REPRODUCTIONREVIEW

Meta-analysis of expression profiles in granulosa cells during

Daulat Raheem Khan1, Éric Fournier1, Isabelle Dufort1, François J Richard1, Jaswant Singh2 and Marc-André Sirard1

1Centre de Recherche en Biologie de la , Département des Sciences Animales, Faculté des sciences de l’agriculture et de l’alimentation, Université Laval, Quebec City, Québec, , and 2Department of Veterinary Biomedical Sciences, Western College of Veterinary Medicine, University of Saskatchewan, Saskatoon, Saskatchewan, Canada Correspondence should be addressed to M-A Sirard; Email: [email protected]

Abstract

Folliculogenesis involves coordinated profound changes in different follicular compartments and significant modifications of their gene expression patterns, particularly in granulosa cells. Huge datasets have accumulated from the analyses of transcriptomic signatures in predefined physiological contexts using different technological platforms. However, no comprehensive overview of folliculogenesis is available. This would require integration of datasets from numerous individual studies. A prerequisite for such integration would be the use of comparable platforms and experimental conditions. The EmbryoGENE program was created to study bovine granulosa cell transcriptomics under different physiological conditions using the same platform. Based on the data thus generated so far, we present here an interactive web interface called GranulosaIMAGE (Integrative Meta-Analysis of Gene Expression), which provides dynamic expression profiles of any gene of interest and all isoforms thereof in granulosa cells at different stages of folliculogenesis. GranulosaIMAGE features two kinds of expression profiles: gene expression kinetics during bovine folliculogenesis from small (6 mm) to pre-ovulatory follicles under different hormonal and physiological conditions and expression profiles of granulosa cells of dominant follicles from post-partum cows in different metabolic states. This article provides selected examples of expression patterns along with suggestions for users to access and generate their own patterns using GranulosaIMAGE. The possibility of analysing gene expression dynamics during the late stages of folliculogenesis in a mono-ovulatory species such as bovine should provide a new and enriched perspective on ovarian physiology. Reproduction (2016) 151 R103–R110

Introduction dominant follicle, whereas the remaining follicles The is a highly dynamic structure, of which undergo atresia (Lussier et al. 1987). the principal functional unit is the follicle. In foetal In contrast to other somatic tissues, granulosa cells , primordial germ cells proliferate during the during folliculogenesis undergo very dynamic and first trimester of gestation and develop into primordial highly coordinated changes. During the late stages follicles by mid-gestation. A primordial follicle is of folliculogenesis, the changes accelerate in all typically 30–40 µm in diameter and each is composed of compartments of the follicle wall (granulosa, cumulus, a partially differentiated (arrested in prophase-1 and theca cells, vascular and inter-cellular stromal of ) enclosed by one layer of specialized somatic components), culminating in the release of a competent cells called follicular or granulosa cells. Further follicle oocyte and the formation of a new tissue called the development begins before birth as small cohorts of . In the developing follicle, acquisition primordial follicles undergo progressive growth and of oocyte competence involves interplay between atresia until puberty. Folliculogenesis progresses in the a multitude of intrinsic and extrinsic factors, which ovary, leading to the formation of a fluid-filled all act to bring about rapid development of distinct cavity called the antrum and the emergence of a highly gene expression profiles in different follicular cells specialized type of granulosa cell called cumulus cells, (Wigglesworth et al. 2014, Khan et al. 2015). This is which are in direct contact with the oocyte (Gougeon particularly apparent in granulosa cells (Sirard 2014), 1996). In mono-ovulatory species such as cattle, one in which gene expression patterns are important not ovule per reproductive cycle is released from a single only for the and luteinization processes but

© 2016 Society for Reproduction and DOI: 10.1530/REP-15-0594 ISSN 1470–1626 (paper) 1741–7899 (online) Online version via www.reproduction-online.org Downloaded from Bioscientifica.com at 09/27/2021 01:44:50PM via free access

10.1530/REP-150594 R104 D R Khan and others also for the developmental competence of the oocyte (Wigglesworth et al. 2015), meta-analysis of different contained therein (Assidi et al. 2008, Hamel et al. 2010). ovarian transcriptomic studies remains scarce. An Interestingly, FSH has been implicated in acquisition of online public collection called the ovarian kaleidoscope oocyte developmental competence, both in vivo and in database or OKdb (http://okdb.appliedbioinfo.net/) vitro (Sirard et al. 2007). In bovine, in vivo experiments provides information on gene expression in different have shown that ovarian super-stimulation with a FSH ovarian cell types and their association with various support for 5 days (endogenous FSH following removal ovarian functions (Hsueh & Rauch 2012). However, of dominant follicle (for 2 days) followed by 3 days of a chronological/dynamic interface of folliculogenesis FSH injections twice a day) followed by no FSH period based on integrated ovarian cell gene expression profiles (called coasting) for 44–68 h yields the best oocyte has yet to be constructed. The principal obstacles quality for subsequent development of (Nivet to achieving this are incomparable technological et al. 2012). The study of granulosa cell transcriptome platforms and experimental conditions in the different dynamics in different physiological contexts therefore studies (Tseng et al. 2012). The integration of such remains crucial to understand the physiology of ovarian studies requires vast knowledge of ovarian physiology tissue as a whole. combined with highly specialized bioinformatics skills. Conventional experimental designs do not provide Based on the availability of several publicly the overall perspective that is essential in order to available transcriptomic analyses generated on a single understand follicular dynamics. The huge amounts technological platform called “EmbryoGENE”, an online of data that have accumulated remains scattered in interface called GranulosaIMAGE (Granulosa Integrative database repositories and require integration and meta- Meta-Analysis of Gene Expression) has been developed. analysis in order to chart overall gene dynamics in this GranulosaIMAGE provides easy consultation of the tissue. The regular manuscript format allows sharing temporal kinetics of gene expression during follicular of 1–2% of the data analysed (i.e. highlighted ), development from small-diameter (>5 mm) follicles to and access to supplemental data, although possible, is pre-ovulatory in different physiological contexts, along difficult to re-analyse. Other than a recent comparison of with dominant follicle gene expression profiles for cumulus and mural granulosa cell transcriptomes in mice various post-partum time intervals and metabolic states.

Figure 1 Summary of GranulosaIMAGE workflow. Data from transcriptome studies conducted by EmbryoGENE network scientists on granulosa cells have been deposited in the ELMA database. Granulosa cells were obtained from follicles at various stages of folliculogenesis from cows in different physiological and metabolic states (top panel). This flow diagram provides the working model of data retrieval from ELMA, its normalization, statistical tests and graphical representation on the GranulosaIMAGE web-based resource.

Reproduction (2016) 151 R103–R110 www.reproduction-online.org

Downloaded from Bioscientifica.com at 09/27/2021 01:44:50PM via free access Gene expression profiles of granulosa cells R105

Programme description and methods Statistical methods GranulosaIMAGE provides a view of the dynamics Data including intensity files and sample annotation of bovine genes and their isoforms by integrating (metadata) from previous studies of granulosa cells 74 microarray datasets generated using the using the EmbryoGENE microarray platform (Robert EmbryoGENE platform (Robert et al. 2011) and a et al. 2011) have been filed in the EmbryoGENE LIMS uniform analysis pipeline. Although these studies and Microarray Analysis (ELMA) database (Robert et al. were conducted independently, they cover collec- 2011). These data are used here to perform meta-analysis tively most stages of development and to generate expression profiles. The working model (Fig. 1). GranulosaIMAGE (http://emb-bioinfo.fsaa. of GranulosaIMAGE is summarized in Fig. 1. Intensity ulaval.ca/granulosaIMAGE/) was thus generated values for any probe (along with their associated as a web resource for easy consultation of the metadata) are retrieved from the ELMA database and relative dynamics of practically any gene in bovine after logarithmic transformation (base 2) of the raw granulosa cells. measured intensity; normalization of the intensity values is performed by subtraction of the background threshold. Studies included The background threshold is defined as the mean of the intensities of the negative control spots on an array The studies included in GranulosaIMAGE database plus twice the standard deviation of these intensities. are summarized in Table 1, which provides principally The resulting relative intensities are then quantile- two types of granulosa gene expression pattern for normalized using the limma bioconductor package each gene. (Ritchie et al. 2015) and plotted on the y-axis for each 1. Time-course gene kinetics during follicular condition. These normalized intensity values are used to development from the small (6 mm) to pre-ovulatory produce the colour scale applied to the different points stage under different hormonal or physiological (the higher the intensity, the darker the point appears) conditions. showing the expression level under each condition. 2. Gene expression in granulosa cells recovered post- The levels of significant difference for each probe partum from dominant follicles of cows synchronized were determined using ad hoc statistical tests comparing using prostaglandins (52 h after synchronization, specific conditions within each experiment. We used for uniformity across samples and to precede the the t-test for experiments with two conditions and endogenous LH surge). a fixed-effect ANOVA for experiments with three or

Table 1 The studies included in GranulosaIMAGE.

Reference Group Title of article Douville & Sirard (2014) 6–9 mm follicles (follicular state effect) Changes in granulosa cells gene expression associated with growth, plateau and atretic phases in medium bovine follicles Girard et al. (2015) >9 mm follicles (follicular state effect) Global gene expression in granulosa cells of growing, plateau and atretic dominant follicles in cattle In vitro culture (FSH effect) Unpublished Guillemin et al. (2015) expression level A genetical genomics methodology to identify genetic markers of a bovine fertility phenotype based on CYP19A1 gene expression Dominant follicle (age effect) Unpublished Nivet et al. (2012) Dominant follicle (coasting effect) FSH withdrawal improves developmental competence of in the bovine model Gilbert et al. (2012) Dominant follicle (time to LH surge) Impact of the LH surge on granulosa cell transcript levels as markers of oocyte developmental competence in cattle GSE69247 Pre-ovulatory follicle (24 h post-LH Age effect) Unpublished Dias et al. (2013a,b) Super-stimulated pre-ovulatory follicle Effect of duration of the growing phase of ovulatory follicles on (24 h post-LH) oocyte competence in super-stimulated cattle Differential gene expression of granulosa cells after ovarian super-stimulation in beef cattle Golini et al. (2014) Dominant follicle (post-partum period effect) Transcriptome analysis of bovine granulosa cells of pre-ovulatory follicles harvested 30, 60, 90 and 120 days post-partum Girard et al. (2015) Dominant follicle (energy balance effect at The effect of energy balance on the transcriptome of bovine 60 ± 5 days post-partum) granulosa cells at 60 days post-partum Dominant follicle (pre-LH surge effect of Unpublished vitamin B9 & B12) Gagnon et al. (2015) Dominant follicle (vitamin B9 and B12 effect Effects of intramuscular administration of folic acid and vitamin B12 irrespective of LH surge) on granulosa cells gene expression in post-partum dairy cows www.reproduction-online.org Reproduction (2016) 151 R103–R110

Downloaded from Bioscientifica.com at 09/27/2021 01:44:50PM via free access R106 D R Khan and others

Figure 2 GranulosaIMAGE homepage. This screenshot illustrates the organization of the homepage containing a brief introduction to the database along with the gene search bar into which the sought probe ID or gene symbol is entered. This page also shows the manner in which a gene search result is displayed and lists the statistical methods used and the publications from which results are included in the database. External links to EmbryoGENE microarray annotation files and the ELMA database are also provided. more conditions. All P-values were then corrected to whether the probe is in the constitutive region or a unified FDR (false discovery rate). In the resulting some untranslated region (UTR) of the genome. graphs, the comparisons with corrected P-values <0.05 Examples: We provide two examples. First, by are highlighted in red. The statistical tests do not take submitting gene symbol AREG (amphiregulin), the into consideration array effects, dye effects, technical results summary shows one probe (EMBV3_31722) replicates, or secondary experimental factors such as that is of constitutive type. Secondly, looking individual animals. For further details, the reference for MAML2 (mastermind-like 2) returns), three articles should be consulted where the above effects probes are listed (EMBV3_10398, EMBV3_29461 have been considered for each study separately. and EMBV3_34221), the second of which is constitutive, whereas the other two are the alternative 3′-UTR sites. Data visualization 2. Probe details: By selecting the probe ID, the search is GranulosaIMAGE charts the intensity profiles of a given directed to the detailed results of the corresponding gene in granulosa cells under different conditions. The probe. A short table summarizes the properties of the GranulosaIMAGE homepage (Fig. 2) provides a brief probe previously mentioned (probe ID, gene symbol description of the use of the programme along with an and probe type) along with gene names and options introduction to the gene search results panel, links to for external links to additional details on the gene of a help section, a description of the statistical methods interest and its sequence information. used and a list of published results included in the Example: Selecting the single probe listed for AREG programme, and the search bar on which the sought (EMBV3_31722) directs the user to additional gene probe ID or gene symbol is entered. A link to the information including the name amphiregulin. EmbryoGENE microarray annotation file Robert( et al. Furthermore, the user may choose to consult external 2011) is provided for consultation of probe IDs with the links via ‘Entrez Gene’ or ‘Refseq’. corresponding gene names, isoforms and symbols. This information becomes very helpful when a GranulosaIMAGE displays the results page in response query returns the probes corresponding to more to submission of a gene symbol. The results page contains than one gene. For example, the entry ADAMTS1 several sets of information, which we present along with (ADAM metallopeptidase with thrombospondin examples later in this article. type 1 motif, 1) returns 12 probes, of which only one (EMBV3_20771) corresponds specifically 1. Results summary: The summary of results contains to ADAMTS1. The others correspond to genes the numbers and IDs of the probes corresponding with similar names or symbols. This is useful as to the entered keyword and also indicates the GranulosaIMAGE thus provides information about symbol of the gene corresponding to each probe and the gene families as well as different gene isoforms.

Reproduction (2016) 151 R103–R110 www.reproduction-online.org

Downloaded from Bioscientifica.com at 09/27/2021 01:44:50PM via free access Gene expression profiles of granulosa cells R107

3. Expression graphs: Normalized intensity values of each condition refer to the mean relative intensity of any gene are displayed graphically. These graphics the probe for this condition. The panels highlighted are divided into two compartments: “follicles over in pink indicate that an ad hoc statistical test on time” showing gene dynamics from the small to pre- the given group revealed a significant difference ovulatory follicle stages in different physiological (P < 0.05). contexts and second dominant follicle gene Examples: AREG is a thoroughly studied gene involved expression patterns in the post-partum period. in ovarian physiology (Conti et al. 2006). The graphical In these graphics, each column contains data from representation of AREG shows that FSH and LH affect a particular experiment (named at the top), whereas expression of this gene (Fig. 3A). FSH stimulation the specific conditions for that experiment are during in vitro granulosa culture produced significant mentioned at the bottom. The red line (at 0) shows differences in AREG intensity. Also, when granulosa the background level. An array value below the were obtained from follicles undergoing FSH coasting red line indicates that the intensity value is smaller in vivo, significant differences inAREG levels were than the background or in other words the gene is observed. Likewise, AREG intensity levels were not expressed under those conditions. These values greatly increased in granulosa cells early (6 h) after the have not been removed from the analysis, in order to LH surge compared with 2 h before or 22 h after. This include all the conditions in an experiment regardless information confirms previous findings published in of gene expression. The small horizontal lines for the literature (Conti et al. 2006, Sugimura et al. 2014,

Figure 3 Integrative meta-analysis in GranulosaIMAGE: Expression dynamics of (A) amphiregulin (AREG) and (B) ADAM metallopeptidase with thrombospondin type 1 motif, 1 (ADAMTS1) under different conditions. Significant changes in gene expression profiles are highlighted in pink. (See the expression graphs section for explanation). www.reproduction-online.org Reproduction (2016) 151 R103–R110

Downloaded from Bioscientifica.com at 09/27/2021 01:44:50PM via free access R108 D R Khan and others

Khan et al. 2015). We note that additional information Construction of customized profiles on AREG dynamics in other contrasts, such as follicular GranulosaIMAGE indicates relative gene expression in developmental phase (growth, plateau and atresia), terms of mRNA abundance under different conditions animal age, ovarian super-stimulation and post- of folliculogenesis. This manner of presenting expression partum period effects, is seldom described in studies of AREG. Although AREG does not change significantly kinetics provides a basis for the construction of in such contrasts, grouping these data in one location customized profiles by users in the form of simpler adds considerable value to GranulosaIMAGE and illustrations. We present here two examples: (1) temporal contributes to understanding the overall physiology of kinetics during different physiological states (growth, ovarian tissue in all these contexts. plateau and atresia), dominance and relevance to LH The example of ADAMTS1 further illustrates this surge (Fig. 4) and (2) gene dynamics at different intervals notion (Fig. 3B). This gene is related closely to and energy states during the post-partum period (Fig. 5). the of extra-membranous domains and activation of AREG in granulosa cells in response to 1. Gene dynamics during folliculogenesis: First, we the LH surge (Sayasith et al. 2013). GranulosaIMAGE present the genes with significant roles in ovarian shows that the dynamics of its response to FSH physiology (e.g. AREG and ADAMTS1) and the stimulation and LH surge are very similar to those FSH and LH receptors (FSHR and LHCGR). The of AREG. GranulosaIMAGE provides the additional abundance of AREG, FSHR and LHCGR transcripts information that expression of this gene is increased remains unchanged from the growth to plateau significantly during follicular growth and atresia in stages, whereas the transition to atresia induces small (6 mm) but not large follicles (>9 mm), and down-regulation of LHCGR (in follicles 6–9 mm in more in pre-ovulatory follicles of older than younger diameter), whereas FSHR and AREG are unaffected. cows. This observation of the effect of age on the An interesting continuous decrease in ADAMTS1 transcript abundance of a gene that is usually studied expression is noted during all of these phases. in association with ovulatory response and post-LH The illustration shows that the dominance stage is surge cumulus expansion is very interesting. characterized by a relatively greater abundance of

Figure 4 Illustration of gene dynamics during folliculogenesis as represented in GranulosaIMAGE. Dynamics of four genes (FSHR, LHCGR, AREG and ADAMTS1) were consulted initially in GranulosaIMAGE and then represented graphically vs follicular developmental stage, either destined to ovulate (grey line) or undergoing atresia (green line). The x-axis values are arbitrary units. Users may consult the dynamics of any gene of interest and may illustrate these versus their preferred follicular parameters (development, super-stimulation and effect of age or various post-partum metabolic states).

Reproduction (2016) 151 R103–R110 www.reproduction-online.org

Downloaded from Bioscientifica.com at 09/27/2021 01:44:50PM via free access Gene expression profiles of granulosa cells R109

Figure 5 Illustration of gene dynamics during the post-partum period as represented in GranulosaIMAGE. The dynamics of four genes (CCNB1, LHCGR, FOXO1 and PTX3) were consulted in GranulosaIMAGE and then drawn to illustrate the effects of (A) 30, 60, 90 and 120 days post- partum and (B) maternal energy status (BHB level) and vitamin supplementation (pre-LH and irrespective of LH surge) at 60 ± 5 days post- partum. The x-axis values are arbitrary units. Users may consult the dynamics of any gene of interest and illustrate these versus a variety of follicular conditions of interest.

LHCGR and a drop in ADAMTS1. The next major In addition, the effects of energy status (BHB level) shift in gene expression profiles is due to the LH (Fig. 5A) and vitamin supplementation are analysed surge. LHCGR transcripts are very abundant 2 h at approximately 60 ± 5 days post-partum (Fig. 5B). before the surge and drop continuously during the Here, we illustrate the expression dynamics of cyclin post-LH period. A transient drop in FSHR mRNA is beta-1 (CCNB1), LH receptors (LHCGR), forkhead seen in the period spanning 2 h before the surge to box O-1 (FOXO1) and pentraxin (PTX3). Transcript 6 h after, followed by an increase measured 22 h after abundance shows that LHCGR and FOXO1 activities the LH surge. We note with interest that AREG and increase significantly from 30 to 60 days post-partum, ADAMTS1 transcripts mirror FSHR expression, with whereas CCNB1 and PTX3 remain unchanged. We transient up-regulation 6 h after the LH surge. note that none of these genes was subject to maternal Illustrations such as Figs 4 and 5 are helpful for energy status (BHB level) 60 days post-partum. At this following ovarian physiological response kinetics in time point, increased expression of LHCGR and PTX3 a variety of different contexts. Known phenomena decreased CCNB1 and no effects on FOXO1 are such as the loss of LHCGR during atresia (Webb et al. observed in response to vitamin supplementation. 2003) suggest the arrest of follicular growth and a shift The transitions from days 60 to 90 and then days 90 towards follicular atresia. Likewise, the acquisition to 120 post-partum denote a steady decline in the of LHCGR during dominance and a further increase expression levels of LHCGR and FOXO1 and relative before the LH surge suggest tight regulation of these up-regulation of PTX3, whereas CCNB1 remains events during the pre-ovulatory phase (Webb et al. unchanged. 2003). Furthermore, the expression of ADAMTS1 and The most striking feature of this illustration is the AREG relative to LHCGR affirms their roles in inducing down-regulation of CCNB1 and up-regulation of cumulus expansion and ovulation following the LH LHCGR at 60 days post-partum (pre-LH group) in surge. Presentation of gene dynamics in relationship response to vitamin supplementation. In fact, CCNB1 with follicular development allows the exploration is involved in cell cycle regulation. Its down-regulation of time-course changes in ovarian physiology and in this group indicates a relatively more differentiated hence the generation of new hypotheses. cell state in which LH receptors are more abundant. 2. Gene dynamics during the post-partum period: Our This diagram supports the conclusions drawn by database contains studies covering granulosa gene Gagnon et al. (2015), suggesting that vitamin supple­ expression profiles from 30 to 120 days post-partum. mentation alters post-partum follicular dynamics. www.reproduction-online.org Reproduction (2016) 151 R103–R110

Downloaded from Bioscientifica.com at 09/27/2021 01:44:50PM via free access R110 D R Khan and others

Conclusions Gagnon A, Khan DR, Sirard MA, Girard CL, Laforest JP & Richard FJ 2015 Effects of intramuscular administration of folic acid and vitamin B12 on The ovarian follicle is a remarkable structure having granulosa cells gene expression in postpartum dairy cows. Journal of diverse functions and highly complex and dynamic Dairy Science 98 7797–7809. (doi:10.3168/jds.2015-9623) Gougeon A 1996 Regulation of ovarian follicular development in primates: physiology. Understanding ovarian physiology is facts and hypotheses. Endocrine Reviews 17 121–155. essential in order to optimize female fertility, and huge Hamel M, Dufort I, Robert C, Leveille MC, Leader A & Sirard MA 2010 amounts of data on ovarian tissues therefore have Genomic assessment of follicular marker genes as predictors for IVF. Molecular 16 87–96. been generated. However, these data are scattered in Hsueh AJ & Rauch R 2012 Ovarian Kaleidoscope database: ten databases that are difficult to dig in, creating a need for years and beyond. Biology of Reproduction 86 192. (doi:10.1095/ novel ways of integrating and presenting the information biolreprod.112.099127) for the purpose of advancing knowledge in this field. We Khan DR, Guillemette C, Sirard MA & Richard FJ 2015 Characterization of FSH signalling networks in bovine cumulus cells: a perspective on present here GranulosaIMAGE, a web-based interface oocyte competence acquisition. Molecular Human Reproduction 21 that provides gene expression profiles of granulosa cells 688–701. (doi:10.1093/molehr/gap079) from a new perspective. It is an interactive, easy-to- Lussier JG, Matton P & Dufour JJ 1987 Growth rates of follicles in the access resource for researchers in the field of ovarian ovary of the cow. Journal of Reproduction and Fertility 81 301–307. (doi:10.1530/jrf.0.0810301) physiology. This is the first step towards integration Nivet AL, Bunel A, Labrecque R, Belanger J, Vigneault C, Blondin P & of various time points of interest in the reproductive Sirard MA 2012 FSH withdrawal improves developmental competence of cycle. For the moment, GranulosaIMAGE presents oocytes in the bovine model. Reproduction 143 165–171. (doi:10.1530/ REP-11-0391) only the transcriptomic data that have been produced Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W & Smyth GK 2015 using EmbryoGENE platform. Although due to intricate limma powers differential expression analyses for RNA-sequencing and technical constraints GranulosaIMAGE does not include microarray studies. Nucleic Acids Research 43 e47. (doi:10.1093/nar/ the data produced by various groups using microarray gkv007) Robert C, Nieminen J, Dufort I, Gagne D, Grant JR, Cagnone G, Plourde D, platforms other than the EmbryoGENE, we look forward Nivet AL, Fournier E, Paquet E et al. 2011 Combining resources to to include RNAseq data in this tool that could be filtered obtain a comprehensive survey of the bovine transcriptome in a relatively more homogenous manner. It also provides through deep sequencing and microarrays. Molecular Reproduction and Development 78 651–664. (doi:10.1002/mrd.21364) a preliminary basis for comparing different follicular Sayasith K, Lussier J & Sirois J 2013 Molecular characterization and tissues such as theca cells, cumulus cells, and oocytes, transcriptional regulation of a disintegrin and metalloproteinase with which are also becoming increasingly available. thrombospondin motif 1 (ADAMTS1) in bovine preovulatory follicles. Endocrinology 154 2857–2869. (doi:10.1210/en.2013-1140) Sirard MA 2014 Toward building the cow folliculome. Animal Reproduction Science 149 90–97. (doi:10.1016/j.anireprosci.2014.06.025) Declaration of interest Sirard MA, Desrosier S & Assidi M 2007 In vivo and in vitro effects of The authors declare that there is no conflict of interest that FSH on oocyte maturation and developmental competence. Theriogenology 68 (Supplement 1) S71–S76. (doi:10.1016/j. could be perceived as prejudicing the impartiality of the theriogenology.2007.05.053) research reported. Sugimura S, Ritter LJ, Sutton-McDowall ML, Mottershead DG, Thompson JG & Gilchrist RB 2014 Amphiregulin co-operates with morphogenetic 15 to increase bovine oocyte developmental competence: effects on gap junction-mediated metabolite supply. Funding Molecular Human Reproduction 20 499–513. (doi:10.1093/molehr/ This study was funded by Natural Science and Engineering gau013) Tseng GC, Ghosh D & Feingold E 2012 Comprehensive literature review Research Council of Canada (NSERC) as part of EmbryoGENE and statistical considerations for microarray meta-analysis. Nucleic Network and was conducted in collaboration with Boviteq Inc. Acids Research 40 3785–3799. (doi:10.1093/nar/gkr1265) Webb R, Nicholas B, Gong JG, Campbell BK, Gutierrez CG, Garverick HA & Armstrong DG 2003 Mechanisms regulating follicular development Acknowledgements and selection of the dominant follicle. Reproduction Supplement 61 71–90. The authors acknowledge all the researchers who participated Wigglesworth K, Lee KB, Emori C, Sugiura K & Eppig JJ 2014 Transcriptomic in the EmbryoGENE program and whose work has served to diversification of developing cumulus and mural granulosa cells in mouse ovarian follicles. Biology of Reproduction 5 23. (doi:10.1095/ construct GranulosaIMAGE. biolreprod.114.121756) Wigglesworth K, Lee KB, Emori C, Sugiura K & Eppig JJ 2015 Transcriptomic diversification of developing cumulus and mural granulosa cells in References mouse ovarian follicles. Biology of Reproduction 92 23. (doi:10.1095/ biolreprod.114.121756) Assidi M, Dufort I, Ali A, Hamel M, Algriany O, Dielemann S & Sirard MA 2008 Identification of potential markers of oocyte competence expressed in bovine cumulus cells matured with follicle-stimulating and/or phorbol myristate acetate in vitro. Biology of Reproduction 79 Received 16 December 2015 209–222. (doi:10.1095/biolreprod.108.067686) First decision 5 February 2016 Conti M, Hsieh M, Park JY & Su YQ 2006 Role of the epidermal growth factor network in ovarian follicles. Molecular Endocrinology 20 Revised manuscript received 18 February 2016 715–723. (doi:10.1210/me.2005-0185) Accepted 15 March 2016

Reproduction (2016) 151 R103–R110 www.reproduction-online.org

Downloaded from Bioscientifica.com at 09/27/2021 01:44:50PM via free access