| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Research Article |
School of Molecular and Biomedical Science,3 University of Adelaide, Adelaide, South Australia 5005, Australia
Institute for Hormone and Fertility Research,4 University of Hamburg, 20246 Hamburg, Germany
Inserm,5 U 625, GERHM; IFR 140; University of Rennes I, Campus de Beaulieu, F-35042 Rennes, France
Male Health Care,6 Schering AG, 13353 Berlin, Germany
Howard Florey Institute,7 University of Melbourne, Victoria 3010, Australia
ABSTRACT
The new peptide hormone insulin-like peptide 3 (INSL3) is a member of the insulin-relaxin family, yet, unlike insulin, it signals through a new G-protein coupled receptor, LGR8, distantly related to the receptors for LH and FSH. INSL3 is produced in large amounts by the Leydig cells of the testis in both fetal and adult mammals. Using a combination of mRNA analysis by RT-PCR, immunohistochemistry, ligand-binding, and/or bioactivity assays, the distribution of LGR8 expression was assessed in testicular tissues and cells and in the epididymis. There was consistent agreement that LGR8 was expressed in meiotic and particularly postmeiotic germ cells and in Leydig cells, though not in Sertoli or peritubular cells. Leydig cells appear to express only a low level of the LGR8 gene product; other transcripts may be present, representing nonfunctional products. Messenger RNA analysis suggested that LGR8 transcripts in germ cells represented mostly full-length forms. LGR8 mRNA was also expressed in the epididymis, though no function can yet be ascribed to this expression. Therefore, the INSL3/LGR8 system represents a further paracrine hormone-receptor system in the testis, which conveys information about Leydig cell status to germ cells, and possibly as part of an autocrine feedback loop.
Leydig cells, relaxin, spermatogenesis, testis
INTRODUCTION
The relaxin-like peptide hormone insulin-like peptide 3 (INSL3; previously known as relaxin-like factor, RLF) is a major new circulating hormone in the male [1, 2]. It is produced almost exclusively by the Leydig cells of the testis, with anorchid men having undetectable circulating levels of the hormone [3]. INSL3 is made by both fetal and adult-type populations of Leydig cells, but only when these have attained their mature phenotype [4]. Studies in postnatal rodents show immunohistochemical staining for INSL3 epitopes only after about Day 20 in the rat [5, 6] and Day 15 in the mouse [4]. Various studies have indicated that the Insl3 gene and protein are expressed constitutively once Leydig cells are mature, thus making INSL3 an excellent marker of Leydig cell differentiation status [4, 5]. It has been used in this context to study changes in testis function in seasonally breeding animals [7, 8], and during aging [9], as well as to identify unequivocally Leydig cells within testis sections [10] and in primary culture [11].
Currently, little is still known about its function. Mice whose Insl3 gene has been deleted are characterized by cryptorchidism and secondary complications [12, 13]. If the cryptorchidism is surgically corrected at birth, there appear to be no further phenotypic consequences of the mutation in the male. In female mice, however, disruption of the Insl3 gene leads to a reduction in fertility and increased apoptosis in both steroidogenic and germ cells [12, 14]. A recent study in rats whose testes were subjected to an apoptogenic insult has suggested that in males, too, INSL3 may be acting in the adult as a survival or antiapoptotic factor vis-a-vis germ cells [15]. In rodents, INSL3 expression in the fetal testis appears to be a target for endocrine disruption by xenoestrogenic substances in the environment, in particular by phthalates [6, 16] or by diethylstilbestrol [17, 18]. A direct link between cryptorchidism and fetal INSL3 expression in humans is still lacking, though circumstantial evidence is increasing.
The receptor for INSL3 has been recently identified as the novel G-protein coupled receptor (GPCR) LGR8, also known as Great [19, 20]. This receptor is structurally closely related to the GPCR LGR7, which is now shown to be the principal receptor for relaxin in all reproductive tissues. Like the relaxin-LGR7 system, INSL3 appears to act through its receptor to induce an intracellular increase in cAMP, both in transfected cells overexpressing cloned Lgr8 [20], and in gubernacular [20] and prostate [21] cells with naturally expressed receptors. Knockout mice (Great mice), whose Lgr8 gene is deleted, have an identical cryptorchid phenotype to that of the Insl3 knockout mice [19, 22], and further breeding experiments [23] have conclusively shown that no other ligand is able to activate LGR8, nor is any other receptor able to respond to INSL3. It has been shown that LGR8 expressed by the cells of the fetal gubernacular ligament, which attaches the testis ventrally to the inguinal body wall, responds to INSL3 by causing a growth and condensation of the ligament, effectively retaining the testis in the inguinal region during the first, so-called transabdominal phase of testicular descent [24].
The INSL3-LGR8 system appears to be evolutionarily relatively modern [25, 26]. There is no evidence for such a system in any submammalian vertebrate or invertebrate, and it would appear to have evolved parallel to the evolution of the relaxin (as opposed to relaxin-3)-LGR7 system in female physiology. Relaxin-3 is now recognized as being the ancestral hormone, is expressed predominantly in the brain, and acts through another class of GPCR (GPCR135) [27]. Thus, although the LGR7/LGR8 subfamily of GPCRs appears to be quite old [28], its association with relaxin and INSL3 seems to have evolved together with the development of classical mammalian traits such as a scrotal testis and viviparity.
Preliminary studies suggest that Lgr8 transcripts are expressed in a wide range of tissues, including the testis [15, 24]. However, there is no detailed knowledge as to the cell types expressing this INSL3 receptor, nor as to whether the transcripts are indeed manifest as the encoded receptor protein. The present study was carried out to characterize the expression of LGR8 within the testis, and makes use of both rodent and human tissues.
MATERIALS AND METHODS
Preparation of Tissues and Cells
The mouse Sertoli cell lines SK11 and SK49, as well as the mouse Leydig tumor cell line MA-10, were cultured as previously described [29]. Primary mouse Leydig cells were prepared from the testes of adult (60-day) NMRI mice also as previously described [11]. Rat gubernacula were dissected from various pre- and postnatal stages of Wistar strain rats and immediately extracted for RNA preparation. Whole rat testes from control Wistar rats and from rats treated with a single bolus of ethane dimethylsulfonate (EDS) were as in Spiess et al. [30]; primary rat Leydig cells and seminiferous tubules were prepared as in Ivell et al. [31]. Various rat germ cell stages were prepared independently on two occasions by differential elutriation. Those ascribed letters G, F, C, N, or L are exactly as in Ivell et al. [32]; others were prepared as in Pineau et al. [33]. The Kitw/Kitw-v mice were obtained from The Jackson Laboratory, Bar Harbor, ME. All other rat or mouse tissues were dissected from freshly killed adult animals and immediately frozen in liquid nitrogen before processing for RNA extraction as below. All procedures were followed in accordance with National Institutes of Health guidelines, and under the approval of the local animal ethics committee.
RT-PCR analysis
Total RNA was prepared from all tissues and cells using the RNAwiz reagent (Ambion), and checked for rRNA integrity by agarose gel electrophoresis. A 2:1 ratio of sharp and clear ethidium bromide stained 28S:18S rRNA bands was observed for all the samples. Single-stranded cDNA was made using Superscript II reverse transcriptase (Invitrogen) primed from oligo(dT), exactly according to the manufacturer's instructions. Semiquantitative RT-PCR was carried out to characterize Lgr8 transcripts from mouse and rat RNA using different sets of oligonucleotide primers as listed in Table 1. All primers were designed, as far as possible, so as to recognize both rat and mouse transcripts equally well. The positions of the different primers and their products are shown in Figure 1. The complete rat Lgr8 cDNA was compiled from the cDNA entry in the Genbank database (accession number NM001012475) extending this 5' of the translation start by looking at the upstream genomic sequence from the rat, assuming a similar exon structure to the mouse gene. The 3' UTR sequence was obtained from the available EST sequences. RT-PCR was carried out for the constitutively expressed transcript for ribosomal protein S27a, exactly as previously described [34]. This gene product has been shown not to vary across a range of tissues and testicular developmental stages [34, 35], and care was taken to use a moderately low cycle number so as to maintain amplification within the linear phase. Finally, RNA from some rat cells and tissues was additionally analyzed by quantitative real-time (q) RT-PCR (qRT-PCR) using the primer combination 82–83 (see below). All PCR reactions for all samples were repeated at least three times; typical results are illustrated. All PCR products were subjected to DNA sequencing to confirm identity, except for the products of mouse Leydig cell RNA with primers R1 and R16, for which the appropriately sized bands, or the regions of gel corresponding to these bands, were eluted and subjected to a heminested PCR using primers R15 and R16, resulting in shorter PCR products of the expected size.
|
|
Semiquantitative RT-PCR conditions were as follows. Each reaction mixture of 20 µl contained 1 µl of the cDNA, 1 µM of each primer, and 1x of premixed TaqTM (TaKaRa ExTaqTM; Takara Bio). For all Lgr8 primers (except for 82/83, R1/R15, and R15/R16; see below) in semiquantitative PCR (performed on an MJ Research PTC 200 thermal cycler), after an initial denaturation at 95°C for 3 min, reactions used a touchdown protocol with annealing temperature descending from 59°C to 53°C over three cycles, followed by 31 cycles at 51°C; denaturation was at 95°C, with elongation at 72°C. All steps lasted 1 min. There was a final elongation step of 72°C for 10 min. For primer pair 82/83, touchdown was applied for two cycles only from 57°C to 55°C, followed by 40 cycles at 53°C. All other conditions were as above. For primer pairs R1/R16 and R15/R16, annealing was at a constant 60°C for 42 cycles, otherwise as above. For S27a, conditions were set at initial denaturation at 95°C for 2 min, followed by 20 cycles of denaturation at 95°C, annealing at 62°C, and elongation at 72°C, all of 20 sec duration. There was a final elongation step of 72°C for 10 min. For the mouse PCR products of primers R1 and R16, reactions were additionally scaled up to 50 µl for cDNA samples from testis, primary Leydig cells, and MA-10 cells, 40 µl of the resulting products were electrophoresed, specific bands were eluted, and then 2 µl of the respective eluates were subjected to a heminested PCR reaction using primers R15 and R16. All PCR products were analyzed by gel electrophoresis on 1.5% agarose gels with ethidium bromide staining, followed by fluorescence digitization using a Bio-Rad GelDoc XR system (Bio-Rad).
For real-time qRT-PCR (employing the Rotor-Gene RG-3000; Corbett Research) using primers 82 and 83, and including the SybrGreen reagent (Syber Premix ExTaqTM; Takara Bio) for visualization, after an initial denaturation at 95°C for 3 min, cycling progressed for a total of 41 cycles (denaturation 95°C, annealing with touchdown from 57 to 53 for three cycles followed by 38 cycles at 53°C, elongation 72°C), with data capture at the end of the elongation cycle. Conditions for S27a were as above. Melt curves were obtained to determine reaction efficiency and effectiveness after a 2-min final incubation at 40°C, with samples being heated at 1°C increments from 55°C to 99°C. Results of qRT-PCR were processed using Q-gene96, a Microsoft Excel-based software application program [36]. The normalized gene expression (NE) for Lgr8 was calculated according to the equation NE = (Etarget)CT target/(Eref)CTref where, Etarget and Eref and CTtarget and CTref are the PCR amplification efficiencies and the threshold cycles (where the fluorescence curve intersects the threshold) of the PCR amplification of Lgr8 and S27a gene products, respectively.
Immunohistochemistry
To analyze LGR8 expression at the protein level, polyclonal antibodies were raised against two peptides derived from the N-terminal ectodomain (L8–1; CLKKNKIHSLPDKVFIK) and the third intracellular loop (L8–2; CSIQKTALQTTEVRN) of the human LGR8 receptor sequence. Each was coupled via the N-terminal cysteine residue to keyhole limpet hemocyanin for immunization following a conventional protocol. Two rabbits were immunized for each immunogen (Pineda Antikoerper-Service). To validate these antibodies, cytospin preparations of HEK293T cells were prepared, into which expression constructs encoding human LGR8, human LGR7, or another GPCR, HE6, had been transiently transfected, exactly as previously described [37]. These cytospin preparations were subjected to immunohistochemistry using the LGR8 antibodies obtained, also as previously described [37]. Rabbits immunized against the L8–2 epitope failed to yield specific LGR8 antisera. Thus, all subsequent work was carried out using only the L8–1 antisera.
The human L8–1 peptide used as immunogen contains two mismatches compared to the equivalent rodent sequence. Therefore, immunohistochemistry was carried out only on human tissue sections. The blocks used for sectioning were a generous gift of Professor Mikhail Davidoff (University Hospital, Hamburg-Eppendorf), and represented testis tissue and epididymal tissue from healthy fertile men. For immunohistochemistry, 8 µm paraffin sections were processed exactly as previously described [37], using the rabbit primary antisera, as well as the negative control preimmune sera from the same animals, at a dilution of 4000:1, and the ABC staining kit (Vector Laboratories). Immunostaining was visualized using the chromogen DAB (Sigma), providing a brown-black color in the sections. These were lightly counterstained with hemalaun to indicate cell nuclei.
Activity of Human INSL3 on Mouse LGR8
HEK293T cells were transfected with pcDNA3.1 vector encoding human or mouse LGR8 (the human and mouse LGR8 constructs were kindly provided by Prof. Aaron Hsueh, Stanford University, and Dr. Anne Riesewijk, NV Organon, The Netherlands, respectively) and plated into 96-well plates for cAMP stimulation as previously described [38]. The cells were incubated with increasing concentrations of synthetic human INSL3 peptide [39] for 30 min, followed by the measurement of cAMP levels in the cell lysates using a commercial kit (Amersham Biotrak). Data are mean ± SEM of three experiments performed in triplicate and have been plotted using the Graphpad Prism program.
Radioligand Binding to MA-10 Mouse Leydig Tumor Cells
RT-PCR as well as immunohistochemistry suggested that Leydig cells are expressing the Lgr8 transcript and its protein product. To confirm this at a more functional level, radioligand binding studies were carried out using the mouse MA-10 Leydig tumor cell line. These cells were chosen to provide greater numbers and uniformity, and hence lower variance and higher sensitivity compared to primary cells. MA-10 mouse Leydig tumor cells were cultured as previously described [5]. First, 2 x 106 cells per well were seeded in poly-L-lysine-coated six-well culture plates (Cat. No. 3046, Falcon) and allowed to adhere by overnight incubation at 37°C. Media were then aspirated, cells were washed once with PBS, and 0.5 ml receptor binding buffer (20 mM HEPES, 1.5 mM CaCl2, 50 mM NaCl, and 0.01% sodium azide, pH 7.5) containing 1% BSA was added per well. After 30 min, the buffer was aspirated and replaced by fresh receptor binding buffer with 1% BSA containing 180000 cpm of [125I] human INSL3 (labeling carried out by Perkin Elmer; specific activity 2200 ci/mmol). This was competed by simultaneous addition of unlabeled human INSL3 at concentrations ranging from 0 to 100 pmol. Incubation was allowed to proceed for 1 h at room temperature (25°C). Medium was then removed, and cells were washed with PBS before being taken up in 1 M NaOH. The residual radioactivity associated with the cells was counted in a gamma counter (LKB 1260). Experiments were repeated four times, and data were analyzed using the Graphpad Prism software. Binding was tested in parallel in HEK293T cells transiently transfected with mouse LGR8 (as above) in poly-L-lysine-coated 24-well culture plates.
Leydig Cell Bioassays to Determine Receptor Functionality
To determine whether the LGR8 receptors putatively identified on Leydig cells are indeed functionally relevant to the cells, various bioassays were carried out using primary mouse Leydig cells. As endpoint for measurement, intracellular cAMP was assessed using a highly sensitive time-resolved fluorescence immunoassay (TRFIA), as described previously [11]. Fifty thousand cells per well were treated for 20 and 45 min with doses of 20 and 50 nM human INSL3 in the presence and absence of a general phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine (IBMX) at 1 nM. As a positive control for the assay, cells were treated with 50 ng/ml hCG as an LH receptor agonist. Parallel experiments were also performed using cGMP and testosterone as endpoints, similarly measuring the products by sensitive TRFIA as previously described [11]. All experiments were carried out three times independently. Finally, MA-10 cells were also subjected to human INSL3 stimulation under identical conditions, measuring cAMP by TRFIA as endpoint.
RESULTS
Expression of Gene Transcripts for the INSL3 Receptor LGR8 in the Testis and Other Tissues
A robust semiquantitative RT-PCR was established for mouse and rat sequences, initially assessing a transcript region encoding part of the extracellular ectodomain of this GPCR (Table 1, primers 82 and 83). This was applied first to assess a series of cDNAs derived from mouse tissues and cells (Fig. 2). Adult wild-type testis and derived seminiferous tubules indicated strong Lgr8 mRNA expression, confirmed using two other primer pairs, U2/84 and R1/R16, corresponding to the 5' end of the transcript and the 7-transmembrane domain, respectively. Primary mouse Leydig cells were also positive, though with a much weaker signal. Sequencing showed that only those products indicated by arrowheads in Figure 2 were specific; all other bands represented unspecific gene products. Confirmation of Leydig cell expression was obtained using the primer pair R1/R16 (Fig. 2B, left panel), applying a larger amount of the PCR product to the gel, with subsequent elution of the appropriately sized specific product (589 bp; Fig. 2B, left panel, black arrowhead) and a further heminested PCR reaction (Fig. 2B, right panel). MA-10 mouse Leydig tumor cells also indicated a weak positive signal, using this nested PCR approach (Fig. 2B, right panel). The two Sertoli cell lines were consistently negative. RNA from Day 10 postnatal testis, as also from the testes of Kitw/Kitw-v animals, showed specific signals of moderately strong intensity. Kitw/Kitw-v mice are azoospermic, having no postmeiotic germ cell stages present, but, unlike the Day 10 testes, they have adult-type mature Leydig cells. Also the epididymis of Kitw/Kitw-v mice indicated moderate Lgr8 gene expression. Taken together, these results suggest Lgr8 mRNA expression at a low level in Leydig cells and in a non-Sertoli cell component of the seminiferous tubules, presumably germ cells. The mouse epididymis also expresses Lgr8 transcripts.
|
The PCR products derived from the ectodomain-encoding region of the Lgr8 gene were mostly represented by two bands. Sequencing showed that the larger product at 
400 bp (primers 82/83) or 670 bp (primers U2/84) represented the full-length transcript, whereas the more rapidly migrating product, at a 75-bp smaller size, was missing exon 3. All other bands were shown upon sequencing to represent products unrelated to Lgr8.
A comparable analysis was then performed using rat cells and tissues (Figs. 3–5). As positive control in this series of analyses, RNA from pre- and postnatal gubernacula was examined, using the primer combination 82/83 (Fig. 3). As expected, there is prominent expression of Lgr8 transcripts, including in the early embryonic stages, when INSL3 is known to act on the gubernaculum to induce testicular descent. Figure 3 also indicates Lgr8 expression in the epididymis, as in the mouse, and also in the spleen. Skeletal muscle and lung were negative in this assay. Analysis of rat testicular tissues and cell types (Fig. 4), using different primer combinations covering most of the Lgr8 transcript, revealed Lgr8 transcript expression in intact adult testis, in adult testes where Leydig cells had been depleted by treatment with EDS supplemented by testosterone, and weakly in primary Leydig cells. Analysis of germ cell stages purified by differential elutriation (Fig. 4) indicated consistent expression in round spermatids, and also possibly in spermatogonia and pachytene spermatocytes. Sertoli cells and peritubular cells (not shown) indicated no Lgr8 gene expression. A second analysis of rat germ cells, also purified independently by differential elutriation, was carried out both by conventional RT-PCR (not shown), providing essentially the same results as before (not shown), and also by qRT-PCR (Fig. 5), using the primer combination 82/83. Note that the gels illustrated in Figure 5 are of the final (saturation) products from the qRT-PCR, and are thus not intended to be quantitatively comparable, but merely serve to illustrate the quality of the final PCR product. The real-time quantitation (Fig. 5), normalized to the qRT-PCR analysis of S27a, confirms the earlier results that spermatids (1n; fractions N and L) express significantly more Lgr8 transcripts than 4n spermatocytes (fractions G and F). Fraction C represents a mixture of 4n and 2n germ cells, and is intermediate in the amount of expressed Lgr8 transcripts. A similar qRT-PCR analysis was carried out using the same primer combination on intact testis, primary Leydig cells, and EDS-treated testes (supplemented with testosterone). The results (not illustrated) showed that intact rat testes were 2.83 ± 0.82 (mean ± SD)-fold enriched in Lgr8 transcripts compared to EDS-treated (supplemented with testosterone) testes, and also 3.12 ± 1.75-fold enriched compared to primary rat Leydig cells.
|
|
|
Figure 4 includes a series of semiquantitative RT-PCR analyses using primer combinations (Table 1 and Fig. 1) together spanning most of the Lgr8 transcript. It has recently been shown [40] that the human LGR8 gene can also be expressed as variantly spliced transcripts. One such variant is missing exon 11. Whereas for the rat, the gubernaculum consistently shows a secondary, faster-migrating band, approximately 75 bp shorter than the major specific product, all other cells and tissues consistently indicated only the presence of a product consistent with the full-length LGR8 receptor. The secondary band from the gubernaculum, as for the mouse, corresponded to a product missing exon 3. All attempts to devise and apply primers from the putative exon 1 sequence in the rat failed to provide any products for any RNA sample (altogether 5 primers were tested from this region), suggesting that the 5' upstream region of the rat Lgr8 transcript may differ from the similar region in the mouse.
Immunohistochemical Analysis of LGR8 Epitopes in Human Testis and Epididymis
Control experiments were carried out to validate the new monotypic polyclonal antibodies raised in rabbits against a peptide from the N-terminal ectodomain of the human LGR8 sequence, corresponding to the region of the first leucine-rich repeat domain encoded by exon 5. These antibodies only recognized HEK293T cells that had been transiently transfected with an expression construct encoding human LGR8, and not untransfected cells or cells transfected with constructs encoding the closely related LGR7 or more distantly related HE6 G-protein coupled receptors (Fig. 6, A–C). We had previously shown that the LGR7- and HE6-expressing cells were indeed producing protein that reacted specifically with their respective antibodies [37]. When applied to tissue sections from human testis, the LGR8 antibodies recognized interstitial Leydig cells (Fig. 6D), and germ cells identifiable as round spermatids (Fig. 6, D and F, arrowheads). Parallel sections treated with the preimmune serum from the same animal were entirely free of any immunostaining (Fig. 6, E and G). Identical results were obtained using a second antibody raised in another rabbit against the same immunogen (not shown). Sections of human epididymis also showed specific immunostaining for LGR8 epitopes within the epithelial cells (Fig. 6H), which were not detectable when applying the preimmune serum (Fig. 6I).
|
Specific Radioligand-Binding to MA-10 Mouse Tumor Leydig Cells
It has previously been demonstrated that synthetic human, sheep, and rat INSL3 have similar activity on human LGR8 [20]. In this study we demonstrate that human INSL3 will also activate mouse LGR8 (pEC50 = 9.4 ± 0.04) with identical activity to that seen on cells expressing human LGR8 (pEC50 = 9.4 ± 0.09) (Fig. 7). Human INSL3 labeled with 125I has been previously shown to be a high-affinity ligand for human LGR8 [40]. It was used here to test for specific binding to MA-10 mouse tumor Leydig cells (Fig. 8), in parallel to cells overexpressing mouse LGR8 (not shown). In a competitive assay using increasing amounts of unlabeled human INSL3, it is clearly demonstrated that intact MA-10 cells possess high affinity binding sites (EC50 0.3 nM) for the hormone (Fig. 8). Importantly, the binding affinity was very similar to that found in mouse LGR8 overexpressing cells (not shown). Calculation of receptor number suggests only about 400 binding sites per cell, presumably representing intact LGR8 (cf. [40]).
|
|
Responsiveness of Cultured Rodent Leydig Cells to INSL3 Application
Because Leydig cells consistently proved to be positive for low level LGR8 expression in all assays so far evaluated, preliminary experiments were carried out to determine whether application of human INSL3 had any influence on specific signal transduction and metabolic pathways. INSL3 was applied at two different doses to primary cultures of mouse Leydig cells, with or without the phosphodiesterase inhibitor IBMX. Repeated experiments were carried out, and altogether three different metabolic endpoints were determined: cAMP, cGMP, and testosterone. For none of these pathways was any effect of INSL3 detected (not shown). The concentrations of human INSL3 tested gave maximal stimulation of cAMP levels in mouse LGR8 expressing cells (Fig. 7), used here as control. Importantly, the positive control hCG, used in all experiments, was able to stimulate cAMP, cGMP, and testosterone (progesterone in the case of MA-10 cells) production in the Leydig cells cultures (not shown).
DISCUSSION
Besides testosterone, Leydig cells have recently been shown to produce and secrete into the circulation another hormone, INSL3, in substantial amounts. In normal men [3, 41, 42] and in rats [43] concentrations of approximately 1 ng/ml (0.17 nM) are detectable in the circulation of sexually mature individuals. Within the testis, concentrations will be much higher [3]. Interestingly, INSL3 appears to be uncoupled from the hypothalamic-pituitary-gonadal (HPG) axis, in that short-term applications of LH (hCG) or testosterone are without effect in vivo [3] or in vitro [5]. INSL3 expression appears to be altered only in relation to the differentiation status of the Leydig cells, being most highly expressed in the normal adult testis, but being reduced, for example, in prepubertal animals [4, 5], in aging rats [9], in sexually inactive seasonal breeders [7, 8], or in dedifferentiated Leydig cell tumors [44]. Although there is evidently a high circulating concentration in adult men, but not in women, one obvious function recognized in male mice, wherein the gene for either the hormone INSL3 or its receptor LGR8 (Great) have been ablated, is control of the first, transabdominal phase of testicular descent within the fetus [12, 13]. Here, INSL3 appears to interact with LGR8 receptors in the gubernacular ligament, linking the testis to the inguinal abdominal wall, causing this to thicken and shorten [19, 45]. The present study was undertaken to identify and postulate possible functions for INSL3 within the adult testis, where local concentrations of hormone are likely to be well above a level sufficient to activate LGR8 receptors; concentrations of >14 ng/ml have been measured for INSL3 in the spermatic vein of male probands [3].
Employing a combination of semiquantitative and real-time PCR, with various primer sets able to distinguish full-length and splice variant Lgr8 transcripts, together with immunohistochemistry and radioligand binding, a consistent picture is presented, although three different species have been used. Full-length putatively functional LGR8 receptors appear to be located on Leydig cells (weak expression) and on postmeiotic (and possibly premeiotic) germ cells, but on no other testicular cell type analyzed. Immunohistochemical staining of human testis sections show both Leydig cell staining and staining in postmeiotic germ cells. The cytoplasmic distribution of the immunostaining would be consistent with LGR8 epitopes being stored before presentation at the cell surface, or more likely after being internalized following receptor desensitization.
The finding of LGR8 expression in germ cells is consistent with the recent preliminary report of Kawamura et al. [15] that INSL3 appears to have a protective, antiapoptotic effect on male germ cells. This study demonstrated Lgr8 mRNA expression and binding of biotinylated human INSL3 to germ cells; however, these authors failed to demonstrate either in Leydig cells. Here we show that there appears to be expression of Lgr8 transcripts and protein in Leydig cells. However, because of the low level of specific gene expression, and the difficulty to devise upstream primers for the rat from the putative exon 1 sequence, we were unable to prove the full-length integrity of these transcripts unambiguously. It should be additionally noted that the L8–1 antibodies, recognizing an exon 5-encoded peptide sequence, would not be able to distinguish full-length LGR8 proteins from those from which upstream exons may be missing but in which nevertheless the correct reading frame is maintained and translated. In support of the idea that many Lgr8 transcripts in the testis may not represent full-length transcripts is the recent in situ Lgr8 mRNA hybridization on rat testis sections, which suggests relatively higher levels of undefined transcript expression within the tubular compartment [24] and which thus appears discordant with the present findings.
This study has made use of several physiological models to deduce patterns of LGR8 expression. Comparing adult testes from wild-type mice with those from mutant Kitw/Kitw-v mice and with normal postnatal Day 10 testes suggests that Lgr8 transcripts are probably also expressed in 2n spermatogonial stages. Also, the quantitative comparisons between rat adult testis, EDS-treated testis, and primary Leydig cells support the view that Leydig cells probably contribute substantially to the total testicular expression of Lgr8 mRNA, although the relative proportions of each cell type in the various preparations also needs to be considered here.
The absence of any effect of INSL3 on several classic Leydig cell metabolic pathways is in some ways disappointing. However, this result is fully consistent with what has already been learned from studies of the expression and regulation of the INSL3 hormone. That is, were INSL3 to have an influence on steroidogenesis, or its precursor signaling molecule cAMP, then this would imply that INSL3 would be able to exert a short-term influence on the HPG axis and vice versa. Yet all results so far show that there is a clear disconnect between INSL3 production by Leydig cells and the HPG axis (see above). Furthermore, because both INSL3 and its receptor appear to be present on the same Leydig cells, it is unlikely that INSL3 would have an effect on those factors influencing Leydig cell differentiation, such as LH. Nor does there appear be any effect on INSL3 production itself, at least in the adult, because INSL3 production levels appear to be maintained very consistently and constitutively in the adult male. One could speculate that INSL3 represents a mechanism to maintain the differentiated state of Leydig cells, because at the high local concentrations likely to be present near Leydig cells, the receptors are probably permanently occupied and hence possibly desensitized. This would need to be tested once an appropriate signaling mechanism has been determined. INSL3 would not appear to be acting in this manner during puberty, because the knockout animals all seem to progress through postnatal development at the same rates as wild-type animals. Finally, it is possible that this autocrine-paracrine system represents another example of what has been described as "biochemical radar" [46], whereby the secreted hormone acts to assess the Leydig cell microenvironment and maintains some form of redundant local homeostasis.
Taken together, these results and speculations would support a role for INSL3 and its receptor in the testis, whereby the hormone homeostatically regulates its own production through an influence on Leydig cell differentiation status, independently of the HPG axis. But its principal role in the testis is to act as a survival factor for germ cells, as suggested by Kawamura et al. [15]. The localization of specific mRNA by RT-PCR in purified testicular cells, particularly of postmeiotic stages, as well as immunohistochemistry fully supports this view. This hypothesis is also fully in accord with findings in the ovary. LGR8 receptors have been tentatively identified on oocytes [15], and a detailed analysis of Insl3–/– female mice shows a reduction in fertility caused by increased apoptosis of follicles and corpora lutea [14]. Furthermore, in bovine ovaries there is a clear association between expression of INSL3 in follicular theca cells and the protection of immunopositive follicles from atresia (apoptosis) [47].
Finally, both immunohistochemical studies in the human and RT-PCR analysis in the rat and mouse indicate there is also expression of LGR8 in the epididymis. What its function might be there is as yet unknown. Future research will need to consider the possible roles of this new Leydig cell-secreted factor in the functioning of other organs in the adult male, especially because its independence of the HPG axis would suggest that it is conveying different information from normal androgen pathways.
ACKNOWLEDGMENTS
We gratefully acknowledge helpful discussions and support from our previous colleagues in the Institute for Hormone and Fertility Research at the University of Hamburg. In particular, we should like to thank Drs. Yvonne Pohnke and Ralph Telgmann for preliminary assistance in the characterization of the antisera, and Dr. Hans-Joachim Paust for sequencing some of the PCR products. We are also very grateful to Drs. Bernard Jegou, Christophe Staub, and Charles Pinaeu (Rennes, France) for help in providing RNA samples of rat gubernaculum and some germ cell stages. The authors would like to thank Dr. John Wade (Howard Florey Institute, University of Melbourne) for provision of human INSL3 peptide, Dr. Yieh Ping Wan (Perkin-Elmer life and analytical sciences) for [125I]-labeled of human INSL3, Tania Ferraro and Sharon Layfield for technical assistance at HFI, and Dr. Anne Riesewijk (Department of Pharmacology, NV Organon, Oss, The Netherlands) and Dr. Aaron Hsueh (Stanford University) for providing the mouse and human LGR8 expression plasmids, respectively.
FOOTNOTES
1 Supported in part by a stipendium of the Graduirtenkolleg 336 of the German Research Council (DFG) to R.J.K.A-I., and subsequently by the small grant scheme of the Faculty of Health Sciences at the University of Adelaide, as well as by funds of BioInnovation SA, Adelaide, to R.I. Studies at the HFI were supported by a National Health and Medical Research Council (NHMRC; 300012) to R.A.D.B. Part of this project was supported by the EDN consortium of the EU 5th framework (QLK4-CT-2002-00603). 
2 Correspondence. FAX: 61 8 8303 3356; ravinder.anand-ivell{at}adelaide.edu.au
Received: 30 September 2005.
First decision: 20 October 2005.
Accepted: 7 February 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  R. A. L. Bayne, T. Forster, S. T. G. Burgess, M. Craigon, M. J. Walton, D. T. Baird, P. Ghazal, and R. A. Anderson Molecular Profiling of the Human Testis Reveals Stringent Pathway-Specific Regulation of RNA Expression Following Gonadotropin Suppression and Progestogen Treatment J Androl, July 1, 2008; 29(4): 389 - 403. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Anand-Ivell, R. Ivell, D. Driscoll, and J. Manson Insulin-like factor 3 levels in amniotic fluid of human male fetuses Hum. Reprod., May 1, 2008; 23(5): 1180 - 1186. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Bay, H. E. Virtanen, S. Hartung, R. Ivell, K. M. Main, N. E. Skakkebaek, A.-M. Andersson, The Nordic Cryptorchidism Study Group, and J. Toppari Insulin-Like Factor 3 Levels in Cord Blood and Serum from Children: Effects of Age, Postnatal Hypothalamic-Pituitary-Gonadal Axis Activation, and Cryptorchidism J. Clin. Endocrinol. Metab., October 1, 2007; 92(10): 4020 - 4027. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Feng, N. V Bogatcheva, A. Truong, B. Korchin, C. E Bishop, T. Klonisch, I. U Agoulnik, and A. I Agoulnik Developmental Expression and Gene Regulation of Insulin-like 3 Receptor RXFP2 in Mouse Male Reproductive Organs Biol Reprod, October 1, 2007; 77(4): 671 - 680. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. K. Amory, S. T. Page, B. D. Anawalt, A. D. Coviello, A. M. Matsumoto, and W. J. Bremner Elevated End-of-Treatment Serum INSL3 Is Associated With Failure to Completely Suppress Spermatogenesis in Men Receiving Male Hormonal Contraception J Androl, July 1, 2007; 28(4): 548 - 554. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. V. Bogatcheva, A. Ferlin, S. Feng, A. Truong, L. Gianesello, C. Foresta, and A. I. Agoulnik T222P mutation of the insulin-like 3 hormone receptor LGR8 is associated with testicular maldescent and hinders receptor expression on the cell surface membrane Am J Physiol Endocrinol Metab, January 1, 2007; 292(1): E138 - E144. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ferlin, A. Garolla, F. Rigon, L. Rasi Caldogno, A. Lenzi, and C. Foresta Changes in Serum Insulin-Like Factor 3 during Normal Male Puberty J. Clin. Endocrinol. Metab., September 1, 2006; 91(9): 3426 - 3431. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
