Human Reproduction Update, Vol.10, No.3 pp.227-239, 2004
© European Society of Human Reproduction and Embryology 2004; all rights reserved
Characterization of the ovarian transcriptome through the use of differential analysis of gene expression methodologies
Division of Reproductive Sciences, Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR 97006 and Department of Obstetrics & Gynecology, Oregon Health & Science University, Portland, OR 97239, USA 1 To whom correspondence should be addressed at: Division of Reproductive Sciences, Oregon National Primate Research Center, Oregon Health & Science University, 505 NW 185th Ave, Beaverton, Oregon 97006, USA. e-mail: henneboj{at}ohsu.edu
 |
Abstract |
|---|
 TOP Abstract IntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
Prior to the development of high-throughput methods for the analysis of differential gene expression, genes required for proper ovarian function were identified on a case-by-case basis. Recently, however, several techniques have been developed that enable investigators to study large-scale changes in gene expression under a variety of experimental conditions. The utilization of these methodologies has led to the identification of a number of novel or previously unappreciated genes that are expressed within distinct cell types in the ovary or at specific stages of the ovarian cycle. This review details the recent use of differential analysis strategies in identifying (i) genes that are expressed exclusively or preferentially in the ovary, (ii) genes that are differentially expressed in isolated ovarian cells in response to hormonal stimulation, and (iii) those genes that are expressed at specific stages of the ovarian cycle. The genes identified through the use of these approaches represent potential targets for designing agents capable of regulating ovarian physiology and thus fertility.
Key words: differential gene expression/ovary
|
Introduction |
|---|
|
TOPAbstractIntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
Steroid and peptide hormone signalling cascades are critical for normal follicular growth, ovulation, and corpus luteum formation. It is well established that the pituitary-derived gonadotrophins, FSH and LH, are required for the initiation of gene expression cascades necessary for ovarian function (Richards et al., 1988
; Richards, 1994). Intrafollicular mediators also play a role in normal ovarian processes and examples include basic fibroblast growth factor (FGF2) (Nilsson et al., 2001), insulin-like growth factors (IGF) (Adashi, 1998), Kit ligand (KITL) (Besmer et al., 1993; Huang et al., 1993), as well as members of the transforming growth factor-ß (TGFß) family (Dong et al., 1996; Durlinger et al., 1999; Liu et al., 1999; Vitt et al., 2000; Lee et al., 2001; Otsuka et al., 2001; Yan et al., 2001; Findlay et al., 2002; Nilsson et al., 2003) (Figure 1). Each of these intraovarian factors, identified on a case-by-case basis, regulates different aspects of follicular development, ovulation, and luteinization.
|
Recent high-throughput genomic approaches have aided efforts to identify novel genes involved in different aspects of ovarian biology. Techniques that allow for the identification of differentially expressed genes have been developed mainly within the last decade. Such techniques, including differential display (DD) RT–PCR, suppression subtractive hybridization (SSH), serial analysis of gene expression (SAGE), and DNA microarray, have been used to identify genes expressed in reproductive tissues under a variety of conditions. The completion of genome sequencing projects, including the human and mouse genomes, has also allowed for the use of in silico or computational means to identify novel genes involved in reproduction. The focus of this review, therefore, is to highlight recent genomic approaches that have led to the discovery of novel or previously unappreciated genes which are expressed selectively within specific cell types that comprise the ovary or at different stages of the ovarian cycle.
|
Ovary/oocyte-selective genes |
|---|
|
TOPAbstractIntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
Prior to the development of modern differential analysis techniques, genes that are expressed selectively in the ovary were identified individually. Examples include TGFß family members (Gdf9 and Bmp15) (McGrath et al., 1995
; Dube et al., 1998); factor in the germline-
(Figla) (Soyal et al., 2000); FSH receptor (Fshr) (Simoni et al., 1997); NACHT, leucine rich repeat and PYD containing 5 (Nalp5; also known as maternal antigen that embryos require or MATER) (Tong et al., 2000); Mos (Propst et al., 1985); as well as the zona pellucida glycoproteins 1–3 (Zp1-3) (Rankin et al., 1996). The critical role these genes play in ovarian physiology and fertility was subsequently established through the generation of individual null mutant mice. From these studies, it was clear that those genes expressed selectively or perhaps exclusively in a specific tissue are likely required for the normal functioning of that organ system. For example, the ovaries of Fshr null mutant mice lack follicular development past the early antral stage of folliculogenesis (Abel et al., 2000). These mice are very similar in phenotype to the FSHß (Fshb) null mutants, indicating that FSH is not required for follicular development in mice up until the antral stage of folliculogenesis (Kumar et al., 1997). Gdf9 null mutants exhibit a block in the development of follicles past the primary stage of folliculogenesis (Dong et al., 1996). Deleting Figla leads to the loss of expression of several important oocyte genes including the Zp1–3 genes (Soyal et al., 2000) and also to the absence of primordial follicles (Soyal et al., 2000). Deletion of the serine/threonine kinase Mos results in oocyte parthenogenesis due to a failure to arrest at the appropriate stages of meiosis (Colledge et al., 1994). Taken together, the results from such studies point to the importance of identifying and characterizing genes that are expressed selectively in the ovary.
With this in mind, the differential analysis technique SSH was used to identify genes which are expressed exclusively or selectively in the mouse ovary and which may therefore be critical components of ovarian function (Hennebold et al., 2000). Genes common to brain, heart, lung, liver, spleen, kidney, muscle and ovaries at the unstimulated, follicular, ovulatory and luteal phases of a simulated estrous cycle were effectively removed by SSH. Of the 340 non-redundant clones isolated by SSH, 83 cDNA were determined by BLASTn analysis to represent putative novel genes. All 83 novel cDNA were subsequently analysed with respect to their selective expression in the ovary (via northern blots containing RNA isolated from 12 non-ovarian tissues including testis), as well as regarding their expression through the course of a simulated estrous cycle induced in immature mice. From this secondary screen, it was determined that 21 of the 83 cDNA were expressed exclusively (at the level of northern blot sensitivity) in the ovary. Of the 21 novel, ovary-selective cDNA, 17 were differentially expressed through the course of a simulated estrous cycle. It was this group of 17 novel cDNA that were the subject of further studies regarding their role in ovarian biology.
Since the SSH technique yields cDNA fragments, full-length sequences were obtained by either rapid amplification of cDNA ends (RACE) or through the screening of an ovarian cDNA library. To date, the full-length cDNA sequence has been obtained for several of the 17 novel cDNA. From the full-length sequences, the coding regions were identified and used to generate the putative amino acid sequences of the corresponding proteins, which were subjected to further in silico analysis to identify potential functional domains. As expected, the identified domains suggest that the proteins are involved in a wide range of biological activities (Table I). Based on domain analysis, these proteins possess functions that range from unknown (steroidogenic tissue acidic mitochondrial protein or Stamp) to potential roles in chromatin compaction or remodelling (H1foo) (Hennebold et al., 2000).
|
H1foo is the best characterized gene product identified from the ovary-selective cDNA library (Tanaka et al., 2001
). Further analysis revealed that the product of the H1foo gene contains several elements that place it firmly in the family of linker type (H1) histones. Protein BLAST analysis revealed mouse H1foo shares a high level of homology between the linker histone B4 of Xenopus laevis and cs-H1 of the sea urchin Parechinus miliaris (Tanaka et al, 2003). Linker histones in general possess a three domain structure that includes a central globular domain thought to be important for binding to the linker DNA that separates nucleosomes, which are formed from the interaction of DNA with an octamer consisting of core histones (H2A, H2B, H3 and H4) (Hayes et al., 1991; Ramakrishnan, 1997; Georgel et al., 2001). The globular domain of the linker histones (including H1foo) lies in the middle of the protein and is flanked by a stretch of amino acids that comprise unordered N-terminal and C-terminal tails. Alignment of mouse H1foo with B4 and cs-H1 revealed an overall identity of 54 and 52% within the globular domain respectively (Tanaka et al., 2001).
In situ hybridization studies (Tanaka et al., 2001) localized the expression of H1foo mRNA exclusively to the oocyte. H1foo mRNA could be detected at all stages of follicular growth with higher levels evident in prophase I (PI) arrested oocytes. Expression of H1foo was absent in all other cell types of the ovary including granulosa and theca cells. These findings were extended to the regulation of H1FOO at the protein level by subsequent immunohistochemical analysis using an antibody that was generated from a peptide corresponding to the globular region of the protein. Western blot and immunofluorescence analysis of H1FOO protein during meiosis revealed that the highest levels of expression occur at the metaphase II (MII) stage of meiosis in mouse oocytes. H1FOO clearly associates with condensed chromosomes of the MII oocyte and the first polar body. Following fertilization, there was a progressive decrease in mouse H1FOO protein levels in the newly formed zygote. By the 4-cell stage of embryogenesis, H1FOO expression was undetectable (Tanaka et al., 2001).
Prior to the completion of oocyte maturation, about the time it becomes competent to resume meiosis, transcription is inhibited. Subsequently following fertilization, transcriptional activity in the zygote resumes and is readily detectable at the 2-cell stage (Schultz, 2002). The observed disappearance of H1FOO during the transition from oocyte to embryo in the mouse correlates with the induction of embryonic gene activation (Tanaka et al., 2001; Schultz, 2002). Results from previous studies have also shown that a minor contribution to the linker histone content of oocytes comes from some of the other somatic linker histone subtypes (HIST1H1A through HIST1H1E and H1FO) (Fu et al., 2003). It appears, however, that H1FOO is by far the most abundant linker histone during the period of transcriptional quiescence in the ovulated mouse oocyte. Collectively, these results suggest that H1FOO may serve to regulate the level of transcription in the oocyte following fertilization, perhaps by acting as a global inhibitor of oocyte gene expression during this period and thus allowing for the switch to embryonic transcription. The means by which H1FOO performs this function is unknown but may be through its ability to regulate the overall compaction of chromatin or via its capacity to interact with and control chromatin remodelling complexes (i.e. SMARC proteins) (Ramachandran et al., 2003).
It is clear, as demonstrated in the previous study (Hennebold et al., 2000), that various molecular approaches can lead to the identification of genes that exhibit a tissue-selective pattern of expression. Recently, however, the development of bioinformatics software and ongoing high-throughput sequencing projects have allowed for the development of computational methodologies that also provide tissue-specific gene expression profiles. Over the past decade there has been an explosion in the amount of DNA sequence information that has been placed into public databases. A significant portion of this information comes from expressed sequence tags EST, partial sequences generated from individual clones of cDNA libraries that were created from a wide range of cell types or tissues. The warehousing of such sequence data now provides scientists the means to identify genes that are expressed selectively in different tissue/cell types using a computational or in silico approach. In the arena of reproductive biology, Rajkovic et al. (2001) capitalized on the inclusion of EST sequences in the Genbank database to identify transcripts that are only found in cDNA libraries derived from unfertilized oocytes. By screening the Unigene collection of mouse cDNA libraries, 3499 sequences were identified that originated from unfertilized oocytes. Of these, 258 were expressed only in oocyte libraries, with 43 being selected for further analysis. By RT–PCR, five were validated as germ cell selective as no expression could be detected in cDNA derived from heart, lung, liver, brain, spleen, stomach, intestines, kidney or uterus. In situ hybridization confirmed the in silico identification of germ cell-selective expression as only the oocytes were positive for expression within the mouse ovary.
Subsequent studies were performed to yield the complete coding sequence for a number of these newly identified novel cDNA. The analysis of the open reading frame (ORF) led to the identification of domains that hinted at the potential role these proteins play in oocytes (Table I). For example, RACE was used to identify two highly homologous genes termed oocyte-specific homeobox (Obox) 1 and 2 (Rajkovic et al., 2002b). The Obox1 and Obox2 cDNA each encode a protein of 204 amino acids that includes a homeodomain of 60 amino acids. Both Obox1 and Obox2 share a significant level of homology (97% identity), whereas they share only a limited degree of homology with other homeobox proteins. In fact, Xenopus distal-less 4 and mouse distal-less homeobox 2 (Dlx2) proteins possess the greatest degree of homology at 44 and 43% respectively (Rajkovic et al., 2002b). Screening of a mouse/hamster radiation hybrid panel revealed that mouse Obox1 was located on chromosome 7 and was organized into six exons. Subsequent screening of a BAC library and the available mouse genome database led to the identification of additional related sequences that share a significant degree of homology to Obox1 and Obox2. From this analysis the Obox group was expanded to include a total of six different Obox genes (Table I).
Molecular studies were then performed to identify the tissues that express the individual Obox family members. RT–PCR analysis revealed that Obox1, 2, 3, 5 and 6 mRNA are selectively expressed in the ovary, whereas Obox4 mRNA is preferentially transcribed in the testis. In situ hybridization studies revealed that Obox1, 2, and 6 mRNA are expressed predominantly in the oocytes of follicles at the primary to antral stages of folliculogenesis. It was noted that the cellular localization and pattern of mRNA expression for Obox1 and 6 parallel that of Gdf9 and Bmp15 (Rajkovic et al., 2002b). In an independent study, Yeh et al. (2002) cloned the gene that corresponds to Obox3, which they termed oocyte-specific homeobox (Ohx) gene. Obox3 expression was detected in the oocyte after the resumption of meiosis, continuing through fertilization and past the 2-cell stage of early embryonic development. Obox3 mRNA expression subsequently declined at later cleavage stages and was absent in the morula. It was also noted that Obox3 expression increased following the administration of hCG in an ovulation induction protocol, suggesting the potential for regulation by LH (Yeh et al., 2002).
Matzuk and colleagues identified a second novel gene from the same in silico screen that encodes a protein also possessing a homeodomain (Suzumori et al., 2002). By RT–PCR analysis it was determined that this novel gene was expressed in the ovaries of newborn mice, and thus subsequently termed newborn ovary homeobox-encoding gene (Nobox; current nomenclature: OG2 homeobox or Og2x). The full-length Og2x cDNA includes a single ORF yielding a protein of 527 amino acids with the homeodomain corresponding to amino acids 136–195. In situ hybridization studies revealed that Og2x mRNA expression was restricted to oocytes at all stages of development (primordial through antral). Expression of Og2x was elevated in ovaries obtained from Gdf9 null mutants relative to their wild type counterparts (Suzumori et al., 2002).
Several genes identified by the in silico approach were found to be expressed both in the ovary and the testis (Rajkovic et al., 2002a; Yan et al., 2002a). One such novel sequence possesses a tripartite structure that included domains related to a cysteine-rich RING finger, a coiled-coiled motif, and a B30.2 motif that shares features with the RING-B30 family of proteins (Rajkovic et al., 2002a). The highest level of homology was with human Ret finger protein-like genes 1–3 (Rfpl1–3). Consequently, this newly identified gene was termed Rfpl4. Generally, proteins containing such RING finger domains are involved in ubiquitin-dependent proteolysis by acting as ubiquitin ligases (i.e. E3 ubiquitin ligases) (Jackson et al., 2000). This activity results in the attachment of ubiquitin to specific proteins, thus targeting their destruction via the proteosome (Jackson et al., 2000). Subsequent RT–PCR, in situ hybridization and immunohistochemical studies revealed that RFPL4 localizes to oocytes and spermatids. High levels of Rfpl4 mRNA expression were observed in oocytes from the primary to the antral stage of follicle development (Rajkovic et al., 2002a). This level of mRNA expression is maintained in all growing oocytes, through to the completion of meiosis, and declines during early embryonic cleavage (Suzumori et al., 2003). In the testis, Rfpl4 gene expression is observed in elongating spermatids (Rajkovic et al., 2002a). Using a yeast two-hybrid system, it was demonstrated that the RFPL4 protein associates with a member of the E2 family of ubiquitin-conjugating enzymes (HR6A), ubiquitin B and cyclin B (Suzumori et al., 2003). Cyclin B is a known target of ubiquitin-dependent degradation via a proteosomal pathway (Tokumoto et al., 1997). Taken together, these findings suggest that RFPL4 may regulate cell cycle progression through meiosis in germ cells by holding in check the machinery responsible for the progression of mitosis.
Other novel genes expressed in the germ cells of both the ovary and testis were identified by Matzuk and colleagues and include Zfp393, a gene that would encode a zinc finger protein and thus may represent a novel transcription factor; as well as Gasz (germ-cell specific with ankyrin, sterile- motif, and bZIP domains), a gene that potentially encodes a protein involved in cytoplasmic signalling (Yan et al., 2002b). Expression of Zfp393 was observed in oocytes at all stages of follicular development. In the germ cells of the testis, Zfp393 mRNA levels were significantly elevated in mice 20 days after birth, a point in spermatogenesis when round haploid spermatids are formed (Yan et al., 2002b). These results suggest that Zfp393 plays a role in regulating the transcription of specific genes involved in germ cell differentiation or development. Likewise, in situ hybridization experiments revealed a significant level of Gasz mRNA in oocytes at all stages of follicular development, from primordial follicles to ovulation, fertilization, and through to the 8-cell stage of embryogenesis. GASZ protein also localizes to the cytoplasm of spermatocytes at the late pachytene stages of meiosis.
Genes that are expressed exclusively in the ovary, such as Gdf9 and Bmp15, are important for the development of the follicle and subsequent maturation of the oocyte (Dong et al., 1996; Yan et al., 2001). In Gdf9–/– mice, primary follicles fail to develop from the ovarian endowment of primordial follicles (Dong et al., 1996). SSH was used, therefore, to identify genes that are differentially expressed in the ovaries of Gdf9 null mutant mice relative to wild type mice (Burns et al., 2003). Following the generation of the cDNA library, individual clones were identified and subjected to BLAST analysis in an attempt to determine their functional significance. Two clones, one possessing homology to a zinc-binding plant homeodomain (PHD) motif (Wu et al., 2003) with the other being highly homologous to Xenopus laevis nucleoplasmin 2 (NPM2) (Burns et al., 2003), were expressed selectively in the ovary and subsequently chosen for further analysis (Table I). The gene encoding the protein with the PHD motif matched several oocyte-restricted EST sequences and was termed zygote arrest 1 (Zar1) (Wu et al., 2003).
The mouse Zar1 gene consists of a 1.4 kbp transcript that encodes a protein of 361 amino acids (Wu et al., 2003). The human ZAR1 gene, in contrast, encodes a 424 amino acid protein that is 59% identical with the mouse protein. The C-terminal regions of the mouse and human ZAR1 protein share 91% identity, suggesting a conserved role for the region that contains the PHD-like motif. Proteins containing PHD motifs have been reported to regulate transcription, modulate chromatin structure, and serve as E3 ligases in ubiquitination reactions (Coscoy et al., 2003; Gozani et al., 2003). In situ hybridization studies revealed that mouse Zar1 gene expression could be easily documented in growing oocytes. This expression was observed through fertilization and in the zygote. By the 2-cell stage of embryogenesis, expression was diminished and all but absent at the 4-cell stage. The expression profile for the ZAR1 protein, determined by using a polyclonal anti-ZAR1 antibody in immunohistochemistry studies, was similar to the mRNA expression profile (Wu et al., 2003).
To better understand the functional role of ZAR1 in the mouse ovary, Zar1 null mutant mice were generated (Wu et al., 2003). Zar1–/– mice develop normally and males are fertile, whereas Zar1–/– females are infertile. At the gross level, the ovaries of Zar1 null mutants appeared normal with all stages of folliculogenesis being represented. It was also noted that several CL were present, indicating successful ovulation. When the ability to complete meiosis was examined, it was noted that oocytes obtained from both wild type and null mutant mice were equivalent. Following fertilization, however, fewer cleavage stage (2-cell stage) null mutant embryos developed with none progressing to the blastocyst stage. It appears that there is a disruption in cell cycle progression in the null mutant embryos, as the cells seem to stall at the S or G2 phase. This block in cell cycle progression was also associated with a block in the expression of the transcription-requiring complex (TRC), which are markers of embryonic gene activation (Conover et al., 1991). These results suggest that Zar1 represents a stored maternal gene that is necessary for a successful oocyte-to-embryo transition (Wu et al., 2003).
The second ovary-selective gene identified from the SSH screen possesses domains suggesting that it is the mouse orthologue of the Xenopus Npm2 gene (Burns et al., 2003). In the frog, this gene serves to promote the removal of sperm protamines and facilitates nucleosomal assembly (Laskey et al., 1993; Gillespie et al., 2000). In the mouse, the Npm2 gene encodes a 207 amino acid protein that shares 61.4 and 81.6% identity with human and rat proteins respectively (Burns et al., 2003). Through both in situ hybridization and immunohistochemical studies, it was determined that mouse NPM2 is restricted in its expression to growing oocytes. Immunofluorescence studies revealed that NPM2 expression is restricted to the nucleus of germinal vesicle (GV) stage oocytes. Following germinal vesicle breakdown (GVBD), the NPM2 protein redistributes to the cytoplasm with levels declining through the 8-cell stage of embryogenesis. Blastocyst expression of NPM2 is low to undetectable.
Npm2–/– mice were generated to better understand the function of this gene in the oocyte/embryo. Female null mutants were infertile despite containing ovaries that appeared grossly normal (Burns et al., 2003). Null mutant ovaries exhibited a normal pattern of folliculogenesis and ovulated the same number of oocytes as wild type mice undergoing an ovulation induction protocol. Further analysis revealed that Npm2 deficiency did not alter fertilization rates, but drastically reduced the number of 2-cell stage embryos that developed when compared to fertilized oocytes from wild type mice. Additional detailed investigations revealed that normal nucleoli did not develop in zygotes generated from Npm2–/– oocytes. It was also noted that after fertilization, Npm2 null mutants exhibited delayed metaphase spindle formation at the pronuclear stage. This delay generally preceded DNA fragmentation and the induction of apoptosis as determined by terminal deoxynucleotidyl transferase-mediated dUTP end labelling (TUNEL). Collectively, these results point to NPM2 as an important oocyte-specific protein involved in the 1–2-cell transition of embryogenesis.
Additional studies have been completed that also provided information regarding the gene expression profiles of oocytes and early embryos. Mention is briefly made of these studies and the techniques that were used in order to characterize the transcriptional potential of oocytes as well as changes in gene expression that occur after fertilization and through early embryonic development. Neilson et al. (2000) provided a baseline characterization of gene expression within human GV stage oocytes by using SAGE. While SAGE provides sequence information that allows for the identification of individual genes, it can also yield data detailing their relative abundance. In a more directed approach, Taft et al. (2002) used an innovative signal sequence trap methodology to identify mouse genes expressed within growing (isolated from 12 day old mice) and fully grown GV stage oocytes (isolated from 22 day old mice). This approach led to the identification of genes encoding proteins destined to be secreted by the oocyte and which, therefore, may be involved in communication with the surrounding cumulus granulosa cells. The genes identified represented both known and novel elements (Taft et al., 2002). Lastly, SSH was used to monitor the changes in gene expression that occur during the oocyte-to-embryo transition (Zeng and Schultz, 2003). RNA was collected from mouse GV stage oocytes and 8-cell stage embryos. The RNA was then used to generate cDNA for the end goal of identifying differentially expressed genes through the use of SSH. As such, this technique led to the identification of numerous genes that are expressed selectively in either the oocyte or 8-cell stage embryo, some of which represent novel or previously unappreciated molecules that may be important for early embryonic development (Zeng and Schultz, 2003).
|
Genes differentially expressed during follicle development or selection |
|---|
|
TOPAbstractIntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
Ovarian follicular growth is the result of a complex genetic program that underlies the activities of the cellular constituents comprising the follicle. Inducing the growth of a cohort of resting primordial follicles and the ultimate selection of a dominant follicle requires the activities of several genes, including endocrine (FSH) and intraovarian (IGF, activins, inhibins, estradiol) factors (Figure 1). Despite the identification and characterization of a handful of key factors involved in this process, the complete profile of genes required for folliculogenesis remains to be elucidated. Also, the molecular events involved in the selection and development of the dominant follicle in preparation for ovulation are unclear. Several studies have used differential analysis techniques in an attempt to bridge this gap in knowledge regarding folliculogenesis.
Experiments utilizing granulosa cells in vitro identified genes that are regulated in response to the addition of specific hormones, and as such, may be involved in the development of the follicle in vivo. In one study, a rat ovarian granulosa (ROG) cell line was used to identify genes that are specifically regulated by FSH (Ko et al., 2003). ROG cells were either untreated or exposed to FSH prior to harvest for RNA extraction. Differential gene expression was subsequently determined by using a DD RT–PCR methodology. Of the differentially expressed genes isolated, the gene encoding µ-crystallin (Crym), also known as cytosolic T3 binding protein or CTBP, was chosen for further study (Kim et al., 1992). From the DD RT–PCR results, Crym mRNA expression was significantly reduced 6 h post-FSH treatment relative to untreated cells. To confirm these results in primary cultures, Crym mRNA expression was measured in granulosa cells isolated from estradiol-primed 24 day old rats. As observed in the ROG cell experiments, Crym mRNA levels also decreased 6 h after FSH stimulation. In situ hybridization studies using randomly cycling adult rats revealed that Crym mRNA was highly expressed only in small growing follicles, primarily in granulosa cells. A positive hybridization signal was not observed in atretic follicles. This pattern of Crym mRNA expression was confirmed by in situ hybridization studies using immature rats treated for different times (from 0 to 48 h) with pregnant mare’s serum gonadotrophin. Crym has been shown to bind efficiently and transfer T3 into the nucleus of thyroid hormone (TH) receptor-expressing cells (Mori et al., 2002). This may be of biological significance with respect to ovarian function as oocytes and granulosa cells have been shown to possess TH receptors (N.G.Wakim et al., 1987, A.N.Wakim et al., 1993; Zhang et al., 1997) and TH has been implicated in the regulation of granulosa cell steroidogenesis in response to gonadotrophins (Datta et al., 1998). TH treatment of mouse granulosa cells has also been shown to inhibit aromatase activity (Cecconi et al., 1999). These results point to the possibility that a decrease in the level of Crym may prevent or limit TH receptor activation when maximal aromatase activity is required for follicular development.
A similar in vitro experiment was performed by Slee et al. (2001) in an attempt to identify genes that are differentially expressed in rat granulosa cells in vitro following treatment with FSH alone or in combination with dihydrotestosterone (DHT). DD RT–PCR analysis led to the identification of six distinct cDNA that were confirmed as differentially expressed by subsequent northern blot experiments. Of these six cDNA, two were chosen for further analysis and included connective tissue growth factor (Ctgf) and lysyl oxidase (Lox). CTGF is a member of the connective tissue growth factor/cysteine-rich 61/nephroblastoma-overexpressed (CCN) family of growth factors, proteins that are immediate early genes induced by serum, growth factors, or oncogenes (Perbal, 2001). In the porcine ovary, CTGF mRNA levels increase in granulosa and theca cells during early antral follicle development (Wandji et al., 2000). Lox, on the other hand, possesses the capacity to regulate the degree of extracellular matrix (ECM) cross-linking through its ability to catalyse the deamination of peptidyl lysine residues (Kagan et al., 1991). A link between CTGF and LOX also exists as it has been shown that CTGF induces LOX activity in fibroblasts (Hong et al., 1999). Northern blot studies demonstrated that the mRNA levels for both genes (Ctgf and Lox) were reduced following the treatment of granulosa cells with FSH. Whereas Ctgf mRNA expression was not influenced by DHT treatment of granulosa cells, Lox mRNA expression was significantly induced by androgen treatment. In situ hybridization studies revealed that Ctgf mRNA expression was found mainly in preantral or early antral follicles. The induction of follicular development to the pre-ovulatory stage was associated with the loss of both Ctgf and Lox expression. The induction of granulosa cell luteinization by the administration of an ovulatory bolus of hCG led to further reductions in Lox mRNA expression.
In addition to in vitro studies analysing changes in granulosa cell gene expression in response to hormonal stimulation, other investigators have utilized in vivo systems and differential analysis of gene expression to identify novel determinants of follicular development and formation of the dominant follicle (Robert et al., 2001; Bédard et al., 2003; Sisco et al., 2003). The findings of such studies are of significance as the molecular mechanisms responsible for the formation of the dominant follicle in monovulatory species are not well defined. Bédard et al. (2003) used DD RT–PCR to identify genes that were differentially expressed in bovine follicles classified by size through ultrasonography as dominant (>8 mm), subordinate (<4 mm) or pre-ovulatory. Six genes were specifically induced in the dominant follicle, with little expression observed in either the subordinate or pre-ovulatory follicle. One differentially expressed cDNA was subsequently used to screen a bovine ovarian cDNA library, yielding eight positive clones. Sequencing the clones led to the identification of an ORF that encodes a protein of 43.8 kDa. Amino acid sequence homology searches revealed that this protein is likely the bovine orthologue of serine or cysteine protease inhibitor, clade E, member 2 (SERPINE2), also known as Glia-derived nexin (GDN) or nexin-1 (PN1) (Silverman et al., 2001). This protein is known to inhibit the activity of several proteases including thrombin and factor XIa (Baker et al., 1980; Stone et al., 1995; Knauer et al., 2000), with thrombin being implicated in proteolytic processes that occur within the ovary (Gentry et al., 2000; Hirota et al., 2003). Northern blot analysis revealed that SERPINE2 is expressed in several tissues including fetal ovary, corpus luteum, uterus, epididymis, and brain. Semi-quantitative RT–PCR validated the DD RT–PCR findings that SERPINE2 mRNA levels are significantly higher in the dominant follicle than in any other ovarian structure. Following the generation of an anti-SERPINE2 antibody, subsequent immunohistochemistry studies documented bovine SERPINE2 expression in the granulosa cells and oocytes of primordial, primary and secondary follicles. In large antral follicles, this protein was mainly associated with granulosa cells and follicular fluid. No change was observed for SERPINE2 protein levels in follicular fluid isolated from subordinate, dominant or pre-ovulatory follicles as determined by western blot analysis.
A similar set of experiments using cattle undergoing synchronized estrous cycles was performed by Sisco et al. (2003) to identify by SSH genes that are expressed preferentially in the dominant follicle. In this study, the dominant follicle was identified by monitoring the level of IGF-binding proteins (IGFBP) within the follicular fluid of the largest follicles. The IGF system promotes and enhances the responsiveness of granulosa cells to FSH (Adashi, 1998). Thus, follicles that possess the highest level of IGF may survive in the face of declining FSH levels. The IGF system, however, does not simply include ligands and receptors, it also contains binding proteins (IGFBP-1–5) that sequester IGF, as well as IGFBP proteases that cleave the BP and thus release biologically active IGF (Clemmons, 1998). Follicular fluid from the largest pre-ovulatory bovine follicle contains the lowest level of IGFBP (Rhodes et al., 2001). The lowest IGFBP content (IGFBP-2, -4 and -5) and largest follicle diameter was consistently associated with the dominant follicle. Based on these findings, the granulosa cells isolated from individual follicles were classified as either dominant or subordinate and used to generate the cDNA needed for SSH.
Following the completion of SSH, 1600 individual clones were isolated and PCR amplified. The individual PCR products were spotted onto nylon membranes and hybridized with radiolabelled cDNA generated from granulosa cells of either dominant or subordinate follicles to screen out false positive clones. Out of the original 1600 clones, 25 were chosen for further study based on their high degree of differential expression. The resultant 25 clones, representing 22 individual genes, were sequenced and subjected to BLASTn analysis to ascertain possible functional significance. Sequence analysis revealed that six of the 22 cDNA lacked homology to any entries within public databases, whereas 16 were found to represent known genes. Quantitative real-time RT–PCR was performed for each of the 22 genes using cDNA generated from dominant or subordinate follicle RNA. Six of the 22 genes showed an increase in mRNA levels in the dominant follicle relative to the subordinate follicle and included ß-actin (ACTB), carboxypeptidase D (CPD), cytochrome P450 family 19 subfamily A1 (CYP19A1 or aromatase), inhibin ßA (INHBA), low density lipoprotein receptor-related protein 8 (LRP8; also known as apolipoprotein E receptor 2 or ApoER2), mitogen activated protein kinase kinase kinase 5 (MAP3K5; also known as apoptosis signal-regulating kinase1 or Ask1), procollagen lysine 2-oxoglutarate 5-dioxygenase-2 (PLOD2), and solute carrier family 29 member 1 (SLC29A1; also known as equilibrative nucleoside transporter 1 or Ent1). Using a more precise timing of bovine follicular development, all six genes were analysed throughout the course of dominant follicle selection. Levels of CPD, CYP19A1, INHBA, LRP8 and MAP3K5 mRNA were higher in the dominant follicle relative to the largest subordinate follicle. SLC29A1 and PLOD2, however, were not differentially expressed between dominant and subordinate follicles. Even before the dominant follicle could be identified by size or IGFBP content, the highest level of CPD, CYP19A1, INHBA and LRP8 mRNA co-localized to the same follicle. The activity of each of these gene products (see Table II for description of function), therefore, may play an important role in the molecular processes associated with the selection/survival of the dominant follicle.
|
Using a very similar approach to Sisco et al., Fayad et al. (2004
) also used SSH to identify genes that are differentially regulated in subordinate and dominant follicles of cattle. Using synchronized estrous cycles, dominant (>8 mm) or subordinate (2–4 mm) follicles were aspirated to collect their cellular contents. This material was then used for RNA extraction and generation of the cDNA necessary for SSH. A total of 298 cDNA selectively expressed in the dominant follicle were identified, of which 42 represented non-redundant genes. Of these 42 cDNA, 22 matched known or previously characterized genes. The remainder represented uncharacterized entities in that they matched EST sequence entries, BAC sequence entries, or did not match any entry within publicly accessible databases. A number of the genes identified as being expressed predominantly in the dominant follicle were the same as those found in the previously described study performed by Sisco et al. (2003) and included LPR8 CPD and INHBA. Genes expressed selectively in the dominant follicle identified in this study and not by Sisco et al. include chondroitin sulphate proteoglycan 2 (CSPG2; also known as versican); nuclear receptor subfamily 5, group A, member 2 (NR5A2); pregnancy-associated plasma protein A (PAPPA); regucalcin (RGN); and tyrosine 3-monoxygenase/tryptophan 5-monoxygenase activator protein epsilon (YWHAE; also known as 14-3-3
) (Table II). In subsequent gene expression experiments, CPD, CSPG2, LRP8, RGN and YWHAE expression was significantly greater in dominant follicles than in subordinate follicles, pre-ovulatory follicles or corpus luteum. These genes, therefore, appear quite selective as markers of the dominant follicle in cattle. NR5A2 and PAPPA exhibited significantly higher expression in dominant and pre-ovulatory follicles, suggesting a potential role in both dominant follicle selection and ovulation (Fayad et al., 2004). |
Ovulation-selective genes |
|---|
|
TOPAbstractIntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
Studies designed to identify genes that are selectively expressed through the peri-ovulatory period have been more forthcoming than the differential analysis of gene expression through the other stages of the ovarian cycle. This is largely due to the work of Drs Espey, Richards, and colleagues. They employed DD RT–PCR to identify genes that are induced following the administration of an ovulatory stimulus (hCG) in gonadotrophin (equine chorionic gonadotrophin; eCG)-primed immature rats. From these studies, >20 genes have been identified that are expressed selectively through the peri-ovulatory period. Recent in-depth reviews discuss the putative role each gene plays in the cellular and molecular processes that occur during ovulation (Robker et al., 2000
; Espey et al., 2002; Richards et al., 2002). Therefore, this review will only briefly discuss the kinetics and cellular site of their expression in general terms.
From the ovulation-selective genes identified, the genes encoding aminolevulinic acid synthase 1 (Alas1) and early growth response 1 (Egr1) were the first to be induced (1–2 h) following the administration of an ovulatory stimulus (hCG) in eCG-primed rats (Espey et al., 2000, 2002). Alas1 and Egr1 are involved in haem biosynthesis and transcriptional regulation respectively (Ferreira et al., 1995; Khachigian et al., 1998). Alas1 and Egr1 mRNA levels peaked 4 h after hCG injection and then subsequently returned to baseline after ovulation. The remaining granulosa cell-expressed genes identified from the DD RT–PCR studies shared a similar pattern of expression, with most showing induction after 2 h, peak levels at 4–8 h, and return to baseline at 12–24 h post-hCG administration. The genes within this category include glutamate-cysteine ligase, catalytic subunit (Gclc); prostaglandin-endoperoxide synthetase 2 (Ptgs2; also known as cyclooxygenase or COX-2); epiregulin (Ereg); adenylate cyclase-activating polypeptide 1 (Adcyap1); tumour necrosis factor- induced protein 6 (Tnfaip6); regulator of G-protein signaling 2 (Rgs2); steroid acute regulatory protein (Star); aldo-keto reductase family 1, member C6 (Akr1c6; also known as 3-hydroxysteroid dehydrogenase or 3-HSD); CD63; and phosphodiesterase 4D, cAMP specific (Pde4d) (Espey et al., 2002). The gene encoding ferredoxin 1 (Fdx1) exhibited a biphasic expression profile in which maximal expression occurred 4 and 72 h post-hCG injection. Such an expression profile suggests a role in events related to ovulation and corpus luteum formation or function. Another granulosa cell-expressed, ovulation-selective gene, a disintegrin-like and metalloprotease with thrombospondin type 1 motif, 1 (Adamts1), differed in its temporal expression profile in that maximal induction was not observed until close to the actual time of follicle rupture (
12 h post-hCG).
Not all ovulation-selective genes identified from the DD RT–PCR screen were localized to granulosa cells, several were expressed selectively in theca and stroma, as well as vascular endothelium of the rat ovary (Espey et al., 2002). The genes expressed within the theca or stromal compartment include tissue inhibitor of metalloproteinases-1 (Timp1), carbonyl reductase 1 (Cbr1), and a G-protein-coupled receptor, the ligand of which has not been identified. The gene encoding pancreatitis-associated protein (Pap) is unique in that its expression was localized to the vascular endothelium of blood vessels in the hilar region of the stroma. In general, maximal expression for these genes was observed at 8 h post-hCG. Timp1 gene expression did, however, extend into the luteal phase (72 h post-hCG) of the simulated rat estrous cycle.
Another differential analysis methodology, DNA array, was also employed to identify ovulation-selective genes in immature rats undergoing a simulated estrous cycle (Leo et al., 2001). In this study, one group of eCG-primed rats was left untreated while the other received an ovulatory bolus of hCG. The ovaries were removed 6 h later, with the isolated RNA being used to generate radiolabelled cDNA for hybridization studies utilizing nylon membranes spotted with 597 known rat genes. To identify genes that are either up- or down-regulated following the induction of ovulation through the administration of hCG, the ratio of relative expression between the control and 6 h post-hCG time-points was determined. Several induced genes identified in the microarray study have been previously implicated in the ovulatory process and include matrix metalloproteinase 2 (Mmp2); Timp1; serine or cysteine proteinase inhibitor, clade E, member 1 (Serpine1; also known as plasminogen activator inhibitor-1 or PAI-1); and urokinase plasminogen activator receptor (Plaur). Additionally, a number of genes implicated in ovarian function were identified whose expression was repressed following the administration of hCG, including the LH receptor (Lhr) and estrogen receptor 1 (Esr1; also known as ER-).
Three of the differentially expressed genes identified were chosen for further analysis and included fatty acid-binding protein 5, epidermal (Fabp5, 4-fold increase post-hCG injection); interleukin (IL)-4 receptor- (Il4ra, 2.7-fold increase post-hCG injection); and prepronociceptin (Pnoc, 2.24-fold increase post-hCG injection). Until this report, these three genes had not been implicated in ovulatory processes. FABP5 is an intracellular transporter of fatty acids, sterols, prostaglandins and bile acids (Zimmerman et al., 2002). With such activities, FABP5 may play a role in sequestering and/or transport of cholesterol to allow for efficient formation of steroid hormones. IL-4R heterodimerizes with a common 
-chain following ligand binding and generates the appropriate intracellular signals for the development of immune responses that are characterized by the production of specific antibody isotypes (IgA) and inhibition of Type 1 T cell cytokine synthesis (i.e. IL-2 and IFN) (Boothby et al., 2001). IL-4 has also been reported to function as a potent anti-inflammatory cytokine and may serve to limit the production/activities of inflammatory cytokines implicated in follicle rupture. Lastly, PNOC, the precursor for nociceptin, has been shown to play a role in the transmission of pain stimuli in the central nervous system (Okuda-Ashitaka et al., 2000). It is not clear at this time what role the product of this gene plays in ovulation.
Northern blot analysis and in situ hybridization were used to further determine the level of mRNA expression and cellular source of expression respectively of Fabp5, Il4ra and Pnoc throughout the rat peri-ovulatory interval (Leo et al., 2001). Both Fabp5 and Pnoc mRNA expression was low prior to hCG stimulation and increased to a maximal level at 9–12 h post-hCG injection. Il4ra, on the other hand, reached a maximum 3 h post-hCG injection and declined steadily thereafter, reaching pre-hCG levels between 12 and 24 h post-hCG administration. By in situ hybridization, it was determined that all three genes (Fabp5, Il4ra and Pnoc) were expressed predominantly in the theca cell layer of antral and pre-ovulatory follicles. Il4ra message was also detected, albeit at significantly lower levels than that of the theca cell layer, in the granulosa cells of pre-ovulatory follicles.
|
Genes differentially expressed during luteal regression |
|---|
|
TOPAbstractIntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
After ovulation, the remnants of the follicle differentiate and hypertrophy. These events, coupled with the invasion of theca cells and newly formed blood vessels, lead to the development of a mature corpus luteum. Once formed, the corpus luteum is an active endocrine gland responsible for the production of progesterone, which in turn prepares the uterus for the implantation and maintenance of pregnancy in the event of successful fertilization (Davis et al., 2002
; Stouffer, 2003). If pregnancy is not established, the corpus luteum eventually undergoes regression. Despite the dynamic processes that occur throughout the life span of the corpus luteum (i.e. development, steroid production, regression), the genetic program responsible for such transitions have received little attention. The studies described below represent the only published attempts to identify genes that are differentially expressed through the course of luteal regression.
In rodents, it is well established that prolactin (PRL) serves to support luteal formation and function (Choudary et al., 1969; McCracken et al., 1999), whereas prostaglandin F2 (PGF2) is a mediator of luteal regression (McCracken et al., 1999; Niswender et al., 2000). It was subsequently shown that the expression of aldo-keto reductase family 1, member C18 (Akr1c18; 20-hydroxysteroid dehydrogenase), an enzyme that converts biologically active progesterone to a compound with weak progestational activity (20-hydroxyprogesterone) (Wiest et al., 1964), is diametrically regulated in the corpus luteum; PRL silences the expression of this gene (Albarracin et al., 1994) whereas PGF2 greatly stimulates its expression (Stocco et al., 2000). In an attempt to identify additional genes that are differentially regulated by luteotrophic PRL and luteolytic PGF2, Stocco et al. (2001) utilized a DNA array approach to identify PRL- or PGF2-dependent genes. To identify genes that are regulated by PRL, RNA was isolated from corpus luteum removed from hypophysectomized pregnant rats (day 4 of pregnancy) treated with either vehicle or PRL. Alternatively, RNA was isolated from corpus luteum of pregnant rats (day 20 of pregnancy) that received either vehicle or PGF2. RNA was converted to cDNA and hybridized with nylon DNA arrays that contained 1176 distinct rat genes. Several genes were either up- or down-regulated in the corpus luteum of pregnancy in response to PRL or PGF2 alone. It was noted, however, that 13 genes exhibited opposing expression profiles in response to PRL or PGF2 treatment (Table III). One of the 13 included Akr1c18, the gene that is differentially expressed in the corpus luteum in response to PRL or PGF2. Other genes that were also induced by PGF2 and inhibited by PRL include prostaglandin F receptor (Ptgfr), phospholipase-C
(Plcd), and transforming growth factor ß1 (Tgfb1). It has been reported that PGF2 has the capacity to regulate the expression of its own receptor (Ptgfr) (Olofsson et al., 1996). The ability of PRL to inhibit the expression of this gene, however, is a novel finding. The induction of Tgfb1 mRNA expression is also of relevance to luteal regression as this cytokine has been shown to induce apoptosis in a number of cell types including hepatocytes, endometrial stromal cells, and germ cells in the developing testis (Olaso et al., 1998; Albright et al., 2003; Chatzaki et al., 2003). Of the 13 genes identified that are inversely regulated by PRL and PGF2, nine were induced by PRL and inhibited by PGF2 (Table III). One of the PRL-induced genes included LH receptor (Lhr), a gene previously shown to be under the regulatory control of PRL and PGF2 in the rodent corpus luteum (Holt et al., 1976; Luborsky et al., 1984; Gafvels et al., 1992; Bjurulf et al., 1994, 1996, 1998). The remaining genes encode proteins with activities not previously implicated in ovarian function. Thus, this study led to the identification of genes whose expression is regulated in an opposing manner depending on the actions of luteotrophic PRL or luteolytic PGF2.
|
In another study, DD RT–PCR was used to identify genes in the bovine corpus luteum that are differentially regulated following the induction of luteolysis through the administration of PGF2
(Sayre et al., 2000). In domestic animals, it is well known that PGF2 serves as a key substance responsible for initiating the demise of the corpus luteum (McCracken et al., 1999; Niswender et al., 2000). In this study, luteal regression was induced at specific stages of the estrous cycle by the administration of PGF2. At specific time-points thereafter, corpora lutea were removed and used to prepare the cDNA template for DD RT–PCR experiments. Six genes were identified as differentially expressed, one being down-regulated and the remaining five being up-regulated following PGF2 treatment. One up-regulated gene, corresponding to bovine IGFBP1, was chosen for additional characterization with regard to its role in PGF2-mediated luteal regression. Using RT–PCR and western blot analyses, it was determined that IGFBP1 mRNA and protein expression both increase in the corpus luteum following PGF2 administration. The PGF2-dependent increase in IGFBP1 expression was associated with a reduction in circulating progesterone levels, indicating the onset of induced luteolysis.
These results suggest that IGFBP1 may play a role in the regulation of PGF2-mediated luteal regression by antagonizing the effects of IGF1. IGF1 mRNA and IGF1 protein expression has been observed throughout the luteal phase of the bovine estrous cycle (Amselgruber et al., 1994; Woad et al., 2000). Also, type I and type II IGF receptors (IGF1R and IGF2R) have been identified in the bovine corpus luteum (Perks et al., 1999; Schams et al., 1999; Woad et al., 2000), the activation of which by IGF1 results in increased secretion of progesterone (Schams et al., 1999). If IGFBP are added to luteal cultures containing IGF1, there is an inhibition of IGF1 effects, indicating that the IGFBP act as physiological IGF antagonists.
|
Conclusions |
|---|
|
TOPAbstractIntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
The recent development of high-throughput differential screening methodologies has given researchers the opportunity to identify genes that are selectively expressed under a variety of conditions or within a given tissue or cell type. These techniques, namely DNA microarray, DD RT–PCR, SAGE and SSH, have led to the identification of numerous novel or previously unappreciated genes that may be important for normal ovarian function. With the completion of several genome sequencing projects and the generation of a ‘complete genome microarray’, it is conceivable that a fairly comprehensive picture of the ovarian transcriptome could develop in the foreseeable future. The next major task will then include determining the role each of these newly identified, differentially expressed target genes plays in ovarian biology. Recent advances in high-throughput, two-hybrid screening systems have allowed for the creation of comprehensive protein–protein interaction maps in several organisms (Ito et al., 2001
; Li et al., 2004). When similar mammalian ‘interactome’ maps are generated, genes identified in differential analysis screens may be rapidly placed into the context of specific molecular pathways. With this information, possible activities of the identified genes as they relate to ovarian function may be inferred and then directly tested. As there will likely be hundreds, perhaps thousands, of genes to be studied in the context of ovarian biology, another major undertaking will be to prioritize these genes for further investigation. This task may be simplified as database and informatics resources develop to provide a centralized source of information regarding the detailed characteristics of the genes found within an organism’s genome. For example, information regarding homology to other genes, cellular/tissue source of expression, regulation of expression, as well as phenotypic data from genetic studies in other model organisms, may aid in determining the relative importance of individual genes or groups of genes in regard to ovarian biology. Lastly, this review points to the wide range of organisms analysed with regard to differential gene expression in the mammalian ovary, with one exception: humans. This is due to the obvious ethical and technical issues related to obtaining the necessary material for research purposes. Gene activities that are conserved in ovarian processes in different species, especially in non-human primate model systems, would serve as obvious targets for further investigation with regard to their relationship in human reproductive health and disease. |
Acknowledgements |
|---|
|
TOPAbstractIntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
The author thanks Dr Mary Zelinski-Wooten and Dr Ted Molskness for their suggestions and critical reading of the manuscript, and Julianne White for her expert help in the formatting and preparation of the manuscript.
|
References |
|---|
|
TOPAbstractIntroductionOvary/oocyte-selective genesGenes differentially expressed...Ovulation-selective genesGenes differentially expressed...ConclusionsAcknowledgementsReferences |
|---|
Abel, M.H., Wootton, A.N., Wilkins, V., Huhtaniemi, I., Knight, P.G. and Charlton, H.M. (2000) The effect of a null mutation in the follicle-stimulating hormone receptor gene on mouse reproduction. Endocrinology 141, 1795–1803.
Adashi, E.Y. (1998) The IGF family and folliculogenesis. J. Reprod. Immunol. 39, 13–19.[CrossRef][ISI][Medline]
Albarracin, C.T., Parmer, T.G., Duan, W.R., Nelson, S.E. and Gibori, G. (1994) Identification of a major prolactin-regulated protein as 20 alpha-hydroxysteroid dehydrogenase: Coordinate regulation of its activity, protein content, and messenger ribonucleic acid