Using a comparative genomics approach, a list of fetal Leydig cell specific candidate genes was identified. As with any candidate gene search, there are likely false-positives and false-negatives in the gene list. From a starting analysis of over 28,000 genes containing 12 known Leydig cell-specific genes, the final list of 54 candidate Leydig cell-specific genes contained 8 of the known Leydig cell-specific genes. Thus, there is a striking enrichment of Leydig cell genes in the candidate gene list. When the expression of five Leydig cell-specific candidate genes with unknown gonadal expression patterns were localized via ISH in GD13 mouse gonads, all showed higher expression in testis compared to ovary. With the exception of Vgll3
, the expression pattern of the localized genes was similar to the known Leydig cell-specific gene Cyp11a1
. Using the same fetal mouse testis cell isolates as used here but not the microarray data from DBP exposed fetal testes or whole fetal testes, Jameson et al., compiled a set of Leydig cell-enriched candidate genes 
; about 25% of the 54 genes in our candidate gene lists are present within the Jameson gene list. While this concordance is significant, overlap between the two lists was reduced by different statistical criteria and the use of additional filters in our protocol, including whole testis and DBP exposure data. In utero
DBP exposure reduces the expression of Leydig cell genes involved in INSL3 production and lipid metabolism (cholesterol, fatty acid and steroid biosynthesis; 
) but does not reduce all Leydig cell-specific genes (such as Stc1
). In this light, the candidate list of 54 genes may be enriched in genes involved in Leydig cell lipid metabolism. However, not all 54 candidate genes were screened for expression in Leydig cells. File S1
shows the genes identified at each step in the selection process. These genes are potential factors involved in fetal Leydig cell differentiation or hormone production, and mutation of some may be linked potentially to disorders of human masculinization.
Two particularly interesting candidate genes are Vsnl1
. VSNL1 is a myristoylated calcium binding protein that regulates intracellular signaling 
, while GRAMD1B is a predicted transmembrane protein with unknown function. By ISH in fetal testis, both appear to be expressed specifically in Leydig cells. Our localization data for Gramd1b
is corroborated by ISH data from genpaint.org (set identification number ES3003); at GD14 in the mouse, Gramd1b
mRNA is expressed only in steroidogenic cells of the adrenal gland and testis. In steroidogenic adrenal cells, Vsnl1
mRNA levels are increased by NR5A1 overexpression 
. Because NR5A1 activity controls fetal Leydig cell steroidogenesis 
may be downstream effectors of NR5A1 activity in fetal Leydig cells.
For functional studies of fetal Leydig cell-specific candidate genes, we focused on genes encoding receptors that may modify Leydig cell steroidogenesis. Of the three candidate genes we studied, only CRHR1 stimulated fetal testis steroidogenesis in rats and mice. Localized Crhr1
mRNA expression in the interstitial compartment of GD13 mouse gonads with a pattern similar to the Leydig cell-specific gene Cyp11a1
suggests CRHR1 is produced specifically in fetal Leydig cells. These Crhr1
ISH data are corroborated by our extensive comparative microarray analysis. When GD17 rat testes were exposed to varying concentrations of the CRHR1 agonist CRH for 24 hours ex vivo
, 10 nM and 100 nM concentrations increased mRNA levels of the steroidogenic genes, Cyp11a1
, and Star
. While a trend of increased steroidogenic gene expression was seen with 1 nM CRH, this increase was not significant. However, an upregulation of testosterone production was observed at 1 nM and higher CRH. These concentration response data approximate the low nM dissociation constant of CRH for CRHR1 in non-testis cell types 
. Similar enhancing effects on steroidogenesis by CRH were seen in GD15 mouse testis ex vivo
cultures. Expression levels of steroidogenic genes increased in testes exposed to 10 and 100 nM CRH. To determine if the increase in steroidogenic mRNA levels was due to CRHR1 receptor activation, GD15 mouse testis was treated with CRH and an antagonist specific for CRHR1, NBI 27914, for 24 hours ex vivo
. Exposure of testis to 10 nM CRH and 10 µM antagonist decreased steroidogenic gene expression when compared to 10 nM CRH alone, indicating CRH stimulates fetal Leydig cell steroidogenesis through CRHR1 activation. UCN1 also activates CRHR1 
, and the ability of UCN1 to elevate fetal rat Leydig cell steroidogenic gene expression provides additional evidence supporting a stimulatory role of CRHR1 in fetal Leydig cell steroidogenesis (). CRH is an agonist for both CRHR1 and CRHR2, but it binds with higher affinity to CRHR1 
mRNA levels were below the level of detection in mouse and rat fetal testes microarrays (data not shown), suggesting CRHR2 is not expressed in fetal rodent testes. From the totality of these data, we conclude that CRH and UCN1 stimulate fetal Leydig cell steroidogenesis through activation of CRHR1.
To date, rat and mouse adult Leydig cells and dispersed Leydig cells from GD20 rats have been used in previous CRH studies examining steroidogenesis. Consistent with our data using fetal rat and mouse testes ex vivo
, CRH stimulated steroidogenesis in MA-10 mouse Leydig cells and primary adult mouse Leydig cells. In postnatal mouse Leydig cells, CRH stimulates Leydig cell steroidogenesis via a mechanism similar to hCG 
. In contrast, CRH does not enhance steroidogenesis in primary adult rat Leydig cells 
. Instead, CRHR1 activation may inhibit hCG-stimulated steroidogenesis in primary adult rat Leydig cells 
; however, Huang et al., 
did not observe this inhibitory effect in adult rat Leydig cells. One study examined the steroidogenic effect of CRH on rat Leydig cells isolated from GD20 animals (just prior to parturition). After isolation, these cells were cultured for an additional four days and then exposed to CRH. Like adult rat Leydig cells from work performed by this lab 
, CRH exposure alone did not affect steroidogenesis, but CRH did inhibit hCG-induced steroidogenesis 
. The reason for the discrepancy between our ex vivo
fetal testis data and those from Ulisse et al. 
are unknown but could be related to differences in fetal age at analysis, the different durations of the culture periods, and/or using dispersed fetal Leydig cells versus intact fetal testes. The data described here are the first to examine intact fetal testes from rat and mouse, and the first to show CRHR1 agonism stimulates fetal testis testosterone production.
Similar to our rodent fetal testis data, CRH exposure of MA-10 cells stimulated steroidogenic gene expression and progesterone production (MA-10 cells do not produce testosterone) 
. The MA-10 cell line was treated with 10 nM CRH for multiple time intervals to determine optimal exposure time (1, 3, 6, or 24 hours). qRT-PCR analysis of Cyp11a1
, and Star
showed increased expression of these genes at multiple time points, with all genes increased at 6 hours. When treated with varying CRH concentrations for 6 hours in culture, steroidogenic mRNA levels showed significant increases at 10 nM and 100 nM CRH. Combined with the fetal mouse testis ISH data showing Crhr1
expression in fetal Leydig cells, the ability of CRH to increase MA-10 steroidogenesis indicates that CRH stimulates fetal testis steroidogenesis by direct activation of Leydig cells.
Multiple organs were screened using qRTPCR to identify potential organ sources for CRH and UCN1 in the rodent fetus (). Ucn1
mRNA was undetectable in all organs screened, but Crh
mRNA levels were detectable in brain and hypothalamus. This is consistent with ISH data from other labs showing Crh
mRNA is present in the hypothalamus of fetal rats beginning around GD17 to GD21 
, and Crh
expression is seen in the human fetal hypothalamus as early as GW12 
. At all gestation ages examined, no Crh
mRNA was observed in the fetal testis. The presence of CRH in amniotic fluid from GD17 male rats and GD15 male mice () suggests the rodent fetus produces CRH peptide during the masculinization programming window. Beginning around GW8–10, expression of CRH by the human placenta also coincides with the start of steroidogenesis in the human fetal testis 
, and CRH is detectable in human amniotic fluid during pregnancy 
. From these data, we conclude that CRH is expressed in the mammalian fetus during the masculinization programming window and that the likely source of CRH in the rodent is the hypothalamus.
Our data show CRHR1 agonism stimulates rodent fetal Leydig cell steroidogenesis under ex vivo
conditions, but the in vivo
significance of these data is unknown. The microarray and ISH data indicate Crhr1
mRNA is expressed in fetal mouse Leydig cells from the early stages of steroidogenesis (GD13 and later). CRHR1 agonists stimulated mouse and rat fetal testis steroidogenesis during the in utero
masculinization programming window. However, male Crhr1
knockout mice are fertile 
, suggesting CRHR1 may not be required for in utero
masculinization of the mouse male reproductive tract. Nonetheless, fetal Leydig cell function has not been analyzed rigorously in Crhr1
knockout mice, and it remains possible that fetal Leydig cell steroidogenesis is reduced in these mice but not to a level that would cause masculinization defects resulting in infertility.
Some aspects of human male reproductive tract masculinization are independent of LHCGR-mediated stimulation of fetal Leydig cell testosterone production. In males with an inactivating LHCGR mutation, two androgen-dependent tissues (the epididymis and vas deferens) still masculinize 
. These data suggest that either basal levels of testosterone production are sufficient to masculinize Wolffian duct-derived tissues or that factors other than LHCGR can stimulate fetal Leydig cell steroidogenesis during the initial period of fetal male masculinization (from approximately GW6 to GW10). One possible mechanism is functional removal of a steroidogenic repressor molecule, as has been suggested to occur in rodent fetal testes 
. Although no human data exist to support such a role, another possible mechanism is steroidogenic stimulation via CRHR1 agonism.