The elaboration of a quality oocyte is integrally linked to the correct developmental progression of cumulus cell phenotype. In humans and nonhuman primates, oocyte quality is diminished with in vitro maturation. To determine the changes in gene expression in rhesus monkey cumulus cells (CC) that occur during the final day prior to oocyte maturation and how these changes differ between in vitro (IVM) and in vivo maturation (VVM), we completed a detailed comparison of transcriptomes using the Affymetrix gene array. We observed a large number of genes differing in expression when comparing IVM-CC and VVM-CC directly but a much larger number of differences when comparing the transitions from the prematuration to the post-IVM and post-VVM states. We observed a truncation or delay in the normal pattern of gene regulation but also remarkable compensatory changes in gene expression during IVM. Among the genes affected by IVM are those that contribute to productive cell-cell interactions between cumulus cell and oocyte and between cumulus cells. Numerous genes involved in lipid metabolism are incorrectly regulated during IVM, and the synthesis of sex hormones appears not to be suppressed during IVM. We identified a panel of 24 marker genes, the expression of which should provide the foundation for understanding how IVM can be improved for monitoring IVM conditions and for diagnosing oocyte quality.
- oocyte quality
- in vitro maturation
life begins in the oocyte. The oocyte accumulates a remarkable and complex macromolecular reservoir that mediates and controls the initial development of each new individual while at the same time eliminating a range of molecules that promote gametogenesis but would otherwise impede early embryogenesis (8, 36, 37, 67). Failure to complete these transitions appropriately yields an oocyte with limited developmental potential, unable to undergo fertilization or activation, or unable to sustain embryogenesis after fertilization and activation.
Because of the key role of the oocyte in early development, understanding the molecular features that confer a high oocyte developmental potential is of paramount importance in reproductive biology. Such information will reveal the underlying mechanisms and processes that drive oogenesis and early embryogenesis. Moreover, such information will provide markers with which to screen and select high-quality oocytes for assisted reproduction in humans and could enhance success in agricultural reproductive biology and species conservation applications.
The oocyte develops in close coordination with the somatic companion cells, particularly the cumulus oophorous cells, which maintain direct contact with the oocyte via transzonal processes and gap junctions until ovulation. Through its secretion of several exogenous factors, the oocyte exerts a commanding role in the overall development of the follicle and the overall differentiation of the cumulus cells (24, 27, 28, 45, 91). In return, the cumulus cells play a key supportive role by providing to the oocyte a range of extracellular and intracellular molecules that sustain oocyte growth, regulate meiotic progression, and serve as essential metabolic precursors, among other functions (21, 39, 72).
This intimate relationship between the oocyte and attached cumulus cells implies that the two cell types develop and differentiate progressively and in concert with one another and that disruption in the development of either cell component could compromise oocyte quality. Accordingly, clues to the molecular determinants of oocyte quality may be sought by examination of the cumulus cell phenotype. Meeting this objective could be advantageous as a noninvasive means of evaluating oocyte quality clinically, and moreover, it could illuminate the mechanisms of successful follicle development and oogenesis. Cumulus cell gene expression has been studied at the mRNA and protein levels recently in several species to gain insight into the relationship between cumulus cells and oocytes (1, 2, 4–7, 16, 56, 74, 76, 92). However, a clear picture of how cumulus cell phenotype relates to oocyte developmental competence has not yet emerged.
In the rhesus monkey, in vivo maturation (VVM) yields oocytes of high quality that can support efficient development, whereas in vitro maturation (IVM) consistently yields oocytes of limited developmental potential (62–65). Oocytes retrieved after 7 days of in vivo FSH stimulation and then signaled to mature in vitro are able to be fertilized and undergo cleavage and in vitro development to blastocyst stage, but long-term development is very limited (65, 66). Modified culture conditions can increase this success (19, 66) and affect gene expression in the oocyte (50), but improvements can be limited to attaining intermediate stages, and successful development to advanced stages or term after IVM remains limited (10, 62, 66). Studies indicate that embryos derived from IVM oocytes in the monkey are deficient in the timely onset of embryonic gene transcription (66).
True IVM, wherein oocytes are given the initial stimulus to mature in an in vitro environment, has also not been very successful in the human. Only a few reports have been published for IVM of human oocytes from nonstimulated cycles in women, leading to term development (9, 17, 60, 78, 90). Oocyte viability and embryonic development after IVM are substantially lower than with VVM. Often, oocytes are retrieved from patients either after failure to mature subsequent to in vivo stimulation or after a low-level dose of human chorionic gonadotropin to initiate the process in vivo (18, 89), and it has been shown that mRNA content with IVM even after short-term human chorionic gonadotropin (hCG) stimulation is profoundly altered compared with in vivo matured oocytes (32).
It is remarkable that the difference between IVM and VVM in the rhesus monkey, yielding such a profound effect on oocyte quality (38, 62–66), is just that the last 24 h of maturation in response to hormonal stimulation occurs within cumulus-oocyte complexes (COCs) that have been removed from the follicular environment. This indicates that follicular signals during the final 24 h of maturation must normally modulate cumulus cell and/or oocyte phenotype in a manner that confers high oocyte quality.
The basis for reduced developmental competence can be sought at the level of either the oocyte or the companion cumulus cells. Recent studies indicated alterations in the transzonal processes between the cumulus cells and oocyte with rhesus monkey IVM (20). We reported recently that the transcriptomes of IVM and VVM metaphase II stage oocytes displayed a very limited degree of difference, with only 59 mRNAs showing significant alterations, most of which were mRNAs that failed to decline in abundance appropriately (38). Here, we examine the transcriptomes of the companion cumulus cells. We found that the transcriptomes of IVM and VVM cells display extensive differences related to divergent patterns of gene regulation during that final 24 h of maturation. The results highlight specific pathways and processes that are regulated aberrantly and point toward specific mechanisms that control the acquisition of high oocyte developmental potential. Moreover, we identified a panel of 24 marker genes, the expression of which was significantly related to maturation condition and which may be useful cumulus cell markers of oocyte quality.
MATERIALS AND METHODS
Collection and lysis of rhesus cumulus cells.
Adult female rhesus macaques were housed at the California National Primate Research Center. All procedures employed to obtain COCs were conducted according to recommendations of the Guide for the Care and Use of Laboratory Animals and the Animal Welfare Act and its amendments. All procedures for maintenance and handling of the animals were reviewed and approved in advance by the Institutional Animal Use and Care Administrative Advisory Committee at the University of California at Davis. Animals were caged individually with a 0600-1800 light cycle and maintained at a temperature of 25–27°C. Animals were allowed to socialize by being housed in pairs during the day from ∼0800 to 1400. Animals were fed a diet of Purina Monkey Chow and water ad libitum, and seasonal produce, seeds, and cereal were offered as supplements for environmental enrichment. Only females with a history of normal menstrual cycles were selected for study.
Females were observed daily for signs of vaginal bleeding, and the 1st day of menses was assigned cycle day 1. Beginning on cycle days 1–4, recombinant human FSH (rhFSH; Organon, West Orange, NJ) was administered (37.5 IU im) twice daily for 7 days total. For IVM experiments, COCs were collected on day 8. To obtain in vivo matured (VVM) COCs, females were given recombinant hCG (1,000 IU Ovidrel; Serono, Rockland, MA) on treatment day 8 in addition to the rhFSH treatment outlined above. COCs were aspirated from follicles at 28–30 h following hCG by ultrasound-guided aspiration (82, 83).
All COCs were collected into Tyrode lactate (TL)-HEPES medium (37°C) containing 0.1 mg/ml polyvinyl alcohol and 5 ng/ml rhFSH (Organon). Aspirates were immediately placed in a heated isolette (37°C), where COCs were retrieved from aspirates and placed in fresh TL-HEPES medium. Immature oocytes were recovered at the germinal vesicle stage and cultured for 24 h in 70-μl drops under oil in chemically defined basal medium (CMRL) (11) containing 10% bovine calf serum (Gem Cell, Woodland, CA), rhFSH, human luteinizing hormone (0.03 IU/ml Pergonal, Ares-Serono), and 1 μg/ml androstenedione (Steraloids, Newport, RI) [also reported as CMRL medium (66)]. Cumulus cells were obtained from IVM oocytes after culture for 24 h and from prematuration (PM) or VVM oocytes only immediately after collection. The cumulus oocyte complexes were placed individually into drops of TL-HEPES medium, and cumulus cells were removed from oocytes by trituration through a small bore pipette. Only cumulus cells from oocytes that had a germinal vesicle were included in the prematuration IVM control group, and only oocytes with one polar body were utilized for the IVM and VVM groups.
RNA isolation, amplification, and array hybridization.
We generated transcriptome profiles of three groups of rhesus monkey cumulus cells (PM-CC, IVM-CC, and VVM-CC) using the Affymetrix Rhesus Genome arrays (Fig. 1). Each array sample represents multiple COCs obtained from one monkey. In total, 11 array samples (4 PM-CC, 3 IVM-CC, and 4 VVM-CC) were collected from 11 different female monkeys. Total RNA was isolated from cumulus cells using the PicoPure RNA isolation kit (MDS Analytical Technologies, Sunnyvale, CA) according to the manufacturer's instructions. Fifty nanograms of total RNA from each array sample containing COC pools from each individual monkey was subjected to two rounds of cDNA synthesis and in vitro transcription labeling to achieve a linear amplification (Eukaryotic Small Sample Target Labeling Assay, Affymetrix GeneChip Expression Analysis Technical Manual) with minor modifications (initial 5-μl volume for annealing and reverse transcription for 30 min at 42°C followed by 30 min at 45°C). The biotin-labeled cRNA samples were fragmented, and 10 μg were hybridized onto Affymetrix Rhesus Genome GeneChip arrays at the University of Pennsylvania Microarray Core Facility. Posthybridization washing, staining and scanning were performed as described in the Affymetrix GeneChip Expression Analysis Technical Manual.
Microarray data analysis.
Probe hybridization intensity data were imported into the Affymetrix Expression Console Software and summarized using the Robust Multichip Analysis algorithm with a global background correction and a quantile normalization (31). To minimize false detections, expression data were filtered on the basis of Present/Absent calls determined by MAS 5.0 algorithm at the default settings (detection P value <0.05 and tau = 0.015). Only those probe sets called “present” in all biological replicates of one of the two groups in comparison were selected for the analysis. To minimize the chance of false positives in subsequent analyses, probe sets with maximum raw intensity values <100 were omitted from the data set. To identify differentially expressed genes, the Significance Analysis of Microarray (81) was performed. Differentially expressed probe sets were identified at a false discovery rate <1%, and Student's t-test was used to select further for the genes with statistical significance (P < 0.01). Full data sets have been deposited in NCBI's Gene Expression Omnibus (22) and are accessible through GEO Series accession no. GSE25288 (http://www.ncbi.nlm.nih.gov/geo/) and at www.preger.org.
Pathway analysis and biological functional analysis were performed using the rhesus array annotations provided in the Ingenuity Pathway Analysis software (Ingenuity Systems, Redwood City, CA). For biological functional analysis, Fisher's exact test was performed with the P value threshold of 0.05 to identify molecular and cellular functional categories with statistical significance.
Real-time quantitative RT-PCR analysis.
A panel of 43 oocyte quality marker candidates was chosen on the basis of a combination of fold change, array signal strength, and biological function. The mRNA expression of these 43 genes was compared between IVM-CC and VVM -CC by the quantitative real-time RT-PCR method. Each sample used for the quantitative (q)RT-PCR assay corresponds to total RNA pooled from multiple COCs obtained from one monkey. For a greater statistical power, total number of IVM-CC and VVM-CC samples was expanded for quantitative (q)RT-PCR from seven to 13 samples, including one additional IVM-CC and five additional VVM-CC samples. In total, nine samples of VVM-CC and four samples of IVM-CC were used for the qRT-PCR assay. Total RNA was isolated using the PicoPure RNA isolation kit (Molecular Devices, Sunnyvale, CA) and subjected to a whole transcriptome amplification using the QuantiTect whole transcriptome kit (Qiagen, Valencia, CA). Custom TaqMan gene expression assays were designed based on rhesus cDNA sequences using Primer Express software version 3.0 (Applied Biosystems, Foster City, CA). The primer and probe sequences of the custom-desgined TaqMan assays are listed in Supplemental Table S1 (Supplemental Material for this article is available online at the AJP-Endocrinology and Metabolism website). Quantitative real-time PCR was performed on ∼100 ng of cDNA for each assay using ABI StepOne Plus instrument according to the manufacturer's recommendations (Applied Biosystems). The mRNA abundance of a target gene was normalized to the endogenous mitochondrial ribosomal protein S18C gene (MRPS18C) for sample-to-sample comparisons, and the relative expression ratio of IVM to VVM groups was obtained by the comparative CT method (44). Statistical analysis was performed between IVM and VVM groups using the Relative Expression Software Tool for the genes assayed by qRT-PCR method (57).
Gene expression values were analyzed for a relationship to oocyte quality. Single- and multiple-gene generalized estimating equations (GEE) were used to calculate the relationship of gene expression to oocyte quality. GEE is a special case of the generalized linear model, taking into consideration the occurrence of a binary outcome (oocyte quality) from multiple events of a single experimental unit.
Overview of the array analysis.
With array data inspected, all quality control parameters were within the acceptable ranges for all 11 array samples (Supplemental Table S2). The average detection rate for PM-CC samples was 47%, corresponding to ∼25,000 probe sets. The detection rates for IVM-CC and VVM-CC were 38 and 41%, respectively. A hierarchical clustering analysis revealed that the biological replicates cluster to their corresponding groups without any apparent outliers (Supplemental Fig. S1). The three-dimensional principal component analysis also showed the clustering of all replicates to their corresponding groups (Supplemental Fig. S2).
We observed that cumulus cells undergo a dramatic change in their transcriptome during the last 24 h of maturation. A total of 3,252 out of 19,076 annotated genes (17%; 4,431 probe sets) on the rhesus array were either up- or downregulated during VVM, and 14.5% (2,772 genes; 3,840 probe sets) of all genes were modulated during IVM. There were more genes downregulated than upregulated during the cumulus cell maturation process (Supplemental Table S3). Of the probe sets differentially regulated during VVM, 74% [2,459 unique genes; 3,284 probe sets (2,776 annotated)] were downregulated, whereas only 26% [793 unique genes; 1,147 probe sets (883 annotated)] were upregulated. A similar relationship was observed during IVM [1,865 genes (2,592 probe sets) downregulated compared with 907 genes (1,248 probe sets) upregulated].
Identification of differentially regulated gene sets.
To assess the biological processes or functions affected by this vast amount of change, our objective was to identify sets of differentially regulated genes and then apply ingenuity pathway analysis (IPA) to understand the biological implications of these differences. Some genes displayed inconsistent results between different probes sets, and their inclusion in both up- and downregulated categories can confound pathway analysis. Moreover, with such large numbers of genes affected, IPA can fail to yield a clear picture of the major pathways affected. To simplify and focus our analysis, 485 genes (1,099 probe sets) with inconsistencies between probes sets were omitted from further consideration, and only the genes that differed in expression at the threshold of twofold or greater were employed for IPA. We began our analysis by categorizing genes as correctly or incorrectly regulated (i.e., same direction of change) and as either up- or downregulated during maturation, comparing IVM-CC and VVM-CC with PM-CC. Our objective for this was to understand the transitions that occur or do not occur during IVM or VVM rather than just comparing the end-stage IVM-CC and VVM-CC to each other. Genes were thus classified by comparing the temporal transitions and identified as either correctly or incorrectly regulated and within these categories either up- or downregulated or oppositely regulated. Once the differences in developmental transitions were characterized, VVM-CC and IVM-CC were also compared directly to each other.
Identification of genes incorrectly regulated during IVM.
Genes that changed in opposite directions between IVM-CC and VVM-CC compared with PM-CC or those changed in only one or the other condition were classified as “incorrectly regulated.” The first two groups of genes incorrectly regulated during IVM included those for which mRNA expression changed (increased or decreased compared with PM-CC) during IVM but not VVM. These genes, designated as “IVM unique” (Fig. 2A), included 1,551 probe sets (1,216 annotated) corresponding to 1,126 unique genes. About one-half of these genes were downregulated, and the other half were upregulated (Supplemental Table S4).
The next two groups of incorrectly regulated genes included those that changed significantly (increased or decreased compared with PM-CC) during VVM but not IVM (VVM unique; Fig. 2A). A total 1,607 unique genes (2,142 probe sets, 1,760 annotated) belonged to the VVM-unique category. Compared with the IVM-unique genes, a higher proportion (1,423 probe sets, 1,112 unique genes) of the “VVM-unique” genes was downregulated during maturation than upregulated (719 probe sets, 495 unique genes; Supplemental Table S5).
A small number of genes (28 probe sets corresponding to 28 unique genes) was regulated in opposite directions (Fig. 2B). Twenty-one of these genes were upregulated during IVM but downregulated during VVM, and seven genes were downregulated during IVM but upregulated during VVM (Supplemental Table S6).
Identification of correctly regulated genes.
Genes that changed in common and in the same direction going from PM-CC to either IVM-CC or VVM-CC were classified as “correctly regulated.” The genes that were correctly regulated during IVM were identified by comparing the probe sets that showed significant changes in gene expression during IVM with those that showed significant changes during normal VVM maturation. These genes may represent a minimally required set of changes for cumulus cell accompanying oocyte maturation. About 4% of the probe sets (2,289 out of 52,865) or 8.6% of all annotated genes on the rhesus array (1,648 out of 19,076) were regulated during both IVM and VVM (IVM + VVM common; Fig. 2A). The majority of these [2,261 probe sets (1,824 annotated), 1,620 unique genes] were regulated in the same direction during IVM and VVM, either increased or decreased in expression during maturation (Fig. 2B and Supplemental Table S7). Among the relatively small pool of upregulated genes (n = 293), serglycin (SRGN, a.k.a. PRG1 or proteoglycan 1) showed the highest upregulation in gene expression, as high as 128-fold.
Direct comparison of IVM and VVM cumulus cells and identification of more incorrectly regulated genes.
The array analyses discussed above identified differences in the transition from PM-CC to IVM-CC and VVM-CC. These analyses may indicate which gene or pathway interactions are misregulated during IVM. The information from these analyses may be critical in aiding us to understand the cause of the poor quality of IVM oocytes. However, it is also necessary to identify the genes that show differential mRNA abundance at the final stage of maturation to understand the overall difference in CC phenotype that is established and to identify markers that may predict oocyte quality.
Interestingly, we found that despite the large number of genes showing incorrect modes of regulation relative to PM-CC, the final outcome of maturation was a much smaller set of differentially expressed mRNAs. There were 558 probe sets (452 unique genes) that showed differential mRNA abundance between IVM and VVM cumulus cells. Of these, 76% were more abundant in IVM cells than in VVM cells (Supplemental Table S8). The disparity between the number of VVM-CC vs. IVM-CC differences and the number of differences in the transitional comparisons with PM-CC indicates that many of the differences seen comparing the transitions are likely to reflect quantitative differences rather than qualitative differences in gene regulation.
An additional group of oppositely regulated genes (89 probe sets corresponding to 68 unique genes) was identified by comparing the transcriptomes from IVM- and VVM-CCs. (Supplemental Table S9). These genes were not identified above as “incorrectly regulated” due to either a small fold change or P value greater than the cutoff setting of 0.01 in the comparison with PM-CC but as significant end-stage differences between VVM-CC and IVM-CC. Among these genes, 58 were upregulated during IVM but downregulated during VVM.
Functional characteristics of genes correctly regulated during in vitro maturation.
Biological functions and pathways significantly overrepresented in genes regulated either correctly or incorrectly during IVM or differing between VVM-CC and IVM-CC (Supplemental Tables S4–S9) were examined. Biofunction analysis of correctly regulated genes revealed that “cell cycle” was a biological function enriched at the highest significance among the genes downregulated during both IVM and VVM (Table 1). Genes positively associated with cell cycle progression (GO:0045786; e.g., AKT1, CDC2, CDC20, CDC25A, CDC25C, CDC45L, CENPF, NUSAP1, PGGT1B) were downregulated, and some genes negatively associated with cell cycle progression (BUB3, ESR1, BCL2, CHMP1A, CDKN1B, ILKAP, BUB1) were also downregulated during maturation. Cell cycle regulatory genes also comprised the most highly affected biological function among the genes upregulated during maturation. The mRNA abundances of CDKN1A and APBB1 genes, which promote cell cycle arrest (GO:0045787), were increased during maturation.
Functional characteristics of genes incorrectly regulated during in vitro maturation.
To understand the biological functions regulated in VVM- and IVM-unique genes, IPA overrepresentation analysis was performed on the genes with the fold change of difference greater than two (Table 2). For VVM-unique genes, DNA replication, recombination, and repair was the most significantly affected functional category. Among IVM-unique genes, cellular growth and proliferation genes were significantly regulated. Differences in pathway regulation were plotted for easier evaluation (Fig. 3). DNA replication, recombination, and repair was most different between IVM and VVM cumulus cells in both the confidence interval and gene number. Eighty-one genes involved in DNA replication, recombination, and repair were either up- or downregulated during VVM, whereas only nine were regulated during IVM. In VVM cumulus cells, most of the genes in this biofunctional category are involved in DNA repair or synthesis and replication. IVM cumulus cells regulated a handful of genes involved in DNA packaging or genome instability (Table 3). For the cell cycle functional category, not only was the total number of affected cell cycle genes greater in VVM cumulus cells than in IVM cells, there was also slightly higher expression of genes that delay cell cycle in VVM cumulus cells compared with IVM cumulus cells (Table 3). IVM and VVM cells differed particularly in the regulation of the genes involved in the G1/S transition of the cell cycle (Table 4). VVM-CC showed a preference in upregulating many genes involved in the cell cycle arrest at the G1/S transition phase, whereas IVM cumulus cells showed the opposite trend, upregulating more genes that promote the G1/S cell cycle progression. Interestingly, although IVM-CC did not regulate the same genes as the VVM-CC, they upregulated a distinct set of G1/S cell cycle arrest genes, such as insulin-like growth factor-binding protein-3 (IGFBP-3) and transforming growth factor-β 1 (TGFβ1). Additionally, IVM-CC successfully downregulated some genes that promote progression at the G1/S transition. This analysis suggests that IVM cumulus cells may utilize pathways or mechanisms distinct from VVM cumulus cells as the oocyte matures. The reduced oocyte developmental competence may result from the fact that these alternate pathways used by IVM-CCs may not be as efficient as the normal pathway used by VVM-CCs for supporting oocyte development, and this may be the reason why IVM cumulus cells yield reduced quality.
The modulation of distinct sets of genes by IVM and VVM cells was also observed in other biofunctional categories. Changes in the “cell death” and “cell growth and proliferation” functions were associated with distinct sets of genes during IVM and VVM (Supplemental Tables S10 and S11), but the overall directionality seemed to be regulated similarly. For example, the percentages of proapoptotic and antiapoptotic genes in IVM-CC were almost the same as those in VVM-CC (Table 3). Genes involved in growth promotion and growth inhibition were likewise distinct but distributed in similar proportions in IVM and VVM cells. These observations suggest that IVM cumulus cells may regulate the maturation process using gene sets distinct from those used by VVM cumulus cells.
Next, to explore the functional relevance of genes most highly altered during IVM, a more stringent fold change criterion (≥10-fold) was applied to the IVM-unique and VVM-unique genes. This identified periostin (POSTN) and FBJ murine osteosarcoma viral oncogene homolog B (FOSB) as molecules prominently upregulated or downregulated, respectively, during IVM (Supplemental Table S12). Many of the genes upregulated at ≥10-fold during IVM encode extracellular matrix proteins (POSTN, EFEMP1, SPON2, LAMB1), membrane proteins (KCNK3, TMEM26, TTMA, THBD) or those related to extracellular matrix proteins (IGFBP-3, TNFAIP6, MMP20). IPA showed that this group of genes is overrepresented in molecular functions such as cellular growth and proliferation, cellular development, and cell-to-cell signaling. There were only four genes downregulated ≥10-fold during VVM. Three out of these four genes are transcription factors. FOSB and EGR3 mRNA expression is maintained at a higher level in VVM cumulus cells compared with IVM cumulus cells (Supplemental Table S12). In addition to FOSB and EGR3, other early transcription factors FOSL2 and JUND mRNAs were downregulated during IVM but maintained at a more steady level during VVM.
The characteristics of the genes differentially regulated during VVM are distinct from those regulated during IVM. For both the genes upregulated and the genes downregulated during VVM with the highest degree of changes in mRNA expression, lipid metabolism was the most highly overrepresented function (Table 5 and Supplemental Table S13). Genes upregulated during VVM by ≥10-fold were enriched in the biological functions of lipid metabolism and small molecule biochemistry. Those genes are HSD11B1, steroidogenic acute regulatory (STAR) protein, EREG, prostaglandin-endoperoxide syhthase 2 (PTGS2), CEBPA, PTGER3, ANXA1, ACSL1, CD36, RORA, PRKCA, IGF1, ABHD5, UGP2, SLC16A10, and CLDN1 (Table 5 and Supplemental Table S13). They are involved in various subcategories of lipid metabolism function, such as the synthesis, metabolism and transport of lipids, β-oxidation of fatty acids, the synthesis and metabolism of cholesterol and prostaglandins, etc. It is also notable that a well-known cumulus cell maturation marker, PTGS2 (prostaglandin-endoperoxide synthase 2), was identified as differentially upregulated during VVM but not during IVM. The genes downregulated during VVM (fc ≥ 10) are connected in a network involved in small molecule biochemistry and lipid metabolism (Supplemental Table S14 and Supplemental Fig. S3). Cytochrome P450 family 19 subfamily A polypeptide 1 (CYP19A1) and hydroxysteroid (11β) dehydrogenase 2 (HSD11B2), for example, interact with sex steroid hormones such as progesterone, β-estradiol, and prolactin (14, 41, 79, 85). Also, CYP19A1 may increase the expression of progesterone receptor (PGR) (42). This is consistent with our finding that PGR mRNA expression was significantly upregulated in IVM-CC, in which CYP19A1 mRNA expression failed to get downregulated. (Supplemental Table S14).
Included in the group of oppositely regulated genes were those involved in RNA posttranscriptional modification (POP5, AARS), carbohydrate metabolism (NUDT5, TKT, PTHLH, SLC37A4), and cell-to-cell signaling interaction (ITGAE, PTHLH). Only 10 of 68 genes were downregulated during IVM but upregulated during VVM. Among these were genes involved in posttranslational modification (SERP1, PTPN1, HERPUD1, NCSTN) and cell morphology (DLG1, EGR2, PTPN1) (Supplemental Table S9).
Identification of oocyte quality markers by qRT-PCR and statistical analysis.
The mRNA expression of 43 candidate markers of oocyte quality was examined by qRT-PCR. The rationale was that VVM oocytes are of the highest quality, whereas IVM oocytes are known to have limited developmental potential; hence, analysis of adherent cumulus cells could reveal markers predictive of ooctye quality. The qRT-PCR results confirmed differential expression for these genes between IVM-CC and VVM-CC (Fig. 4). Univariate GEE analysis showed that the expression levels of 24 of the 43 genes tested were significantly related to oocyte quality (IVM vs. VVM) (genes with asterisks in Fig. 4, Supplemental Table S15). Thus any of these genes may be an informative marker of oocyte quality. Changes in gene expression are thus associated with increases or decreases in predicted oocyte quality as specified by the odds ratios given. Moreover, multiple genes from this large panel can be employed together to further strengthen the predictive power. Three genes (NEK6, AQP11, and IGF1) were identified by stepwise multiple GEE to be predictive in combination (Supplemental Table S16). The following equation may be used to predict the probability of having a high-quality oocyte: probability (high-quality oocyte) = e5.608 + 0.645[NEK6] + 0.100[AQP11] − 2.17[IGF1]/1 + e5.608 + 0.645[NEK6] + 0.100[AQP11] − 2.17[IGF1].
Our transcriptome comparisons of PM-CC, IVM-CC, and VVM-CC yielded a number of surprising results that have significant implications for understanding the biology of oocyte maturation and its relationship to attendant oocyte quality. First, we found that a large number of genes (558 probe sets, 452 genes, 4.5% of total genes detected) differ in expression when IVM-CC and VVM-CC are compared directly. A much larger number of differences are seen when the two developmental transitions, PM-CC to IVM-CC vs. PM-CC to VVM-CC, are compared, indicating that many more genes display quantitative differences in regulation than is revealed just by comparing IVM-CC and VVM-CC. Overall, a total of 2,830 genes were incorrectly regulated (1,607 VVM-unique + 1,127 IVM-unique + 96 oppositely regulated), which is more than 20% of the genes detected on the array. This result highlights the need to consider developmental transitions rather than just end points when evaluating divergent processes and also emphasizes a striking disruption in cumulus cell development during the terminal 24-h maturation period when this is undertaken in vitro. The extensive degree of difference observed between IVM-CC and VVM-CC contrasts with the minor degree of difference reported for the oocytes (38). This suggests that reduced oocyte quality may not be reflected highly at the level of oocyte mRNA expression. This may be because key changes in the oocyte reside at the posttranslational level, whereas changes in cumulus cell phenotype that affect oocyte quality are manifested at the mRNA level.
Second, several aspects of the array data indicate truncation or delay in the normal pattern of gene regulation. A higher proportion of VVM-unique genes is normally downregulated during VVM but not downregulated during IVM, indicating that IVM-CC are delayed or defective in downmodulating gene expression. The group of VVM- and IVM-unique genes might have resulted in part from a failure of IVM cumulus cells in timely gene regulation or mRNA degradation. We observed deficient expression of some transcription factors. The expression of these early transcription factors is required for maturation and terminal differentiation of mural granulosa cells (30, 59, 69). Thus, IVM-CC may be deficient in factors necessary to complete the maturation process. Among the genes with the greatest expression differences, the cumulus cell differentiation markers FOSB, FOSL2, JUND, PTGS2, and EGR2 failed to be regulated correctly during IVM. This suggests that IVM-CC are not attaining a phenotype identical to VVM-CC.
An increase in PTGS2, STAR, HAS2, and GREM1 gene expression is associated with the cumulus cell expansion process in the mouse, and this increase is thought to be regulated by GDF9 secretion from the oocyte (23, 33, 52, 68). Although HAS2 and GREM1 mRNAs are expressed here, there does not appear to be a high degree of upregulation associated with either IVM or VVM. In contrast, our rhesus cumulus cell microarray results show that PTGS2 and STAR are two of the genes with the highest relative increase in mRNA levels in VVM-CC compared with the PM-CC. These genes are not increased in IVM-CC, yet cumulus expansion does occur in IVM COCs (51). The association between cumulus expansion and genes that are regulated by GDF9 in the mouse does not appear to be as direct in the primate. However, higher PTGS2, HAS2, and GREM1 mRNA levels have been associated with better-quality oocytes and embryos in humans (46).
Third, we find evidence of remarkable compensatory changes in gene expression during IVM. Although many individual genes differ in regulation between IVM-CC and VVM-CC, this is offset by up- or downregulation of other genes related to specific cell functions. We observed this particularly for cell cycle arrest at the G1/S phase transition, cell death, and cell growth and proliferation categories. This indicates that IVM-CC and VVM-CC may arrive at similar physiological states but that these similar states likely differ significantly at the level of individual regulatory gene expression. The consequences of such an altered, albeit similar, cumulus cell physiological state for the oocyte remain to be determined.
Many of the genes associated with cell cycle progression have been studied in cell types other than those in the ovarian follicle, which could partially explain some of the apparent differences found in this study. It is known that cell cycle progression decreases at ∼24 h post-hCG in vivo (26), the time in which the VVM samples were obtained for this study. Furthermore, it has been demonstrated that cumulus cell morphology is different during IVM and VVM (20), which may also be reflected in the gene expression patterns of these two groups of cumulus cells.
Fourth, we found that among the genes affected in cumulus cells by IVM are those that likely contribute to productive cell-cell interactions between cumulus cell and oocyte. Our analysis of the IVM oocyte transcriptome likewise pointed to likely changes in cell-cell interactions (38). Here, we observe changes in the expression of growth factors and other ligands, their receptors, and other interactive cell surface molecules. For example, the DLG1 gene (discs, large homolog 1) is responsible for the structural integrity of adherens junctions and tight junctions. The fact that IVM cumulus cells failed to upregulate this gene confirms again the compromised cell-to-cell interaction observed in IVM oocytes and cumulus cells. Early growth response 2 (EGR2) is another early response gene that is regulated downstream of MEK-ERK pathway (12, 34). The mRNA expression of EGR2 in VVM cumulus cells tends to be maintained high, but in IVM cumulus cells the mRNA abundance for the EGR2 gene was downregulated 18-fold (Supplemental Table S9).
The SRGN mRNA is less highly upregulated with IVM. The SRGN (PRG1) protein is best known for its role in formation of mast cell secretory granules and a granule-mediated apoptosis via the association with granzyme B and perforin (47, 58). SRGN also binds to hyaluronic acid as well as its receptor CD44 and can inhibit the adhesion of cells to extracellular matrix proteins (25, 77). Serglycins influence many aspects of cellular processes such as matrix formation, cell-cell and cell-matrix adhesion, and cell proliferation and migration, and they are also implicated in ovarian development, ovulation, and fertilization (61). The fact that proteoglycan synthesis is regulated by gonadotropins suggests the involvement of these molecules in folliculogenesis (48). Also, it has been suggested that serglycins bind circulating lipoproteins to regulate their cellular uptake and concentration in follicular fluid (86). Previous studies showed that higher LDL concentrations in follicular fluid are associated with decreased oocyte fertilization rate (84). However, the direct association or mechanisms of serglycins in steroidogenic activity of cumulus cells have not yet been fully understood.
Additional evidence of defective cumulus cell-oocyte interaction was noted when comparing the actin and tubulin labeling of the transzonal processes (TZP) of cumulus cells from rhesus monkey COCs at various times following hCG in vivo or in vitro (20). The IVM COCs exhibited a remodeling of the tubulin component of the TZPs and retained the TZPs longer, yet the timing of nuclear maturation was significantly advanced in IVM oocytes. In contrast, the VVM COCs did not exhibit TZP remodeling, and the number of TZPs decreased within 3 h of hCG administration. These microarray results provide insight into the underlying basis for the aberrant responses of the IVM cumulus cells.
A fifth major finding is that many of the genes that are misregulated to the greatest degree during IVM are involved in lipid metabolism. Lipid metabolism is implicated as a key regulation pathway in cumulus cell maturation. Steroidogenesis and prostaglandin synthesis are the two main mechanisms that must be carefully regulated for COCs to achieve complete maturation. Cytochrome P450 genes are key genes involved in the maturation of mural cells. Among the various cytochrome P450 genes, CYP19A1 in particular was incorrectly regulated during IVM. The CYP19A1 mRNA abundance was significantly downregulated during VVM but not during IVM. The CYP19A1 gene is implicated in steroidogenesis, particularly the synthesis of estrogen. It is well known that estrogens play a critical role in regulating homeostatic pathways and influence fertility, lipid metabolism, and other key biological pathways (70). During follicular development, the transcription of CYP19A1 in ovarian mural granulosa cells is upregulated by FSH and prostaglandin E2 (15, 53), but later, with the surge of LH/hCG, CYP19A1 gene expression must be rapidly downregulated. The downregulation of CYP19A1 gene expression in mural cells is a characteristic of ovarian follicular differentiation (3, 49). The failure of IVM-CC to downregulate CYP19A1 expression further indicates that IVM cumulus cells are likely compromised in their terminal differentiation.
Numerous genes involved in lipid metabolism are incorrectly regulated during IVM (Fig. 5). For example, LDLR and STAR genes are responsible for regulating the transfer of cholesterol. LDL receptor (LDLR) is a cell surface protein that takes up cholesterol-carrying LDL molecules (13), and STAR mediates the transfer of cholesterol to the inner mitochondrial membrane from outer membrane (71). The expression of these two mRNAs was upregulated during VVM but not IVM. Additional enzymes involved in various steps in lipid metabolism were misregulated during IVM (Fig. 5). The CYP11A1 and HSD3B2 mRNAs, encoding products involved in the breakdown of cholesterol to progesterone, are overexpressed during IVM, as was the mRNA for PGR. Progesterone plays a critical role in granulosa cell differentiation via the interaction with specific receptors (PGRs). The reduction of PGR gene expression is associated with better embryo quality in humans (29); therefore, the overexpression of PGR and the enzymes involved in the progesterone production in IVM-CC may decrease the developmental competence of IVM oocytes and embryos. The association of a higher level of PGR mRNA with COCs that will yield less competent embryos is consistent with the observation that higher-quality human embryos came from COCs with reduced levels of PGR (29). This classic nuclear PGR does not appear in mouse cumulus cells, although it is present in the granulosa cells of preovulatory follicles after the hCG stimulus (75), and similar transient expression in response to LH has also been demonstrated in the rat (54, 55). The rhesus cumulus cell microarray results appear to delineate another significant difference between primates and rodents.
Additionally, our data suggest that the synthesis of sex hormones is not suppressed during IVM. The enzymes that convert androstenedione into testosterone or estradiol, HSD17B1 and CYP19A1, were downregulated during VVM but not during IVM. The IVM procedure seems to allow the continued production of androgens that is shut down during VVM. The abnormal regulation of lipid metabolism in IVM cumulus cells is also evidenced in the glucocorticoid and mineralocorticoid pathways. Glucocorticoids are steroid hormones that play an important role in the maintenance of physiological processes. Glucocorticoids mediate their functions by interacting with their receptors. Cortisol (or hydrocortisone) is the active form of glucocorticoids that can bind and activate glucocorticoid receptors, whereas corticosterone is the inactive form. HSD11β1 predominately generates active cortisol from inactive cortisone, whereas HSD11β2 converts active cortisol to inactive cortisone (Fig. 5). Glucocorticoids activated by HSD11β1 can upregulate the expression of PTGS2; both HSD11B1 and PTGS2 mRNAs were upregulated during VVM but not IVM.
While this article was in preparation, another study reported on changes in mRNA expression detected in whole follicles during oocyte maturation and ovulation (88). That study observed many of the changes in gene expression that we report here for cumulus cells. Our data thus pinpoint changes specifically associated with the cumulus cell. Moreover, our study particularly discloses the difference between in vitro and in vivo maturation. By comparing the developmental transitions from PM-CC to either IVM-CC or VVM-CC, our study highlights striking differences in gene regulation, such as delays in some gene expression changes and compensatory changes seen in other changes. Additionally, our study highlights the key interactions of cumulus cells with the follicular environment during the last 24 h of maturation, which are not apparent in analysis of whole follicles.
The identification of a panel of 24 genes, the expression of which correlates highly with protocols yielding high- or low-quality oocytes, may provide the foundation for understanding how IVM can be improved and for monitoring IVM conditions to recognize such improvements. Additionally, the availability of this panel of markers may be applicable for screening and identifying high- or low-quality oocytes. All or a subset of these markers could be adopted into a readily applicable PCR- or protein expression-based assay that could be applied to small clumps of cumulus cells isolated from individual COCs. This could allow clinical selection of the highest-quality oocytes in situations that require fertilized embryos to be transferred (35) and may provide an invaluable means of evaluating oocyte quality after in vitro oocyte culture and IVM, for example, in the context of oncofertility programs (87), or efforts to manipulate germinal vesicle stage oocytes in emerging assisted reproduction methods involving microsurgery (40, 43, 73, 80). These markers may also provide a means of optimizing reproductive biology approaches to enhance production of valuable agricultural species or enhance species conservation approaches. Evolutionary conservation of most of these genes (Supplemental Table S17) indicates the feasibility of application across species.
This research was supported by grants from the National Institutes of Health, National Center for Research Resources: RR-15253 (K. E. Latham), RR-00169 (California National Primate Research Center), and RR-025880 (C. A. VandeVoort). This project was funded in part under a grant with the Pennsylvania Department of Health to Z. Obradovic. The Department specifically disclaims responsibility for any analyses, interpretations, or conclusions.
No conflicts of interest, financial or otherwise, are declared by the authors.
We acknowledge the technical assistance of Dana Hill in collecting material for this study.
- Copyright © 2011 the American Physiological Society