Systematic determination of differential gene expression in the primate corpus luteum during the luteal phase of the menstrual cycle

Abstract

The molecular and cellular processes required for development, function, and regression of the primate corpus luteum (CL) are poorly defined. We hypothesized that there are dynamic changes in gene expression occurring during the CL life span, which represent proteins and pathways critical to its regulation. Therefore, a genomic approach was utilized to systematically identify differentially expressed genes in the rhesus macaque CL during the luteal phase of natural menstrual cycles. CL were collected between d 3-5 (early stage), d 7-8 (mid), d 10-12 (mid-late), d 14-16 (late), or d 18-19 (very-late) after the midcycle LH surge. From the early through very-late stages, 3234 transcripts were differentially expressed, with 879 occurring from the early through late stages that encompass the processes of luteinization, maintenance, and functional regression. To characterize gene changes most relevant to these processes, ontology analysis was performed using the list of 879 differentially expressed transcripts. Four main groups of related genes were identified with relevance to luteal physiology including: 1) immune function; 2) hormone and growth factor signaling; 3) steroidogenesis; and 4) prostaglandin biosynthesis, metabolism, and signaling. A subset of genes representing each of the four major categories was selected for validation of microarray results by quantitative real-time PCR. Results in mRNA levels were similar between the two methodologies for 17 of 18 genes. Additionally, protein levels for three genes were determined by Western blot analysis to parallel mRNA levels. This database will facilitate the identification of many novel or previously underappreciated pathways that regulate the structure and function of the primate CL.

Figure 1

Patterns of Gene Expression in the Rhesus Macaque CL during the Luteal Phase of the Natural Menstrual Cycle Panel A is a dendrogram that illustrates the relatedness between CL stages in terms of gene expression. The more branching that occurs, the more unrelated groups are. The scale bar is an arbitrary measurement of Euclidean distance. Panel B is a heat map for all the probe sets from the early through late stages that were differentially expressed using P < 0.01 (>2-fold change; ANOVA). Each column represents a stage of the luteal phase, and each row is a probe set corresponding to an individual transcript. Arrows indicate the approximate point where the pattern of gene expression changes. Less common patterns are observed in regions 1 and 3. Region 2 contains probe sets that demonstrated a relatively steady increase from the early through late stages, and region 4 contains the probe sets that had a relatively steady decrease in expression from the early through late stages of the luteal phase. Min, Minimum; Max, maximum.

Figure 2

Differentially Expressed Immune Function Genes in Rhesus Macaque CL through the Luteal Phase Panel A contains a heat map of representative immune function genes identified as differentially expressed (>2-fold change; ANOVA, P < 0.01). Each column is a stage of the luteal phase, and each row represents a single gene. All genes represented on this heat map were identified as differentially expressed using data from the early through late stages only (indicated by the bracket above the columns), although their corresponding expression levels in the very-late stage are shown as well. In panel B, the validation of microarray results by Q-PCR is displayed for three immune function genes. Microarray results are in black columns on the left axis, and Q-PCR results are in white columns on the right axis. Uppercase letters denote significant differences for microarray data (P < 0.05), and lowercase letters denote significant differences for Q-PCR data. Min, Minimum; Max, maximum.

Figure 3

Differentially Expressed Hormone and Growth Factor Signaling Genes in Rhesus Macaque CL through the Luteal Phase Representative differentially expressed genes involved in hormone and growth factor signaling are displayed in the heat map in panel A. All genes represented on this heat map were identified as differentially expressed (>2-fold change; ANOVA, P < 0.01) using data from the early through late stages only (indicated by the bracket above the columns), although their corresponding expression levels in the very-late stage are shown as well. Panel B contains microarray and Q-PCR results for RLN1, CPE, and PRLR. Uppercase letters denote significant differences for microarray data (P < 0.05), and lowercase letters denote significant differences for Q-PCR data. Min, Minimum; Max, maximum.

Figure 4

Steroidogenic Gene Expression during the Rhesus Macaque Luteal Phase Genes involved in steroidogenesis are displayed in the heat map in panel A. A single asterisk denotes differential expression using data from the early through very-late stages (>2-fold change; ANOVA, P < 0.01), whereas genes with a double asterisk were differentially expressed (>2-fold change; ANOVA, P < 0.05) when mRNA levels from the early through late stages only were compared (indicated by the bracket above the columns). Nondifferentially expressed genes associated with steroidogenesis are included for completeness. Panel B contains Q-PCR validation of microarray results for LHCGR, HSD3B2, and CYP19A1. Uppercase letters denote significant differences for microarray data (P < 0.05), and lowercase letters denote significant differences for Q-PCR data. Min, Minimum; Max, maximum.

Figure 5

Expression of Genes Associated with Prostaglandin Biosynthesis, Metabolism, and Signaling during the Rhesus Macaque Luteal Phase Genes involved in prostaglandin biosynthesis, metabolism, and signaling are displayed in the heat map in panel A. Genes with a single asterisk after their name met the criteria for differential expression (>2-fold change; ANOVA, P < 0.01) using data from the early through very-late stages, whereas genes with a double asterisk were differentially expressed (>2-fold change; ANOVA, P < 0.05) when analyzed from the early through late stages only (indicated by the bracket above the columns). Nondifferentially expressed genes associated with prostaglandin biosynthesis, metabolism, and signaling are included for completeness. Panel B contains Q-PCR and microarray results plotted next to each other for PTGS2, PTGES, PTGER3, and PTGFR. Uppercase letters denote significant differences for microarray data (P < 0.05), and lowercase letters denote significant differences for Q-PCR data. Min, Minimum; Max, maximum.

Figure 6

Differential Expression of PRLR Protein Panel A is a representative Western blot using samples pooled from CL collected at each stage. The approximate molecular weight for each form of PRLR is indicated. The lower image is β-tubulin (TUBB) expression on the same membrane to demonstrate equivalent protein loading. Panel B contains densitometry results for each of the three isoforms of PRLR detected. Levels of PRLR from individual CL (n = 4/stage) were normalized to β-tubulin, and the resultant ratio was analyzed by ANOVA followed by comparison between groups using the SNK test. Columns with different letters are significantly different (P < 0.05).

Figure 7

Differential Expression of HSD3B2 Protein Panel A is a Western blot using samples pooled from CL collected at each stage. The upper image is probed for HSD3B2 with the last three lanes containing pooled mid-late, late, and very-late stage samples that were probed with primary antibody that had been preabsorbed with immunizing peptide. The lower band of the doublet disappeared after preabsorption of the primary antibody, but not the upper band, indicating that only the lower band corresponds to HSD3B2. The lower image is β-tubulin (TUBB), which was used as a loading control. Levels of HSD3B2 from individual CL (n = 4/stage) were normalized to β-tubulin, and the resultant ratio was analyzed by ANOVA followed by comparison between groups using the SNK test. Columns with different letters are significantly different (P < 0.05).

Figure 8

Differential Expression of CYP19A1 Protein Panel A is a Western blot using samples pooled from CL collected at each stage. The upper image is CYP19A1, and the lower is β-tubulin (TUBB), which served as a loading control. Levels of CYP19A1 from individual CL (n = 4/stage) were normalized to β-tubulin, and the resultant ratio was analyzed by ANOVA followed by comparison between groups using the SNK test. Columns with different letters are significantly different (P < 0.05).