Skip to main content
NCBI home page
Endocrinology logoLink to Endocrinology
. 2008 Oct 23;150(3):1521–1529. doi: 10.1210/en.2008-1201

Dynamic Changes in Gene Expression that Occur during the Period of Spontaneous Functional Regression in the Rhesus Macaque Corpus Luteum

Randy L Bogan 1, Melinda J Murphy 1, Jon D Hennebold 1
PMCID: PMC2654732  PMID: 18948396

Abstract

Luteolysis of the corpus luteum (CL) during nonfertile cycles involves a cessation of progesterone (P4) synthesis (functional regression) and subsequent structural remodeling. The molecular processes responsible for initiation of luteal regression in the primate CL are poorly defined. Therefore, a genomic approach was used to systematically identify differentially expressed genes in the rhesus macaque CL during spontaneous luteolysis. CL were collected before [d 10–11 after LH surge, mid-late (ML) stage] or during (d 14–16, late stage) functional regression. Based on P4 levels, late-stage CL were subdivided into functional-late (serum P4 > 1.5 ng/ml) and functionally regressed late (FRL) (serum P4 < 0.5 ng/ml) groups (n = 4 CL per group). Total RNA was isolated, labeled, and hybridized to Affymetrix genome microarrays that contain elements representing the entire rhesus macaque transcriptome. With the ML stage serving as the baseline, there were 681 differentially expressed transcripts (>2-fold change; P < 0.05) that could be categorized into three primary patterns of expression: 1) increasing from ML through FRL; 2) decreasing from ML through FRL; and 3) increasing ML to functional late, followed by a decrease in FRL. Ontology analysis revealed potential mechanisms and pathways associated with functional and/or structural regression of the macaque CL. Quantitative real-time PCR was used to validate microarray expression patterns of 13 genes with the results being consistent between the two methodologies. Protein levels were found to parallel mRNA profiles in four of five differentially expressed genes analyzed by Western blot. Thus, this database will facilitate the identification of mechanisms involved in primate luteal regression.

Genes differentially expressed during spontaneous functional regression in the rhesus macaque corpus luteum are identified, which in turn will further our understanding of primate luteolysis.

The corpus luteum (CL) is a transient endocrine gland that forms from cells remaining in the follicle after ovulation. It is responsible for the synthesis of progesterone (P4) during the luteal phase of the menstrual cycle and early pregnancy. Timely regression of the CL (luteolysis) at the end of nonfertile cycles is necessary to allow the initiation of the next ovarian cycle, whereas luteolysis must be prevented if conception occurs. Therefore, understanding the factors controlling luteolysis is of considerable importance. The first event in luteolysis is a decrease in P4 synthesis (functional regression), which is then followed by the degradation and structural remodeling of the CL (structural regression) (1,2). Functional and structural regressions are distinct events, and it appears that functional regression is obligatory to allowing structural regression to proceed due to the luteoprotective effects of P4. However, the molecular processes responsible for initiation of functional regression and subsequent structural regression in the primate CL are poorly defined.

There is reduced sensitivity to LH as the CL ages (3,4,5), and increasing pulses of LH or a more potent LH-like signal [chorionic gonadotropin (CG)] are needed to maintain the primate CL by the end of the normal luteal phase (6). Thus, functional regression in the primate involves diminished P4 synthesis without significant decreases in LH or LH receptor levels (7). However, the mechanisms responsible for reduced LH sensitivity are not known. The absolute requirement of the primate CL for LH to maintain P4 synthesis (8) is unique because a large percentage of P4 production occurs independently of LH in domestic animal species (9), possibly due to an enhanced level of basal protein kinase A activity (9,10). Moreover, although it has been demonstrated that prostaglandin F2α (PGF2α) released from the uterus initiates functional regression in rodent and domestic animal species (8,11), the functional lifespan of the primate CL is not extended by hysterectomy (12,13). Numerous studies have demonstrated that intraluteal infusion of PGF2α causes luteolysis in monkeys and women (14,15,16,17,18,19), however, administration of a prostaglandin synthesis inhibitor to monkeys beginning at the midluteal phase inhibited P4 secretion and shorted CL lifespan rather than delaying luteolysis, as might have been predicted (20). Thus, the source and physiological role for PGF2α in spontaneous luteolysis of primates, if any, have not been determined (8,11). Collectively, these studies indicate that gonadotropin and prostaglandin-induced regulation of CL function and lifespan is very different between primate and nonprimate species, and studies to understand the control of luteolysis in primates are needed.

In addition to reduced responsiveness to LH, the resultant loss in P4 synthesis may further promote luteolytic events because accumulating data indicate that P4 itself mediates many processes necessary for maintaining the structure and function of the CL (2). Treatment of monkeys with the 3β-hydroxysteroid dehydrogenase (HSD3B) inhibitor trilostane during the midluteal phase inhibited P4 secretion and shortened CL lifespan, despite restoration of normal adrenal function after cessation of treatment (21). Another steroid ablation/progestin replacement study demonstrated that P4 mediates many of the effects of CG to rescue the CL from luteolysis during early pregnancy (22). Given that the first event in luteolysis is a cessation of P4 production and that the subsequent loss in P4-mediated signaling is likely a prerequisite for structural regression, we hypothesized that there are differences in gene expression just before and immediately after the onset of spontaneous functional regression that may represent proteins/pathways responsible for the loss of P4 producing potential in the primate CL. Previously, we identified differentially expressed genes occurring in CL collected during the early, mid, mid-late, late, and very late stages of the luteal phase of normal menstrual cycles (23). The late stage corresponds to the period when functional regression occurs. In this previous microarray study, some of the CL within this stage had not yet undergone functional regression, as indicated by high-serum levels of P4 at CL collection (>5 ng/ml in some cases). Therefore, we hypothesized that if late-stage CL were analyzed based on their functional state (as indicated by serum levels of P4), gene changes more closely associated with functional regression, and separate from downstream changes during structural regression, could be delineated. To perform the current study, additional CL were collected between d 14 and 16 (late stage) after the midcycle LH surge, and divided into nonregressed (P4 > 1.5 ng/ml) and functionally regressed (P4 < 0.5 ng/ml) groups (n = 4 CL per group). Total RNA was used for microarray analysis with Affymetrix rhesus macaque genome DNA microarrays, thereby allowing for the complete analysis of the macaque transcriptome. Using gene expression data from mid-late stage CL (d 10–12) as the baseline, changes in mRNA levels immediately preceding (high P4 late stage) and after (low P4 late stage) functional regression were determined.

Materials and Methods

CL collection and hormone analyses

All protocols involving animals were approved by the Institutional Animal Care and Use Committee. Regularly cycling female rhesus macaques between 5 and 10 yr of age were used for these studies. Animal care and tissue collection were performed as described previously (23). Starting 6 d after onset of menses, serum samples were assayed daily for estradiol and P4 concentrations by a specific electrochemoluminescent assay (IMMULITE 2000 (Siemens Healthcare Diagnostics, Deerfield, IL) in the Endocrine Services Laboratory at the Oregon National Primate Research Center (24). The first day of low-serum estradiol (<100 pg/ml) after the midcycle estradiol surge is denoted d 1 of the luteal phase (6). The CL were collected between d 10 and 12 (mid-late stage, n = 4 CL), or 14–16 (late stage, n = 8 CL) of the luteal phase. The late-stage CL were further divided into two groups (n = 4 CL per group) based on serum concentrations of P4. CL from animals with serum levels of P4 more than 1.5 ng/ml were considered nonregressed, and CL from animals with serum levels of P4 less than 0.5 ng/ml were considered to have already undergone functional regression. Serum levels of LH were determined by the Oregon National Primate Research Center Endocrine Services Laboratory using a specific RIA as previously described (25,26).

Microarray analysis

Total RNA was extracted using TRIzol reagent (Invitrogen Corp., Carlsbad, CA) according to the manufacturer’s instructions and further purified using RNeasy spin columns (QIAGEN, Inc., Valencia, CA). Final RNA concentrations and purity were determined by spectrophotometry, and the integrity of RNA samples was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Palo Alto, CA). Microarray assays using GeneChip Rhesus Macaque Genome Arrays (Affymetrix, Inc., Santa Clara, CA; > 47,000 transcripts represented) were performed by the Affymetrix Microarray Core of the Oregon Health & Science University Gene Microarray Shared Resource. A detailed description of microarray assay procedures and subsequent generation of processed image files has been reported previously (23). See supplemental data, which is published as supplemental data on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org, for all information necessary for Minimum Information About Microarray Experiments compliance. The data discussed in this publication have been deposited in the National Center for Biotechnology Information’s Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) and are accessible through Gene Expression Omnibus Series accession no. GSE12807.

The processed image files (.cel) were normalized across arrays using the robust multichip average algorithm (27) and log transformed (base 2), thus allowing direct comparison of probe set values between all samples used in the experiment. After normalization, GeneSifter (VizX Labs, Seattle, WA) microarray expression analysis software was used to identify differentially expressed transcripts. With the mid-late CL serving as the baseline reference, transcripts that exhibited a greater than 2-fold change (ANOVA, P < 0.05; Benjamini and Hochberg correction for false discovery rate) were considered differentially expressed. In pairwise comparisons, transcripts that were more than 2-fold different between groups (t test, P < 0.05) were considered differentially expressed. The GeneSifter microarray analysis software only determines whether there is a statistically significant change relative to the baseline but does not specify at which stage(s) this occurs. Therefore, to facilitate in-depth comparison of the microarray results, Affymetrix Expression Console version 1.0 software was used to generate raw data files (.chp) that were normalized with the robust multichip average algorithm and log (base 2) transformed. Selected transcripts were then analyzed by one-way ANOVA and pairwise differences determined with the Student-Newman-Keuls test (P < 0.05).

The list of differentially expressed genes generated by GeneSifter was divided into groups based on similarities in expression patterns. Ontology reports (molecular function) were generated from the various groups and assigned a z score by GeneSifter, with a score greater than 2.0 and a minimum of five genes being considered significant (28). Ontologies with more than 40 genes were considered too general. Significant ontologies were divided into primary and sub-ontologies. An ontology was defined as a sub-ontology if the majority of its genes were included in another significant ontology that had a greater total number of differentially expressed genes.

TaqMan quantitative PCR (Q-PCR)

Generation of cDNA and Q-PCR procedures were performed as described previously (23). Primer and TaqMan MGB probe sequences for the LH receptor (LHCGR), HSD3B Δ5/Δ4 isomerase 2 (HSD3B2), cytochrome P450, family 19, subfamily A, polypeptide 1 (CYP19A1), microsomal prostaglandin E2 (PGE2) synthase-1 (PTGES), PGF2α receptor (PTGFR), and hydroxyprostaglandin dehydrogenase 15-(NAD) (HPGD) have been reported (23). Forward and reverse primer, as well as MGB-probe sequences for the steroidogenic acute regulatory protein (STAR), low-density lipoprotein receptor (LDLR), scavenger receptor type B class 1 (SCARB1), ATP-binding cassette (ABC) family members A1 and G1, and apolipoprotein subfamily L (APOL) members 2 and 4 are provided in supplemental Table 1.

Relative concentrations of the target genes were normalized to 18s rRNA levels, and the ratios were log transformed before statistical analysis. Data were analyzed using one-way ANOVA, followed by pairwise comparisons with the Student-Newman-Keuls test, and differences were considered statistically significant at P < 0.05.

Western blot analysis

Whole cell homogenates were prepared from frozen (−80 C) CL as described previously (23). Antibodies against STAR (Affinity BioReagents, Inc., Golden, CO; catalog no. PA1-560), HSD3B2 and β-tubulin (TUBB) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA; catalog nos. sc-30821 and sc-9104, respectively), CYP19A1 (Novus Biologicals, Littleton, CO; catalog no. NB100-1596), and PTGES and HPGD (Cayman Chemical Co.; Ann Arbor, MI; catalog nos. 160140 and 160615, respectively) were purchased commercially.

Either 25 μg (HSD3B2), 30 μg (STAR, PTGES, and HPGD), or 40 μg (CYP19A1) of total protein was resolved on either 10% (HSD3B2) or 4–15% gradient Tris-HCl gels. Western blot procedures were similar to those described previously (23). Antibodies were used at the following concentrations: anti-STAR, 0.65 μg/ml; anti-HSD3B2, 0.3 μg/ml; anti-CYP19A1, 0.8 μg/ml; anti-PTGES, 0.4 μg/ml; and anti-HPGD, 2 μg/ml. Specificity was determined by estimating molecular mass via comparison of bands with molecular mass markers, as well as preventing band detection by preabsorbing the primary antibody with immunizing peptide if available (HSD3B2, PTGES, and HPGD) (23,29), or exclusion of primary antibody if no peptide was available (CYP19A1). The specificity of the anti-STAR antibody has been determined previously (10). After detection of target protein, membranes were reprobed for TUBB (0.2 μg/ml) to serve as a loading control.

Films were scanned, and densitometry analysis was performed using Quantity One version 4.3.1 software (Bio-Rad Laboratories, Inc., Hercules, CA). The background-adjusted volume of each band was normalized to TUBB, and data were log transformed if necessary for statistical normalization. The ratio of target protein to TUBB for each stage was analyzed by one-way ANOVA, followed by pairwise comparison using the Student-Newman-Keuls test with differences considered significant at P < 0.05.

Results

Serum concentrations of P4 and LH

There was no significant difference between mid-late and functional late CL in serum concentrations of P4 (4.4 ± 0.6 and 5.4 ± 1.2 ng/ml, respectively). However, functionally regressed late CL had significantly (P < 0.05) lower serum concentrations of P4 (0.4 ± 0.1 ng/ml) compared with all other groups. There were no significant differences between groups in serum levels of LH (1.0 ± 0.5 ng/ml mid-late; 2.1 ± 0.7 ng/ml functional late; and 0.7 ± 0.2 ng/ml functionally regressed late), indicating that the loss of P4 secretion in functionally regressed late CL is not due to differences in LH levels at CL collection.

Identification of differentially expressed genes throughout functional regression of the macaque CL

From the mid-late through functionally regressed late stages, 681 transcripts met the criteria for differential expression (>2-fold change; ANOVA, P < 0.05). After performing pairwise comparisons, 305 transcripts (>2-fold change; t test, P < 0.05) were differentially expressed between mid-late and functional late CL, of which 233 were increased, and 72 were decreased in expression. There were 314 differentially expressed transcripts in the pairwise comparison between functional and functionally regressed late-stage CL, with 102 increasing and 212 decreasing in expression.

Cluster and scatter-plot analysis of the three experimental groups indicated that the mid-late and functional late CL were the most similar in terms of gene expression profiles, with the functionally regressed late CL being the most unrelated (supplemental Fig. 1). There were three primary patterns of expression among the 681 transcripts that were differentially expressed from the mid-late through functionally regressed late stages: 1) genes whose mRNA levels increased sequentially from the mid-late through the functionally regressed late stage (382 transcripts); 2) transcripts that decreased sequentially from the mid-late through the functionally regressed late stage (218 transcripts); and 3) genes whose mRNA levels increased from mid-late to nonregressed late CL, then decreased in functionally regressed late CL (81 transcripts) (Fig. 1).

Figure 1.

Figure 1

Open in a new tab

Three general patterns of expression among differentially expressed genes. The 681 differentially expressed transcripts from the mid-late (ML) through functionally regressed late (FRL) stages were divided based on general expression patterns as depicted in these three heat maps. Group 1 includes transcripts that underwent an incremental increase in expression from the mid-late through functionally regressed late stages (382 transcripts). Group 2 includes transcripts that decreased in expression from the mid-late through functionally regressed late stages (218 transcripts). Group 3 includes transcripts that increased from the mid-late to functional late (FL) stages, then decreased in expression by the functionally regressed late stage (81 transcripts). Max, Maximum; Min, minimum.

Molecular function ontology reports were generated for each group of genes (based on expression pattern) using GeneSifter software (supplemental Table 2). For the sequentially increasing transcripts (group 1), there were three primary ontologies (transcription factor activity, GTP binding, molecular adaptor activity) and five sub-ontologies that were significant. There were three primary ontologies (oxidoreductase activity, cytoskeletal protein binding, protein serine/threonine kinase activity) and four sub-ontologies that were significant among the sequentially decreasing gene group (group 2). In the transcripts that increased in expression from the mid-late to nonregressed late CL, followed by a decrease in functionally regressed late CL (group 3), one primary (catalytic activity) and two sub-ontologies were identified (supplemental Table 2).

Differentially expressed genes found by pairwise comparison of mid-late vs. functional late, or functional late vs. functionally regressed late stages were also analyzed for significant ontologies (supplemental Table 3). In the mid-late vs. functional late pairwise comparison, differentially expressed transcripts that increased in expression outnumbered those that decreased in expression by a more than 3:1 ratio. Of those transcripts that increased in expression, there were three primary ontologies (hydrolase activity, molecular/signal transducer activity, receptor binding) and nine sub-ontologies that were significant among these genes. In the functional late vs. functionally regressed late pairwise comparison, differentially expressed transcripts that decreased in expression outnumbered those that increased in expression by a more than 2:1 ratio. For those transcripts that decreased in expression, there were two primary ontologies (oxidoreductase activity, cytoskeletal protein binding) and two sub-ontologies that were significant among these genes.

Validation of microarray data

A total of 13 genes was selected for quantification of mRNA levels by Q-PCR to validate the microarray results. Because the late-stage CL were divided based on whether or not functional regression had occurred (serum levels of P4 at CL collection), genes encoding steroidogenic proteins were singled out for further analysis. There were five genes that met the criteria for differential expression from the mid-late through functionally regressed late stages whose gene products are involved in LH-stimulated steroidogenesis (P4 or estradiol) (Fig. 2A). Of these, four were validated with regard to mRNA levels by Q-PCR (Fig. 2B). Statistically significant differences between groups were identical as determined by both microarray and Q-PCR methodologies for all four genes. With the exception of HSD3B2, mRNA levels for all of the differentially expressed genes related to LH-stimulated steroidogenesis did not change significantly from the mid-late to functional late stage but had a significant decrease in expression from the functional late to functionally regressed late CL groups as determined by microarray. For HSD3B2 gene expression, all three experimental groups were significantly (P < 0.05) different from one another.

Figure 2.

Figure 2

Open in a new tab

Steroidogenic gene expression throughout the period of functional regression in the rhesus macaque CL. A, Genes involved in steroidogenesis are displayed in the heat map. Bold lettering indicates that mRNA for the corresponding gene met the criteria for differential expression (>2-fold change; ANOVA, P < 0.05). B, Q-PCR validation of microarray results for LHCGR, STAR, HSD3B2, and CYP19A1. Data were analyzed by ANOVA followed by comparison between groups using the Student-Newman-Keuls test. Uppercase letters denote significant differences (P < 0.05) for microarray data, and lowercase letters denote significant differences for Q-PCR data. FL, Functional late; FRL, functionally regressed late; Max, maximum; Min, minimum; ML, mid-late.

Several differentially expressed genes encode proteins involved in cholesterol uptake, transport, and efflux. Six of these were selected for validation of microarray results by Q-PCR (Fig. 3). Q-PCR results were highly similar to the microarray findings in almost all cases. The primary exception was LDLR whose mRNA levels did not change from the mid-late to functional late stage as determined by microarray, whereas Q-PCR detected a significant (P < 0.05) 2-fold decrease from the mid-late to functional late CL. The LDLR and SCARB1 lipoprotein receptors both had their lowest mRNA levels in functionally regressed late CL (≥3- and 4-fold decreases, respectively, as determined by microarray and Q-PCR). In contrast, significant (P < 0.05) increases in mRNA levels were observed for ABCA1 and ABCG1, proteins that are involved in cholesterol efflux (30,31), with their highest levels observed in functionally regressed late CL (≥ 4-fold total increases for both genes) (Fig. 3). In addition, APOL2 and APOL4 mRNA levels significantly (P < 0.05) increased from mid-late to functional late CL (≥2 and 3-fold increases, respectively).

Figure 3.

Figure 3

Open in a new tab

Divergent expression of cholesterol uptake, transport, and efflux genes. A, Genes involved in cholesterol uptake, transport, or efflux are displayed in the heat map. Bold lettering indicates that mRNA for the corresponding gene met the criteria for differential expression (>2-fold change; ANOVA, P < 0.05). B, Q-PCR validation of microarray results for LDLR, SCARB1, APOL2, APOL4, ABCA1, and ABCG1. Data were analyzed by ANOVA followed by comparison between groups using the Student-Newman-Keuls test. Uppercase letters denote significant differences (P < 0.05) for microarray data, and lowercase letters denote significant differences for Q-PCR data. FL, Functional late; FRL, functionally regressed late; Max, maximum; Min, minimum; ML, mid-late.

Due to the known role of prostaglandins in initiating luteolysis in many nonprimate species, we searched for differentially expressed genes that encode proteins involved in prostaglandin biosynthesis, metabolism, and signaling. Two genes, HPGD and PTGES, were among the 681 differentially expressed transcripts from the mid-late through functionally regressed late stages. Moreover, by pairwise comparison, PTGFR also increased in expression between mid-late and functional-late stages, then decreased from functional to functionally regressed late stages. The mRNA levels of HPGD, PTGES, and PTGFR were quantified by Q-PCR (Fig. 4). Both microarray and Q-PCR detected a significant (P < 0.05) 3-fold or more decrease in HPGD mRNA from functional to functionally regressed late CL, whereas there was a significant (P < 0.05) 2-fold or more decrease in PTGES mRNA from the mid-late to functional-late CL. There was a significant (P < 0.05) 3-fold or more decrease in PTGFR mRNA from functional to functionally regressed late CL, as determined by both methodologies (Fig. 4).

Figure 4.

Figure 4

Open in a new tab

Q-PCR validation of microarray expression data for differentially expressed prostaglandin synthesis, metabolism, and signaling genes. Quantification of mRNA levels for HPGD (top panel), PTGES (middle panel), and PTGFR (bottom panel) by microarray and Q-PCR methodologies. Data were analyzed by ANOVA followed by comparison between groups using the Student-Newman-Keuls test. Uppercase letters denote significant differences (P < 0.05) for microarray data, and lowercase letters denote significant differences for Q-PCR data.

Corresponding protein levels of differentially expressed mRNAs

The corresponding protein levels for each of five differentially expressed mRNAs were determined by Western blot analyses. Among the steroidogenic proteins, levels of STAR, HSD3B2, and CYP19A1 were determined (Fig. 5). Concentrations of the mature form of STAR (30 kDa) were not different between mid-late and functional late CL; however, there was a significant (P < 0.05) 4.6-fold decrease in STAR in functionally regressed late CL. All experimental groups were significantly different (P < 0.05) from one another with regard to HSD3B2 protein. There was a 1.9-fold decrease from mid-late to functional late CL, with HSD3B2 levels being nearly nondetectable in functionally regressed late CL. There was no significant difference in CYP19A1 between mid-late and functional late CL, however, there was a significant (P < 0.05) 5.4-fold decrease in CYP19A1 levels in functionally regressed compared with functional-late CL (Fig. 5).

Figure 5.

Figure 5

Open in a new tab

Differential expression of steroidogenic proteins. A, Western blots using samples pooled from CL collected at each stage. The bottom image is TUBB, which served as a loading control. In the HSD3B2 image, only the lower band of the doublet corresponds to HSD3B2 as determined previously (23). B, Levels of STAR, HSD3B2, and CYP19A1 from individual CL (n = 4 per stage) were normalized to TUBB, and the resultant ratio was analyzed by ANOVA followed by comparison between groups using the Student-Newman-Keuls test. Columns with different letters are significantly different (P < 0.05). FL, Functional late; FRL, functionally regressed late; ML, mid-late.

There were two bands (∼25 and 33 kDa) detected by the anti-HPGD antibody that may represent two isoforms of HPGD (29 kDa predicted molecular mass) because both bands disappeared upon preabsorption of the antibody with its immunizing peptide (data not shown). Levels of either band did not change significantly across groups (Fig. 6). Levels of PTGES were significantly (P < 0.05) higher in mid-late CL than either functional or functionally regressed late CL (minimum 3.3-fold decrease).

Figure 6.

Figure 6

Open in a new tab

Differential expression of proteins involved in prostaglandin E2 synthesis and metabolism. A, Western blots using samples pooled from CL collected at each stage. The bottom image is TUBB, which served as a loading control. B, Levels of HPGD and PTGES from individual CL (n = 4 per stage) were normalized to TUBB, and the resultant ratio was analyzed by ANOVA followed by comparison between groups using the Student-Newman-Keuls test. Columns with different letters are significantly different (P < 0.05). FL, Functional late; FRL, functionally regressed late; ML, mid-late.

Discussion

The current study details the generation of a database containing gene expression changes occurring immediately before and after the onset of functional regression in the rhesus macaque CL. Luteolysis is arguably the most poorly understood process associated with CL life span in primates. Functional regression appears to be the key event in this process because CG, which is essential for establishing pregnancy in primates, stimulates P4 synthesis via the LHCGR (8,11), steroid ablation prevents CG-mediated rescue of the CL during simulated early pregnancy (22), and numerous luteotrophic or antiluteolytic processes that may be modulated specifically by P4 have been reported (2). With regard to the latter, P4 suppresses expression of certain proteases (32), and inhibits immune-mediated luteolytic events (33). As a result of our initial analysis of differential gene expression throughout the normal luteal phase (23), it was clear that the most dynamic changes in gene expression were associated with the period encompassing functional and structural regression (late to very late-stage transition). Thus, because functional regression is likely the key event in luteolysis, it became apparent that a more-detailed analysis of gene expression profiles uniquely associated with this process was necessary.

As expected, dynamic changes in gene expression were observed throughout spontaneous functional regression of the macaque CL. There were three general patterns of expression found among differentially expressed transcripts occurring from the mid-late through functionally regressed late stages, including: 1) genes whose mRNA levels increased sequentially from the mid-late through the functionally regressed late stage; 2) transcripts that decreased sequentially from the mid-late through the functionally regressed late stage; and 3) genes whose mRNA levels increased from mid-late to nonregressed late CL, then decreased in functionally regressed late CL. We predict that transcripts increasing in expression from the mid-late to functional-late CL represent genes that encode pro-luteolytic proteins. On the other hand, transcripts that decrease in expression from the mid-late to functional late luteal phase are predicted to encode luteotrophic proteins necessary for CL function.

Overlapping ontologies among differentially expressed genes revealed mechanisms or cellular processes that potentially facilitate luteolysis of the primate CL. The transcription factor activity ontology, as well as transcriptional repressor activity and transcriptional corepressor activity sub-ontologies were significant among genes that increased in expression (supplemental Table 2). These findings indicate that there may be a general suppression of transcription after functional regression of the primate CL. This is consistent with our previous microarray findings that there was more differential gene expression during the late to very late (CL collected a few days after functional regression occurred, d 18–19 after LH surge) stage transition than between all earlier stages of the luteal phase combined, with the majority of the genes in the late to very late stage transition exhibiting decreased mRNA levels (23). In addition, when performing pairwise comparisons, we found that the majority of differentially expressed transcripts increase in expression from the mid-late to functional late stages. In contrast, the majority of differentially expressed transcripts from the functional to functionally regressed late CL decreased in expression. This indicates that there is an initial increase in expression of what are most likely pro-luteolytic genes before the onset of functional regression, followed by a suppressive pattern of global gene expression after functional regression occurs. This is further evidence that functional regression is the key event in luteolysis because the predominant suppression of gene transcription from the late to very late stages that was noted in our previous study (23) may have been a consequence of the initial increase in expression of genes associated with transcriptional repression as observed in this study.

Another significant ontology among genes with increased levels of expression from the mid-late to late luteal phase (supplemental Table 2) included molecular adaptor activity (z score = 5.26). Genes in this ontology facilitate signaling from cell surface receptors by controlling receptor trafficking (e.g. trafficking protein, kinesin binding 1) (34) or by serving as important structural and functional constituents of large signaling complexes (e.g. tyrosine kinase receptors). For example, Grb2-associated binding protein-1 has been identified as a docking protein in epidermal growth factor and insulin receptor signaling (35), whereas chimerin 2 (36) and intersectin 2 (37) may serve as guanine nucleotide exchange factors that activate Ras-related guanosine 5′-triphosphatases (GTPases). Interestingly, the third primary ontology that was significant among genes increasing in expression from the mid-late to late luteal phase was GTP binding, which had a related sub-ontology of GTPase activity (z score = 4.09). Most of these genes are members of, or are related to, the RAS family, which are proteins that transduce signals from different tyrosine kinase activating receptors. Collectively, the increase in expression of genes associated with molecular adaptor and GTP binding ontologies may facilitate enhanced signaling from certain receptors that are necessary for luteolysis of the primate CL.

Among the three primary ontologies that were significant in the list of differentially expressed genes that decreased in expression from the mid-late to functionally regressed late CL (supplemental Table 2), oxidoreductase activity had the highest z score and the greatest total number of genes. Several of the genes in this ontology belong to the cytochrome P450 family. Two are associated with P4 and estradiol synthesis (HSD3B2 and CYP19A1), as well as biosynthesis of their cholesterol precursor (HMGCR, SQLE). Thus, as expected, several genes associated with steroid or cholesterol biosynthesis decrease in expression during functional regression.

For the genes that increased in expression from the mid-late to functional late stages and then decreased in expression in the functionally regressed late stage, catalytic activity was the only significant primary ontology identified, with hydrolase and peptidase activity sub-ontologies also being noted (supplemental Table 2). It may not be surprising that an increase in expression of hydrolytic and proteolytic genes occurs before functional regression, however, the most-likely explanation of the role they may play in functional regression is not immediately clear. Because some of their gene products can be extracellularly located (e.g. ADAMTS14) (38,39), there may be modification or degradation of cell surface receptors, extracellular ligands, or extracellular matrix constituents that are necessary for maintenance of CL function. Other genes in this group encode proteins that may be involved in immune-related processes that occur during luteolysis. For example, endoplasmic reticulum aminopeptidase 1 and 2 proteins act in a concerted mechanism to cleave peptides for subsequent presentation by major histocompatibility complex type-1 molecules (40). In any case, it seems logical that hydrolytic and proteolytic gene expression is consistent with a pro-luteolytic function.

When ontology analysis was performed on differentially expressed genes identified through pairwise comparisons, many ontologies were the same as those identified through the multiple comparison procedure, although they were not always due to differential expression of the same genes (e.g. GTPase activity, supplemental Tables 2 and 3). A few novel ontologies were discovered through pairwise comparisons (e.g. molecular/signal transducer activity and receptor binding activity), and most of the genes associated with these ontologies were related to immune function. Our finding that an increase in expression of many immune function-related genes occurs before the decrease in P4 synthesis supports the idea that the immune system plays an active role in causing functional regression of the primate CL, similar to reports from nonprimate species (33).

It was not surprising that several genes related to LH-stimulated steroidogenesis were differentially expressed. In general, expression of steroidogenic genes closely paralleled serum levels of P4. All of the differentially expressed steroidogenic genes decreased in expression, and in most cases involved a dramatic decrease from functional to functionally regressed late CL with no difference in mRNA levels between mid-late and functional late CL. Corresponding changes in protein levels were also detected for STAR, HSD3B2, and CYP19A1. Some of these changes are similar to what has been previously reported in primates. In humans, STAR and CYP19A1 mRNA expression is decreased during the late luteal phase (7). In the current study, STAR and CYP19A1 mRNA and protein decreased in functionally regressed late CL, but not functional late CL. In monkeys, functional regression induced by either PGF2α infusion or LH withdrawal causes decreases in LHCGR and HSD3B2 mRNA and protein (41), which is consistent with our findings during spontaneous functional regression.

Decreases in steroidogenic protein levels may be responsible for the reduced sensitivity to LH that is characteristic of the regressing primate CL. However, the factor(s) inhibiting steroidogenic gene expression is not known, and it is unclear whether these decreases are causative or an effect of functional regression. It has previously been reported that after LH withdrawal in the primate, decreases in luteal mRNA levels for steroidogenic enzymes do not occur until 24 h after the decline in P4 synthesis (42). Another possible mechanism underlying the loss of P4 production in the primate CL may involve a restriction of cholesterol availability through inhibited lipoprotein uptake (LDLR, SCARB1), decreased de novo synthesis of cholesterol (HMGCR, SQLE), and increased cholesterol efflux from the CL (ABCA1, ABCG1).

In summary, the current study provides a comprehensive analysis of gene expression across the entire macaque transcriptome during functional regression of the CL. Microarray results were compared with Q-PCR data, and the corresponding protein levels were determined by Western blot for selected genes. These procedures validated the microarray data because only minor differences were noted (e.g. LDLR Q-PCR results, HPGD Western blot results). The process of luteolysis in primates is poorly understood. Therefore, further research to delineate how luteolysis occurs in the primate is needed, and this publicly accessible database (http://www.ncbi.nlm.nih.gov/geo/; accession no. GSE12807) can be a valuable resource for these investigations.

Supplementary Material

[Supplemental Data]
en.2008-1201_index.html (2.1KB, html)

Acknowledgments

We thank the following core facilities at the Oregon National Primate Research Center: Endocrine Services (Dr. Francis Pau), Molecular and Cellular Biology (Drs. Eliot Spindel and Yibing Jia), and the Division of Animal Resources. We also thank the Oregon Health and Science University Affymetrix Microarray Core (Dr. Chris Harrington) for help with the microarray studies.

Footnotes

This research was supported by National Institutes of Health Grants U54 HD55744 (to J.D.H.), R01 HD42000 (to J.D.H.), RR00163 (to J.D.H.), and T32 Training Grant HD007133 (to R.L.B.).

Disclosure Statement: The authors have nothing to disclose.

First Published Online October 23, 2008

Abbreviations: ABC, ATP-binding cassette; APOL, apolipoprotein subfamily L; CG, chorionic gonadotropin; CL, corpus luteum/corpora lutea; FRL, functionally regressed late; GTPase, guanosine 5′-triphosphatase; HPGD, hydroxyprostaglandin dehydrogenase 15-(NAD); HSD3B, 3β-hydroxysteroid dehydrogenase; LDLR, low-density lipoprotein receptor; ML, mid-late; P4, progesterone; PGF2α, prostaglandin F2α; Q-PCR, quantitative PCR; SCARB1, scavenger receptor type B class 1; STAR, steroidogenic acute regulatory protein; TUBB, β-tubulin.

References

  1. Hoyer PB 1998 Regulation of luteal regression: the ewe as a model. J Soc Gynecol Investig 5:49–57 [DOI] [PubMed] [Google Scholar]
  2. Stouffer RL 2003 Progesterone as a mediator of gonadotrophin action in the corpus luteum: beyond steroidogenesis. Hum Reprod Update 9:99–117 [DOI] [PubMed] [Google Scholar]
  3. Brannian JD, Stouffer RL 1991 Progesterone production by monkey luteal cell subpopulations at different stages of the menstrual cycle: changes in agonist responsiveness. Biol Reprod 44:141–149 [DOI] [PubMed] [Google Scholar]
  4. Cameron JL, Stouffer RL 1982 Gonadotropin receptors of the primate corpus luteum. II. Changes in available luteinizing hormone- and chorionic gonadotropin-binding sites in macaque luteal membranes during the nonfertile menstrual cycle. Endocrinology 110:2068–2073 [DOI] [PubMed] [Google Scholar]
  5. Eyster KM, Ottobre JS, Stouffer RL 1985 Adenylate cyclase in the corpus luteum of the rhesus monkey. III. Changes in basal and gonadotropin-sensitive activities during the luteal phase of the menstrual cycle. Endocrinology 117:1571–1577 [DOI] [PubMed] [Google Scholar]
  6. Duffy DM, Stewart DR, Stouffer RL 1999 Titrating luteinizing hormone replacement to sustain the structure and function of the corpus luteum after gonadotropin-releasing hormone antagonist treatment in rhesus monkeys. J Clin Endocrinol Metab 84:342–349 [DOI] [PubMed] [Google Scholar]
  7. Devoto L, Kohen P, Vega M, Castro O, Gonzalez RR, Retamales I, Carvallo P, Christenson LK, Strauss JF 2002 Control of human luteal steroidogenesis. Mol Cell Endocrinol 186:137–141 [DOI] [PubMed] [Google Scholar]
  8. Stouffer RL 2006 Structure, function, and regulation of the corpus luteum. In: Knobil E, Neill JD, eds. The physiology of reproduction. 3rd ed. New York: Raven Press; 475–526 [Google Scholar]
  9. Diaz FJ, Anderson LE, Wu YL, Rabot A, Tsai SJ, Wiltbank MC 2002 Regulation of progesterone and prostaglandin F2α production in the CL. Mol Cell Endocrinol 191:65–80 [DOI] [PubMed] [Google Scholar]
  10. Bogan RL, Niswender GD 2007 Constitutive steroidogenesis in ovine large luteal cells may be mediated by tonically active protein kinase A. Biol Reprod 77:209–216 [DOI] [PubMed] [Google Scholar]
  11. Niswender GD, Juengel JL, Silva PJ, Rollyson MK, McIntush EW 2000 Mechanisms controlling the function and life span of the corpus luteum. Physiol Rev 80:1–29 [DOI] [PubMed] [Google Scholar]
  12. Castracane VD, Moore GT, Shaikh AA 1979 Ovarian function of hysterectomized Macaca fascicularis. Biol Reprod 20:462–472 [DOI] [PubMed] [Google Scholar]
  13. Ranney B, Abu-Ghazaleh S 1977 The future function and fortune of ovarian tissue which is retained in vivo during hysterectomy. Am J Obstet Gynecol 128:626–634 [DOI] [PubMed] [Google Scholar]
  14. Zelinski-Wooten MB, Stouffer RL 1990 Intraluteal infusions of prostaglandins of the E, D, I, and A series prevent PGF2 α-induced, but not spontaneous, luteal regression in rhesus monkeys. Biol Reprod 43:507–516 [DOI] [PubMed] [Google Scholar]
  15. Auletta FJ, Speroff L, Caldwell BV 1973 Prostaglandin F-2 induced steroidogenesis and luteolysis in the primate corpus luteum. J Clin Endocrinol Metab 36:405–407 [DOI] [PubMed] [Google Scholar]
  16. Auletta FJ, Kamps DL, Pories S, Bisset J, Gibson M 1984 An intra-corpus luteum site for the luteolytic action of prostaglandin F2 α in the rhesus monkey. Prostaglandins 27:285–298 [DOI] [PubMed] [Google Scholar]
  17. Wentz AC, Jones GS 1973 Transient luteolytic effect of prostaglandin F2α in the human. Obstet Gynecol 42:172–181 [PubMed] [Google Scholar]
  18. Kirton KT, Pharriss BB, Forbes AD 1970 Luteolytic effects of prostaglandin F2a in primates. Proc Soc Exp Biol Med 133:314–316 [DOI] [PubMed] [Google Scholar]
  19. Bennegard B, Hahlin M, Wennberg E, Noren H 1991 Local luteolytic effect of prostaglandin F2 α in the human corpus luteum. Fertil Steril 56:1070–1076 [PubMed] [Google Scholar]
  20. Sargent EL, Baughman WL, Novy MJ, Stouffer RL 1988 Intraluteal infusion of a prostaglandin synthesis inhibitor, sodium meclofenamate, causes premature luteolysis in rhesus monkeys. Endocrinology 123:2261–2269 [DOI] [PubMed] [Google Scholar]
  21. Duffy DM, Hess DL, Stouffer RL 1994 Acute administration of a 3 β-hydroxysteroid dehydrogenase inhibitor to rhesus monkeys at the midluteal phase of the menstrual cycle: evidence for possible autocrine regulation of the primate corpus luteum by progesterone. J Clin Endocrinol Metab 79:1587–1594 [DOI] [PubMed] [Google Scholar]
  22. Duffy DM, Stouffer RL 1997 Gonadotropin versus steroid regulation of the corpus luteum of the rhesus monkey during simulated early pregnancy. Biol Reprod 57:1451–1460 [DOI] [PubMed] [Google Scholar]
  23. Bogan RL, Murphy MJ, Stouffer RL, Hennebold JD 2008 Systematic determination of differential gene expression in the primate corpus luteum during the luteal phase of the menstrual cycle. Mol Endocrinol 22:1260–1273 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Young KA, Chaffin CL, Molskness TA, Stouffer RL 2003 Controlled ovulation of the dominant follicle: a critical role for LH in the late follicular phase of the menstrual cycle. Hum Reprod 18:2257–2263 [DOI] [PubMed] [Google Scholar]
  25. Zelinski-Wooten MB, Slayden OD, Chwalisz K, Hess DL, Brenner RM, Stouffer RL 1998 Chronic treatment of female rhesus monkeys with low doses of the antiprogestin ZK 137 316: establishment of a regimen that permits normal menstrual cyclicity. Hum Reprod 13:259–267 [DOI] [PubMed] [Google Scholar]
  26. Molskness TA, Hess DL, Maginnis GM, Wright JW, Fanton JW, Stouffer RL 2007 Characteristics and regulation of the ovarian cycle in vervet monkeys (Chlorocebus aethiops). Am J Primatol 69:890–900 [DOI] [PubMed] [Google Scholar]
  27. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP 2003 Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4:249–264 [DOI] [PubMed] [Google Scholar]
  28. Doniger SW, Salomonis N, Dahlquist KD, Vranizan K, Lawlor SC, Conklin BR 2003 MAPPFinder: using Gene Ontology and GenMAPP to create a global gene-expression profile from microarray data. Genome Biol 4:R7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Bogan RL, Murphy MJ, Stouffer RL, Hennebold JD 2008 Prostaglandin synthesis, metabolism, and signaling potential in the rhesus macaque corpus luteum throughout the luteal phase of the menstrual cycle. Endocrinology 149:5861–5871 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Schmitz G, Langmann T, Heimerl S 2001 Role of ABCG1 and other ABCG family members in lipid metabolism. J Lipid Res 42:1513–1520 [PubMed] [Google Scholar]
  31. Kaminski WE, Piehler A, Wenzel JJ 2006 ABC A-subfamily transporters: structure, function and disease. Biochim Biophys Acta 1762:510–524 [DOI] [PubMed] [Google Scholar]
  32. Young KA, Stouffer RL 2004 Gonadotropin and steroid regulation of matrix metalloproteinases and their endogenous tissue inhibitors in the developed corpus luteum of the rhesus monkey during the menstrual cycle. Biol Reprod 70:244–252 [DOI] [PubMed] [Google Scholar]
  33. Pate JL, Landis Keyes P 2001 Immune cells in the corpus luteum: friends or foes? Reproduction 122:665–676 [DOI] [PubMed] [Google Scholar]
  34. Gilbert SL, Zhang L, Forster ML, Anderson JR, Iwase T, Soliven B, Donahue LR, Sweet HO, Bronson RT, Davisson MT, Wollmann RL, Lahn BT 2006 Trak1 mutation disrupts GABA(A) receptor homeostasis in hypertonic mice. Nat Genet [Erratum (2006) 38:389] 38:245–250 [DOI] [PubMed] [Google Scholar]
  35. Holgado-Madruga M, Emlet DR, Moscatello DK, Godwin AK, Wong AJ 1996 A Grb2-associated docking protein in EGF- and insulin-receptor signalling. Nature 379:560–564 [DOI] [PubMed] [Google Scholar]
  36. Leung T, How BE, Manser E, Lim L 1994 Cerebellar β 2-chimaerin, a GTPase-activating protein for p21 ras-related rac is specifically expressed in granule cells and has a unique N-terminal SH2 domain. J Biol Chem 269:12888–12892 [PubMed] [Google Scholar]
  37. Pucharcos C, Estivill X, de la Luna S 2000 Intersectin 2, a new multimodular protein involved in clathrin-mediated endocytosis. FEBS Lett 478:43–51 [DOI] [PubMed] [Google Scholar]
  38. Colige A, Vandenberghe I, Thiry M, Lambert CA, Van Beeumen J, Li SW, Prockop DJ, Lapiere CM, Nusgens BV 2002 Cloning and characterization of ADAMTS-14, a novel ADAMTS displaying high homology with ADAMTS-2 and ADAMTS-3. J Biol Chem 277:5756–5766 [DOI] [PubMed] [Google Scholar]
  39. Bolz H, Ramirez A, von Brederlow B, Kubisch C 2001 Characterization of ADAMTS14, a novel member of the ADAMTS metalloproteinase family. Biochim Biophys Acta 1522:221–225 [DOI] [PubMed] [Google Scholar]
  40. Saveanu L, Carroll O, Lindo V, Del Val M, Lopez D, Lepelletier Y, Greer F, Schomburg L, Fruci D, Niedermann G, Van Endert PM 2005 Concerted peptide trimming by human ERAP1 and ERAP2 aminopeptidase complexes in the endoplasmic reticulum. Nat Immunol 6:689–697 [DOI] [PubMed] [Google Scholar]
  41. Duncan WC, Illingworth PJ, Young FM, Fraser HM 1998 Induced luteolysis in the primate: rapid loss of luteinizing hormone receptors. Hum Reprod 13:2532–2540 [DOI] [PubMed] [Google Scholar]
  42. Benyo DF, Ravindranath N, Bassett S, Hutchison J, Zeleznik AJ 1993 Cellular aspects of corpus luteum function in the primate. Hum Reprod 8(Suppl 2):102–106 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

[Supplemental Data]
en.2008-1201_index.html (2.1KB, html)
en.2008-1201_1.pdf (81KB, pdf)
en.2008-1201_2.pdf (7.5KB, pdf)
en.2008-1201_3.pdf (99.7KB, pdf)
en.2008-1201_4.pdf (153.6KB, pdf)
en.2008-1201_5.pdf (84.4KB, pdf)
en.2008-1201_ML_to_FRL_multiplecomparison.xls (199KB, xls)
en.2008-1201_ML_vs_FL_pairwise.xls (88KB, xls)
en.2008-1201_FL_vs_FRL_pairwise.xls (89KB, xls)

Articles from Endocrinology are provided here courtesy of The Endocrine Society

RESOURCES