Insulin-Like 3 Exposure of the Fetal Rat Gubernaculum Modulates Expression of Genes Involved in Neural Pathways

Abstract

Insulin-like 3 (INSL3) signaling directs fetal gubernacular development and testis descent, but the actions of INSL3 in the gubernaculum are poorly understood. Using microarray gene expression profiling of fetal male rat gubernaculum explants exposed to 10 or 100 nM INSL3, significant changes in expression were identified for approximately 900 genes. Several of the genes showing the largest inductions regulate neuronal development or activity, including Pnoc (34-fold), Nptx2 (9-fold), Nfasc (4-fold), Gfra3 (3-fold), Unc5d (3-fold), and Crlf1 (3-fold). Bioinformatics analysis revealed BMP and WNT signaling pathways and several gene ontologies related to neurogenesis were altered by INSL3. Promoter response elements significantly enriched in the INSL3-regulated gene list included consensus sequences for MYB, REL, ATF2, and TEF transcription factors. Comparing in vivo gene expression profiles of male and female rat fetal gubernaculum showed expression of the Bmp, Wnt, and neurodevelopmental genes induced by INSL3 was higher in males. Using quantitative RT-PCR, the microarray data were confirmed, and the induction of Bmp3, Chrdl2, Crlf1, Nptx2, Pnoc, Wnt4, and Wnt5a mRNA levels were examined over a range of INSL3 concentrations (0.1–100 nM) in male and female gubernaculum. In both sexes, an increasing gene expression response was observed between 0.1 and 10 nM INSL3. These data suggest that INSL3 signaling in the fetal gubernaculum induces morphogenetic programs, including BMP and WNT signaling, and support other rodent data suggesting a role for these pathways in development of the gubernaculum.

Insulin-like 3 (INSL3) promotes fetal gubernacular development via expression of secreted factors involved in the regulation of neuronal growth and function.

INTRODUCTION

During mammalian development, the testis normally descends into the scrotum, but the human condition of undescended testis (cryptorchidism) is relatively frequent, occurring in 2–4% of male births [1]. If the testis fails to descend, spermatogenesis is disrupted, leading to subfertility. Cryptorchidism is also associated with other conditions of male reproductive maldevelopment, including hypospadias and testis cancer. The causes of human cryptorchidism are not well understood, but roles for genetic polymorphisms and environmental chemical exposures are postulated [2].

In mammals, testis descent occurs during both fetal and neonatal life and is controlled by testis-produced hormones (insulin-like 3 [INSL3] and androgen) that drive male-specific development of a ligament called the gubernaculum [3, 4]. At its cranial pole, the gubernaculum is attached to the testis/epididymal complex and enlarges in response to testis hormones, creating a space within the inguinal canal that facilitates testis descent. In male fetal rats, the bulk of the fetal gubernaculum, called the bulb, is composed of a core of mesenchymal cells surrounded by striated muscle that is continuous with the abdominal wall musculature and innervated by the genitofemoral nerve. The gubernaculum develops in both sexes and is histologically similar until Gestational Day (GD) 16 in the rat fetus but enlarges in males only during GDs 16–20 [5–8]. Exposure of GD 17 rat gubernacula to INSL3 and androgen in vitro stimulates cellular proliferation and development of both mesenchymal and muscle layers within the gubernaculum [7, 9]. The only receptor for INSL3 is a G protein-coupled receptor called relaxin/insulin-like peptide family receptor 2 (RXFP2) [10], and Rxfp2 mRNA and protein expression is detected in the muscle layer of the fetal gubernaculum [11]. In mice with an Insl3 or Rxfp2 null mutation, the male gubernaculum fails to enlarge normally, and testes remain within the abdominal cavity [12–14]. Conversely, transgenic overexpression of INSL3 in female mice causes abdominal translocation of the fetal ovary and male-like development of the gubernacular bulb [15, 16]. These data show that INSL3-dependent signaling via RXFP2 within the fetal gubernaculum critically controls the initial phase of testis descent.

The only reported effects of INSL3 signaling in the gubernaculum are production of the second-messenger cAMP, histological changes in muscle and mesenchyme, and tissue enlargement, caused in part by increased cell proliferation [7, 9, 16]. The goal of the present study was to identify downstream pathways that are activated by INSL3-RXFP2 signaling in the gubernaculum that may provide insight regarding genetic factors contributing to cryptorchidism. We hypothesized that INSL3 would induce a morphogenetic program in the fetal gubernaculum based upon changes in gene expression, and this hypothesis was tested by examining the gene expression profile of the male and female fetal rat gubernacular bulb following in vitro exposure to INSL3.

MATERIALS AND METHODS

Animals

Breeding colonies of Long-Evans rats were maintained at the Alfred I. duPont Hospital for Children in a facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International, and animal care and use was approved by the Institutional Animal Care and Use Committee of the hospital. Rats were provided food (LabDiet Rat Chow 5021; PMI Nutrition International) and water ad libitum and housed in polycarbonate cages with pine-shaving bedding in a room with a 12L:12D photoperiod and controlled temperature (70 ± 2°C) and humidity (35–70%). Following overnight cohabitation, females fertilized by males were identified by sperm in vaginal lavage fluid. When sperm were detected, this day was designated GD 0. On GD 17, between 1200 and 1500 h, dams were euthanized by carbon dioxide asphyxiation and pups by decapitation.

Gubernaculum Tissue Culture and INSL3 Exposure

The culture system was modeled after that described by Emmen et al. [7]. Immediately after microdissection, two GD 17 gubernacular bulbs per biological replicate were cultured on a Millicell CM membrane (catalog no. PICM03050; Millipore Corp.) in six-well polystyrene culture plates (catalog no. 3506; Corning, Inc.) containing 1 ml of medium per well. All medium reagents were obtained from Invitrogen Corp. and included Dulbecco modified Eagle medium/Ham F12 (catalog no. 21041–025), 2% charcoal stripped fetal bovine serum (catalog no. 12676–011), 1× insulin-transferrin-selenium-X supplement (catalog no. 51500–056), and 1× antibiotic-antimycotic (catalog no. 15240–096). Cultures were maintained at 37°C in a humidified chamber with 5% carbon dioxide. For female gubernacula, tissue was incubated immediately with medium containing no hormone (control) or mouse INSL3 (catalog no. 035–43; Phoenix Pharmaceuticals, Inc.). This recombinant mouse INSL3 has been shown to stimulate cAMP production in cells via RXFP2 [10]. Because GD 17 male rat gubernacula have been exposed to INSL3 in vivo for approximately 2 days [11], male gubernacula were cultured for 24 h in medium containing no INSL3 to reduce any preexisting, INSL3-dependent gene expression. Gubernacula were then cultured for 24 h with either basal or INSL3-containing medium. The INSL3 concentrations employed (0.1–100 nM) bracketed the dissociation constant of INSL3 for its receptor (RXFP2) [17–19] and the estimated minimum in vivo exposure of the fetal gubernaculum based on serum INSL3 levels at GD 17 of approximately 0.5–1 nM (unpublished observations). Local levels of INSL3 likely are higher based on experimental data showing that unilateral castration of rabbit fetuses inhibits ipsilateral, but not contralateral, development of the gubernaculum [20].

Microarrays

Group (biological replicate) sizes for the microarrays were six control replicates, five 10 nM INSL3 replicates, and six 100 nM INSL3 replicates. These two concentrations were chosen because they elicited maximal stimulation of cAMP production in primary gubernacular cells and Rxfp2-transfected cells [17]. Total RNA was isolated using an RNeasy Micro Kit (Qiagen), with genomic DNA digested on the column. RNA quality was verified with an Agilent 2100 Bioanalyzer, and RNA Integrity Numbers of greater than nine were obtained. Next, 150 ng of total RNA were labeled using a GeneChip 3′ IVT Express Kit (Affymetrix, Inc.). Using a GeneChip Scanner 3000 7G System (Affymetrix, Inc.), 15 μg of biotin-labeled cRNA were hybridized to GeneChip Rat Genome 230 2.0 microarrays (Affymetrix, Inc.) containing approximately 31 000 probesets. After washing and staining, microarrays were scanned to obtain raw signal intensities for each probe, and these were background-corrected, normalized, and summarized using Guanine Cytosine Robust Multiarray Analysis and the affylmGUI Bioconductor package within R (http://www.bioconductor.org/). Microarray data were deposited in the Gene Expression Omnibus of the National Center for Biotechnology Information (NCBI; www.ncbi.nlm.nih.gov/geo/) and assigned accession number GSE19658. Principal component analysis was performed on the normalized expression data and showed that the INSL3 exposure replicates grouped together with the exception of two replicates in the 10 nM INSL3 group (data not shown). Compared to all the other samples, these two samples also had low cRNA yields (data not shown). Therefore, these two replicates were excluded from further analysis.

Gene profiling methodology for the GD 17.5 male and female gubernacula comparison used GeneChip RAE230A microarrays (Affymetrix, Inc.) containing approximately 16 000 probesets. Raw data acquisition and samples for these arrays have been described previously [21].

Quantitative RT-PCR

Gubernaculum starting material for these experiments was distinct from gubernacula used for microarray analysis. Total RNA purification, cDNA generation, Taqman-based PCR, and data analysis using the delta-delta threshold cycle method were performed as described previously [21]. Expression of target genes was normalized to Gapdh expression and quantified relative to expression in a fetal rat embryo total RNA preparation (catalog no. AM7928; Applied Biosystems/Ambion). Prevalidated Taqman assays (Applied Biosystems, Inc.) were as follows: Bmp3, Rn00567346_m1; Chrdl2, Rn01510694_m1; Crlf1, Rn01419973_m1; Nptx2, Rn01756377_m1; Pnoc, Rn00564560_m1; Wnt4, Rn00584577_m1; Wnt5a, Rn00575260_m1; and Gapdh, Rn99999916_s1.

Microarray and Quantitative RT-PCR Statistical Analysis

Before statistical analysis of the microarray data, an expression filter was applied to exclude probesets showing a log2 expression value of less than 3.5 in all replicates of at least one group. Using the LIMMA package in Bioconductor (http://www.bioconductor.org/), probesets with a significant expression change after INSL3 exposure were identified using a linear model approach with false-discovery rate (FDR) correction. Probesets with an FDR-corrected P value of less than 0.05 were considered to be significantly changed.

Significance testing of quantitative RT-PCR (qRT-PCR) data was performed with GraphPad Prism 5.0 software and used one-way ANOVA and Dunnett post-hoc tests. P values of less than 0.05 were deemed to be significant.

Bioinformatics

Bioinformatics pathway analysis was performed on lists of genes (i.e., probesets) with a significant change in expression after INSL3 exposure. To be included in the analysis, a gene had to have an FDR-corrected P value of less than 0.05 in both the 10 and 100 nM exposure groups. Various gene lists were analyzed: 1) all significant probesets, 2) significant probesets with an expression fold-change greater than or equal to an absolute value of 1.5, and 3) significant probesets with an expression fold-change greater than or equal to an absolute value of 2. Pathway analysis was conducted using both Ingenuity Pathway Analysis (IPA: version 8.0; Ingenuity Systems, Inc.) and the Database for Annotation, Visualization, and Integrated Discovery (DAVID; version 2008) [22]. Both IPA and DAVID provide complementary pathway analysis: IPA uses a proprietary knowledge base, whereas DAVID considers ontologies from the Gene Ontology (GO) project and pathways from the Kyoto Encyclopedia of Genes and Genomes (KEGG). Before importing gene lists into IPA, probesets not annotated to a Human Genome Organization gene symbol by Affymetrix were annotated by performing a Basic Local Alignment Search Tool search with public domain software provided by the National Center for Biotechnology Information and using the probeset target sequence. Gene symbols were assigned to a probeset when a mammalian gene with a RefSeq mRNA accession had an alignment score to the target sequence of greater than or equal to 200. Using gene symbols as identifiers, gene lists were imported into IPA and DAVID [22]. Within IPA, statistically significant molecular pathways enriched in the gene list were determined from the IPA library of canonical pathways, and an FDR-corrected Fisher exact test P value of less than 0.05 was considered to be significant. For DAVID analysis, the 16 504 probesets expressed in control and/or hormone-exposed gubernacula were used as the gene population background, annotations were limited to Rattus norvegicus, default analysis parameters were employed, and GO annotation terms and KEGG pathways with a Benjamini-corrected Fisher exact test P value of less than 0.05 were considered to be significant.

Promoters of differentially expressed genes (identified for both 10 and 100 nM INSL3) were analyzed for enrichment of transcription factor binding sites using Promoter Analysis and Interaction Network (PAINT; v3.9; www.dbi.tju.edu/dbi/tools/paint) [23]. Affymetrix probeset IDs were converted to mouse NCBI Gene IDs using NetAffx and NCBI Homologene databases. DNA sequences for 665 Gene IDs were in the PAINT database. These sequences, which included 2 kb upstream of each transcription start site, were examined for transcription binding sites defined in the Transfac 7.0 Public 2005 database using PAINT v3.9. Binding site enrichment was determined by comparison to all mouse promoters in the PAINT database, and an FDR-corrected P value of 0.1 or less was considered to be significant.

RESULTS

INSL3-Dependent Microarray Gene Profiling and Functional Pathway Analysis

Gene expression microarray analysis was used to identify mRNA transcripts in fetal male rat gubernaculum modulated by exposure to INSL3. Male GD 17 gubernacular bulb cultures exposed to 10 or 100 nM INSL3 for 24 h were examined. Of the approximately 31 000 queried probesets, 3482 displayed significantly altered expression in response to at least one INSL3 concentration (Supplemental Tables S1 and S2; all Supplemental Data are available online at www.biolreprod.org). Similar numbers of probesets with significant expression changes were observed at 10 or 100 nM INSL3 (2167 and 2302 probesets, respectively). Only 20 probesets were significantly different between 10 and 100 nM INSL3 exposures, and these differences were generally in magnitude of expression, not in direction of change. Compared to 100 nM INSL3, the 10 nM INSL3 gene list contained nearly twice as many genes with a 1.5-fold or greater expression decrease, but most (∼67%) of these decreased genes were not significantly altered by 100 nM INSL3. Using qRT-PCR analysis of a separate series of INSL3-exposed tissue, we failed to confirm the reduced microarray expression of five genes (Axin2, Ccnd2, Gsk3b, Lrp6, and Sfrp1) that were identified only in the 10 nM INSL3 group. However, we were able to corroborate the expression change of all seven analyzed genes significantly altered by both 10 and 100 nM INSL3 (see below). Therefore, downstream bioinformatics analyses of the microarray data were limited to those genes with significant changes in expression after both 10 and 100 nM INSL3 exposure. This list contained a total of 989 probesets, representing approximately 900 unique genes (Supplemental Table S2). Increased expression was observed for 714 probesets, and of these, 180 were upregulated at least 2-fold. The robustness of this gene list was exemplified by an excellent correlation of the direction of expression change observed between 10 and 100 nM exposures: All 989 probesets showed the same direction of expression change at both 10 and 100 nM INSL3.

Within the INSL3 upregulated gene list, Pnoc (encoding the secreted neuropeptide nociceptin) displayed the greatest INSL3-induced expression change (34-fold by 10 nM INSL3), and a gene encoding a prohormone convertase known to process prepronociceptin (Pcsk2) also was highly upregulated (9-fold by 10 nM INSL3). In gubernacular cells, INSL3 binding to RXFP2 increases production of cAMP [18], and in other tissues, cAMP elevates the Icer form of Crem [24] and phosphodiesterase mRNA levels [25]. Consistent with cAMP production in our culture model, mRNA levels of Crem and three cAMP phosphodiesterases (Pde10a, Pde8a, and Pde4d) were upregulated between 2- and 13-fold. A number of genes with large fold-change increases are involved in neurogenesis or modulation of synaptic transmission, including several secreted ligands and receptors for secreted ligands (Table 1).

TABLE 1.

Neurogenesis and synaptic transmission genes upregulated at least 2-fold by INSL3.

Pathway analysis of the INSL3 microarray data revealed molecular signaling and cellular differentiation programs potentially downstream of INSL3 signaling in the gubernaculum. The INSL3-responsive gene list was examined with pathway analysis software using all significantly changed genes or gene subgroups with a 1.5- or 2-fold expression change. IPA canonical signaling pathways significantly enriched in the gene lists are shown in Table 2, and genes with significant expression changes in each pathway and images of the pathways can be found in Supplemental Table S3 and Figures S1–S3. Many of the observed pathways were detected because they incorporated Wnt and/or Bmp signaling. On their own, Wnt/β-catenin and Bmp canonical signaling pathways were significantly changed by INSL3 exposure. Within the Wnt and Bmp pathways, mRNAs of several ligands, including Wnt2, Wnt4, Wnt5a, Wnt9a, and Wnt16 and Bmp1, Bmp2, Bmp3, Bmp5, and Bmp7, were upregulated. One of the most highly induced genes encodes a BMP antagonist (Chrdl2; increased 13-fold by INSL3). When only genes with a relatively high expression fold-change of greater than 1.5 or 2 are included in the analysis, the cAMP-mediated signaling and BMP signaling pathways became statistically significant. The same pathways were called significant only if genes with increased expression were used in the analysis, and only the WNT/β-catenin canonical pathway was detected when genes with decreased expression were examined (data not shown). Therefore, genes induced by INSL3 were largely responsible for all IPA canonical pathway significance calls.

TABLE 2.

Pathway analysis of INSL3-responsive genes using Ingenuity Pathways Analysis (IPA).^a

In addition to IPA, we extracted biological meaning from the microarray gene expression data using DAVID and PAINT. Table 3 shows selected statistically significant GO annotation terms and KEGG pathways derived from the INSL3-regulated gene list using DAVID. All GO terms and KEGG pathways with a corrected P value of less than 0.05 and associated genes are found in Supplemental Table S4. Prominent within the list of annotation terms were those related to neurodevelopment, including the higher-level ontology terms nervous system development and neurogenesis as well as subcategories such as neuron differentiation, generation of neurons, and axonal guidance. Another potential functional role of INSL3 may be cholesterologenesis, because genes in the GO term sterol metabolic process were significantly modulated by INSL3; several of these genes were induced between 1.5- and 2-fold and are involved in cholesterol biogenesis (Cyp51, Hmgcr, Idi1, Lss, Sc4mol, Sc5dl, Srebf1, and Srebf2). Like the IPA analysis, the significant KEGG pathway basal cell carcinoma consisting of WNT pathway factors was enriched significantly. GO terms related to subcellular locations that were over-represented included extracellular space and plasma membrane. The subset of gubernaculum genes induced by INSL3 produced similar significantly enriched GO terms through DAVID analysis (data not shown). Using PAINT, transcription factor consensus sequences over-represented within promoters of genes modulated by INSL3 were determined. When all INSL3-responsive genes, only upregulated genes, or only downregulated genes were analyzed, PAINT revealed that consensus binding sequences for four transcription factors (ATF2, TEF, REL, and MYB) were significantly enriched in INSL3-modulated genes (Table 4 and Supplemental Table S5). The number of genes with identified promoter consensus sites for ATF2, TEF, REL, and MYB were 7, 8, 83, and 83, respectively. By microarray analysis, Tef expression was increased after INSL3 exposure.

TABLE 3.

Selected DAVID gene functional classifications of INSL3 altered gene list.^a

TABLE 4.

Transcription factor binding site analysis using PAINT of INSL3-responsive genes.^a

Because high INSL3 levels are found in male compared to female fetuses, we predicted that many of the genes modulated by INSL3 in the gubernaculum culture model would show significant expression differences in vivo between male and female gubernacula. Furthermore, a large number of genes increased by INSL3 in culture should show increased expression in male compared to female gubernacula. To examine these predictions, in vivo microarray gene expression data from GD 17.5 male and female gubernacula (Supplemental Table S6) were compared with genes meeting two criteria from the gubernaculum INSL3 exposure experiment: 1) a significant expression change after both 10 and 100 nM INSL3 exposure, and 2) a 2-fold or greater change following 10 nM INSL3 exposure. Because of the different microarray chips used in each experiment, only those probesets meeting these two criteria and present on both microarrays were compared. Of 108 probesets increased by INSL3 and tested in both experiments, seven showed higher expression in female gubernacula, 51 were not significantly different between male and female gubernaculum, and nearly half (n = 50) displayed higher expression in male gubernaculum (Supplemental Table S7). Similar results were observed for genes decreased by INSL3 exposure. Of 35 probesets decreased by INSL3 exposure and also examined in male and female gubernacula in vivo, nine were not significantly different between male and female gubernaculum, five had higher expression in male gubernaculum, and more than half (n = 21) displayed reduced expression in male gubernaculum. Thus, fetal gubernaculum genes responsive to INSL3 exposure in vitro generally exhibit similar differential expression between male and female fetal gubernaculum in vivo.

Microarray Data Confirmation and Concentration Response of INSL3-Induced Gene Expression

To confirm the INSL3 microarray data and examine the timing and INSL3 concentration response of gene expression in fetal gubernaculum, a set of rat gubernacula not used in the microarray experiments were exposed to INSL3, and the expression of seven genes was analyzed by qRT-PCR. These seven genes were chosen because they displayed increased expression in male compared to female gubernaculum and are involved in Bmp signaling, Wnt signaling, or neurodevelopmental pathways identified via bioinformatics analyses. Using the same 24-h exposure model as for the microarray experiment, male gubernacula were incubated with INSL3 concentrations ranging from 0.1 to 100 nM. Overall, excellent concordance was found between the microarray and qRT-PCR data (Fig. 1). The gene expression response to INSL3 became saturated at 10 nM INSL3 for most genes, although Wnt4 levels peaked at 1 nM INSL3. Whereas trends were found for increased expression of Pnoc, Wnt4, Bmp3, and Chrdl2 at 0.1 nM INSL3, the responses did not reach statistical significance using a one-way ANOVA and Dunnett post-hoc test. Using just the control and 0.1 nM values, an unpaired t-test produced P values of less that 0.05 for these four genes (data not shown).

FIG. 1.

After incubation for 24 h in the presence of INSL3 concentrations ranging from 0 (control) to 100 nM, male fetal gubernacular bulb gene expression was examined by Taqman-based qRT-PCR. Each graph shows mRNA levels, with the gene identified above the graph. Below each graph, the fold-change (FC) is given for each gene after 24-h exposure to 10 or 100 nM INSL3 as determined by microarray expression analysis. Data are presented as the mean ± SD (n = 7–11). An asterisk indicates a significant difference from control (P < 0.05) as determined by one-way ANOVA and Dunnett post-hoc test.

To determine if a shorter INSL3 exposure modulated gubernaculum gene expression, mRNA levels were examined by qRT-PCR after a 6-h incubation with INSL3 (Fig. 2). In comparing the 6- and 24-h exposures, the responses were gene-specific. After 6 h, no significant expression changes were observed for Crlf1, and Wnt5a levels were increased at 10 nM, but not at 100 nM, INSL3. For Wnt4, Bmp3, and Chrdl2, the concentration response and level of induction were similar at both time points. Finally, Nptx2 and Pnoc induction levels at 6 h were approximately twice the levels observed at 24 h.

FIG. 2.

After incubation for 6 h in the presence of INSL3 concentrations ranging from 0 (control) to 100 nM, male fetal gubernacular bulb gene expression was examined by Taqman-based qRT-PCR. Each graph shows mRNA levels, with the gene identified above the graph. Data are presented as the mean ± SD (n = 5–6). An asterisk indicates a significant difference from control (P < 0.05) as determined by one-way ANOVA and Dunnett post-hoc test.

Transgenic expression of INSL3 in fetal female mice causes male-like gubernacular development and transabdominal descent of the ovary [15, 16], suggesting responsiveness of the female gubernaculum to INSL3. Likewise, histological development of the gubernaculum is similar in males and females, but enlargement occurs in males, but not in females, from GD 16 to GD 20 [5–7]. We used female GD 17 gubernacula to provide comparative gene expression data using an anatomically similar organ not previously exposed to INSL3 that is morphologically similar to the male. Overall, female gubernaculum gene expression responses after 24 h of exposure were similar to the responses in males, but for Bmp3, Chrdl2, and Wnt4 expression, the female tissue appeared to show increased sensitivity to INSL3 (Fig. 3).

FIG. 3.

After incubation for 24 h in the presence of INSL3 concentrations ranging from 0 (control) to 100 nM, female fetal gubernacular bulb gene expression was examined by Taqman-based qRT-PCR. Each graph shows mRNA levels, with the gene identified above the graph. Data are presented as the mean ± SD (n = 6–7). An asterisk indicates a significant difference from control (P < 0.05) as determined by one-way ANOVA and Dunnett post-hoc test.

DISCUSSION

Using high content analysis, we examined the gene expression response of cultured fetal rat gubernaculum to INSL3 exposure during its early outgrowth phase. Several results demonstrated the robustness of the microarray data. First, statistical significance and fold-change values were similar for the seven genes examined by both microarray and qRT-PCR. This was observed even though separate biological replicates were used for each technique. Second, data from INSL3-exposed female gubernacula corroborated expression data from male gubernacula. Third, the consensus binding site of a cAMP-responsive transcription factor (ATF2) was over-represented in promoters of genes regulated by INSL3, and genes in the IPA cAMP-mediated signaling canonical pathway were significantly affected by INSL3 exposure. Finally, the microarray gene expression responses following exposures to 10 and 100 nM INSL3 were nearly identical and mirrored INSL3-induced cAMP production at these two concentrations [17–19].

Gubernaculum microarray gene expression and cAMP production data possess similar INSL3 concentration responses. Stimulation of cAMP production downstream of INSL3/RXFP2 binding in Rxpf2-transfected cells or primary neonatal gubernaculum cells begins at ligand concentrations of 0.1 nM and becomes maximal at approximately 10 nM [17–19]. Although not true for every gene examined, 0.1 nM INSL3 induced Bmp3, Chrdl2, Pnoc, and Wnt4 gene expression in male gubernaculum to levels just above control, and expression of most genes became maximal at 10 nM INSL3. This saturation event was observed in both qRT-PCR and microarray expression data. By microarray analysis, mRNA levels of only 20 genes were significantly different between 10 and 100 nM INSL3, even though each INSL3 concentration altered expression of more than 1000 genes. Putting these INSL3 concentration-response data in an in vivo context is difficult, because to our knowledge, no reports of INSL3 protein levels in fetal gubernaculum or serum have been published. However, Insl3 mRNA levels are high, and INSL3 immunoexpression is readily detected in fetal rat testis [11]. In pubertal male rat serum, INSL3 protein peaked at a concentration of approximately 1 nM [26], and our preliminary data suggest similar concentrations in GD 17 male rat serum (Ivell and Barthold, unpublished observations).

To relate the in vitro INSL3-dependent gene expression results with potential INSL3-dependent gene expression during in vivo gubernacular development, we compared genes altered by INSL3 exposure in vitro and genes differentially expressed in male and female fetal gubernaculum in vivo. Although fetal males produce much higher levels of INSL3 compared to females [27], linking sexually dimorphic gubernaculum gene expression in vivo to INSL3 alone is confounded by the additional effects of androgen on gubernacular development [7, 9]. Nevertheless, a striking conservation of gene expression patterns was observed between gubernaculum INSL3 exposure in vitro and male and female gubernacula in vivo. Of genes stimulated at least twofold after INSL3 exposure in vitro and also compared between male and female gubernacula in vivo, approximately half (71/143) showed a similar response in both magnitude and direction of expression change. From these data, we conclude the following: 1) In vitro INSL3 exposure of gubernaculum explants recapitulates INSL3-dependent processes occurring in gubernaculum in vivo, and 2) many of the gene expression differences observed between male and female tissue in vivo are the result of higher INSL3 levels in males.

Although INSL3 increases cAMP production in gubernaculum through a G protein-dependent pathway, little else is known about the INSL3 signaling mechanism or its regulation. Our microarray expression data provide information about potential additional downstream INSL3 targets related to cAMP signaling. Pnoc, a cAMP-regulated gene [28], shows a rapid and robust response to INSL3 exposure in both male and female gubernaculum. Crem/Icer is another cAMP-inducible gene that represses cAMP-mediated signaling [24], and its expression was induced more than twofold by INSL3 exposure. INSL3 also regulated gubernaculum expression of phosphodiesterases, which can negatively regulate cellular cAMP levels and compartmentalize cAMP signaling processes to subcellular domains [29]. In vitro INSL3 exposure of the gubernaculum induced at least a twofold expression change of three phosphodiesterase mRNAs (Pde10a, Pde8a, and Pde4). In fact, Pde10a induction was among the highest observed, at more than 10-fold. Comparison of male and female gubernaculum gene expression corroborated this observation, because Pde10a expression was 13-fold higher in male compared to female GD 17 gubernaculum. These results support the hypothesis that increased phosphodiesterase, particularly Pde10a, and Crem/Icer expression in response to INSL3 exposure modulates INSL3-dependent signal transduction and gubernacular development.

Pathway analysis of gubernaculum INSL3-responsive genes suggested INSL3 mediates gubernacular development by inducing production of extracellular ligands and glycoproteins. Several IPA canonical pathways significantly over-represented in the INSL3 gene list were growth factor or morphogenetic signaling pathways, including HGF, WNT, and BMP signaling pathways. The mRNA expression of ligands driving these pathways was induced and included Hgf as well as a number of BMP family members (Bmp1, Bmp2, Bmp3, Bmp5, and Bmp7) and WNT family members (Wnt2, Wnt4, Wnt5a, Wnt9a, and Wnt16). Supporting a crucial role of both WNT and BMP signaling in gubernacular development are studies showing mutations in INSL3-regulated genes encoding secreted modulators of WNT and BMP signaling produce males with high intra-abdominal testes, including mouse mutations in Bmp5 [30], the BMP antagonist Nog [31], the WNT pathway modulators Sfrp1 and Sfrp2 [32], and Wnt5a [32]. Cryptorchid phenotypes in WNT signaling pathway mutants appeared despite apparently normal testicular hormone production, and both WNT and BMP pathway mutants phenocopy genetic deletion of Insl3 or Rxfp2 [12–14]. DAVID bioinformatics analysis identified the GO cellular component term extracellular space (mainly comprised of secreted proteins) as significantly enriched by INSL3-responsive genes. In developing female mice transgenically overexpressing INSL3, gubernaculum enlargement similar to that in males and transabdominal migration of the ovary similar to that of the testis are observed. In our hands, the gene expression responses after INSL3 exposure of fetal male and female gubernaculum were similar. Collectively, these data suggest that an important downstream event of INSL3 during fetal gubernacular development is activation of Wnt and Bmp pathways via production of pathway ligands.

In addition to secreted factors and their signaling pathways, a major pathway identified by bioinformatics analysis of the microarray data was nervous system development. Using DAVID analysis of the INSL3-regulated gene lists, a number of related GO biological process annotations, such as neuron differentiation, generation of neurons, and neuron projection morphogenesis, were called significant. With both DAVID and IPA, the axonal guidance pathway was identified. Although these data suggest a role for INSL3 in neuronal development and/or function during the outgrowth phase of the fetal gubernaculum, INSL3 may have no direct effect on gene expression in nerve cells in our in vitro model of denervated gubernacula. Because RXFP2 is immunolocalized to gubernacular muscle cells [11], INSL3 may regulate trophic interactions between muscle cells and developing axons. Several genes with products that can regulate neuronal development or function were induced at least 2-fold by INSL3, including Wnt family members [33], Bmp family members [33, 34], Crlf1 [35, 36], Gda [37], Nptx2 [38, 39], Sema3g [40], Nfasc [41, 42], and Unc5d [43]. The gene with the largest fold-change was Pnoc (34-fold), encoding the neuropeptide nociceptin, and the gene encoding a protease capable of prepronociceptin activation in vivo (Pcsk2) also was increased (6-fold) [44]. Nociceptin is regulated by cAMP, stimulates neuronal differentiation, and is expressed in and modulates release of calcitonin gene-related peptide (CGRP) in sensory neurons [28, 45, 46]. Interestingly, afferent CGRP release is postulated to stimulate motility of the gubernaculum during testis descent via receptors localized to muscle [47, 48], and sensory neuron number and CGRP content is altered in a congenitally cryptorchid rat strain [49].

In conclusion, novel microarray gene expression data and bioinformatics analyses describing molecular outcomes of INSL3 exposure in fetal gubernaculum are presented. Our results showed that INSL3 modulates the expression of approximately 900 genes with a concentration response similar to that of cAMP induction. Bioinformatics analysis suggests that a major biological outcome of INSL3 exposure includes transcriptional activation of neurodevelopmental genes. Based upon the known expression of the INSL3 receptor in gubernacular muscle, we hypothesize that INSL3 may contribute to neuromuscular signaling within the gubernaculum, possibly by secretion of muscle-derived neurotrophic factors. These data provide further insight regarding downstream pathways that contribute to gubernacular development and subsequent testis descent.