Deletion of the Igf1 Gene: Suppressive Effects on Adult Leydig Cell Development

Abstract

Deletion of the insulin-like growth factor 1 (Igf1) gene was shown in previous studies to result in reduced numbers of Leydig cells in the testes of 35-day-old mice, and in reduced circulating testosterone levels. In the current study, we asked whether deletion of the Igf1 gene affects the number, proliferation, and/or steroidogenic function of some or all of the precursor cell types in the developmental sequence that leads to the establishment of adult Leydig cells (ALCs). Decreased numbers of cells in the Leydig cell lineage (ie, 3β-hydroxysteroid dehydrogenase–positive cells) were seen in testes of postnatal day (PND) 14–90 Igf1^−/− mice compared with age-matched Igf1^+/+ controls. The development of ALCs proceeds from stem Leydig cells (SLCs) through progenitor Leydig cells (PLCs) and immature Leydig cells (ILCs). The bromodeoxyuridine labeling index of putative SLCs was similar in the Igf1^−/− and Igf1^+/+ mice. In contrast, the labeling index of PLCs was reduced in the Igf1^−/− mice on each day of PND 14 through PND 35, and that of more mature Leydig cells (referred to herein as LCs, a combination of ILCs plus ALCs) was reduced from PND 21 through PND 56. In Igf1^−/− mice that received recombinant IGF-I, the labeling indices of PLCs and LCs were similar to those of age-matched Igf1^+/+ mice, indicating that the reductions in the labeling indices seen in the PLCs and LCs of the Igf1^−/− mice were a consequence of reduced IGF-I. On each day of PND 21 through PND 90, testicular testosterone concentrations were significantly reduced in the Igf1^−/− mice, as were the expressions of testis-specific mRNAs involved in steroidogenesis, including Star, Cyp11a1, and Cyp17a1. The increased expression of the gene for 5α-reductase (Srd5a1) in adult Igf1^−/− testes suggests that the depletion of Igf1 might suppress or delay Leydig cell maturation. These observations, taken together, indicate that the reduced numbers of Leydig cells in the adult testes of Igf1^−/− mice result at least in part from altered proliferation and differentiation of ALC precursor cells, but not of the stem cells that give rise to these cells.

The differentiation of adult Leydig cells (ALCs) from Leydig precursor cells is a prerequisite for the onset of the increased circulating testosterone levels that accompany puberty. The cells that give rise to ALCs in rats and mice first become apparent by postnatal day (PND) 11–14 as spindle-shaped cells in the testis interstitium that express 3β-hydroxysteroid dehydrogenase (3βHSD). By virtue of their expression of 3βHSD, these cells, which have been termed progenitor Leydig cells (PLCs), are considered to have entered the Leydig cell lineage (Ge et al, 1996). PLCs and immature Leydig cells (ILCs), the intermediate stages of development, primarily produce 5α-reduced androgens because of their relatively higher expression of the androgen-metabolizing enzymes 5α-reductase 1 (SRD5A1) and 3α-hydroxysteroid dehydrogenase (AKR1C9) compared with testosterone-producing ALCs (Ge and Hardy, 1998).

Stem Leydig cells (SLCs) have been shown to be present in neonatal testes, and it has been shown that it is these cells that give rise to PLCs (Ge et al, 2006). The SLCs were shown to be spindle-shaped, 3βHSD-negative, luteinizing hormone (LH) receptor–negative, and platelet-derived growth factor receptor α (PDGFRα)-positive cells that are able to expand in number during prolonged culture in vitro without differentiation. When cultured in a differentiation medium containing a mixture of growth factors, SLCs were induced to express proteins indicative of Leydig cell differentiation, including LH receptor, steroidogenic acute regulatory protein (StAR), P450 cholesterol side chain cleavage enzyme (CYP11A1), 3βHSD, and 17α-hydroxylase/20-lyase (CYP17A1). Moreover, the cells expanded their numbers and differentiated in vivo when injected into the interstitial compartment of ALC-depleted rat testes (Ge et al, 2006).

The proliferation and differentiation of SLCs and PLCs are integrally involved in producing the normal adult population of the testosterone-producing ALCs. It is well established that ALC steroidogenic function is regulated predominantly by LH (Cooke, 1996). Growth factors, prostaglandins, and steroids also have been shown to play regulatory roles in Leydig cell development and function (Saez and Lejeune, 1996). Deletions of PDGFRα or desert hedgehog genes were shown to result in reduced numbers of Leydig cells (Clark et al, 2000; Gnessi et al, 2000). Deletion of the insulin-like growth factor 1 (Igf1) gene resulted in reduced numbers of Leydig cells as compared with age-matched wild-type mice (Baker et al, 1996). In previous studies, IGF-I was shown to stimulate the proliferation of PLCs and ILCs, and deletion of the Igf1 gene to cause reduced numbers of ALCs at PND 35 (Wang et al, 2003; Wang and Hardy, 2004). However, these previous studies did not address which cell types in the Leydig cell lineage might be affected by IGF-I, nor the effect of the Igf1 gene deletion on ALC gene expression and steroidogenic function in the adult mouse.

In the present study, we hypothesized that IGF-I plays an important regulatory role in the developmental sequence leading to the ALC population at least in part through its effect on the proliferation and/or differentiation of one or more of the Leydig cell precursor cells.

Materials and Methods

Animals and Treatments

Igf1^−/− mice were produced as described previously (Baker et al, 1996). In brief, adult heterozygotes (Igf1^+/−) with a predominantly MF1 × 129/Sv background and bearing the targeted Igf1 gene deletion (provided by Dr Argiris Efstratiadis, Columbia University, New York, New York) were mated to produce Igf1^−/− mice. The genotypes of the Igf1^−/− and Igf1^+/+ male mice were ascertained by multiplex polymerase chain reaction (PCR) of DNA samples obtained from tail tips (Baker et al, 1996). Mice were observed twice daily (morning and evening) for litters. The day of birth was considered as PND 1. All animal procedures were approved by the Institutional Animal Care and Use Committee of Rockefeller University.

Mice were sacrificed on PND 14, 21, 28, 35, 49, 56, and 90 (n = 12 per group for each of Igf1^−/− and Igf1^+/+ mice). At 2 hours before they were killed, 6 of the 12 mice of each age from the Igf1^−/− and Igf1^+/+groups received an intraperitoneal injection of bromodeoxyuridine (BrdU; 40 μg/g body weight; Boehringer Mannheim, Indianapolis, Indiana; Wang et al, 2003). Testes were collected from the other 6 mice per group for assay of intratesticular testosterone and of 5α-reduced androgens, and for total RNA analysis (see below). Another group of Igf1^−/− mice received either human recombinant IGF-I (Intergen Company, Purchase, New York; 0.8 μg/g body weight/d, n = 6) or vehicle (n = 6) by subcutaneous injection every 12 hours for 2 days, from PND 26 to PND 28, the peak time of proliferation of Leydig cell precursor cells (Ge et al, 1996). The IGF-I dose used for this study, 0.8 μg/g body weight/d, was shown to be effective in stimulating cell proliferation in a previously published study (Wang et al, 2003). BrdU injections to these mice were administered 2 hours prior to their sacrifice. The animals were killed at PND 28, 2 days after injection of IGF-I.

Identification of SLCs and Cells in the Leydig Cell Lineage

Mice were anesthetized by IP injection of sodium pentobarbital (25 mg/100 g body weight; Abbott Laboratories, North Chicago, Illinois). Testes were fixed with Bouin by whole-body perfusion via the left ventricle of the heart. The testes were stored in the fixative overnight and subsequently embedded in paraffin for immunocytochemical analysis. Cells in the Leydig cell lineage were identified by their staining for PDGFRα and for 3βHSD. Immunohistochemical staining was performed using the ABC Elite kit (Vector Laboratories, Burlingame, California). For BrdU staining (Wang et al, 2003), antigen retrieval was carried out by microwave irradiation for 10 minutes in 10 mM (pH 6.0) citrate buffer, and endogenous peroxidase was blocked with 0.5% H₂O₂ in methanol for 30 minutes. The sections were then incubated with a monoclonal anti-BrdU antibody (1:1000; RPN 202; Amersham Biosciences, Little Chalfont, United Kingdom) for 30 minutes at room temperature. Antibody bound to the nuclei was visualized with diaminobenzidine (Cat. No. sk-4100; Vector Laboratories), and the labeled nuclei were stained black by adding a nickel solution to the chromogen. After washing, the sections were double-labeled by incubation with a 3βHSD polyclonal antibody diluted 1:3000 (provided by Dr Van Luu-The, Laval University, Quebec, Canada) for 1 hour at room temperature. The antibody-antigen complexes were visualized with diaminobenzidine alone, resulting in brown cytoplasmic staining in positively labeled Leydig cells (Wang et al, 2003). The sections were counterstained with Mayer hematoxylin. In control experiments, sections were incubated with nonimmune rabbit immunoglobulin G (IgG; 3βHSD) or mouse IgG (BrdU) using the same working dilution as the primary antibody.

Computer-Assisted Image Analysis and Stereology

Absolute numbers of cells were determined by stereology, using the fractionator method as described previously (Hardy et al, 1989). In brief, 20 randomly selected fields in each of 3 nonadjacent sections per testis were captured using a Nikon Eclipse E800 microscope (Nikon, Inc, Melville, New York) equipped with a 40 × objective and a SPOT RT digital camera (model 2.3.0; Diagnostic Instruments Inc, Michigan City, Indiana) and interfaced to a computer. The images that were displayed on the monitor represented areas of 0.9 mm² of testis. Interstitial cell numbers were estimated using image analysis software (Image-Pro Plus; Media Cybernetics, Silver Spring, Maryland). The total number of cells was calculated by multiplying the number of cells counted in a known fraction of the testis by the inverse of the sampling probability.

The identifications of the different interstitial cell types (peritubular myoid cell, SLC, PLC, ILC/ALC, and endothelial) were based on their well-established nuclear morphologies, locations, and staining characteristics, as described in detail previously (Hardy et al, 1989) and also briefly in “Results.” The BrdU labeling indices for putative SLCs, PLCs, and LCs (ILCs plus ALCs) were determined as the number of BrdU-labeled cells of a given type divided by the total number of labeled plus unlabeled cells of that type, multiplied by 100. More than 200 cells of each type were counted per testis.

Testosterone and 5α-androstane-3α, 17β-diol Concentrations

The prepubertal testis produces 5α-androstane-3α,17β-diol (3αDIOL) as its primary product, and the mature testis primarily produces testosterone (Ge and Hardy, 1998). Consequently, the levels of 3αDIOL and testosterone are reflective of the maturity of the Leydig cells. Testes of Igf1^+/+ and Igf1^−/− mice (6 mice per group) were homogenized to extract steroids, and both 3αDIOL and testosterone were measured as previously described (Cochran et al, 1981). In brief, whole testes (prepubertal) or testis pieces were weighed. The steroids were extracted with methanol and the recovery rate determined by adding 1000 cpm testosterone or 3αDIOL. Testosterone and 3αDIOL concentrations were measured by radioimmunoassay as previously described (Ge and Hardy, 1998). Interassay and interassay variations were between 7% and 8%.

Quantitative PCR

Total RNAs were extracted from the testes of Igf1^−/− and Igf1^+/+ mice in Trizol, according to the manufacturer’s instructions (Invitrogen, Carlsbad, California). The 12 genes that were analyzed, and the primers used for their analysis, are listed in the Table. The RNA was reverse transcribed using random hexamers and MMLV reverse transcriptase (Promega, Sunnyvale, California). The cDNA synthesis step and real-time PCR were performed as previously described (Ge et al, 2007). The RNA levels for the housekeeping gene ribosomal protein S16 (Rps16) were assayed in all samples as an internal control. Measurements of mRNA were normalized using a robust global normalization algorithm. All control crossing threshold (Ct) values were corrected by the median difference in all samples from Rps16. All samples were then normalized by the difference from the median Ct of the 3 corrected control gene Ct levels in each sample. Levels of mRNA were calculated relative to Rps16 mRNA. Quantitative PCR was carried out in a 25-μL volume with SYBR Green PCR Mix (PE Applied Biosystems, Foster City, California). Reactions were carried out and fluorescence was detected on a GeneAmp 5700 system (PE Applied Biosystems).

Table.

Primer sequences and PCR size

Gene Name	Symbol	Primers	Size (bp)	Function
Kit oncogene	Kit	Forward: 5′ATTGGCTTTGTGGTCGCA3′ Reverse: 5′GGCACTTGGTTTGAGCATCT3′	308	Receptors for growth factor and hormone
Luteinizing hormone receptor	Lhcgr	Forward: 5′GAGAAGCGAATAACGAGACG3′ Reverse: 5′TTAGCCAAATCAACACCCTAA3′	197
Scavenger receptor class B, member 1	Scarb1	Forward: 5′GCCAGCGTGCTTTTATGA3′ Reverse: 5′CCGTTCCATTTGTCCACC3′	216	Cholesterol transporting in Leydig cells
Steroidogenic acute regulatory protein	Star	Forward: 5′AAAAGACACGGTCATCACTCA3′ Reverse: 5′CCACCCCTTCAGGTCAATAC3′	267
Cytochrome P450, family 11, subfamily a1	Cyp11a1	Forward: 5′GCACTTTGGAGTCAGTTTACATC3′ Reverse: 5′AGGACGATTCGGTCTTTCTT3′	180	Testosterone biosynthesis in Leydig cells
Cytochrome P450, family 17, subfamily a1	Cyp17a1	Forward: 5′CCAGGACCCAAGTGTGTTCT3′ Reverse: 5′CCTGATACGAAGCACTTCTCG3′	547
3β-hydroxysteroid dehydrogenase isoform 6	Hsd3b6	Forward: 5′TCCTCTGCCCCTGCTCTA3′ Reverse: 5′TCTGCTTTGCTTCCTCCC3′	196
5α-reductase isoform 1	Srd5a1	Forward: 5′CACATCCTGCGGAATCTGA3′ Reverse: 5′TGCTGCCTCGCTCTGGT3′	200	Testosterone metabolism
Kit ligand	Kitl	Forward: 5′ATGAAGAAGACACAAACTTGGATTA3′ Reverse: 5′GATCTCCTTGGTTTTGACAAGAGGA3′	84	Sertoli cell–secreted ligand
Insulin-like growth factor 3	Insl3	Forward: 5′CGCTGCTACTGATGCTCCT3′ Reverse: 5′ACAGGTCTTGCTGGGTGC3′	338	Leydig cell–secreted ligand
Proliferating cell nuclear antigen	Pcna	Forward: 5′GATGCCGTCGGGTGAAT3′ Reverse: 5′TCTATGGTTACCGCCTCCTC3′	179	Cell proliferation
Follicle stimulating hormone receptor	Fshr	Forward: 5′CTGGCATTCTTGGGCTCG3′ Reverse: 5′GGGCGGAATCTCGGTCA3′	101	Sertoli cell marker

Statistical Analysis

Data are expressed as mean ± SEM. Leydig cell numbers, labeling indices, and gene expression levels between Igf1^−/− and Igf1^+/+ mice were compared by Student’s t test. The normality of data was checked. Analysis of variance (ANOVA) was used to determine if there were group differences in labeling indices among the Igf1^−/−, Igf1^+/+, and Igf1^−/− plus IGF1 groups. If P < .05 was identified by ANOVA, the Newman-Keuls test was used to determine differences between groups.

Results

Cells in the Leydig cell lineage, identified by their staining for 3βHSD, were enumerated in the testes of PND 14–90 Igf1^−/− and Igf1^+/+ mice by stereological analysis (Figure 1). The total number of 3βHSD-positive cells (denoted “Leydig cell number”) was significantly lower in Igf1^−/− than in age-matched Igf1^+/+ mice at each age examined from PND 14 to PND 90. At PND 90, there were about 3 million 3βHSD-positive cells per testis in Igf1^+/+ mice, but only about 0.5 million per testis in Igf1^−/− mice. Notably, there was a 4-fold increase in 3βHSD-positive cell numbers from PND 21 to PND 28 in the Igf1^+/+ testes, whereas there was little increase during this time period in Igf1^−/− testes, indicative of a decrease in the proliferation of PLCs, the predominant cell type during this period (Ge et al, 1996).

Figure 1.

Numbers of 3β-hydroxysteroid dehydrogenase (3βHSD)-positive cells (denoted “Leydig cell number”) in testes of postnatal day (PND) 14–90 Igf1^+/+ (●) and Igf1^−/− (○) mice, as determined by stereological analysis. Shown are means ± SEM, n = 14 each point.

As reported previously (Ge et al, 1996), the development of ALCs proceeds from SLCs through PLCs and ILCs. The putative SLCs were characterized as spindle-shaped, PDGFRα-positive, and 3βHSD-negative cells. The PLCs also are spindle-shaped and PDGFRα-positive, but 3βHSD-positive. It is the latter characteristic that distinguishes them from putative SLCs in sections. The ILCs and ALCs are PDGFRα-positive round cells that are more intensively 3βHSD-positive than the PLCs. ILCs and ALCs were not easily distinguishable from each other, and therefore were evaluated together and referred to as LCs. Fetal Leydig cells, round cells with prominent cytoplasmic lipid inclusions that tended to cluster together, were seen at all ages. There were few such cells at any age, and particularly so in the adult.

To determine whether the reduced numbers of 3βHSD-positive cells in the adult Igf1^−/− mice were related to changes in the proliferation of some or all of the precursor cell types, labeling indices (percentage of BrdU-labeled cells) were determined for putative SLCs, PLCs, and LCs in Igf1^−/− and Igf1^+/+ mice that had been administered BrdU (Figure 2). The labeling index of the putative SLCs decreased progressively with postnatal age in the testes of both Igf1^+/+ and Igf1^−/− mice, with no statistical differences in age-matched Igf1^+/+ and Igf1^−/− mice at any age from PND 14 to PND 90. In contrast, Igf1 perturbation had significant effects on the proliferation of PLCs and LCs. With respect to PLCs, 15%–20% were BrdU-positive in the Igf1^+/+ testes at PND 14 and 21, but few labeled cells were seen in the Igf1^−/− testes. By PND 28, the percentage of labeled PLCs had decreased to 5% in the Igf1^+/+ testes, but a significantly lower percentage was labeled in the Igf1^−/− testes. By PND 35, very few PLCs were labeled in either the Igf1^+/+ or the Igf1^−/− testes. With respect to LCs, about 5% were labeled on each day of PND 21, 28, 35, and 56 in the Igf1^+/+ testes, but none were labeled in the Igf1^−/− testes. These observations, in combination with the observed reduced numbers of 3βHSD-positive cells seen in Igf1^−/− compared with Igf1^+/+ testes at all ages examined, are consistent with the conclusion that the reduced numbers of Leydig cells in the adult testis are due, at least in part, to the significantly reduced proliferation of PLCs and LCs, but not to changes in the ability of SLCs to proliferate.

Figure 2.

Labeling indices of putative stem Leydig cells (SLCs), progenitor Leydig cells (PLCs), and LCs (immature Leydig cells [ILCs] plus adult Leydig cells [ALCs]). Postnatal day (PND) 14–90 Igf1^+/+ and Igf1^−/− mice were injected with bromodeoxyuridine (BrDU). Double labeling with BrdU and 3β-hydroxysteroid dehydro-genase (3βHSD) was performed to identify the cell types. Mean ± SEM, n = 6. *, significant difference between age-matched Igf1^−/− and Igf1^+/+ at P < .05.

If, as suggested by the data presented in Figure 2, reduced IGF-I is the cause of the changes in the proliferation of PLCs and LCs and thus of reduced LC numbers, the administration of IGF-I to Igf1^−/− mice should overcome the effect of deleting the Igf1 gene. To test this, Igf1^−/− mice were injected with recombinant IGF-I from PND 26 to PND 28, an interval chosen because IGF-I peaks in Igf1^+/+ testes from PND 21 to PND 28 (Ge et al, 1996). Mice also received an injection of BrdU 2 hours before they were killed on PND 28. As seen in Figure 3A, neither deletion of the Igf1 gene nor IGF-I administration altered the SLC labeling index in the Igf1^−/− mice. In contrast, whereas the Igf1 deletion significantly decreased the labeling indices for the PLCs (Figure 3B) and LCs (Figure 3C), the labeling indices for both cell types were maintained at wild-type (Igf1^+/+) levels in Igf1^−/− mice administered IGF1 for 2 days.

Figure 3.

Effects of IGF-I administration to mice on the labeling indices of putative stem Leydig cells (SLCs) (A), progenitor Leydig cells (PLCs) (B) and LCs (immature Leydig cells [ILCs] plus adult Leydig cells [ALCs]) (C) in Igf1^−/− testes. Three groups are compared: Igf1^+/+ mice, Igf1^−/− mice, and Igf1^−/− mice administered recombinant IGF-I from postnatal day (PND) 26 to PND 28. All mice received an injection of bromodeoxyuridine (BrdU) 2 hours before they were killed on PND 28. Shown are means ± SEM, n = 6. Similar letter designates no significant difference between groups (P >.05).

The reduced numbers of 3βHSD-positive cells in the testes of Igf1^−/− mice suggested that testicular steroid production would differ between Igf1^−/− and Igf1^+/+ mice. Figure 4 shows the results of measuring testosterone concentration per mg testis (Figure 4A) or per testis (Figure 4B) in Igf1^+/+ and Igf1^−/− mice from PND 14 to PND 90. As expected, the testosterone concentration in the testes of Igf1^+/+ mice increased considerably from PND 28 to PND 35 and again from PND 35 to PND 49, reaching its peak at PND 49 and remaining high through PND 90. On each day of PND 35, 49, 56, and 90, testicular testosterone concentration was significantly reduced in the testes of Igf1^−/− mice compared with age-matched Igf1^+/+ mice. 3αDIOL, the primary androgen in immature testes, peaked in Igf1^+/+ mice at PND 35 and then was reduced thereafter through PND 90 (Figure 4C and D). The ratio of 3αDIOL to testosterone, a surrogate for relative numbers of relatively immature to mature cells, peaked in Igf1^+/+ mice at PND 35 and then was reduced thereafter through PND 90 (Figure 4E). In contrast, the 3αDIOL to testosterone ratio increased after PND 35 in Igf1^−/− mice (Figure 4F), suggesting delayed developmental maturation in the Igf1^−/− mice.

Figure 4.

Testicular testosterone (A, B) and 5α-androstane-3α,17β-diol (3αDIOL) (C, D) in Igf1^+/+ (black bars) and Igf1^−/− (white bars) mice, expressed as ng per mg testis (A, C) or ng per whole testis (B, D). (E) Ratio of 3αDIOL to testosterone (T) in wild-type mice; (F) ratio of 3αDIOL to testosterone in Igf1^−/− mice. Shown are means ± SEM, n = 4–9. Letters a, b, and c denote significant difference between age-matched Igf1^+/+ and Igf1^−/− mice at P < .05.

Reductions in the expression of testis-specific mRNAs also were associated with deletion of the Igf1gene. As seen in Figure 5, LH receptor (Lhcgr), ckit (Kit), scavenger receptor class B member 1 (Scarb1), steroidogenic acute regulatory protein (Star), CYP11A1 (Cyp11a1), 3βHSD6 (Hsd3b6), and CYP17A1 (Cyp17a1) all had lower steady-state levels in the testes of Igf1^−/− mice at PND 56 compared with age-matched Igf1^+/+ controls. Neither kit ligand (Kitl) nor proliferating cell nuclear antigen (Pcna) was affected by the deletion. Interestingly, insulin-like growth factor 3 (Insl3), a marker for mature Leydig cells, was significantly reduced, consistent with the suggestion above that there might be delayed maturation of Leydig cells in Igf1^−/− mice. Moreover, the gene for 5α-reductase (Srd5a1) was at significantly higher levels in the Igf1^−/− testis at PND 56 compared with the Igf1^+/+ testis, which also is consistent with delayed maturation of the adult cells and thus a higher 5α-reduced androgen (3αDIOL) to testosterone ratio (Figure 4F). The observation of reduced expression of the follicle-stimulating hormone receptor gene (Fshr) in the Igf1^−/− mice suggests that Sertoli cells also may be affected by deletion of the Igf1 gene.

Figure 5.

Quantitative polymerase chain reaction (PCR) analysis of known Leydig cell and Sertoli cell genes in testes from Igf1^+/+ (black bars) and Igf1^−/− (white bars) mice. Mean ± SEM, n = 6. * indicates significant difference at P < .05; **, P < 0.01; ***, P < 0.001.

Discussion

Although LH, the primary regulator of Leydig cell function, is required for Leydig cell development (Ge and Hardy, 2007), the initial proliferation of SLCs has been shown to occur in the absence of LH (Teerds et al, 1989). Among its other functions, LH has been shown to stimulate IGF-I secretion and also to upregulate IGF-I receptor gene expression in Leydig cells (Lin et al, 1988; Cailleau et al, 1990; Nagpal et al, 1991). These observations prompted our comparison of the development of ALCs in the testes of Igf1^−/− and Igf1^+/+ mice.

Using stereological analysis, we found significantly increased numbers of 3βHSD-positive cells in the testes of PND 21–28 wild-type mice, but not in mice in which the Igf1 gene was deleted. Increases were seen in the numbers of 3βHSD-positive cells in the testes of the Igf1^−/− mice on days thereafter. However, the increases were dramatically reduced from those seen in the Igf1^+/+ mice, and therefore there were significantly fewer 3βHSD-positive cells in the testes of adult Igf1^−/− mice than in the wild-type mice. These results were consistent with those of a previous study of the effects of targeted gene deletion of IGF-I, which demonstrated diminished numbers of Leydig cells, as well as lower circulating testosterone levels, in peripubertal (35-day-old) males (Wang and Hardy, 2004). The current and previous observations suggest that IGF-I plays a role in the establishment of the normal numbers of ALCs.

To address the mechanism by which deletion of IGF-1 results in reduced numbers of Leydig cells, we examined the proliferation of Leydig cell precursors by BrDU staining. Significantly reduced proliferation of PLCs and LCs (ILCs plus ALCs) was seen in PND 14–90 Igf1^−/− as compared with Igf1^+/+ mice. In striking contrast, the proliferation of putative SLCs was not affected by the deletion of the Igf1 gene. These results suggest that the decreased numbers of ALCs in Igf1^−/− adult testes result at least in part from the reduced proliferation of PLCs and ILCs in response to Igf1 depletion. This conclusion is consistent with previous in vitro studies, which also suggested that IGF-I stimulates the proliferation and differentiation of rat Leydig cell precursors (Khan et al, 1992; Ge and Hardy, 1997). However, given that IGF-I has been reported to have antiapoptotic properties (Colón et al, 2007), it is possible that increased apoptosis also might contribute to reduced Leydig cell numbers in the Igf1^−/− mice. Administering IGF-I to the Igf1^−/− mice prevented the reduced proliferation of PLCs and LCs seen in mice that did not receive IGF-I. Thus, although there may be growth factors in addition to IGF-I (eg, transforming growth factor α) that might be involved (Khan et al, 1992), it seems clear that IGF-I is critical for the proliferation of Leydig cell precursor cells and thus for normal Leydig cell numbers and function. These conclusions are consistent with the marked increase in steroidogenesis shown to occur when growth hormone–deficient Snell dwarf mice, which have low circulating IGF-I levels, were administered IGF-I (Dombrowicz et al, 1992).

The data are consistent with a direct effect of IGF-I on Leydig cell precursors. However, IGF-I effects also might be indirect, through other testicular cells. IGF-I is reported to localize to peritubular cells and spermatocytes in addition to Leydig cells (Handelsman et al, 1985; Lin et al, 1986a,b; Dombrowicz et al, 1992), and its receptor to spermatocytes, Sertoli cells, and Leydig cells (Vannelli et al, 1988). We found that Fshr, produced by Sertoli cells, is reduced significantly in Igf1^−/− mouse testes, suggesting altered Sertoli cell function. Additionally, IGF-I is proproliferative and antiapoptotic for Sertoli cells (Froment et al, 2007). Thus, one possibility is that the loss of IGF-I might lead to the reduced secretion of growth factors by Sertoli cells, thereby indirectly affecting the development of Leydig cells. As yet, this possibility is speculative.

Taken together, the results presented herein indicate that the reduced numbers of Leydig cells in the adult testes of Igf1^−/− mice is a consequence, at least in part, of altered proliferation and perhaps of differentiation of ALC precursor cells, but not of the stem cells that give rise to these cells. Regardless of which pertains, the data predict the possibility of relatively immature cells in the adult testis. Indeed, our observations of reduced testosterone levels, altered testicular 3αDIOL to testosterone ratios, reduced expression of genes encoding proteins and enzymes involved in steroidogenesis (Star, Cyp11a1, and Cyp17a1), and increased levels of 5α-reductase in the testis, taken together, suggest that deletion of the Igf1 gene, and thus of IGF-I, indeed, results in increased numbers of relatively immature cells in the adult testes of Igf1^−/− mice. It should be noted that increased 5α-reductase mRNA, measured in whole testis, occurred despite the reductions in Leydig cell numbers, indicating that the differences in mRNA levels between control and Igf1^−/− mice are indicative of altered Leydig cell steroidogenic function in these mice, and perhaps of the low testosterone production. The mechanism by which IGF-I affects the expression of steroidogenic and metabolizing enzymes is not clear.