Dynamic changes in fetal Leydig cell populations influence adult Leydig cell populations in mice

Abstract

Testes contain two distinct Leydig cell populations during development: fetal and adult Leydig cells (FLCs and ALCs, respectively). ALCs are not derived from FLCs, and it is unknown whether these two populations share common progenitors. We discovered that hedgehog (Hh) signaling is responsible for transforming steroidogenic factor 1-positive (SF1⁺) progenitors into FLCs. However, not all SF1⁺ progenitors become FLCs, and some remain undifferentiated through fetal development. We therefore hypothesized that if FLCs and ALCs share SF1⁺ progenitors, increased Hh pathway activation in SF1⁺ progenitor cells could change the dynamics and distribution of SF1⁺ progenitors, FLCs, and ALCs. Using a genetic model involving constitutive activation of Hh pathway in SF1⁺ cells, we observed reduced numbers of SF1⁺ progenitor cells and increased FLCs. Conversely, increased Hh activation led to decreased ALC populations prepubertally, while adult ALC numbers were comparable to control testes. Hence, reduction in SF1⁺ progenitors temporarily affects ALC numbers, suggesting that SF1⁺ progenitors in fetal testes are a potential source of both FLCs and ALCs. Besides transient ALC defects, adult animals with Hh activation in SF1⁺ progenitors had reduced testicular weight, oligospermia, and decreased sperm mobility. These defects highlight the importance of properly regulated Hh signaling in Leydig cell development and testicular functions.—Barsoum, I. B., Kaur, J. Ge, R. S., Cooke, P. S., Yao, H. H.-C. Dynamic changes in fetal Leydig cell populations influence adult Leydig cell populations in mice.

Leydig cells, the testis-specific cell type discovered by Franz Leydig in 1850, produce androgens and other factors responsible for the masculinization of individuals. Two distinct populations of androgen-producing Leydig cells appear during testis development in most mammals: fetal Leydig cells (FLCs) and adult Leydig cells (ALCs) (1). The FLCs in mice arise in the testicular interstitium between embryonic day 12.5 (E12.5) and E13.0, ∼1 d after the appearance of Sertoli cells (1–4). Sertoli cells trigger differentiation of steroidogenic factor 1-positive (SF1⁺) progenitor cells into FLCs via paracrine regulation (1, 2, 5–8). The intercellular Notch signaling pathway is also involved in FLC establishment and maintenance (9). The FLC population increases dramatically during embryonic development despite the fact that differentiating FLCs are mitotically inactive (1, 10), suggesting that expansion of FLC populations results from differentiation of progenitor cells, rather than cell division of existing FLCs. The SF1⁺ cells in gonadal primordia are the primary source of FLCs (11), but other sources such as neighboring mesonephros (12), migrating neural crest cells (13), and cells from the coelomic epithelium (14, 15) or interstitium (16) are potential contributors also. At the end of fetal life and during the first 2 postnatal weeks in rodents, FLCs are gradually replaced by ALCs (1, 17), but the definitive source of the progenitor cells for ALCs has not yet been conclusively identified.

The involvement of the hedgehog (Hh) pathway in FLC development was first revealed by the identification of a role of desert hedgehog (Dhh), one of the 3 secreted Hh ligands in mammals (4, 18, 19). When Dhh was inactivated in mouse embryos, fetal testes developed fewer FLCs and exhibited abnormal testis cord organization. Later in prepubertal and adult life, testes of Dhh^−/− mice had severely impaired spermatogenesis and 92% of the males were infertile and lacked ALCs (4, 6, 18, 20, 21). The role of Dhh in Leydig cell differentiation is also conserved in rats. Rats with a spontaneous missense mutation in Dhh exhibited a reduced number of FLCs and a lack of typical spindle-shaped ALCs, similar to the phenotype of Dhh^−/− mice (22). In humans, point mutations in the DHH gene have been linked to intersex problems involving both mixed and pure gonadal dysgenesis (23–25). The pure gonadal dysgenesis cases, for example, are XY (genetically males) with bilateral rudimentary streak-like gonads and retention of female internal reproductive tract organs and external genitalia. These data demonstrate a conserved role of DHH in fetal testis development in both humans and rodents, with a subsequent effect on adult testis function and fertility.

Between birth and puberty, ALCs arise in the interstitium from unknown progenitor cells and become the major source of androgens that control differentiation of the male reproductive tract and spermatogenesis. ALCs are not derived from FLCs (1, 26), and the origin and the molecular events that control ALC differentiation are not clearly understood. Park et al. (27) showed that Sf1 haplodeficiency in Dhh^−/− mice results in a total blockage of FLC and ALC development, raising a previously unexamined hypothesis that these two Leydig cell populations share a common progenitor pool (i.e., SF1⁺ cells). In this study, we tested this hypothesis by altering the SF1⁺ progenitors and FLC numbers during fetal life via an increased activation of the Hh pathway. The effects of altering the allocation of progenitor and FLC populations on ALCs and testis functions were analyzed.

MATERIALS AND METHODS

Generation of Sf1/Cre;Smo^YFP animals

The Smo^YFP mouse (stock no. 005130; Jackson Laboratory, Bar Harbor, ME, USA) contains a Drosophila smo gene fused with a YFP construct at the Smo C terminus, in the Gt(ROSA)26Sor locus (28). The Smo^YFP gene in the construct contains a point mutation, W539L, rendering it constitutively active. Genotyping information for Smo^YFP allele was provided by the Jackson Laboratory (mutant allele: forward 5′-AAGTTCATCTGCACCACCG-3′ and reverse 5′-TGCTCAGGTAGTGGTTGTCG-3′; wild-type allele: forward 5′-CGTGATCTGCAACTCCAGTC-3′ and reverse 5′-GGAGCGGGAGAAATGGATATG-3′). For amplification, the cycle of 94, 67 and 72°C was repeated 35 times. The expression of the Smo^YFP fusion gene is normally blocked by an upstream STOP fragment flanked by loxP sites. When combined with a Cre recombinase-expressing strain, successful Cre-mediated excision removes the STOP fragment and activates constitutive expression of Smo^YFP (Supplemental Fig. S1). The Smo^YFP mice were crossed with Sf1/Cre transgenic mice, in which Cre recombinase is under the control of the Sf1 promoter (29).

The Smo^YFP females were housed with Sf1/Cre males and checked for vaginal plugs each morning. Detection of a vaginal plug was considered as E0.5. Embryos were genotyped following harvest using described protocols, and their gonads were collected. For adult analysis, test (Sf1/Cre;Smo^YFP) and control (Smo^YFP only) males were collected at postnatal day 35 (PN35) and PN56. All experiments were repeated ≥3 times. All procedures described were reviewed and approved by the Institutional Animal Care and Use Committee at the University of Illinois and the National Institute of Environmental Health Sciences and were performed in accordance with the Guiding Principles for the Care and Use of Laboratory Animals.

Immunohistochemistry and histology

Fetal testes were collected at different ages, fixed in 4% paraformaldehyde at 4°C overnight, and stored in methanol at −20°C. When needed, samples were rehydrated from methanol back to PBS and then processed through sucrose gradient in PBS (10, 15, 20, and 20% sucrose/OCT, 1:1) and embedded in 20% sucrose/OCT (1:3) on dry ice. The frozen blocks were cryosectioned at 8–10 μm. Adult testes were collected at PN35 and PN56, fixed and embedded as above, and then sectioned at 4 μm. Antigen retrieval was performed by boiling the sections on slides in 0.1 M citrate buffer (pH 6.0) for 10 min. Following blocking, sections were incubated at 4°C in different concentrations of primary antibodies: rabbit anti-3-β-hydroxysteroid dehydrogenase (3βHSD; 1:1000) and rabbit anti-SF1 (1:500), both from Dr. Ken Morohashi (National Institutes of Natural Science, Okazaki, Japan), rabbit anti-CYP17 (1:100) from Dr. Dale B. Hales (Southern Illinois University, Carbondale, IL, USA), and rabbit anti-KI67 (AB833, 1:1000; Abcam, Cambridge, MA, USA). Secondary antibodies used were FITC-, rhodamine- or Cy3-conjugated donkey anti-rabbit (all 1:200; Jackson ImmunoResearch, West Grove, PA, USA), and Alexa Fluor 555-conjugated goat anti-rabbit (A-21428, 1:500; Life Technologies, Grand Island, NY, USA). Sections were counterstained with mounting medium containing DAPI before visualization. Fluorescent images were captured using a Fast1394 QImaging Camera (QImaging, Surrey, BC, Canada) installed on a Leica Dmi 4000B microscope (Leica, Solms, Germany). For immunostaining with 2 different primary antibodies from the same species, we used the tyramide amplification (NEL701A, TSA Fluorescein System; PerkinElmer, Waltham, MA, USA) with the first primary antibody, following the protocol described by the manufacturer, as detailed in Büki et al. (30). For histological analysis, testes, epididymides, and seminal vesicles from PN35 and PN56 mice were fixed in 4% paraformaldehyde and paraffin embedded. Sections (4 μm) were stained with hematoxylin and eosin (H&E) following standard protocols.

Serum collection and hormone assays

Blood was collected from PN35 and PN56 mice by cardiac puncture immediately following euthanasia. Blood was allowed to clot and then centrifuged, and the serum was collected and stored at −20°C. Hormone analyses were performed at the University of Virginia Center for Research and Reproduction Ligand Assay and Analysis Core Laboratory (Charlottesville, VA, USA). Testosterone was measured using a commercially available solid-phase RIA kit (Coat-a-Count Total Testosterone Kit; Siemens Medical Solutions Diagnostics, Tarrytown, NY, USA). The sensitivity of the assay was 0.1 ng/ml; intra- and intercoefficients of variation were 3.6 and 7.0%, respectively. Luteinizing hormone (LH) and follicle-stimulating hormone (FSH) concentrations were quantitated using a Milliplex Map Rat Pituitary Kit (Millipore, Billerica, MA, USA). Samples were run in duplicate. The sensitivity of assay was 2.4 ng/ml for FSH and 0.24 ng/ml for LH; intra- and intercoefficients of variations were 6.7 and 16.9% for FSH and 6.9 and 17.2% for LH.

Quantification of SF1⁺ and 3βHSD⁺ cells

One gonad per animal was cryosectioned and processed for sequential immunohistochemistry. For each testis, every fourth section (30 μm apart) was collected and immunostained for both SF1 and 3βHSD. Fluorescent images were captured as described above. Cells positive for SF1, 3βHSD, or both were counted in each section using Image-Pro Discovery software (Media Cybernetics, Silver Spring, MD, USA).

Sperm preparation and computer-assisted sperm analysis (CASA)

Sperm preparation was performed as described previously (31). Briefly, right cauda epididymides were excised, rinsed with medium HS (containing, in mM: 135 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 5 glucose, 30 HEPES, 10 lactic acid, and 1 pyruvic acid, pH 7.4), and then transferred to 1 ml of a swim-out/capacitation medium [HS medium supplemented with 15 mM NaHCO₃ and 5 mg BSA (fraction V)/ml]. Sperm were allowed to swim into the medium from multiple incisions during 10 min incubation at 37°C, 5% CO₂. The resulting sperm suspension was thoroughly mixed and diluted 10-fold in swim-out/capacitation medium. The diluted sperm suspension was subjected to CASA in a 2X-CEL chamber (Hamilton-Thorne Research, Beverly, MA, USA) using the integrated visual optical system (IVOS) motility analyzer (Hamilton-Thorne Research). The default mouse analysis settings of IVOS were used to measure total sperm numbers, the percentage motile sperm, and other velocity parameters. Ten arbitrary and independent fields were analyzed for each sperm suspension.

Stereological analysis of Leydig cell numbers in juvenile and adult testes

Absolute numbers of Leydig cells were compared in testes of Sf1/Cre;Smo^YFP and control mice at PN35 and PN56. The Leydig cell number was determined by stereology, using the fractionator method described originally for rats (32) and subsequently adapted for mice (33). This technique involves the use of 20 randomly selected fields in each of 3 nonadjacent sections per testis that were captured using a Nikon Eclipse E800 microscope (Nikon, Melville, NY) equipped with a ×40 objective and a SPOT RT digital camera (model 2.3.0; Diagnostic Instruments, Michigan City, IN, USA) and interfaced to a computer. The images that were displayed on the monitor represented areas of 0.9 mm² of testis. Leydig cell numbers were estimated using image analysis software (Image-Pro Plus; Media Cybernetics). Total numbers of cells were calculated by multiplying the number of Leydig cells counted in a known fraction of the testis by the inverse of the sampling probability.

Statistical analysis

A 2-tailed t test was used to compare the Hh-activated and normal control groups. A value of P ≤ 0.05 was considered significant.

RESULTS

Increased activation of the Hh pathway in SF1⁺ cells alters the dynamics of FLC populations

Activation of the Hh pathway is required for transforming the SF1⁺ progenitor cells into steroidogenic FLCs (4, 8, 18). We therefore hypothesized that activation of the Hh pathway (referred to as Hh activation hereafter) above the endogenous level could increase FLC populations in testes. We first characterized the developmental sequence of events of FLC development in control fetal testes by coimmunostaining for SF1 and the steroidogenic enzyme 3βHSD. Expression of SF1 identifies somatic progenitor cells, including Sertoli and Leydig cells, whereas 3βHSD marks differentiating FLCs committed to steroidogenesis. At E13.0, SF1 was expressed mostly by Sertoli cells in testis cords (outlined in Fig. 1A) and by only a few cells in the interstitium of control testes. SF1⁺ cells in the interstitium were negative for 3βHSD staining at this time (referred to as SF1⁺/3βHSD⁻ cells; Fig. 1A, yellow arrows). After ∼12 h (E13.5), in addition to the SF1⁺/3βHSD⁻ cells, a new population of cells positive for both SF1 and 3βHSD appeared in the interstitium (SF1⁺/3βHSD⁺ cells; Fig. 1C, white arrowheads). By E16.5, a third population of interstitial cells that lacked SF1 but remained 3βHSD⁺ was seen in control testes (SF1⁻/3βHSD⁺; Fig. 1E, white arrows). At PN1, all 3 populations (SF1⁺/3βHSD⁻, SF1⁺/3βHSD⁺, and SF1⁻/3βHSD⁺) were present in interstitium of control testes (Fig. 1G). Based on the sequence of their appearance, we proposed that SF1⁺/3βHSD⁻ cells in interstitium are progenitors of steroidogenic cells. Between E13 and E13.5, these progenitors differentiate into FLCs as they become steroidogenically active (SF1⁺/3βHSD⁺). Around E16.5, some FLCs lose SF1 expression and become terminally differentiated (SF1⁻/3βHSD⁺).

Figure 1.

Chronological appearance of different populations of interstitial cells in testes of control and Hh-activated (Sf1/Cre;Smo^YFP) mice during fetal and perinatal development. Frozen sections of control (A, C, E, G) and Hh-activated testes (B, D, F, H) were immunostained for SF1 (nuclear green) and 3βHSD (cytoplasmic red). Insets B1, C1, D1, E1, F1, G1, G2, H1, and H2 are magnifications of numbered boxed areas in corresponding images. Yellow arrows indicate SF1⁺/3βHSD⁻ cells, white arrowheads indicate SF1⁺/3βHSD⁺ cells, and white arrows indicate SF1⁻/3βHSD⁺ cells. Each image is representative of testes from ≥3 mice.

Next, we investigated whether increased Hh activation alters the dynamics of FLC appearance and differentiation. At E13, in contrast to control testes where only the SF1⁺/3βHSD⁻ population was present, Hh-activated (Sf1/Cre;Smo^YFP) testes contained both SF1⁺/3βHSD⁻ and SF1⁺/3βHSD⁺ populations (Fig. 1B). After ∼12 h (at E13.5), Hh-activated testes contained significantly more SF1⁺/3βHSD⁺ cells than controls (Fig. 1C, D). At E16.5, the third population (SF1⁻/3βHSD⁺; Fig. 1F) began appearing in testis interstitium, and by PN1, Hh-activated testes contained significantly more SF1⁺/3βHSD⁺ cells in the periphery and more SF1⁻/3βHSD⁺ in the center of the testes compared with controls (Fig. 1H1, H2, respectively).

To quantify differences in cell populations between control and Hh-activated testes, we enumerated 3 cell types in testis sections (see Materials and Methods). At all developmental stages examined, the numbers of SF1⁺/3βHSD⁻ progenitor cells in control testes were significantly higher than those in Hh-activated testes (Fig. 2A). Conversely, SF1⁺/3βHSD⁺ and SF1⁻/3βHSD⁺ populations were lower in the control compared with Hh-activated testes at all ages (Fig. 2B, C). The total number of all 3 populations in Hh-activated testes was not significantly different from that in control testes (Fig. 2D).

Figure 2.

Dynamic changes and percentage distribution of 3 cell populations identified in Fig. 1. A–D) Graphs represent the average number of SF1⁺/3βHSD⁻ cells (A), SF1⁺/3βHSD⁺ cells (B), SF1⁻/3βHSD⁺ cells (C), and 3 cell types (D) combined in testes at stages E13.5, E16.5, E18.5 and PN1. Black lines represent control testes; pink lines represent Hh-activated testes. *P < 0.05, **P < 0.01, ***P < 0.001 between control and Hh activated testis. E) Pie charts represent the percentage of 3 populations in testes at E13.5, E16.5, E18.5, and PN1.

This dynamic shift from the steroidogenically inactive SF1⁺/3βHSD⁻ progenitor population to steroidogenically active populations (SF1⁺/3βHSD⁺ and SF1⁻/3βHSD⁺) was further highlighted by the percentage of each population at individual stages (Fig. 2E). In control testes, the percentage of the steroidogenically inactive SF1⁺/3βHSD⁻ progenitor cells decreased over time from 49% at E13.5 to 30% at E18.5 and then rebounded after birth to 41% (Fig. 2E). In contrast, the percentage of SF1⁺/3βHSD⁺ FLCs gradually increased from 47 to 55% during fetal life and decreased after birth to 35%. The terminally differentiated SF1⁻/3βHSD⁺ population followed the same trend as the SF1⁺/3βHSD⁺ FLC population in control testes. In Hh-activated testes, the percentage of SF1⁺/3βHSD⁻ progenitor cells (6–17%) was significantly reduced at all ages compared with controls (Fig. 2E). In contrast, the SF1⁺/3βHSD⁺ FLC population was increased dramatically to 72–79% prenatally compared with 47–55% in controls and remained higher than the control after birth (44 vs. 35% at PN1). The percentage of terminally differentiated SF1⁻/3βHSD⁺ cells in Hh-activated testes was unchanged prenatally but became higher postnatally compared with control testes (24 vs. 41% at PN1). This observation indicates that increased activation of Hh in the SF1⁺ population promotes differentiation of progenitors (SF1⁺/3βHSD⁻) into steroidogenic FLCs (SF1⁺/3βHSD⁺ and SF1⁻/3βHSD⁺). Consequently, the progenitor population was significantly reduced due to increased Hh activation.

Functional competence of ALCs is not altered in Hh-activated testes

To investigate whether the ALC population is affected by increased Hh activation in the SF1-positive progenitor population, we examined expression of CYP17, a marker for immature Leydig cells and ALCs by immunohistochemistry (Fig. 3). Leydig cell differentiation was obvious in both control and Hh-activated testes at both PN35 (prepubertal) and 56 (young adulthood). However, quantification of Leydig cell numbers revealed a significant reduction of immature Leydig cells at PN 35 in Hh-activated testes compared with controls (47%; P<0.05; Fig. 3E). The numbers of ALCs at PN56 showed a trend toward a reduction in Hh-activated testes but the difference was not statistically significant (Fig. 3E). We also measured serum concentrations of testosterone, LH, and FSH and found no significant differences between Hh-activated and control mice (Table 1).

Figure 3.

Effects of increased activation of the Hh pathway on development of ALCs. A–D) Immunohistochemical staining of CYP17 identifies immature Leydig cells at PN35 and ALCs at PN56 in testis sections from control (A, C) and Hh-activated mice (B, D). Each image is representative of testes from ≥3 different mice. Scale bars = 20 μm. E) stereological enumeration of CYP17⁺ Leydig cells in control and Hh-activated testes at PN35 and PN56. Data are presented as means ± se (n=3). *P < 0.05.

Table 1.

Serum hormone concentrations in control and Hh-activated mice at PN35 and PN56

Age	Genotype	Testosterone (ng/ml)	LH (ng/ml)	FSH (ng/ml)
PN35	Control	0.5 ± 0.2 (7)	1.84 ± 0.78 (4)	104.3 ± 21.0 (4)
	Hh activation	1.0 ± 0.9 (5)	1.77 ± 0.95 (3)	120.2 ± 14.8 (5)
PN56	Control	0.7 ± 0.4 (6)	2.91 ± 1.78 (3)	112.7 ± 14.0 (6)
	Hh activation	1.2 ± 0.6 (7)	2.64 ± 1.24 (3)	96.7 ± 20.2 (6)

Data are presented as means ± se. Values in parenthesis represent sample size.

Effect of increased activation of Hh pathway on male reproductive organs

We next examined the effect of increased Hh activation on development of male reproductive organs when these animals reached adulthood. Visual examination revealed a grossly normal appearance of testes, epididymides, and seminal vesicles in Hh-activated mice (Fig. 4A). However, Hh-activated testes were smaller (Fig. 4A), and testis wet weights (Fig. 4B) and testis wet weight adjusted by body weight (data not shown) were significantly reduced both at PN35 and 56. Epididymal wet weights of Hh-activated mice were also significantly lower than controls at PN35 (37%; P<0.05; Fig. 4C). The epididymis of Hh-activated mice also appeared smaller and weighed less than the controls at PN56, though differences were not significant (Fig. 4A, C, respectively). In contrast, the wet weights of seminal vesicles did not differ between the 2 groups at either time point (data not shown).

Figure 4.

Development of testes and accessory organs in control and Hh-activated mice. A) Male reproductive tracts from PN56 control and Hh-activated mice (arrows indicate testes). B) Testis wet weights in control and Hh-activated mice at PN35 and PN56. C) Epididymal wet weights in control and Hh-activated mice at PN35 and at PN56. D) Total and motile sperm counts in control and Hh-activated mice at PN56. Data are presented as means ± se. *P < 0.05.

Lowered sperm count and decreased sperm mobility in the Hh-activated testes

Histological examination revealed active spermatogenesis in testes of Hh-activated mice and the presence of all germ cell types, including the most differentiated spermatozoa, at PN35 (Fig. 5A, B). All stages of spermatogenesis were also present in Hh-activated testes at PN56, and no apparent histological differences were observed compared with controls (Fig. 5C, D). This observation was corroborated by histological examination of epididymides from Hh-activated mice at PN56, which revealed mature sperm in the cauda epididymis of Hh-activated mice (Fig. 5F). However, increased numbers of sloughed round germ cells were observed in the cauda epididymis of these mice, along with an increased luminal diameter of cauda epididymis. Mature sperm numbers were reduced by 66% in cauda epididymis of Hh-activated mice compared with controls at PN56 (P<0.05; Fig. 4D). In addition, sperm motility in Hh-activated mice was significantly reduced, with only 67% of mature spermatozoa motile in the Hh-activated mice vs. 82% in controls (Fig. 4D).

Figure 5.

Spermatogenesis in control and Hh-activated mice at PN35 and PN56. A–D) Testis sections of control (A, C) and Hh-activated mice (B, D) at PN35 and P56 were stained with H&E. Higher magnifications of images are shown in insets. E, F) H&E stain of the cauda epididymis from PN56 control (E) and Hh-activated mice (F). Arrows indicate sloughed round spermatids. Scale bars = 100 μm.

DISCUSSION

We report for the first time the dynamic transition of SF1⁺ progenitor cells (SF1⁺/3βHSD⁻) into FLCs (SF1⁺/3βHSD⁺ and SF1⁻/3βHSD⁺) in mouse fetal testes. In addition, an increased activation of the Hh pathway in the SF1⁺ progenitor population not only changes the distribution of SF1⁺ progenitors and FLCs during fetal life but also affect the ALC population in adulthood. The nonsteroidogenic SF1⁺/3βHSD⁻ progenitor population first appears in the fetal mouse testis, followed by the steroidogenic SF1⁺/3βHSD⁺ and SF1⁻/3βHSD⁺ FLC populations. We propose that these 3 interstitial cell populations represent 3 different stages of FLCs: progenitor FLCs (SF1⁺/3βHSD⁻), differentiating FLCs (SF1⁺/3βHSD⁺), and terminally differentiated FLCs (SF1⁻/3βHSD⁺). SF1 is a marker for the somatic cell precursors in the gonadal primordium, and inactivation of SF1 causes gonadal agenesis and loss of all somatic cell lineages in the gonad. Sertoli cell-derived DHH maintains SF1 expression and triggers steroidogenesis in the FLC lineage (SF1⁺/3βHSD⁺). The terminally differentiated FLCs (SF1⁻/3βHSD⁺) have not been reported before, and how these cells lose SF1 expression as they mature is not understood. Our findings demonstrate that an increased activation of the Hh pathway in SF1⁺ progenitor cells leads to the premature appearance of differentiating SF1⁺/3βHSD⁺ FLCs while reducing the SF1⁺/3βHSD⁻ progenitor cell population. The decrease in progenitor cells and increase in FLCs numbers likely result from increased differentiation of progenitor cells into FLCs, rather then increased proliferation (refs. 1, 10 and Supplemental Fig. 2).

At the end of fetal life and during the first 2 wk after birth in rodents, FLCs are gradually replaced by ALCs (1, 17). The definitive origin of ALC progenitor cells has not been conclusively identified. Despite their common function in hormone production, ALCs are not derived from preexisting FLCs (1, 26). Rather, ALCs arise from a progenitor population in the interstitium of adult testes. The presence of ALC progenitor (stem) cells was first discovered in experiments in rats involving treatment with ethane dimethanesulfonate (EDS) that specifically eliminated the ALC population. Strikingly, a recovery of the ALC population occurred after EDS treatment (34). The potential stem cells for ALCs were later identified as steroidogenically inactive, spindle-shaped cells that expressed receptors for platelet-derived growth factor receptor-α, leukemia inhibitory factor, and LH (35). The EDS experiments demonstrate the presence of ALC stem cells that are able to replenish the population after chemical eradication. However, it remained unclear from where the stem cell population was originally derived. In this study, we tested the possibility that a common progenitor cell pool in embryonic testes could be a source for both FLCs and ALCs. We hypothesized that forcing embryonic progenitor cells to differentiate into FLCs would decrease the numbers of cells available for the ALC population. The progenitor cells in the fetal testes were induced to become FLCs by an increased activation of the Hh pathway (8). We then assessed whether changes in progenitor cell allocation in fetal life affected the number and function of ALCs. Indeed, we found that increased Hh activation forced differentiation of a majority of the SF1⁺ progenitor cells into FLCs, resulting in a reduction in numbers of SF1⁺ progenitor cells at birth. Depletion of SF1⁺ progenitor cells in the Hh-activated model resulted in significantly decreased numbers of ALCs at PN35, thereby suggesting that fetal SF1⁺ cells are a potential source of ALCs. Interestingly, a partial recovery in ALC numbers was observed at PN56. This partial recovery could result from expansion of the ALC population through stimulatory effects of LH or recruitment of ALCs from the peritubular stem cell population (35, 36).

An active, although abnormal, first wave of spermatogenesis was observed in Hh-activated mice. Despite reduced ALC numbers, these animals had normal serum testosterone concentrations and development of androgen-sensitive organs, such as epididymides and seminal vesicles, indicating functional competence of Leydig cells in Hh-activated testes. Serum LH and FSH concentrations were also within the normal range. Hence, endocrine deficiencies are unlikely to be the cause of decreased spermatogenesis in Hh-activated testes. Similar yet more severe spermatogenic defects were found in mice where the Hh pathway was activated ubiquitously (37).

Our model only activates the Hh pathway in cells derived from the SF1⁺ lineage, i.e., Sertoli and Leydig cells in testes, adrenal cortical cells, gonadotrophs of the anterior pituitary, and cells in the ventromedial hypothalamus. We have examined all other SF1⁺ organs in the Hh-activated mice and have found no abnormality in adrenals or in pituitary and hypothalamus (unpublished results), consistent with the normal serum LH and FSH concentrations. Based on these findings, we speculate that increased activation of the Hh pathway in the SF1⁺ cells in the testis could contribute to subtle changes in the interaction between the somatic cells and germ cells, therefore leading to decreased sperm motility and increased numbers of prematurely sloughed round germ cells. In the normal adult mouse testis, the signaling components of the Hh pathway are expressed in the seminiferous tubules (38). Chemical inhibition of the Hh pathway caused disorganization of seminiferous tubules in culture (39). These observations along with our findings demonstrate the importance of a proper Hh signaling in spermatogenesis.

In summary, our results reveal the presence of 3 distinct cell populations critical for development of FLCs and ALCs. Manipulation of the Hh pathway in the progenitor cell population alters the dynamics of these populations during fetal life and consequently the numbers of ALCs in the adult. Increased Hh activation also resulted in defects in spermatogenesis. Further studies such as genetic lineage tracing experiments are required to understand how some SF1⁺ progenitor cells remain quiescent during fetal stages and whether they are indeed progenitors for ALCs.