirmed (GGT and Thiols groups in plasma samples; Triglycerides in vaginal swabs) in this experiment. Moreover, LFI doesn’t seem to be a good predictive marker of the fertility taking into account that LFI values are not different between Pregnant and Not Pregnant females. Fertility is a very complex trait which is probably affected by a numerous of other parameter not taken into account in this trial like the genetic background of the animals, the effect of the male, the management of the animals by the famers (oestrus detection method, insemination delay, feeding management).
Impacts of genetic improvement tools: in this task, the LFI trait was analysed together with milk production in a genomic selection perspective to assess the achievable genetic gain for novel fertility traits in dairy cattle. A sample of 96 Holstein-Friesian cows were genotyped at high density for 777,962 SNP, with an average call-rate of 99.65%. The minimum per-sample call-rate was 0.88: therefore, no animal was removed because of low-call rate. The minimum per SNP call-rate was 0: 3,561 SNPs with call-rate < 85% were discarded (~0.5%). The residual missing SNP genotypes (still about 150,000 missing data points) were imputed. After imputation, 1747 SNPs with duplicate or unassigned position, and 136,241 monomorphic SNPs were removed, leaving 636,413 available for subsequent analyses. As phenotypes, milk production at 28 DIM and the LFI were measured on the animals. A linear mixed model approach was used for BLUP estimation of genetic parameters and (genomic) breeding values. The covariance structure between observations was modelled building upon genetic similarities.
A negative genetic correlation between LFI and milk yield was estimated from this data, while the phenotypic correlations is practically zero. The predictive abilities which represents the accuracy of genomic predictions for the two traits have been calculated (0.608 and 0.723 respectively for milk production and LFI). The interest is to breed for cows that produce milk while keeping functional (i.e. maintain a healthy reproductive performance). Therefore, metabolically efficient milk production and reproductive performance are the broad breeding goal, which is not easy to measure on individual cows. LFI and milk yield can be used instead as breeding criteria related with functionality/reproduction (the former) and production (the latter). The relationships of the breeding criteria with the breeding goal and between them, are all relevant for the construction of the selection index for the improvement of the breeding goal. Thus, the expected response to selection can be estimated through the well-known ”breeder's equation”. Relevant terms for the response to selection are the intensity of selection, the accuracy of the selection index and the genetic standard deviation of the trait. Different scenarios were simulated, varying the selection intensity from 10% to 25%, and the relative emphasis placed on LFI and milk production (from 1/2 to 2/1). The results show, as expected, that response to selection is higher when the selection intensity is higher. Also, when higher emphasis is placed on one of the two components, the response to selection is faster compared to a more balanced objective where both components are improved simultaneously. This also is expected and well known: placing higher emphasis on one trait gives faster and higher gains, compared to the slower more balanced selection.
Validation of biomarkers of oocyte and embryo quality: in this task, two embryo culture media with low and high pyruvate concentration was evaluated on the development of in vitro produced embryos of two different breeds and origin: embryos produced from Holstein donors up to 12 months of age and older than 12 months (data reported in D7.5) and embryos produced from adult cows of the Chianina breed. The choice of the 2 breeds has been made taking into consideration also the objective of testing the protocols on animals that were considerably different in their physiological status. This study was carried out in a period of approximately 2 years and included 18 Chianina oocyte donor cows and 10 different Chianina bulls were used for IVF. The animals were subjected to the Ovum Pick Up procedure (recovery of oocytes by ultrasound guided follicular aspiration) at intervals of approximately 1 to 3 weeks. No hormonal treatments were applied on donors’ cows at any time in order not to interfere with the ovarian physiology. The oocytes were in vitro fertilized and the derived zygotes of each cow were cultured alternatively in low and high pyruvate in order to test the different culture protocols on all the donors included in the study. In total 104 and 139 OPU sessions were performed on 18 donor Chianina cows and the oocytes retrieved from the OPU sessions were divided in the 2 culture groups, low pyruvate and high pyruvate. The number of oocytes recovered was 1607 and 2219 and after IVF 1015 and 1565 fertilized zygotes respectively were obtained. In total 438 and 887 embryos were obtained from the 2 culture groups. The embryos were evaluated on the basis of the criteria of the International Embryo Transfer Society being G1 and G2 grade the embryos of highest quality (G1: more than 85% of the embryonic mass is viable; G2 more than 50% of the embryonic mass is viable) and G3 grade the embryos of lowest quality (G3: less than 50% of the embryonic mass is viable). The total percentage of embryos developed was calculated on the number of fertilized zygotes and resulted significantly higher for the high pyruvate group compared to the low pyruvate group (56,68% vs 43,15%; chi square test, p<0,05).
Therefore the results of this field study confirm the findings reported in D7.5 because the increase of pyruvate supplementation during culture has proven beneficial for embryo development both in Holstein and Chianina cows. In conclusion, the transfer of the protocols developed in WP7 to WP9 has been successful and represents an improvement in the results that can be obtained applying assisted reproductive technologies, such as Ovum Pick Up, IVF and embryo culture, under field conditions.
Validation of recessive lethal mutation previously identified on French bovine population: in this task, it was investigated the effect on fertility and on embryonic lethality in French populations of 71 candidate lethal recessive mutations previously identified (D5.3). These variants were not available on the routinely used beadchip and imputation to whole genome sequence data using 1000 bull genome population was needed. None of the variants was found to have a significant effect on fertility traits in French population but lack of observed effect may be due to imputation errors. Therefore, we investigated if candidate mutation located either close to an identified QTL or close to a region displaying deficit in homozygote. Out of the 71 candidate mutations, 34 met at least one of the criteria and three fulfilled the two conditions. Further investigations have to be performed in order to validate candidate mutation. Our main option is to include them in our genotype assay and search for a deficit in homozygous at SNP level. Lack of homozygous at SNP level will confirm lethality of the mutation. Analyses can also be redone with future run of 1000 genome. Increase in total number of sequenced animal should lead to higher imputation accuracy and therefore a more powerful QTL detection.
The reproductive performance is one of the major issues in dairy cattle farming and breeding. After the decline of the conception rate at first service, the inclusion of fertility in dairy selection programs has brought about an inversion of the negative trend: in the last 5-10 years, the decline in fertility has stopped, and some initial signs of recovery begin to be noticed. Still, despite the encouraging progress, a lot of work remains to be done, and this involves both basic and applied research. This workpackage integrated 5 tasks aiming to improve this key traits by (1) developing some new biomarkers of NEB and fertility and simulating their integration into breeding schemes; (2) improving in vitro embryo production protocols to increase the intensity of selection and optimize the genetic gain and (3) investigating some putative recessive lethal mutations to avoid part of embryonic losses. Some of these attempts have been successful and others a little bit less but, in any case, additional knowledge has been acquired, useful for the understanding of the complex fertility trait which still remain a great challenge for the breeding industry.
WP 10 Project Dissemination
The www.fecund-project.eu domain has been registered. A WordPress website has been created, populated with project information, and adjusted with a graphical template specific for the FECUND project. The website was created and maintained by PTP presenting project information, consortium details, news and events updates and the electronic versions of the project communication materials. The website contains a direct link to the website of the European project PROLIFIC, closely related with FECUND.
Two project leaflets were produced and distributed both as printed and digital copies. Both leaflets were designed and printed as a double-fold A4 document (size 21x29.7 cm). The two leaflets have similar designs and the main layout and final texts were decided by successive improvements of originally proposed drafts with the Project Management Committee (PMC). The layout is based on the fonts and colours of the FECUND logo. FECUND leaflet 1 provides information on the project background and challenges in dairy cattle fertility and reproduction, the interdisciplinary approach that FECUND is based on, objectives, outputs and potential applications, partnership with PROLIFIC and partners. It was first printed as 1,000 copies in M17. A second and adjusted version was printed as another 1,000 copies in M31. The leaflet 1 was widely distributed whenever relevant, e.g. for scientific events, meetings with stakeholders or mass media both as printed and digital versions via the project partners during the course of the project. Leaflet 1 is available at FECUND Project Website digitally as a high-quality pdf suitable for printing, http://www.fecund-project.eu/?page_id=12.
FECUND leaflet 2 provides information on the project background, key innovations, and impacts followed by the project results grouped as the “genetic model”, “metabolic model” and the “high technologies”. It was published in M45 and 200 copies were printed for the distribution via the partners, during FECUND PROLIFIC Joint Final Conference, other scientific events, meetings with stakeholders. The digital version was disseminated through the different social media channels during the project. The leaflet text was adjusted at the end of the project at M48. Leaflet 2 is available at FECUND Project Website digitally as a high-quality pdf suitable for printing, http://www.fecund-project.eu/?page_id=12 .
Four external project newsletters were designed and published electronically to approximately 400 stakeholders during the course of the project. It provided information on project news, events, results, and publications. The newsletters are available at FECUND Project Website as a high-quality pdf suitable for printing, http://www.fecund-project.eu/ .
Newsletters were sent to a number of different stakeholder groups consisting of academics, breeding industry, farmer representatives, national bodies and policy makers selected from EFFAB and FABRE-TP mailing lists which contain many of the interested industry and academic organisations. The newsletters were also distributed to other EU projects active in the same area such as PROLIFIC. The newsletters were also promoted through EFFAB and FABRE-TP social accounts (Twitter, LinkedIn). The Newsletter mailing list was created to be an active one by enabling subscriptions from the website.
First of the stakeholder events was a joint FECUND-PROLIFIC final conference organised at M45 targeting the scientific community. It was organised jointly with the related EU Project PROLIFIC and the aim of the joint conference was to integrate the results of the two projects to move improvements for fertility in cattle forward. The Joint Final Conference was organised as a satellite workshop to the 20th Annual European Society for Domestic Animal Reproduction (ESDAR) Conference 2016, held on October 30, 2016 in Lisbon, Portugal. The title of the workshop was “Innovations to improve dairy cow fertility”.
The conference was promoted through the project website and EFFAB communicated the event both on EFFAB and FABRE TP website and social media accounts (Twitter and LinkedIn). CONAFE (La Confederación de Asociaciones de Frisona Española) and ESDAR have also communicated about the conference to their network. The main promotion material was the dedicated flyer for the conference designed by EFFAB and disseminated by all parties involved in promoting the event. Fecund partners have also promoted the conference within their networks. The promotion has continued during the conference with live tweeting and sharing updates from EFFAB and FABRE TP social media accounts. A total of 34 tweets with nearly 5,000 views were shared during the conference.The Conference participation was 63 people from 22 countries (16 EU countries). There was a total of 13 speakers, 5 from FECUND, 6 from PROLIFIC, the EU Project Officer and a keynote speaker. All the abstracts and presentations made during the Joint FECUND/PROLIFIC Final Conference are accessible from FECUND (http://www.fecund-project.eu/?p=281) and PROLIFIC (http://www.euprolific.eu/) websites.
The second event was a live webinar broadcasted at M48. The aim of the webinar was to disseminate and communicate the results of the researches carried out in scope of the FECUND project to facilitate the knowledge and technology transfer to the stakeholders, in particular the breeding industry, end-users and the knowledge institutes for future research possibilities. The FECUND Stakeholder Webinar was held on January 17, 2017 using GoToWebinar software. The title of the Webinar was “Improving dairy cattle fertility: Results from the FECUND Project”. Several promotion materials were designed by EFFAB including a Save-the-Date flyer and a webinar program flyer. The FECUND stakeholder webinar was promoted both by EFFAB, PTP (as coordinator) and the FECUND project partners. It was promoted through the project website and EFFAB communicated the event both on EFFAB and FABRE TP website and social media accounts (Twitter and LinkedIn). A mailing list consisting of selected breeding industry, policy makers and knowledge institute representatives (more than 1,500 in total) was created by EFFAB. ATF (Animal Task Force) and EAAP (European Federation of Animal Science) have also communicated the webinar to their network and promoted the event on their websites and newsletters. The participation was 63 people from 20 countries (16 EU countries). There was a total of 5 speakers from FECUND project. All the presentations made during the FECUND Stakeholder Webinar and the webcast of the webinar are accessible from FECUND (http://www.fecund-project.eu/?page_id=258) website. The webcast is on EFFAB YouTube Channel (https://www.youtube.com/watch?v=vaoqkjBT5FM) and has been promoted by EFFAB, FABRE TP and FAO. The webinar recording has already been viewed 215 times so far and is still being viewed.
Potential Impact:
The FECUND project helped to reduce negative fertility trends in dairy cattle and hence promote economically sustainable and ethical livestock production, by creating new knowledge to define specific biological processes associated with early reproductive function. There was still only rudimentary knowledge of the early developmental events around and shortly after conception affecting embryo developmental competence, implantation and the establishment of pregnancy. FECUND used advanced molecular techniques to dissect the interaction between the mother and the embryo during the early phases of development and start defining the signals and responses. A major impact of FECUND was to create new physiological knowledge to unravel the complexity of communication between the mother and the embryo that is necessary in early reproductive phases for establishing a pregnancy.
With increasing milk production many cows are unable to maintain energy balance during the transition period as they establish their lactation, when they have to mobilise body reserves to compensate for the shortfall in dietary intake of energy. Some high producing cows are able to cope with the transition period with little impact on fertility, while others do not cope. FECUND investigated the biological impacts of Negative Energy Balance on early reproduction from oocyte and early embryonic development to the responses of the mother during early reproductive events and by investigating postpartum cows at peak lactation vs postpartum non-lactating cows vs heifers to create knowledge on reproductive physiology. The new knowledge obtained from reproductive physiology provided new phenotypes that will define a specific biological processes associated with early reproductive function. These new traits could be used in genetic selection schemes for reproductive traits. FECUND built on the knowledge created to search for surrogate traits and correlated biomarkers to predict the novel traits, that can be assayed in easily sampled material, such as blood or vaginal swabs. These surrogate markers will be more easily applied in the field to collect sufficient data and hence construct accurate estimated breeding values for the new traits, which can then be used in selection programmes.
Fecund has identified biomarkers that will predict animals that respond badly during the transition period which would enable farmers to identify cows at high risk of developing energy deficiency and to devise improved strategies to avoid impaired fertility. Biomarkers for energy balance risk have been identified and could also be used to produce genomic Estimated Breeding Values for fertility with higher accuracy for genetic selection and will enable farmers to devise strategies to address the energy deficiency for high risk cows. This would give the owners of such data a significant competitive edge.
Sperm survival and vitality, which is essential for good fertility, is likely to be influenced by the female reproductive environment. FECUND investigated the influence of factors in the uterus and oviduct on sperm vitality. Such information will have an impact for farmers, or the insemination services allowing them develop strategies to improve the female environment, e.g. by prior treatment, or by addition of beneficial factors identified by FECUND at, or shortly before, insemination. A major impact of FECUND for management will be the development of biomarkers for sub-optimal female reproductive tract environment at AI.
Embryo based reproductive techniques are being increasingly used to improve the rate of genetic gain. FECUND has integrated the knowledge created on early development to improve in vitro reproductive protocols. A major impact of FECUND for assisted reproduction is the improvement of culture techniques and protocols to produce high quality oocytes and embryos for transfer. Coupling with advanced breeding methods with genetic markers will select the best embryos for transfer and hence achieve even more rapid genetic progress. FECUND provided knowledge to optimise assisted reproduction protocols for the in vitro production of embryos. Genotyping embryos led to the identification of high genetic merit candidate embryos from which to select. Using this technology will also contribute to the preservation of genetic diversity. In addition deriving oocytes from young animal by in vitro maturation then in vitro fertilisation will reduce generation time. FECUND integrated the knowledge created on the molecular genetics and genomics, epigenetics, proteomic and metabolomics of early development to improve in vitro reproductive protocols. A major impact of FECUND for assisted reproduction is the improvement of culture techniques and protocols to produce high quality oocytes and embryos for transfer.
Outputs from the FECUND will have an impact at two levels: 1) those that can be implemented immediately by individual organisations or associations to achieve a competitive advantage for them and their clients or members, such as genetic markers, and 2) those that will need a collective approach at national and international levels to bring about benefits, such as improved trait recording. Collectively, these impacts contributed to the sustainability of farming and hence rural economies and employment, and will ensure food security with reduced cost in Europe and beyond.
Immediate applications from FECUND include validated genetic markers for fertility traits (e.g. lethal recessives) which will be immediately available for genotyping and application in gene assisted selection of bulls and cows by the breeding companies and the national associations. To apply project results to improved genetic selection, the outputs need to be implemented at national and international levels. To achieve this, national and international breeding associations and international structures have been involved. Before implementation of new recording protocols a cost/benefit analysis would be required (as carried out in WP9) and the mode of introducing the novel trait recording scheme established using agreed international standard procedures for the data collection, recording and exchange.
Fecund partners ensured the widest dissemination of the project results to stakeholders, scientists, end-users and farmers by means of several focused activities: an updated project website for the promotion of results and events, 2 dedicated workshops addressed to scientists and stakholders, targeted newsletters and printed leaflets.
List of Websites:
The website address is the following: http://www.fecund-project.eu/
Relevant contact details - Fecund Coordinator:
Dr. Filippo Biscarini
Fondazione Parco Tecnologico Padano
via Einstein - località Cascina Codazza
26900, Lodi (LO), Italy
Tel: +39 3407499754
e-mail: filippo.biscarini@ptp.it
The FP7 European research project Fecund ended in January 2017. For 48 months, the research partners and SMEs addressed the needs of the cattle industry: the problems of reproductive success in the cow using a systems biology approach to deliver targets for genomic selection of fertility, the improvement of biotechnologies for reproduction and the identification of molecular indicators to support early stages of reproduction were investigated. The project structure was defined by 9 scientific work-packages (WP1-9), one transfer and dissemination WP (WP10) and a management WP (WP11).
WP1 was focussed on the establishment of two animal models to independently examine the effects of (i) the metabolic status of the cow and (ii) genetics on the factors involved in establishing a pregnancy. Animals in both models were used to generate follicular fluid, oocytes, embryos, oviduct tissue and fluid, conceptuses, endometrial tissue and uterine lumen fluid which were analysed in subsequent WPs. In WP2 transcriptomic and proteomic analysis were carried out to investigate biological systems directly involved in the formation of oocytes. In WP3 advanced techniques were developed for sampling. Analysis on oviduct cells, sperm transport and fertilization and embryos were performed.In WP4 the question whether metabolic condition or genetic merit for fertility affects the conceptus-endometrial crosstalk in dairy cattle was addressed. WP5 addressed novel approaches to identify candidate recessive lethal mutations, identify variations within key genes involved in early reproductive biology identified in previous WPs.
WP6 carried out activities to identify robust biomarkers predictive of the reproductive status of the cow in biological fluids easy to obtain as plasma and vaginal swabs. WP7 managed to improve the techniques for assisted reproduction taking advantage of the novel findings derived from the animal models developed in previous WPs and testing novel hypothesis reported in recent literature in order to improve oocyte and embryo developmental competence. WP8 provided the bioinformatic structure to manage the data produced and the tools for primary data analysis.
WP9 evaluated the data and technological advances achieved in earlier WPs and developed or tested direct applications that can be used by the industry in the short term to start reversing the negative reproductive trend. WP10 covered the main dissemination tasks through a public website, periodic newsletters, leaflets and workshops.
Main objectives achieved by FECUND were to:
carry out appropriate scientific studies focused on early embryonic development to create new knowledge; develop improved ways of defining, measuring and recording fertility traits including the identification of direct and surrogate measurements of physiological variables;identify phenotypes associated with metabolic profiles and to correlate them with low or high fertility; identify biomarkers associated with early aspects of reproduction that can be used to quantify the new phenotypes;identify genetic loci associated with reproductive performance in the breeding populations; add functional validation to genetic information using system biology studies; contribute to the improvement of reproductive biotechnologies for efficient livestock production; disseminate the information to the wider cattle breeding industry and to end users.
The collaboration and interaction among partners of the consortium led to the achievement of the objectives of the project.
Project Context and Objectives:
FECUND is focussed on fertility of dairy cattle, where the greatest reproductive problems exist and where the financial benefits of improved reproduction efficiency are greatest. The project addressed the reproductive biology of ovulation through implantation which is the period which is associated with greatest reproductive losses. FECUND worked together with the PROLIFIC project, which was funded under the same KBBE call for proposals, to address the most important reproductive problem facing cattle production at a European level.
Concomitant with intensive selection for increased milk yield, reproductive performance of dairy cows has declined in recent decades, in part due to an unfavourable genetic relationship between these traits. Today the decline in fertility has become a major concern of farmers and the dairy industry. The causes of this decline in reproductive success are multi-factorial (Pretheeban et al., 2009) with higher milk yield being shown to have an inverse relationship with the pregnancy rates observed in dairy cows (Faust et al., 1988). Embryonic loss is a major cause of this reduced pregnancy rate (Hansen, 2002). Successful implantation and establishment of pregnancy is dependent on good quality embryos, an appropriate endometrial environment, and the successful interaction between the embryo and the endometrium (Chen et al., 1999). While selection for productivity has had a negative impact on fertility and some genetic loci have been shown to have opposing effects on production and fertility, other loci act on only one or other of these traits. Understanding the relationship between metabolism and fertility and the interaction with genetics will allow producers to optimise productivity and reproductive success.
Good reproductive performance is crucial for economic sustainability as well as animal welfare. However, reproductive success is a highly complex trait and fertility problems in dairy cattle include an increase in postpartum failure to return to cyclicity, poor expression of oestrus, defective oogenesis, poor embryo survival and susceptibility to uterine infections. Both genetic and management changes can be envisaged that may improve reproductive success but these could compromise productivity. Therefore, it is important to understand the interaction between the various facets of reproduction and fertility to devise appropriate strategies to manage and improve reproductive success.
The challenges facing improvement in dairy cattle fertility can be addressed through two interacting factors: biological changes in dairy cattle, resulting from genetic selection, and changes in dairy herd management. The economic environment under which the dairy industry operates leaves little margin for error before dairy farming becomes unsustainable. Human and economic factors differ among countries, regions and individual herds, and so local management decisions may impact directly on fertility and profitability. Improvements in the definition and measurement of fertility in relation to the changing biological characteristics of dairy cattle would help to optimise these management decisions, and have direct economic benefits for the farmer. After the successful first calving each cow is expected to calve at an interval that maximizes the economic output from milk production. However, in the face of declining reproductive efficiency it is difficult to maintain the desired reproductive period, which results in early culling of low fertility cows, with the associated need to breed an increased number of replacements. Indeed the herd life of a modern Holstein cow is little more than 3 lactations (Oltenacu and Broom, 2010), in part caused by reproductive problems. Reduced age at the first calving has, to some extent, decreased the cost of breeding replacements, however, the short herd life is a major concern and large economic burden.
Metabolic demands placed on the cow to sustain a high level of milk production impact negatively on the reproductive function of postpartum cows. Factors other than fertility per se contribute to the decreasing reproductive efficiency of dairy herds. Among these are changes in expression and detection of oestrous behaviour, which results in mistimed insemination. Inappropriate cow management post partum during a period of high metabolic stress amplifies the problem. Nevertheless, the underlying fertility of dairy cows has declined and new knowledge is required to develop successful management systems to meet the metabolic and nutritional needs of highly productive cattle. In addition, housing and social needs of increasingly productive animals impact on their inherent capacity to achieve and maintain pregnancy.
Fertility is a complex trait affected by the environment, management and genetics. Efficiency of selecting for improved fertility using both traditional genetic selection and now genomic selection relies on having accurate fertility data; however, traditional measures of fertility have low heritability making selection of sires based on genetic merit for fertility traits unreliable. This is partly because indicator traits have been used that do not accurately reflect the true underlying biological variation. Furthermore, reproductive success is a trait that is expressed late in life, and so good information on fertility is not available when breeding decisions have to be made. Advanced, technically complex and expensive recording technologies that could be used to improve the measurement of fertility in experimental situations cannot be applied in a commercial setting. Novel approaches and indicator traits are therefore needed to measure fertility. These will be used to identify genetic loci and gene variants to improve selection, to assist cow management and to improve assisted reproductive systems.
The FECUND project addressed the problems of reproductive success in the cow using a systems biology approach to deliver targets for genomic selection of fertility, improve biotechnologies for reproduction and identify molecular indicators of the ability of the cow to support early stages of reproduction, when most embryonic loss occurs.
The overarching objectives of FECUND were therefore:
• to carry out appropriate scientific studies focused on early development to create new knowledge;
• to develop improved ways of defining, measuring and recording fertility traits including direct and surrogate measurements of physiological variables;
• to identify phenotypes associated with metabolic profiles and to correlate them with low or high fertility;
• to identify biomarkers associated with early aspects of reproduction that can be used to quantify the new phenotypes;
• to identify genetic loci associated with good reproductive performance in the breeding populations;
• to add functional validation to genetic information using system biology studies;
• to contribute to the improvement of reproductive biotechnologies for efficient livestock production;
• to disseminate the information to the wider cattle breeding industry and to end users.
Realising these objectives was essential to underpin the sustainability and competitiveness of the European cattle industry.
Project Results:
The Fecund consortium achieved all the planned objectives and deliverables for the project. The consortium has been very active and collaborative to perform the research, and then transfer and disseminate scientific outcomes to researchers and scientific community, policy makers and stakeholders.
WP1: Animal Models
Pregnancy establishment in mammals, including dairy cows, involves several key checkpoints along the developmental axis including the follicle/oocyte, oviduct/early embryo and uterus/conceptus. WP1 was focussed on the establishment of two animal models to independently examine the effects of (i) the metabolic status of the cow and (ii) genetics including genetic merit for milk production on the factors involved in establishing a pregnancy in order to address whether dairy cow infertility is primarily a genetic or metabolic problem. As such, WP1underpinned WPs 2, 3 and 4 as it was responsible for the generation of tissues and samples which were then distributed amongst the partners and used in the subsequent three WPs.
To breed and manage cows under well defined and controlled experimental conditions in order to produce information, biological tissues and fluids from animals of divergent metabolic and genetic status were collected for analysis in subsequent WPs.
The specific objectives are to:
• establish optimised experimental protocols and write SOPs for the experimental work and sample collection
• identify, source and breed suitable animals for the studies
• collect data and experimental materials following the standard operating procedures
• to provide the samples and data to partners for the execution of work in subsequent WPs
The Energy Balance model was based on a comparison of lactating and nonlactating postpartum cows and maiden heifers. Forty in-calf primiparous Holstein-Friesian heifers and 20 non-pregnant Holstein-Friesian heifers with a similar Economic Breeding Index were purchased and enrolled into the study. All cows calved between February and May 2013. At calving, cows were randomly assigned to one of two groups 1) lactating (n = 20) or 2) non-lactating (n = 20). From calving, animals in the lactating group were milked twice per day (0700 and 1600 h), while those in the non-lactating group were dried off immediately after calving (i.e. never milked. Pre-calving, all animals were fed 30 kg grass silage/head/day with pregnant heifers getting an additional 3 kg concentrates/head/day. Post calving, dry cows received ad lib grass silage plus 4 kg concentrates/day, while lactating cows received 24 kg maize silage/16 kg grass silage plus 7 kg concentrates. Beginning two weeks prior to expected calving date, all animals were weighed, body condition score (BCS) recorded and blood sampled twice weekly until Day 85 postpartum. Similar measurements were recorded for the non-pregnant heifers over the same period. A thorough metabolic analyses of all animals was carried out by measuring concentrations of NEFA, BHB, IGF-1, insulin, and glucose.
The Genetic model comprised three groups of heifers (n=45) to assess the impact of genetics on variations in fertility which were used to generate samples to be be used in subsequent WPs. The groups were:
- 15 Prim Holstein heifers with high genetic merit for fertility (FERTI+)
- 15 Prim Holstein heifers with low genetic merit for fertility (FERTI-)
- 15 Montbéliarde heifers, a French local breed with a good fertility (positive control group)
Animals were similarly characterised to those in the Energy Balance model.
Animals in both models were used to generate follicular fluid, oocytes, embryos, oviduct tissue and fluid, conceptuses, endometrial tissue and uterine lumen fluid which were analysed in subsequent WPs.
SOPs were generated for the collection and distribution of these tissues and biological fluids to optimize downstream data collection.
WP2: Follicular fluid protein profiles
WP 2 investigated the biological systems that are involved in the formation and development of oocytes. Therefore cumulus cells, oocytes and follicular fluid generated in Work package 1 were analysed. Proteome analysis in follicular fluid is generally biased by the extremely high abundance of a few individual proteins (e.g. serum albumin, complement c3). Nevertheless, more than 200 protein analytes could be addressed in the samples of both models. Concerning follicular fluid of the metabolic model, 251 proteins were identified, and quantitative differences in the follicular fluid proteome were detected using LC-MS/MS followed by statistical and bioinformatic analyses: Compared to heifers, 4 proteins showed reduced levels in dry cows and 6 proteins in lactating cows. Another six proteins differed between lactating and dry animals. In follicular fluid samples of the genetic model, 247 proteins were identified. Compared to high fertility heifers (LFH), the levels of 3 proteins were significantly enhanced in low fertility heifers. The comparison of low fertility heifers to Montbeliard revealed three proteins enhanced and three decreased in LFH. In the comparison high fertility heifers (HFH) to Montbeliard, 1 protein was less abundant and four were more abundant in HFH.
Thus, the data revealed clear differences in the follicular fluid protein composition between the relevant groups of both animal models. The affected proteins represent mainly the biological processes, metabolic process, response to stimulus and immune system process. According to GO analysis the most prevalent molecular functions of these proteins are catalytic activity, e.g. glutathione peroxidase 3 and serotransferrin, binding and enzyme regulator activity with the involvement of lactadherin, alpha-2-macroglobulin and C8G protein.
RNA expression profiles in cumulus cells: Cumulus cells (CC) are involved in oocyte nutrition and protection and essential for oocyte metabolism, thus influencing oocyte quality. Using NGS technology (RNAseq), we compared RNA expression profiles in cumulus cells of Holstein-Friesian and Monbeliarde animals in metabolic and genetic models. In the CC from Heifers and Lactating cows, 187 DEG enriched were evidenced DEGs were enriched in GO related to lipoprotein binding and steroidogenesis, transcription factor (TF) and protein kinase activity. Thus, NGS analysis of bovine cumulus cells in metabolic model evidenced a strong impact of lactation on transcriptional activity in cumulus cells by the regulation of a large number of genes in particular those involved in intracellular lipid metabolism. Within the genetic model, the highest number of DEGs were found in the comparison of Montbéliard and Fertiplus (205 DEG : 37 higher in Montbéliard, 168 higher in Fertiplus), while the comparison of Montbéliard and Fertiminus revealed only 102 DEG. Interestingly, these 102 DEGs were able to discriminate Montbeliarde and FertiMin groups using PCA, whereas no clear discrimination was observed between Monbéliarde and FertiPlus. The data indicate a strong impact of lactation on transcriptional activity in cumulus cells by the regulation of a large number of genes in particular those involved in intracellular lipid metabolism.
RNA expression profiles in oocytes: Oocytes accumulate the RNA during their growth period for using maternal factors during transcription silencing through oocyte maturation and first embryo cleavages. RNA expression profiles in the immature oocytes of Holstein-Friesian and Montbeliarde animals in metabolic and genetic models was compared using NGS technology (RNAseq). In the metabolic model, oocytes of dry and lactating cows showed 8 DEG, and Oocytes from Dry cows and Heifers showed two DEGs, whereas Heifers and Lactating cows showed 51 DEG: 19 genes were up-regulated and 32 down-regulated in lactating cows. These genes are involved in cellular biosynthesis and transcription regulation. GO terms were enriched in genes of apoptosis and TGF-beta signalling. In the genetic model, 36 DEGs were found between the oocytes from FertiPlus and FertiMinus heifers: 26 genes were up-regulated and 10 down-regulated in FertiPlus. A comparison of Montbéliarde and FertiPlus Holstein oocytes revealed 202 DEGs, 83 up-and 119 down-regulated in Montbéliarde. 68 DEGs were detected by comparison of Montbéliarde and FertiMinus oocytes, 28 were up-regulated and 40 down-regulated in Montbéliarde. DEGs were enriched (p<0.05) in molecular function GO ‘”nuclear acid binding”.
In summary, NGS analysis of bovine oocytes in metabolic model showed that oocyte transcriptomes were moderately affected by the different energy balance due to lactation. In the genetic model, oocytes from the Monbéliarde heifers showed the most discriminative transcriptome profile in comparison to FertiMinus Heifers, which may at least partly explain the difference in their reproductive performance.
Factors influencing oocyte quality: Cumulus cells transcriptome analysis in different models revealed that differently expressed genes (DEG) were involved in particular in pathways related to lipid metabolism with a high expression level of genes involved in fatty acid (FA) synthesis, transport, elongation, degradation and oxidation. Therefore we analysed whether the factors related to FA metabolism may influence oocyte quality: by RT PCR of thirty genes at different times of IVM culture, by expression at the protein level by Western blotting and immunofluorescence, and by mass spectrometry to establish lipid fingerprints. Functional studies using pharmacological modulators of lipid metabolism demonstrated that during IVM (i) C75 neither affected oocyte maturation rate nor viability of CC and oocytes (DO and CEO); (ii) Mildronate had no impact on oocyte/CC viability but significantly diminished oocyte maturation rate in dose-dependent manner (CEO); (iii) Etomoxir dose-dependently blocked meiosis progression in both CEO and DO and affected CC viability, and in consequence, oocyte survival. Mass spectrometry analysis revealed significant differences between lipid fingerprints in CC before and after IVM, and numerous differentially abundant lipid species were determined (p<0.05). Moreover, CC showed significant accumulation of lipid droplets throughout IVM as shown by microscopy analysis.
In addition, it could be shown that signalling pathways were affected by lipid metabolism regulators in oocytes and cumulus/granulosa cells: in oocytes, etomoxir significantly decreased MAPK 3/1 phosphorylation after 6h IVM, and in cumulus/granulosa cells, this agent decreased phosphorylation of MAPK14 (alias P38) and SMAD1/5/8 (p<0.05) but not pMAPK3/1. In summary, functional studies based on in vitro maturation of bovine oocyte-cumulus complexes demonstrated a crucial role of fatty acid metabolism in CC on oocyte maturation and survival. Numerous genes known to be expressed only in lipogenic tissues (adipose, liver, muscles) were investigated in cumulus cells for the first time.
The data suggest that an active regulation of lipid metabolism promotes the successful completion of oocyte maturation and affects oocyte developmental competence.
Effects of selected factors on oocyte quality: Data derived in FECUND describe significant differences in the composition of the follicular fluid of postpartum cows induced by lactation, which could potentially influence the quality of the oocyte ovulated and subsequent fertility. Using these data generated from our unique model of postpartum dry and lactating cows and maiden heifers we have examined the direct effect of certain follicular fluid components on oocyte quality in vitro. Expression level of several key genes involved in lipid metabolism was measured by real-time PCR. Analysis of gene expression in cumulus cells after 22h IVM showed that 100 µM docosahexaenoic acid (DHA) induced a three-fold increase in the FA transporter CD36 and a two-fold decrease in FA synthase (FASN) gene expression in the cumulus cells of corresponding oocytes, whereas lower DHA concentrations had no effect. We could show that administration of a low physiological dose of DHA (1 µM) during IVM may have beneficial effects on oocyte developmental competence in vitro with no effect on nuclear maturation or lipid metabolism in surrounding cumulus cells. In contrast, high DHA concentration (100 µM) had negative effect on embryo development and FA synthesis in cumulus cells. Based on data from Task 2.1 highlighting differences in components of follicular fluid in maiden heifers and postpartum dry and lactating cows, we tested the effect of Stearic acid, Linoleic acid, Linolenic acid, Myristic acid, Docosahexaenoic Acid, and Eicosapentaenoic acid during IVM on oocyte cleavage and blastocyst yield following IVF. At the concentrations tested, exposure of COCs to stearic acid during in vitro maturation significantly decreased cleavage and development of bovine oocytes which is consistent with the published literature on the effect of this fatty acid on oocyte quality. There was no apparent effect of the other fatty acids tested.
The results highlight that some of the differences described between lactating and nonlactating animals may be implicated in the subfertility observed in postpartum lactating dairy cows.
Epigenetic competence of oocytes: the epigenetic modification of the oocyte is important for correct programming of subsequent development. Our data showed that chromosomal compartment of oocytes collected from prepubertal sheep, although genomically stable, is hypomethylated and that impaired DNA methylation machinery persists also in blastocysts developed from those oocytes.
In the bovine metabolic model, serum analysis revealed reduced concentration of homocysteine in lactating cows compared to both not lactating ones and maiden heifers and increased concentration of folate and vitamin B12 in lactating cows compared to not lactating ones. These results confirm the adequacy of three groups as metabolic models. When oocytes collected from the three groups were subjected to qPCR for DNMT1, DNMT3A and DNMT3B genes, no significant differences were observed between maiden heifer and not lactating cows. However, deregulated expression of DNMT1 and DNMT3A was observed in oocytes from lactating cows. These results suggest dysfunction of DNA methylation machinery in a critical developmental window, which will affect the development of embryos and consequently of the ongoing pregnancy.
WP3: Oviduct-embryo interactions
The oviduct, narrow duct between ovary and uterus is hosting several events that are determinant for the success of reproduction. The oviduct is capturing the cumulus oocyte complex at ovulation and drives the female gamete to the site of fertilization. Sperm cells are stored at the uterotubal junction, transported toward the oocyte and activated to allow fertilization. Oviduct is also the site of early development during the first five days, during which the activation of embryonic genome occurs and transports the growing embryo to the uterine cavity for implantation. More than 50% of reproductive failures in cattle, occur during the first days after insemination, indicating that they may take place in the oviduct, highlighting the importance of this organ. However, the oviduct being poorly accessible, it represents a kind of black box in which it is very difficult to know what happened (failure of fertilization, low oocyte quality, early embryo death...). In this context, the FECUND project allowed to gain a high amount of valuable information that will help understanding how the oviduct is regulating early reproductive events and how the genetic and metabolic status of the female may affect this finely tuned environment.
To study oviduct physiology and functions, oviduct samples have been collected after slaughter at Day 3 of a synchronized estrous cycle of heifers (n=5), lactating (n=4) and dry (n=5) cows in the energy balance model in Ireland and from Montbéliarde (n=3), high (n=5) and low (n=4) fertility Holstein cows of the genetic models in France. Oviduct have been dissected and processed to flush the fluid for proteomic and sperm survival function studies and scrap the epithelial cells for transcriptomic studies. Oviduct ipsi- and contralateral side from ovulation, as well as isthmus and ampulla segments of each oviduct were processed separately. In parallel, oviduct have been collected in France at slaughterhouse and classified as preovulatory, postovulatory, early luteal, mid luteal and late luteal phase according to ovarian morphology (follicles and corpus luteum). In addition, to evaluate the impact of variable maternal environment on the set up of gene expression of early developing embryos, in vitro fertilized 2-4 Cell stage embryos were cultured in vitro for 5 days or transferred in the oviduct of females from the different groups for the same period of time, using an innovative laparoscopic approach. The transcriptomic and proteomic analysis have been done at LMU and INRA (proteomic of estrous cycle fluid). The analyze of sperm survival supporting activity of the fluid was done in PTP. The Fecund project clearly showed that oviduct provides a finely tuned environment suitable for sperm survival, fertilization and early development. This environment is under the control of genetic parameters and seems poorly sensitive to variations of the metabolic status of the female, providing steady conditions for the success of early reproductive events. The big data set generated by the project is still under analysis and will be opened to the scientific community.
Advanced techniques developed for sampling: progress in performing tissue, fluid and oocyte sampling including in vivo culture of in vitro derived embryos, embryo collection and embryo transfer have been made. These procedures have been standardized and used in both animal models, metabolic and genetic model. Using this design it was possible to generate extensive molecular genetic and metabolic profiles of pathophysiological events during early reproduction including follicles, oviducts and uterine horns.
Study focusing on the in vivo collection of oviductal fluid: another study has been designed in order to establish a method for in vivo collection of oviductal fluid. This fluid accurately provides the physiological environment timely adjusted to all reproductive processes and requirements, which occur in the Fallopian tube. Since the oviduct provides a very low amount of fluid in which gametes enter the oviduct and accomplish fertilization, and in which embryos develop and migrate many efforts have been performed to access the oviduct for fluid collection. However, these attempts were very invasive and it cannot be ruled out a biased effect on the analytical outcome.
A novel method for collection of tubal proteins and fluid in heifers was established. For this purpose beads were pre-tested in vitro for its binding capacity, normally used for affinity chromatography, selected the ones fulfilling the best criteria, transferred them into the oviduct of heifers and performed re-collection one hour later. A second group of animals served for the collection of tubal fluid. For this purpose a flushing cannula was introduced into the oviduct and flushed the ampulla three times using 0.5 to 0.8 ml medium. Oviducts obtained from slaughtered animals have also been flushed and served as controls. Both techniques have been used in animals at Day 1 and Day 3 of the estrous cycle which accords with two different activity status of the oviduct. Re-collected beads and flushing samples were either used to measure the total protein content or were prepared for mass spectrometer analysis (LC-MS/MS). The experiment showed that both, transferring beads and flushing of the oviduct can be successfully used to collect proteins and the total fluid, respectively. It has been shown that both methods provided enough total protein for analysis. In total, 3165 proteins could be identified together in all groups. According to the physical characteristic of each collection procedure, group-specific protein clusters have been found. This method can be frequently used in cows, it is minimally invasive and does not need much time to accomplish. This is the first study aiming at the collection of tubal protein and fluid in heifers using endoscopy.
As technically developed and demonstrated, this method can now be used in cows for a detailed screening of proteins as well as total oviductal fluid containing many components such as lipids, epithelial cells and immune cells, exosomes etc.
It also suggested to combine these collection methods with in vivo embryo development using single or superovulated animals or by in vivo culture techniques in order to perform tubal sampling according to a precise embryo cleavage stage.
Once, more knowledge has been gained about the functional groups of proteins which might have a pivotal role in embryo development, the bead technology can be improved and refined using beads with specific physical surface characteristics to selectively bind and identify these proteins.
In total it is concluded that within the FECUND project a new method has been described and developed. This method provides new experimental designs which might be very helpful in finding ovidcutal components necessary for a better understanding of processes in early reproduction including in vitro production of embryos.
Oviduct transcriptomic: a very high number of DEGs was observed in the oviduct for Montbéliarde compared to Holstein (both F+ and F-), but very few DEGs were recorded between F+ and F- genotypes. Inverted situation was recorded for the endometrial tissue with few DEGs between breeds and higher number of DEG between F+ and F-. Considering the metabolic model, the number of oviduct DEGs (ipsilateral isthmus) was lower between lactating and dry cows (n=15) than between heifers and lactating or dry cows (n=127 and 223, respectively). This indicates that the breed (Holstein vs. Montbéliarde) and the physiological status (Heifers vs. Cows) are more important than fertility genotype and energy balance in determining the oviduct gene expression profile. The complete lists of DEGs are available in deliverable 3.1. They highlight the importance of several pathways in the oviduct regulation, such as steroids biosynthesis, complement and coagulation cascades.
Oviduct proteomic: the high throughput protein contents of bovine oviduct fluid along the estrous cycle has been published for the first time within Fecund project, showing the fine tuning of protein secretion and pointing out protein candidates specifically expressed by the time of ovulation, fertilization and early development, representing interesting candidates for further studies of oviduct function (Lamy et al., 2016). In addition, it has been shown that many of the oviduct secreted proteins are packed in extracellular vesicles and that these vesicles may have important functions in the regulation of early development (see WP7). In total, 1,740 and 1,902 proteins were identified by nanoLC-MSMS in the NEB and genetic models, respectively. The comparative analysis revealed differentially expressed proteins (DEP) between isthmus and ampulla regions of the oviduct, highlighting the functional specialization of these two segments. These DEP will be interesting candidates to follow up for a better understanding of oviduct functions. A relatively low number of DEP was observed between the groups of the NEB model, while this number was slightly higher between the groups of the genetic model (Figure 4). The full lists of DEP are available in deliverable 3.2. Promising biomarker candidates have been identified by gene ontology analysis of the abundance altered proteins, which revealed that most proteins are assigned to the term metabolic process, such as Peroxiredoxin 2 (PRDX2), SERPIND1 or Glutathione S-transferase theta-1 (GSTT1), which were found to differ significantly among the groups in both models.
Sperm transport and survival: the first approach was to test the effects of the oviductal fluids collected from both models on sperm motility and kinetics parameters, using Computer Assisted Semen Analysis (CASA) system, to reveal if effects observed in vivo can be mimicked in vitro. The motility parameters of sperms decreased significantly during co-incubation time in all the groups. However, this decrease was significantly faster for semen co-incubated in SOF medium alone (P<0.001) than for those incubated in oviduct supplemented medium, whatever the group in both models. Overall the fluids collected in the ampulla seem to have a more positive effect on sperm motility and kinetics parameters conservation, compared to the fluids collected in isthmus. The second approach consisted in the use of a microendoscopic imaging system (Cellvizio) that INRA has developed in sheep to follow the progression of fluorescent spermatozoa in genital tract in vivo, after cervical insemination. The objective was to transfer this technology in cow and compare the migration of fluorescent labeled spermatozoa from the same bull in the genital tracts of cows for both models. A fluorochrome providing stable bright fluorescence staining of the whole spermatozoa without affecting their viability was selected. Fluorescent sperms were inseminated through the cervix and followed in the uterus at 30 min and 3h after AI by insertion of the Cellvizio 1.5mm probe by the same way. The spermatozoa, motile and immotile, could be clearly detected and tracked 30 min after insemination. In contrast, very few spermatozoa were found in the uterine horns 3 hours after AI. The collection of the female genital tract after the slaughter of one cow after imaging revealed that the spermatozoa were absent from the body of the uterus but stored in the highest part of the uterine horns, the utero-tubal junction, which cannot be reached by the probe. Interestingly, inseminating fluorescent beads with sperms allowed to show that beads are rapidly drawn back to the vagina after insemination, evidencing an adverse flow that may also evacuate dead sperms. In conclusion, the method was adapted to observation of uterine cavity in cattle, however, the quick migration of sperms to the uterotubal junction impaired the use of the technique to compare the fate of sperms in the different models.
Embryo transcriptomic: given that oviducts of lactating cows are less able to support early development than oviducts of heifers or dry cows, this task was designed to see if variable maternal environment can modify embryo genome activation expression pattern. Therefore, in vitro fertilized embryos were transferred at day 2 after fertilization into the oviducts of either lactating and nonlactating (dry) Holstein cows and heifers (Metabolic Model) and Holstein heifers with high (F+, n=4) or low (F-, n=4) EBV for fertility compared to Montbéliarde heifers (n=4) as an additional control (Genetic Model). The embryos were recovered at day 4 and investigated by RNA-sequencing (LMU) and compared to in vitro developed ones. Using a unique model of endoscopic embryo transfer coupled with next generation sequencing, this study has demonstrated the influence of metabolic status and estimated breeding value on the evolution of embryonic genome activation in bovine embryos. The effect of metabolic status (heifer vs postpartum lactating and nonlacting cow) was more extreme than that observed for the genetic model where high vs. low estimated breeding values for fertility have only a small effect on the transcriptome embryos that were cultured for two days in their oviducts.
Embryo proteomic: to study the effect of the oviduct on the early embryo, in vitro-produced 4- to 8-cell embryos (Day 2) were transferred into the oviducts of females from the metabolic and genetic models. Two days later, on Day 4 post estrus, the embryos were recovered from the oviducts and were analyzed using a multiplexed “Selected Reaction Monitoring” (SRM) assay, together with D2 embryos and D4 in vitro cultured ones. This multiplexed SRM assay was applied for the relative and absolute quantification of 27 proteins (see deliverable 3.5) in each of the pools of embryos, deriving from the different animals. The proteins were chosen based on previous results obtained in our group, demonstrating that they play an important role and can serve as markers for the different stages in early embryonic development. To absolutely quantify the 27 proteins, two stable isotope-labeled proteotypic peptides (isotopically labeled at incorporated C-terminal Arg or Lys) per protein, i.e. 54 peptides in total, were used as internal standards. Absolute protein concentrations could be determined for 19 proteins in embryos from the metabolic model and 20 proteins in embryos from the genetic model. Proteins could be quantified down to the attomole range per embryo. We observed a clear influence of the embryo environment on individual proteins (in vivo vs. in vitro). We also observed statistically significant differences in protein concentrations between groups of females in both models (GSTM3, HSPB1 and CLU in the metabolic model, and ASTL, KPNA7 and TACC3 in the genetic model).
WP4: Uterus and conceptus
In this work package we addressed the question whether metabolic condition or genetic merit for fertility affects the conceptus-endometrial crosstalk in dairy cattle. This was analyzed by molecular profiling of endometrium and conceptus. In addition, uterine fluid was analysed for proteome changes in a holistic manner.
Previous studies have reported differences in the endometrial response of lactating cows compared to non-lactating cows to the major embryonic embryonic pregnancy recognition signal, interferon tau (IFNT) . Moreover, the endometrium is capable of distinguishing between conceptuses with different developmental trajectories i.e. conceptuses produced by AI compared to in vitro and cloned embryos, and can modify its transcriptomic response accordingly.
Thus, we tested hypothesis that metabolic changes in circulation that are associated with lactation would impede the ability of the uterus to support adequate conceptus elongation and thus the ability the conceptus to provide the appropriate pregnancy recognition signal to endometrium to establish an environment suitable for implantation to begin. In addition, we tested if differences in the estimated breeding index (EBI) for fertility are associated with significant transcriptome changes in the endometrium at the time of implantation. Therefore, endometrium samples from Holstein heifers with high (Fertiplus) or low (Fertiminus) EBI for fertility were investigated by RNA-sequencing, including corresponding samples from Montbéliard heifers as an additional control.
Effects on the endometrium: in total, 256 differentially expressed genes (DEGs) were revealed by comparing endometrial transcriptome profiles of heifers vs. non-lactating cows (FDR 5%, 131 increased and 125 decreased in heifers compared to non-lactating cows, respectively). The comparison of heifers with lactating cows revealed 238 DEGs, 141 of which were increased and 97 were decreased in expression. At a FDR of 10%, 28 DEGs were identified 8 of which increased and 20 decreased in endometria from lactating compared to non-lactating cows on Day 19 of pregnancy. Overall, there was a substantial overlap (92 genes) between the DEGs of the comparisons heifer vs. lactating and heifer vs. non-lactating cows, while there was little overlap found for the DEGs between endometria from lactating and non-lactating cows with the comparisons to the heifer group. Functional annotation clustering showed that among the DEG with lower expression in heifers compared to non-lactating cows functional categories such as ‘vesicle’, ‘signal transduction’, ‘steroid metabolic process’, and ‘carbohydrate binding’ were overrepresented. Categories ‘regulation of cell migration’, ‘cell death’, and ‘oxidoreductase’ were enriched among DEG with higher expression in heifers. Corresponding to the overlap of the DEGs for heifers vs. lactating cows and heifers vs. non-lactating cows, respectively, similar functional categories were found as overrepresented. In addition, DEG with lower expression in heifers included genes with potential transcription factor binding sites for HFH3, FREAC7, and RSRFC4 and ‘cell adhesion’, while DEG with higher expression in heifers were enriched in the functional categories ‘fatty acid metabolic process’, ‘defence response’, and ‘lymphocyte activation’. For the DEGs in endometrium of lactating vs. non-lactating cows, only three overrepresented functional annotation clusters were obtained: ‘response to hormone stimulus’, ‘cell migration’, and genes with potential transcription factor binding sites for IRF7, MYOD, and OCT.
In addition to the holistic approach by RNA sequencing, we analysed the expression of a selection of endometrial genes known to be regulated by the presence of the conceptus (PLET1, SOCS6) and by interferon tau (STAT1, RSAD2, SOCS1, SOCS3) or involved in progesterone (FOXL2, SCARA5) prostaglandin (PTGS2) and oxidative stress (CAT, SOD1, SOD2) molecular pathways by RT-PCR. Variance analyses revealed a significant impact of the maternal metabolic status on FOXL2 (P<0.002; DRY/LACT fold-ratio=2.1) and SOD2 (P<0.04; DRY/LACT fold-ratio=0.8) mRNA expression. Based on these findings, the contribution of FOXL2 to endometrial physiology has been investigated using primary cultures of bovine endometrial cells.
Furthermore, we tested if differences in the estimated breeding index (EBI) for fertility are associated with significant transcriptome changes in the endometrium at the time of implantation. Therefore, endometrium samples from Holstein heifers with high (Fertiplus) or low (Fertiminus) EBI for fertility were investigated by RNA-sequencing, including corresponding samples from Montbéliard heifers as an additional control. Interestingly, the highest number of differentially expressed genes were found in the comparison of Fertiplus and Fertiminus (276 higher in Fertiplus, 418 higher in Fertiminus), while the comparison of these two groups with Montbéliard revealed fewer differentially expressed genes.
In conclusion, this study demonstrated the influence of parity and metabolic stress on the transcriptome of the endometrium independently of the quality of the transferred embryos. Compared to the high fertility uterine environment of heifers, cows exhibited significant transcriptome changes of the endometrium transcriptome at day 19. In addition, a number of transcriptome differences were found between endometrium samples from non-lactating vs. lactating cows suggesting that metabolic stress in the latter group affects gene expression in the endometrium and eventually hampers its ability to respond physiologically to pregnancy recognition signals released by the conceptus. In addition to effect of metabolic condition on gene expression in the endometrium this study showed that high vs. low estimated breeding values for fertility have a clear effect on the endometrium transcriptome at day 19 of pregnancy, which may – at least in part – explain the difference in reproductive performance.
Effects on the uterine fluid: in addition to candidate transcripts, differentially abundant proteins in the uterine fluid of the different recipient groups may serve as biomarkers for fertility. Therefore, this work package included a differential proteome analysis of uterine fluid recovered from the different recipient groups. For the metabolic model (lactating cows, dry cows, heifers) the proteins with the most pronounced abundance differences were:
1. Dry cows vs. heifers
a. More abundant in dry cows: Lysozyme C, milk isozyme; Histone H3.3; Histone H2B; Histone H4 (Fragment); Histone H2A.J
b. Less abundant in dry cows: Ran-specific GTPase-activating protein; CBX3 protein; Prothymosin alpha; Trophoblast Kunitz domain protein 4; Density-regulated protein
2. Dry cows vs. lactating cows
a. More abundant in dry cows: Aminopeptidase N; Superoxide dismutase [Cu-Zn]; Phosphatidylethanolamine-binding protein 4; Ganglioside GM2 activator 2; Phospholipase A(2)
b. Less abundant in dry cows: Apolipoprotein A-IV; Density-regulated protein; Glycerol-3-phosphate dehydrogenase [NAD(+)]; Protein HP-25 homolog 1; Prothymosin alpha
Compared to metabolic condition, high vs. low estimated breeding value for fertility had an even more pronounced effect on the proteome of uterine fluid. Out of 358 proteins showing differences in abundance between Fertiplus and Fertiminus, the ten most prominent were Growth/differentiation factor 8; Tubulointerstitial nephritis antigen-like; Lactoferrin (Fragment); Platelet-activating factor acetylhydrolase; Polymeric immunoglobulin receptor (less abundant in Fertiplus); and NID2 protein; GFPT1 protein; Perilipin; Thymosin beta-10; CALML5 protein (more abundant in Fertiplus).
Consequences for the conceptus: lactation in high yielding dairy cows results in elevated non-esterified fatty acids (NEFA) and beta hydroxy butyrate (BHBA) in response to a deficit between energy intake and energy output as well as decreased circulating concentrations of insulin, insulin-like growth factor 1 (IGF1) and glucose, creating a compromised environment for embryo development.
This study tested the hypothesis that the uterine environment of lactating cows would affect conceptus gene expression. In order to isolate effects of lactation on the conceptus from those on the oocyte and early embryo we used embryo transfer of good quality embryos recovered from nulliparous Holstein heifers. The specific aim of the study was to characterize the transcriptome of the bovine conceptus at the initiation of implantation in a cohort of age-matched postpartum primiparous dairy cows that were either milked postcalving (i.e. lactating) or were dried off immediately at calving (i.e. never milked, non-lactating) as well as a group of contemporaneous heifers.
Principal Component Analysis (PCA) of transcriptomic data revealed moderate separation between the overall expression patterns of conceptuses recovered from lactating cows compared to maiden heifers. PCA of gene expression patterns of conceptuses from non-lactating cows vs. heifers or from non-lactating cows vs. lactating cows revealed no clear separation of these groups. Comparison of conceptus transcriptome profiles from lactating cows vs. heifers revealed 269 DEG, 100 with higher and 169 with lower expression in conceptuses recovered from lactating cows. Gene ontology analysis of these DEGs revealed 37 biological processes falling into six broad categories: (i) the regulation of biological processes; (ii) cellular and metabolic processes; (iii) cell component/cell cycle regulation; (iv) cell death and apoptosis; (v) developmental processes and (vi) mitochondria.
In conclusion, this part of the work package demonstrated the influence of metabolic stress on the transcriptome of grade one embryos. Compared to the high fertility uterine environment of heifers, lactation significantly modifies the transcriptome of a conceptus at day 19. The lack of differences between non-lactating cows and heifers indicated that it is not just an artefact of the source of conceptus i.e. generated by AI compared to superovulation. Moreover, even with no differences in the transcriptome of the conceptus e.g. non-lactating and heifer, here does seem to be differences in the utilization of amino acids (detected by a parallel study of amino acid concentrations in the uterine luminal fluid) i.e. the metabolic activity of the conceptus. It is unclear if these differences in conceptus transcriptome and amino acid utilization are a result of conceptus ‘plasticity’ in response to the lactating environment of have a definite effect or consequence for pregnancy outcome.
Potential practical use of the results:candidate genes identified by expression profiling of endometrium and conceptus provide an interesting source for the identification of genetic polymorphisms (e.g single nucleotide polymorphisms, SNPs) which are associated with fertility and metabolic stability. The candidate genes discovered within this work package can be used by breeding associations for SNP discovery and validation studies as outlined in Ref. [6]. In addition, proteins in uterine fluid of animals with different fertility may serve as putative biomarkers for reproductive status.
WP5: Population genetics
This work package addressed novel approaches to identify candidate recessive lethal mutations, identify variations within key genes involved in early reproductive biology identified in previous WPs. Preliminary results have been validated and tested at the population level: the activity brought to the identification and localisation of putative recessive lethal mutations, the identification of variations in candidate genes identified in earlier WPs and the validation of the effects of the recessive lethal and candidate genes variations on fertility.
To create a list of candidate recessive lethal loci from gene sequence analysis, two complementary approaches were used: 800K SNP genotypes of 1009 progeny tested Italian Holstein sires were analysed to identify putative recessive loci and exome sequences of 20 Italian Holstein sires having extreme values for male and female fertility EBVs. A total of 191 variants in 163 genes were identified as strong candidate deleterious variants for lethal recessive mutations.
Genetic variants, falling within sequences of candidate genes/loci and identified in previous WPs as having an effect on dairy cattle fertility, have been identified. A pipeline of analysis has been realised, to ensure reproducibility, modularity and scalability of the experiment.
First, from an input list of 745 genes 604 have been annotated and positioned on the UMD3.1 genome. Next, polymorphisms falling within candidate regions have been identified in the whole-genome sequence data coming from the 1000 bull genomes project (http://www.1000bullgenomes.com/).
In all the 604 regions, a total count of 758,684 polymorphisms have been identified. The overall distribution of the polymorphisms per gene indicates a very skewed right distribution, as expected.
The Minor Allele Frequency (MAF) for all polymorphisms was calculated. 23,415 variant sites out of 758,684 showed MAF = 0 and were therefore monomorphic. The MAF showed a mean value of 0.059732 (std=0.112234) and a median value of 0.004156.
A genotyping assay for candidate recessive lethal/deleterious polymorphisms coming from the results obtained in deliverable 5.1 and polymorphisms within fertility candidate genes/QTLs has been designed, resulting in a total of 71 potentially tested combinations of variations/haplotypes. Such 71 potential deleterious variants (including variations and haplotypes) were fully assessed to have a complete picture of the genomic effects, their frequencies, the flanking sequences (useful to design the probes in the array), and other variations present in the flanking regions which could difficult the hybridization of the probes. In addition, a further analysis on these 71 variants considering all the information collected was performed to target a reduced subset of the most promising candidates to run the tests on the population. Two variants (DelVar_6 and DelVar_9) were identified as the strongest candidates, considering the complete co-segregation in the population of 17 individuals with both HD and exome data. Candidate variants from D5.2 obtained from whole-genome sequence data were also processed: results are analogous to those for D5.1.
Most promising candidates (DelVar_6 and DelVar_9) were selected for large scale testing in the population in about 5200 Italian Holstein bulls and cows. The frequency of the deleterious alleles in the dataset agreed with the one found in the WES and haplotype data analyses (0.08%). In the WES and SNPchip data we didn’t find homozygotes for the deleterious allele (variant or haplotype). The population data didn’t confirm this behavior: 32 (0.62%) and 37 (0.72%) homozygotes for the deleterious were found in DelVar_6 and DelVar_9, respectively. Also the Hardy-Weinberg equilibrium test on the population data didn’t confirm the haplotype results. The two DelVars resulted in fact in Hardy-Weinberg equilibrium. The p-value of the test was 0.9987 and 0.5435 for DelVar_6 and DelVar_9 respectively. the access to the “1000 bull genome” dataset permitted to use the 450 Holstein animals with WGS as test population for all the 67 variants. Out of the 42 variants with no observed homozygotes in WES, only 18 were confirmed in the 1000 bull genome dataset.
The chi-square test of Hardy-Weinberg equilibrium identified 11 variants with p-value below 0.05. Two of them (DelVar_15 and DelVar_28/29) had a p-value under the threshold of 0.05 corrected for multi-testing (0.00075).
These two variants were further assessed mining the 1000 bull data of other breeds, in particular Brown Swiss (BRO: 105 animals), Simmental type (SIM, 274), Jersey (JER, 66), Angus (ANG, 141) and Hereford (HER, 41). DelVar_28/29 was in Hardy-Weinberg equilibrium in all breeds investigated. DelVar_15 was significantly out of equilibrium in all other breeds investigated and had no homozygotes with the exception of a single one found in Simmental (0.36% of the total SIM dataset) but we can’t exclude a sequencing error or allelic dropout at this nucleotide.
WP 6: Surrogate markers and indicative traits
The WP6 has carried out activities to identify robust biomarkers predictive of the reproductive status of the cow in biological fluids easy to obtain as plasma and vaginal swabs. The activities have been organized in four tasks. The first 3 tasks have been focused on the identification of biomarkers related to fertility status in blood samples collected from animals belonging to WP1;in vaginal swabs; in blood samples and vaginal swabs collected during the negative energy balance (NEB) of cows in their transition phase until the first insemination.
In the last task, the potential markers found in plasma and vaginal fluid as measure of fertility have been validated in a greater population.
The main results obtained in the WP can be summarized in the following points:
Correlated markers for fertility in blood: the purpose of this activity was to discover less invasive biomarkers in blood serum to avoid complex procedures, often invasive, and expensive assays. Blood samples were collected from animals after the insemination up to day 19 for microRNA, proteins and metabolites analysis. Overall, 51 animals of WP1 belonging to tasks 1.2 (Impact of energy balance; n=27) and 1.3 (impact of genetic merit; n=24) have been analysed. Blood samples obtained from task 1.2 belonged to three groups: LACT (n=10), cows that calved normally and had a standard lactation; DRY (n=11), cows that calved normally but dried off immediately after calving; MH (n=6), heifers of breeding age. Blood samples obtained from task 1.3 belonged to three groups: HOLST- (n=8), Holstein heifers with low reproductive indices; HOLST+ (n=8), Holstein heifers with high reproductive indices; MONT (n=8), Montbeliarde heifers breed with a higher level of reproductive success. All animals of both experiments were synchronized with the same procedure, and LACT and DRY show a estrous around 70 days in milk.
The miRNAs in plasma at d19 showed important results in the metabolic model, but not in the genetic model, where only heifers have been used. Some of the miRNAs with different concentrations between high and low fertility groups in the metabolic model are linked with the immune tolerance and the bta-mir140 (up-regulated in pregnant cows) seems a good candidate as an early biomarker of fertility.
The metabolic indices measured at the insemination time in the two models of the project gave the possibility to clarify that: (i) cows dried off immediately after calving or heifers in comparison with lactating cows differed in some functions of the liver (lower concentrations of cholesterol, index of the synthesis of the lipoproteins, GOT, total bilirubin), which reflects a lower metabolic activity of this organ. (ii) pregnant vs non-pregnant cows that follow a standard lactation showed a better liver functionality (higher concentrations of cholesterol and albumin) and lower concentrations of inflammatory markers (positive acute phase proteins: as haptoglobin and ceruloplasmin; globulin); (iii) pregnant vs non-pregnant heifers showed (in both models, but not in HOLST+) lower concentrations of cholesterol and albumin, likely in relationship to the different energy status.
The candidates of fertility emerged from this task include parameters of the liver functionality and inflammation for lactating dairy cows; parameters related to the energy status and markers of long duration of inflammation for heifers.
Biomarkers for fertility in vaginal samples: this task aimed at establishing a sampling procedure of vaginal swabs for the analysis of biomarkers (i.e. metabolites, miRNA) in the vaginal fluid; (ii) setting up methods to measure biomarkers (i.e. metabolites, proteins and miRNA) in samples; (iii) collecting samples from dairy cows in different physiological conditions to correlate the biomarkers measured in the vaginal fluid with measures of fertility and to identify some putative biomarkers to be monitored in a new dedicated experiment (Task 6.4).
The procedure to collect the vaginal fluid has been defined and dedicated methods for the determination of the potential candidates of fertility status in this fluid have been developed. After setting up the analysis for miRNA, possible biochemical markers have been investigated, demonstrating the possibility to detect the following parameters: glucose, NEFA, BHB, total cholesterol, triglycerides, urea, creatinine, AST/GOT, GGT, ALP; lactate dehydrogenase, total bilirubin, Ca, Mg, P, haptoglobin, ceruloplasmin, Zn, globulin, total protein, albumin, paraoxonase, advanced oxidation protein products (AOPP), total antioxidant (ferric ion reducing antioxidant power or FRAP) and thiol groups. Biochemical and miRNA profiles demonstrated a very light presence of molecules in vaginal samples collected from dry cows, while showed a good variability of many molecules in the samples collected at 30 days in milk. These results confirmed the possibility to use these biomarkers as possible candidate of the reproductive traits. To verify this we realized samples of vaginal fluids at 30 days after calving and 19 days after the first insemination in 20 cows submitted to the procedure of estrus synchronization (the same used in the WP1). Within this population, 10 subjects have been selected in accordance to the success (5 cows) or failure (5 cows) of the insemination. The pregnant in comparison to the non pregnant cows showed significant lower concentration of glucose, triglycerides, lactate dehydrogenase, urea, phosphorus and thiol groups and also statistical differences in several miRNA (i.e. bta-mir-2411; bta-mir-150; bta-mir-342; bta-mir-574; bta-mir-98; bta-mir-204; bta-mir-34c; bta-mir-221). Some miRNAs are specific for the high fertile group (bta-mir-451; bta-mir-92a-2; bta-mir-92a-1), others are specific for the hypo fertile group (novel:18_15941; Novel:1_1781; Novel:has-miR-4505; Novel:has-miR-4467; bta-mir-200b; Novel:4_36799) while others are common to both groups (bta-mir-486; Novel:17_14769; bta-mir-215; Novel:9_45060; Novel:13_7606; bta-mir-192; bta-mir-143; bta-mir-126). Nevertheless, none of the significant miRNAs detected in vaginal swabs resulted significant at plasma level. This demonstrates that identified miRNAs are tissue-specific and depending also on the time point analysed.
Markers of NEB in transition cows: for this activity blood samples from the dry period to the first insemination have been collected to assess a wide inflammo-metabolic profile and to select some potential reliable biomarkers of fertility status for a further confirmation in field conditions (WP9). The experiment has been done in a well-controlled herd on 52 subjects (22 primiparous and 30 multiparous). Cows have been classified in accordance to the success (PR, 15 cows, 40% primiparous) or failure (37 cows) of the insemination (AI), done after the estrus synchronization (the same of WP1). Non pregnant cows after the first AI were divided in two further groups: pregnant within 250 days in milk (LatePR, 24 Subjects; 46% primiparous), non pregnant or repeated breeders (e.g. pregnant with more than 3 AI and over 200 days in milk; NoPR, 13 Subjects, 38% primiparous). In this experiment, we have found some biomarkers that statistically differ among fertile (PR) and less fertile (LatePR) or infertile (NoPR) cows.
The most promising plasma biomarkers resulted: (i) indexes of the prolonged effects of the inflammation (high concentrations of ceruloplasmin and globulin, lower level of the ratio albumin/globulin) during the transition period and also around the first AI; (ii) some negative acute phase proteins, during the periparturient period and in some cases also around the first AI. The more fertile cows showed higher concentration of cholesterol (index of lipoproteins), retinol (index of Retinol Binding Protein), total bilirubin (index of enzymes necessary to their clearance); (iii) some molecules with anti-inflammatory effects, as tocopherol and beta-carotenes, which remained higher in more fertile cows after calving and around the first AI. In opposite, the concentration of the total antioxidants and the indexes of the oxidative stress (both ROM and AOPP) do not show significant differences among the groups; (iv) concentrations of energy balance indicators (NEFA and BOHB, respectively at two or four weeks after calving) which showed the lowest concentration in more fertile cows in comparison to less fertile or infertile (but not at the time of the first AI); (v) concentrations of phosphatase alkaline and some minerals (mainly calcium and phosphorous), both around the calving and around the time of first AI. In particular, the lower concentrations of plasma phosphatase alkaline and phosphorus and the higher one of calcium are correlated with conditions of worsened fertility. All these biomarkers appear good candidates for the detection of cows with poor fertility in field conditions and seem able to anticipate the identification of cows that need specific treatments to restore the normal physiology of the reproductive apparatus.
Validation: to validate previous results we collected vaginal swabs after the calving, from the same cows involved in task 6.3 experiment in order to select the most reliable biomarkers of fertility status for a further confirmation in field conditions (WP9). On the same 52 subjects enrolled in task 6.3 two vaginal swabs were taken at 28 days in milk and 19th days after the AI for the measurements of parameters identified in previous activities. Among the wide biochemical profile investigated, the most promising biomarkers of fertility status in the vaginal swabs collected at 30 days in milk include indexes of negative energy (cholesterol, NEFA and BOHB) and protein (urea and creatinine) balance, indices of mineral (Ca, P, Mg, Zn) metabolism, and indices of the inflammation (globulin, ceruloplasmin, total bilirubin), all lower in fertile vs less fertile cows. The indices of oxidative stress do not seem so important in this context. Otherwise, only two biomarkers in vaginal swabs collected at 19 days after first AI (around 90 days in milk) have the potential to identify subjects with low fertility, and in particular zinc and phosphatase alkaline, with lower or higher concentrations respectively in fertile vs less fertile cows. In the whole, these information appear to have a practical meaning when the checks occur at 30 days in milk. For this reason the investigation to confirm the results obtained in this WP and planned in the WP9 on an higher number of cows belonging to more herds, should be realized only around 30 days in milk.
WP7 Reproductive Biotechnologies
With the recent development of animal genomics cattle breeding companies are in the process of designing new breeding schemes by combining reproductive biotechnologies with genomic information. Their objective is to make selection decisions based on genomic estimated breeding values (GEBVs), as they are becoming rapidly available through widespread genotyping of cattle herds, rather than having to wait for phenotypic data from progeny and performance testing. GEBVs integrate genomic and pedigree information and can be calculated very early in the life of the animals, well before they reach the reproductive age. For this reason the potential of taking full advantage of genomic selection relies on reproductive biotechnologies capable of deriving progeny from young heifers and bulls. Therefore the objective of this workpackage has been the improvement of techniques for assisted reproduction taking advantage of the novel findings derived from the animal models developed in previous workpackages and testing novel hypothesis reported in recent literature in order to improve oocyte and embryo developmental competence.
Improving in vitro culture
To improve in vitro bovine embryo culture conditions the specific effect of oviduct factors on early embryo gene expression were tested. Therefore we compared the transcriptomes of embryos developed in vitro either in co-culture with Bovine Oviduct Epithelial Cells (BOEC) (Cordova, Perreau et al. 2014; Schmaltz-Panneau, Cordova et al. 2014) or with another feeder cell type also known to improve bovine embryo early development the kidney epithelial VERO cells (Menck, Guyader-Joly et al. 1997). Because co-culture with BOEC implies to use a high oxygen concentration (20%), we also compared embryo gene expression in 20% or 5% oxygen. We used the new custom 60K bovine microarray developed during the first part of FECUND project. This array displays 26 055 transcripts (corresponding to 23 926 genes) which represents 97.4% of the Ensembl annotated transcripts. Embryonic gene expression was analysed at two different stages (16-cell and Day 8 blastocyst stages). While high level of oxygen induced a significant decrease of blastocyst proportion (but did not change cleavage rates), no impact of oxygen level was observed on embryo transcriptome at both developmental stages. No significant difference of cleavage, 16-cell and blastocyst rates was observed in presence of feeder cells when compared to 5% of oxygen. No impact of feeder cell type was observed on bovine early development kinetic and rates. Regarding to embryonic gene expression, very few genes were directly impacted by the feeder cell types. However, comparing each co-culture system to culture in absence of any feeder revealed greater proportions of differentially expressed genes. Nevertheless, no significant difference in impacted biological pathways identified by Ingenuity Pathway Analysis Software was observed. These regulated pathways are mainly related to cell cycle, free radical scavenging and glucose and lipid metabolisms. Thus, it seems that there is no major impact of feeder cell types (BOEC or VERO) on bovine early development in terms of kinetic, developmental rate and transcriptome.
After investigation the dialogue between the developing embryo and bovine oviduct epithelial cells (BOEC) in an in vitro co culture model showing a beneficial effect of the cells on development rate and quality (see previous report), and considering that many genes involved in vesicles transport were influenced by energy balance in COC and oviduct, we searched for the presence of exosomes in oviduct fluid and in BOEC culture medium. Extracellular sesicles were found in both cases and were analysed for proteic contents, showing that these vesicles are able to transport protein known to be present in oviduct fluid (ex OVGP1). We showed that exosomes are able to cross the zona pellucida and to be internalized by embryo cells and the addition of in vivo exosomes to the embryo culture medium (SOF) had a positive effect on embryo development rate and quality, showing that the effect of BOEC on development may be at least partially mediated by extracellular vesicles.
Another aspect addressed in the project was to investigate the different metabolic requirement of oocytes derived from young heifers versus adult cows followed by the field trial corresponding to deliverable 7.5. The study was carried out in a period of approximately 1 year and during this period the animals were allocated to the 2 groups considering their age at the time of oocyte collection. For this purpose a database for data recording was constructed that included the date of birth of each animal in order allocate the embryo production data accordingly, meaning that the same animals were at first in group 1 (OPU performed when age was below 12 months) and later they were moved in group 2 (OPU performed when age was above 12 months). In total 392 OPU session were performed on over 50 donor animals, divided in the 2 age groups. The number of embryos obtained was 480 of which 359 were transferred in recipient females giving rise to 177 pregnancies. A significant difference was observed between the low pyruvate and high pyruvate medium only in the group of donors older than 12 months and only for the percentage of embryo calculated on the fertilized oocytes (14.94 vs 23.21 chi square test p<0.05). A general tendency for lower pregnancy rate was observed in the young age group (heifers aged less than 12 months) versus the older age groups. Therefore the indication from this experiment is that most likely the culture requirements for donors aged more than 12 months are different from those of younger heifers and that increasing the concentration of pyruvate in the culture media is beneficial only for older donors.
Improvement of oocyte competence in vitro: in this activity we compared two different in vitro maturation systems testing the working hypothesis that the developmental competence of the oocytes can be increased by a period of prematuration during which the oocytes are blocked in meiotic arrest by exposure to cAMP. Resumption of meiosis occurs after a drop in oocyte cAMP levels, which also regulate gap junctional communication between the oocyte an the cumulus cells. It is thought that cAMP levels are regulated with two different mechanisms: the diffusion through gap junctions and the inhibition of its degradation. The diffusion of cAMP from the granulosa and cumulus cells to the oocyte and the consequent increase of cAMP levels prevent oocyte maturation. In the oocyte the degradation of cAMP is catalysed by the specific type 3 phosphodiesterase (PD3A) which is inhibited by cyclic Guanosine Monophosphate (cGMP). Several studies demonstrated that somatic cell derived cGMP enters the oocyte through the gap junctions, thus inhibiting PDE3A and maintaining meiotic arrest. On the background of this knowledge the modulation of intracellular cAMP amount was obtained in this study acting on three different molecular targets through the use of three different meiotic inhibitors: 3-isobutyl-1-methylxanthine (IBMX), forskolin (FSK) and cilostamide. FSK raises the levels of cAMP by stimulating adenylate cyclase activity.IBMX is a methylated xanthine derivate, which acts both as a competitive nonselective PDE inhibitor and as a nonselective adenosine receptor antagonist. The effect of the exposure of COCs to IBMX is the raise of intracellular cAMP and the subsequent activation of PKAs in the cumulus cells and in the oocyte. Cilostamide is a specific inhibitor of PDE3 and is used to selectively modulate cAMP levels in the oocyte. The aim of the present study was to develop a defined in vitro maturation system and to compare two different protocols for pre-maturation of oocytes both modulating cAMP levels. The assessment of the experimental conditions was done first evaluating kinetics and maturation rate and secondly following fertilization evaluating embryo development to blastocyst. Two different prematuration treatments were compared with two control groups. Maturation kinetics was affected by the treatment indicating a delay in the experimental groups up to 24h of maturation while at 30h of maturation no differences were observed. A similar trend was observed during embryo development where the prematuration groups were developing at slightly lower pace in the fist days after fertilization but at the blastocyst stage, 7 days after IVF, no difference was aobserved between the groups. This finding indicate that a prematuration strategy could be used successfully in a practical commercial context to provide more flexibility in the field operation and allowing to plan the fertilization timing according to the field and laboratory logistics.
Validation of culture supplements that influence methylation:Oocyte maturation process includes the establishment of proper epigenetics marks, fundamental to ensure successful pregnancy. Epigenetic maturation of the oocyte depends on one-carbon-metabolism (OCM), whose key elements (cobalamin, folate) are crucial cofactors for the transfer of methyl groups onto chromatin. Commercially available IVM-media only partially supports oocyte metabolic requirements, and thus may negatively affect epigenetic maturation and further development. Of relevance, cobalamin, one of the OCM cofactors normally present in follicular fluid, is missing in IVM-media. Aim of this work was to investigate if cobalamin supplementation of IVM media affects sheep oocyte developmental competence in term of subsequent embryo development and epigenetic pattern. Data demonstrated that Cobalamin supplementation improves oocytes’ competence and partially corrects epigenetic defects (DNMTs expression, genome-wide methylation profile) in pre-implantation embryos and placental tissues at 20-days of pregnancy. Moreover, early placental vascularization, in terms of vessels maturity and vasculogenetic factors’ expression, was improved by the treatment. Data have provided useful information for the improvement of in vitro maturation condition and epigenetic competence of IVP embryos.
WP8: Bioinformatics and statistical analyses
This work package provided the bioinformatic structure to manage the data produced and the tools for primary data analysis. Analyses were carried out by combining and mining data to identify candidate genetic loci, putative biomarkers and novel traits that can be used to predict fertility and to examine the contributions of genetic and metabolism to fertility problems. Shortly, a project database to manage and store project data has been created, bioinformatics tools and pipelines were used to analyse and mine data across WPs and thus investigate genetic and metabolic impact on fertility.
First, a Fecund database and genome browser was put in place to store and allow user-friendly retrieval of information from the other WPs. Bioinformatics pipelines and tools were made available via a bioinformatics platform. In particular software tools for mass spectrometry based protein identification (i.e MASCOT, Andromeda), for quantitative proteomics (Scaffold, MaxQuant), data interpretation (Cytoscape, ClueGO/CluePedia), RNA-Seq and Methyl-Seq data analysis (Trimmomatic, TopHat2, Bismark, ...) Metabolomics (Agilent Chemstation MSD) enabled cutting-edge approaches to process the genomics and proteomics data and ensured high quality analyses and outputs.
Using phenotypic and genotypic data of Montbéliarde and Holstein cows, a genome-wide association study for milk production and cow fertility has been performed. On the same line of thoughts, from the GWAS results we have identified SNP with a significant effect on both fertility and milk production. These SNP were then tested for association in another cattle populations comprising Holstein cows from dairy herds in Northern Italy. In total, 16 associations from the GWAS on French Holsteins and Montbeliarde cattle were validated in an independent Italian Holstein population. These indicate SNP and related genomic regions with a signifcant effect on both milk production and cow fertility. The next step was to pursue these signals further in a functional analysis (gene ontologies and metabolic pathways) to uncover genes and biological functions behind fertility and milk production in dairy cattle.
Differentially expressed genes from transcriptomics experiments in WPs 2, 3 and 4 have been used to explore the gene ontologies and molecular pathways implicated in milk production and fertility. Gene expression was studied in five tissues related to the reproductive function: the cumulus cells, the endometrium, the oocytes, the oviduct and the embryo. A bioinformatics pipeline to automate and standardise gene functional analysis (analysis of gene ontologies and metabolic pathways) was developed for the purpose. The pipeline comprises several steps related to two major groups:
•gene ontololgies (GO terms) + gene set enrichment (wherever possible)
•pathway analysis (metabolic pathways)
Additionally, there were steps for conversion of gene IDs, names and symbols (e.g. to go from SwissProt proteins codes to gene symbols) and for plotting/summarizing results. The pathway comprises scripts in R, Bash and Python, plus using external software when needed. The following biological databases are queried to retrieve ontologies and pathways: Ensembl www.ensembl.org/, NCBI (www.ncbi.nlm.nih.gov) KEGG www.genome.jp/kegg/, Reactome (www.reactome.org/). A list of 4,416 differentially expressed genes was obtained from experiments in WPs 2, 3 and 4. The Gene Ontologies that were retrieved from biological databases are categorized in three groups: biological processes (BP), cell compartments (CC) and molecular functions (MF). The most represented GO terms, for each category, associated to the genes detected in Fecund's experiments were summarized in D8.4. Metabolic pathways that were found to be associated to fertility have been also investigated and reported.
WP9: Field application of validated biomarkers
FECUND has produced a vast amount of information that have added considerably to knowledge of early reproduction. The objectives of the WP 9 were to evaluate the data and technological advances achieved in earlier WPs and to develop or test direct applications that can be used by the industry in the short term to start reversing the negative reproductive trend. The activities performed, results obtained and perspectives open in this WP are summarized below.
Validation of putative biomarkers of Negative Energy Balance and fertility outcomes: candidate biomarkers have been identified in WP6 to predict, in early post-partum, the severity of the NEB (D6.3) and reproductive outcomes at the 1st AI (D6.4). The objective of the tasks 9.1 and 9.2 was to validate the NEB and fertility biomarkers on a higher number of dairy cows in early post-partum. This kind of biomarkers could be useful for farmers for the optimisation of their animal’s management and also to define potential new phenotypes for breeding companies in order to refine actual traits.
Plasma samples have been collected on 155 cows of different breeds, in early post-partum, at 3±1 and 28±2 days in milk (DIM). Vaginal swabs have been collected at 28±2 DIM for 52 out of previously cited females. Body Condition Score (BCS) was registered at each time point (note from 0 to 5). The biochemical determination in plasma samples included the list established in the D6.4 and according to the procedures reported in the D6.3. The biochemical determination in vaginal swabs included the list of the putative biomarkers identified in D6.4 according to the procedures reported in the same document. Complementary data have been extracted from the French national database containing all the milking and reproduction information for each animal.
Cows have been grouped in subjects with low, moderate or severe NEB in accordance to the value of the BCS variation between 3±1 and 28±2 DIM: “low NEB”, “moderate NEB” and “severe NEB” respectively for a variation of 0 (no variation of BCS during this 1st month of lactation); [0;-0.5] and [-0.5;-1.0]. Moreover, cows have been classified in “Pregnant AI1” or “Not pregnant AI1” at the 1st AI based on the non-return in heat 90 days after the 1st AI post-partum. A cow present in the herd 90 days after the 1st AI and not re-inseminated was classified in the “Pregnant AI1” group whereas a cows inseminated twice or more was classified in “not Pregnant AI1”. All the cows not inseminated or inseminated once within a delay shorter than 90 days post 1st AI have been classified in “Not determinable” group and excluded from the analysis.
It was observed that the variation of BCS during the 1st month of lactation can reflect the severity of NEB which significantly impact the fertility success of the 1st AI with 70.5% of pregnancy rate for “Low NEB” females; 52.2% for “Moderate NEB” females and 16.7% for “Severe NEB” females respectively. Thus, this parameter could be considered as a good predictive marker of fertility outcomes but can be difficult to measure in routine on the field without specific monitoring devices.
Blood and vaginal biomarkers were assessed in this trial to have an indirect measure of the severity of NEB of each female. Even if it could be difficult to imagine a massive and systematic collection of vaginal swabs on post-partum females, the measurements made on that fluid highlighted some markers (Urea, Total protein, Globulins, AST/GOT, GGT, Haptoglobin and LDH) significantly different according to the NEB severity groups of females. Easier to obtain, measurements in blood highlighted some others biomarkers (Sodium, Zinc, Cholesterol, Ceruloplasmin, Albumins, total Bilirubin). Finally, the synthetic criteria Liver Functionality Index (LFI) developed by Bertoni and Trevisi is significantly different between females with low, moderate or severe NEB and therefore represents a good biomarker of NEB severity. As biomarkers the determination of a threshold to identify post-partum females with severe NEB with the maximum accuracy is still needed. In this trial, the low number of females presenting a severe NEB doesn’t allow to perform this kind of analysis. Thus, further experiments are still needed to go towards on field application allowing farmers to identify critical females and adapt their management.
Concerning the validation of fertility biomarkers, few of the metabolic parameters suggested in the D6.3 (plasma) and D6.4 (vaginal swabs) have been confirmed (GGT and Thiols groups in plasma samples; Triglycerides in vaginal swabs) in this experiment. Moreover, LFI doesn’t seem to be a good predictive marker of the fertility taking into account that LFI values are not different between Pregnant and Not Pregnant females. Fertility is a very complex trait which is probably affected by a numerous of other parameter not taken into account in this trial like the genetic background of the animals, the effect of the male, the management of the animals by the famers (oestrus detection method, insemination delay, feeding management).
Impacts of genetic improvement tools: in this task, the LFI trait was analysed together with milk production in a genomic selection perspective to assess the achievable genetic gain for novel fertility traits in dairy cattle. A sample of 96 Holstein-Friesian cows were genotyped at high density for 777,962 SNP, with an average call-rate of 99.65%. The minimum per-sample call-rate was 0.88: therefore, no animal was removed because of low-call rate. The minimum per SNP call-rate was 0: 3,561 SNPs with call-rate < 85% were discarded (~0.5%). The residual missing SNP genotypes (still about 150,000 missing data points) were imputed. After imputation, 1747 SNPs with duplicate or unassigned position, and 136,241 monomorphic SNPs were removed, leaving 636,413 available for subsequent analyses. As phenotypes, milk production at 28 DIM and the LFI were measured on the animals. A linear mixed model approach was used for BLUP estimation of genetic parameters and (genomic) breeding values. The covariance structure between observations was modelled building upon genetic similarities.
A negative genetic correlation between LFI and milk yield was estimated from this data, while the phenotypic correlations is practically zero. The predictive abilities which represents the accuracy of genomic predictions for the two traits have been calculated (0.608 and 0.723 respectively for milk production and LFI). The interest is to breed for cows that produce milk while keeping functional (i.e. maintain a healthy reproductive performance). Therefore, metabolically efficient milk production and reproductive performance are the broad breeding goal, which is not easy to measure on individual cows. LFI and milk yield can be used instead as breeding criteria related with functionality/reproduction (the former) and production (the latter). The relationships of the breeding criteria with the breeding goal and between them, are all relevant for the construction of the selection index for the improvement of the breeding goal. Thus, the expected response to selection can be estimated through the well-known ”breeder's equation”. Relevant terms for the response to selection are the intensity of selection, the accuracy of the selection index and the genetic standard deviation of the trait. Different scenarios were simulated, varying the selection intensity from 10% to 25%, and the relative emphasis placed on LFI and milk production (from 1/2 to 2/1). The results show, as expected, that response to selection is higher when the selection intensity is higher. Also, when higher emphasis is placed on one of the two components, the response to selection is faster compared to a more balanced objective where both components are improved simultaneously. This also is expected and well known: placing higher emphasis on one trait gives faster and higher gains, compared to the slower more balanced selection.
Validation of biomarkers of oocyte and embryo quality: in this task, two embryo culture media with low and high pyruvate concentration was evaluated on the development of in vitro produced embryos of two different breeds and origin: embryos produced from Holstein donors up to 12 months of age and older than 12 months (data reported in D7.5) and embryos produced from adult cows of the Chianina breed. The choice of the 2 breeds has been made taking into consideration also the objective of testing the protocols on animals that were considerably different in their physiological status. This study was carried out in a period of approximately 2 years and included 18 Chianina oocyte donor cows and 10 different Chianina bulls were used for IVF. The animals were subjected to the Ovum Pick Up procedure (recovery of oocytes by ultrasound guided follicular aspiration) at intervals of approximately 1 to 3 weeks. No hormonal treatments were applied on donors’ cows at any time in order not to interfere with the ovarian physiology. The oocytes were in vitro fertilized and the derived zygotes of each cow were cultured alternatively in low and high pyruvate in order to test the different culture protocols on all the donors included in the study. In total 104 and 139 OPU sessions were performed on 18 donor Chianina cows and the oocytes retrieved from the OPU sessions were divided in the 2 culture groups, low pyruvate and high pyruvate. The number of oocytes recovered was 1607 and 2219 and after IVF 1015 and 1565 fertilized zygotes respectively were obtained. In total 438 and 887 embryos were obtained from the 2 culture groups. The embryos were evaluated on the basis of the criteria of the International Embryo Transfer Society being G1 and G2 grade the embryos of highest quality (G1: more than 85% of the embryonic mass is viable; G2 more than 50% of the embryonic mass is viable) and G3 grade the embryos of lowest quality (G3: less than 50% of the embryonic mass is viable). The total percentage of embryos developed was calculated on the number of fertilized zygotes and resulted significantly higher for the high pyruvate group compared to the low pyruvate group (56,68% vs 43,15%; chi square test, p<0,05).
Therefore the results of this field study confirm the findings reported in D7.5 because the increase of pyruvate supplementation during culture has proven beneficial for embryo development both in Holstein and Chianina cows. In conclusion, the transfer of the protocols developed in WP7 to WP9 has been successful and represents an improvement in the results that can be obtained applying assisted reproductive technologies, such as Ovum Pick Up, IVF and embryo culture, under field conditions.
Validation of recessive lethal mutation previously identified on French bovine population: in this task, it was investigated the effect on fertility and on embryonic lethality in French populations of 71 candidate lethal recessive mutations previously identified (D5.3). These variants were not available on the routinely used beadchip and imputation to whole genome sequence data using 1000 bull genome population was needed. None of the variants was found to have a significant effect on fertility traits in French population but lack of observed effect may be due to imputation errors. Therefore, we investigated if candidate mutation located either close to an identified QTL or close to a region displaying deficit in homozygote. Out of the 71 candidate mutations, 34 met at least one of the criteria and three fulfilled the two conditions. Further investigations have to be performed in order to validate candidate mutation. Our main option is to include them in our genotype assay and search for a deficit in homozygous at SNP level. Lack of homozygous at SNP level will confirm lethality of the mutation. Analyses can also be redone with future run of 1000 genome. Increase in total number of sequenced animal should lead to higher imputation accuracy and therefore a more powerful QTL detection.
The reproductive performance is one of the major issues in dairy cattle farming and breeding. After the decline of the conception rate at first service, the inclusion of fertility in dairy selection programs has brought about an inversion of the negative trend: in the last 5-10 years, the decline in fertility has stopped, and some initial signs of recovery begin to be noticed. Still, despite the encouraging progress, a lot of work remains to be done, and this involves both basic and applied research. This workpackage integrated 5 tasks aiming to improve this key traits by (1) developing some new biomarkers of NEB and fertility and simulating their integration into breeding schemes; (2) improving in vitro embryo production protocols to increase the intensity of selection and optimize the genetic gain and (3) investigating some putative recessive lethal mutations to avoid part of embryonic losses. Some of these attempts have been successful and others a little bit less but, in any case, additional knowledge has been acquired, useful for the understanding of the complex fertility trait which still remain a great challenge for the breeding industry.
WP 10 Project Dissemination
The www.fecund-project.eu domain has been registered. A WordPress website has been created, populated with project information, and adjusted with a graphical template specific for the FECUND project. The website was created and maintained by PTP presenting project information, consortium details, news and events updates and the electronic versions of the project communication materials. The website contains a direct link to the website of the European project PROLIFIC, closely related with FECUND.
Two project leaflets were produced and distributed both as printed and digital copies. Both leaflets were designed and printed as a double-fold A4 document (size 21x29.7 cm). The two leaflets have similar designs and the main layout and final texts were decided by successive improvements of originally proposed drafts with the Project Management Committee (PMC). The layout is based on the fonts and colours of the FECUND logo. FECUND leaflet 1 provides information on the project background and challenges in dairy cattle fertility and reproduction, the interdisciplinary approach that FECUND is based on, objectives, outputs and potential applications, partnership with PROLIFIC and partners. It was first printed as 1,000 copies in M17. A second and adjusted version was printed as another 1,000 copies in M31. The leaflet 1 was widely distributed whenever relevant, e.g. for scientific events, meetings with stakeholders or mass media both as printed and digital versions via the project partners during the course of the project. Leaflet 1 is available at FECUND Project Website digitally as a high-quality pdf suitable for printing, http://www.fecund-project.eu/?page_id=12.
FECUND leaflet 2 provides information on the project background, key innovations, and impacts followed by the project results grouped as the “genetic model”, “metabolic model” and the “high technologies”. It was published in M45 and 200 copies were printed for the distribution via the partners, during FECUND PROLIFIC Joint Final Conference, other scientific events, meetings with stakeholders. The digital version was disseminated through the different social media channels during the project. The leaflet text was adjusted at the end of the project at M48. Leaflet 2 is available at FECUND Project Website digitally as a high-quality pdf suitable for printing, http://www.fecund-project.eu/?page_id=12 .
Four external project newsletters were designed and published electronically to approximately 400 stakeholders during the course of the project. It provided information on project news, events, results, and publications. The newsletters are available at FECUND Project Website as a high-quality pdf suitable for printing, http://www.fecund-project.eu/ .
Newsletters were sent to a number of different stakeholder groups consisting of academics, breeding industry, farmer representatives, national bodies and policy makers selected from EFFAB and FABRE-TP mailing lists which contain many of the interested industry and academic organisations. The newsletters were also distributed to other EU projects active in the same area such as PROLIFIC. The newsletters were also promoted through EFFAB and FABRE-TP social accounts (Twitter, LinkedIn). The Newsletter mailing list was created to be an active one by enabling subscriptions from the website.
First of the stakeholder events was a joint FECUND-PROLIFIC final conference organised at M45 targeting the scientific community. It was organised jointly with the related EU Project PROLIFIC and the aim of the joint conference was to integrate the results of the two projects to move improvements for fertility in cattle forward. The Joint Final Conference was organised as a satellite workshop to the 20th Annual European Society for Domestic Animal Reproduction (ESDAR) Conference 2016, held on October 30, 2016 in Lisbon, Portugal. The title of the workshop was “Innovations to improve dairy cow fertility”.
The conference was promoted through the project website and EFFAB communicated the event both on EFFAB and FABRE TP website and social media accounts (Twitter and LinkedIn). CONAFE (La Confederación de Asociaciones de Frisona Española) and ESDAR have also communicated about the conference to their network. The main promotion material was the dedicated flyer for the conference designed by EFFAB and disseminated by all parties involved in promoting the event. Fecund partners have also promoted the conference within their networks. The promotion has continued during the conference with live tweeting and sharing updates from EFFAB and FABRE TP social media accounts. A total of 34 tweets with nearly 5,000 views were shared during the conference.The Conference participation was 63 people from 22 countries (16 EU countries). There was a total of 13 speakers, 5 from FECUND, 6 from PROLIFIC, the EU Project Officer and a keynote speaker. All the abstracts and presentations made during the Joint FECUND/PROLIFIC Final Conference are accessible from FECUND (http://www.fecund-project.eu/?p=281) and PROLIFIC (http://www.euprolific.eu/) websites.
The second event was a live webinar broadcasted at M48. The aim of the webinar was to disseminate and communicate the results of the researches carried out in scope of the FECUND project to facilitate the knowledge and technology transfer to the stakeholders, in particular the breeding industry, end-users and the knowledge institutes for future research possibilities. The FECUND Stakeholder Webinar was held on January 17, 2017 using GoToWebinar software. The title of the Webinar was “Improving dairy cattle fertility: Results from the FECUND Project”. Several promotion materials were designed by EFFAB including a Save-the-Date flyer and a webinar program flyer. The FECUND stakeholder webinar was promoted both by EFFAB, PTP (as coordinator) and the FECUND project partners. It was promoted through the project website and EFFAB communicated the event both on EFFAB and FABRE TP website and social media accounts (Twitter and LinkedIn). A mailing list consisting of selected breeding industry, policy makers and knowledge institute representatives (more than 1,500 in total) was created by EFFAB. ATF (Animal Task Force) and EAAP (European Federation of Animal Science) have also communicated the webinar to their network and promoted the event on their websites and newsletters. The participation was 63 people from 20 countries (16 EU countries). There was a total of 5 speakers from FECUND project. All the presentations made during the FECUND Stakeholder Webinar and the webcast of the webinar are accessible from FECUND (http://www.fecund-project.eu/?page_id=258) website. The webcast is on EFFAB YouTube Channel (https://www.youtube.com/watch?v=vaoqkjBT5FM) and has been promoted by EFFAB, FABRE TP and FAO. The webinar recording has already been viewed 215 times so far and is still being viewed.
Potential Impact:
The FECUND project helped to reduce negative fertility trends in dairy cattle and hence promote economically sustainable and ethical livestock production, by creating new knowledge to define specific biological processes associated with early reproductive function. There was still only rudimentary knowledge of the early developmental events around and shortly after conception affecting embryo developmental competence, implantation and the establishment of pregnancy. FECUND used advanced molecular techniques to dissect the interaction between the mother and the embryo during the early phases of development and start defining the signals and responses. A major impact of FECUND was to create new physiological knowledge to unravel the complexity of communication between the mother and the embryo that is necessary in early reproductive phases for establishing a pregnancy.
With increasing milk production many cows are unable to maintain energy balance during the transition period as they establish their lactation, when they have to mobilise body reserves to compensate for the shortfall in dietary intake of energy. Some high producing cows are able to cope with the transition period with little impact on fertility, while others do not cope. FECUND investigated the biological impacts of Negative Energy Balance on early reproduction from oocyte and early embryonic development to the responses of the mother during early reproductive events and by investigating postpartum cows at peak lactation vs postpartum non-lactating cows vs heifers to create knowledge on reproductive physiology. The new knowledge obtained from reproductive physiology provided new phenotypes that will define a specific biological processes associated with early reproductive function. These new traits could be used in genetic selection schemes for reproductive traits. FECUND built on the knowledge created to search for surrogate traits and correlated biomarkers to predict the novel traits, that can be assayed in easily sampled material, such as blood or vaginal swabs. These surrogate markers will be more easily applied in the field to collect sufficient data and hence construct accurate estimated breeding values for the new traits, which can then be used in selection programmes.
Fecund has identified biomarkers that will predict animals that respond badly during the transition period which would enable farmers to identify cows at high risk of developing energy deficiency and to devise improved strategies to avoid impaired fertility. Biomarkers for energy balance risk have been identified and could also be used to produce genomic Estimated Breeding Values for fertility with higher accuracy for genetic selection and will enable farmers to devise strategies to address the energy deficiency for high risk cows. This would give the owners of such data a significant competitive edge.
Sperm survival and vitality, which is essential for good fertility, is likely to be influenced by the female reproductive environment. FECUND investigated the influence of factors in the uterus and oviduct on sperm vitality. Such information will have an impact for farmers, or the insemination services allowing them develop strategies to improve the female environment, e.g. by prior treatment, or by addition of beneficial factors identified by FECUND at, or shortly before, insemination. A major impact of FECUND for management will be the development of biomarkers for sub-optimal female reproductive tract environment at AI.
Embryo based reproductive techniques are being increasingly used to improve the rate of genetic gain. FECUND has integrated the knowledge created on early development to improve in vitro reproductive protocols. A major impact of FECUND for assisted reproduction is the improvement of culture techniques and protocols to produce high quality oocytes and embryos for transfer. Coupling with advanced breeding methods with genetic markers will select the best embryos for transfer and hence achieve even more rapid genetic progress. FECUND provided knowledge to optimise assisted reproduction protocols for the in vitro production of embryos. Genotyping embryos led to the identification of high genetic merit candidate embryos from which to select. Using this technology will also contribute to the preservation of genetic diversity. In addition deriving oocytes from young animal by in vitro maturation then in vitro fertilisation will reduce generation time. FECUND integrated the knowledge created on the molecular genetics and genomics, epigenetics, proteomic and metabolomics of early development to improve in vitro reproductive protocols. A major impact of FECUND for assisted reproduction is the improvement of culture techniques and protocols to produce high quality oocytes and embryos for transfer.
Outputs from the FECUND will have an impact at two levels: 1) those that can be implemented immediately by individual organisations or associations to achieve a competitive advantage for them and their clients or members, such as genetic markers, and 2) those that will need a collective approach at national and international levels to bring about benefits, such as improved trait recording. Collectively, these impacts contributed to the sustainability of farming and hence rural economies and employment, and will ensure food security with reduced cost in Europe and beyond.
Immediate applications from FECUND include validated genetic markers for fertility traits (e.g. lethal recessives) which will be immediately available for genotyping and application in gene assisted selection of bulls and cows by the breeding companies and the national associations. To apply project results to improved genetic selection, the outputs need to be implemented at national and international levels. To achieve this, national and international breeding associations and international structures have been involved. Before implementation of new recording protocols a cost/benefit analysis would be required (as carried out in WP9) and the mode of introducing the novel trait recording scheme established using agreed international standard procedures for the data collection, recording and exchange.
Fecund partners ensured the widest dissemination of the project results to stakeholders, scientists, end-users and farmers by means of several focused activities: an updated project website for the promotion of results and events, 2 dedicated workshops addressed to scientists and stakholders, targeted newsletters and printed leaflets.
List of Websites:
The website address is the following: http://www.fecund-project.eu/
Relevant contact details - Fecund Coordinator:
Dr. Filippo Biscarini
Fondazione Parco Tecnologico Padano
via Einstein - località Cascina Codazza
26900, Lodi (LO), Italy
Tel: +39 3407499754
e-mail: filippo.biscarini@ptp.it