| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Article |
a Université Paris 7 and INSERM-INRA U 418, Tour 33/43, case 7126, 75251 Paris Cedex 05, France
b INSERM-INRA U 418, Hôpital Debrousse, 69222 Lyon, France
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The Sertoli cells are the first to differentiate. These large cells adhere to one another and surround the germ cells to form the seminiferous cords [3]. This occurs from 13.5 to 14.5 days post-conception (dpc) in the rat. The Sertoli cells continue proliferating until 3 wk after birth, resulting in a considerable increase in the diameter and length of the seminiferous cords (reviewed in [4]).
The germ cells migrate from extra-embryonic areas to the genital ridge while dividing. They are called gonocytes once they have reached the gonadal anlage and are enclosed in the emerging seminiferous cords. The gonocytes proliferate into the seminiferous cords until fetal Day 17.5, after which they become quiescent until postnatal Days 2–3, when they resume mitosis to give the first spermatogonia on 6 days postpartum (dpp) [5]. We have shown that both the fetal and neonatal periods of mitosis of gonocytes are also periods of apoptosis for these cells [6].
Leydig cells differentiate from mesenchymal cells in the interstitial compartment [7]. The fetal testis begins to produce testosterone on Day 15.5 post-conception [8, 9], and fetal type Leydig cells undergo functional regression from fetal Day 18.5 onwards [9–11]. Adult-type Leydig cells differentiate after the second postnatal week.
Many factors contribute to the regulation of these developmental steps. We have focused our attention on retinoic acid (RA), since it is an important regulator of fetal development of many organs, including the lung and kidney [12, 13]. Furthermore, it is well known that vitamin A deficiency also leads to an arrest of spermatogenesis in adults [14, 15]. A comparable phenotype was obtained by inactivation of some types of RA receptors (RARs), such as RAR
[16] and RXRß [17].
However, few data are available on the role of RA in the development of the fetal and neonatal testis. Certain elements of the RA transduction system, such as cellular retinoic acid-binding protein (CRABP) I and II, have been found in the fetal rat testis [18], and recent studies using reverse transcription (RT)-polymerase chain reaction (PCR) and/or immunohistochemistry demonstrated that RARs (, ß, 
) and RXRs (, ß, ) are expressed in fetal rat testis [19, 20]. Only three effects of RA on the development of the fetal and neonatal testis have been discovered until now. The first one concerns the formation of the seminiferous cords. It was shown that addition of RA to 14.5-day-old fetal testis in culture affected the deposition of laminin and fibronectin in the basement membrane [21]. Recently, Cupp et al. [19] confirmed and extended this finding by showing that RA inhibits the seminiferous cord formation in cultured 13.5-dpc testis. A second effect of RA, which we recently observed, is its potentiality to decrease the number of gonocytes in dispersed testicular cells from 3-dpp neonates [20]. The third effect of RA is its capacity to inhibit the incorporation of tritiated thymidine in testicular cells dispersed from 0-dpp neonates and cultured in the presence of growth stimulators, but this result is poorly informative since the affected cell types have not been identified in this study [19]. Consequently, our knowledge of the effect of RA on the Sertoli cell development is limited to morphological analyses of a short period (13.5–14.5 dpc), and there are no data on the effect of retinoids on the development of Leydig and germ cells during the fetal life.
Since we and others have shown that organotypic cultures of the fetal and neonatal testis are good systems for studying the development of Sertoli cells [22], gonocytes [23], and Leydig cells [24], we have used this in vitro system to investigate the effect of retinoids on the cellular and functional development of these three cell types. Explants of 3 ages were used: 14.5 dpc, when all the cell types proliferate and testosterone production increases; 18.5 dpc, when the gonocytes are quiescent and testosterone production begins to decrease; and 3 dpp, when gonocytes resume mitotic and apoptotic activity and testosterone production declines.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Female Sprague-Dawley rats from Charles River (Saint Aubin Les Elbeuf, France) were housed under a controlled photoperiod (lights-on 0600–2000 h) and fed a commercial diet (U.A.R., Villemoisson sur Orge, France) with tap water ad libitum. Males were caged with females for the night. Since the estimated time of ovulation and fertilization was 0200 h, the day following an overnight mating was counted as 0.5 dpc. Pregnant rats were anesthetized by an i.p. injection of 4 mg/ml sodium pentobarbital (Sanofi, Libourne, France) on 14.5 and 18.5 dpc, and the testes were removed aseptically from the fetuses under a binocular microscope and immediately explanted in vitro. Natural birth occurred between Day 21.5 at 1300 h and Day 22.5 at 1800 h. Since precise timing of postnatal development was desired, only pups born between 1800 h on 21.5 dpc and 0800 h on 22.5 dpc (i.e., 70% of the neonates) were kept. Fetal Day 22.5 was counted as Day 0 dpp for all newborns. The number of neonates in each litter was standardized at eight pups. Neonates were killed by cervical dislocation on Day 3 postpartum, and their testes were immediately removed.
Chemicals and Solutions
The culture medium was Ham's F-12/Dulbecco's modified Eagle's medium (1:1; Gibco, Grand Island, NY) containing 0.35% glutamine (Flow Laboratories, McLean, VA) and 80 µg/ml gentamicin (Gentalline; Schering-Plough, Levallois-Perret, France). Ovine (o) LH (NIH.LH S19; 1.01 NIH.LH.S1 U/mg) was a gift from Dr. A.F. Parlow (NIDDK, Bethesda, MD). Recombinant human (h) FSH (12 000 IU/mg) was a gift from Dr. B. Mannaerts (Organon International, Oss, The Netherlands) [25]. All-trans retinol (RE) and all-trans RA were purchased from Sigma (St. Louis, MO). Anti-3ß-hydroxysteroid dehydrogenase (3ßHSD), anti-anti-Müllerian hormone (AMH), and anti-cAMP antibodies were generously provided by Dr. G. Defaye (INSERM U244, Grenoble, France), Dr. B. Vigier (INRA, Jouy-en Josas, France) and Dr. J.M. Saez (INSERM U 418, Lyon, France), respectively.
Organ Cultures
Testes were cultured on Millipore (Bedford, MA) filters (pore size: 0.45 µm) as previously described [24]. Briefly, intact 14.5-dpc fetal testes were placed on filters. Older testes were cut into small pieces (4 pieces for 18.5 dpc, 18 pieces for 3 dpp), and all the pieces from the same testis were placed on a single Millipore filter. The filter bearing the pieces of testis was floated on 0.4 ml (14.5-dpc testes) or 1.5 ml culture medium (later ages) in tissue culture dishes and incubated at 37°C, in a humidified atmosphere containing 95% air:5% CO2, for 3 or 5 days. The medium was changed every 24 h. The response to retinoid was measured by comparing one testis cultured in medium containing RA or RE with the other testis from the same fetus cultured in medium alone (control).
At the end of the culture period, 100 ng/ml oLH, or 200 mIU/ml recombinant hFSH, or 1 mM 5-bromo-2'-deoxyuridine (BrdU) was added to all the media for the last 3 h of the culture. For cellular analyses, the whole explant was fixed for 2 h at 4°C in Bouin's fluid (except for TUNEL [terminal deoxynucleotidyl transferase (TdT)-mediated dUTP nick end-labeling], which needed fixation with buffered 4% formaldehyde) and embedded in paraffin, and 5-µm sections were cut.
Identification and Numeration of the Gonocytes
The procedure was that previously described [23]. All the serial sections from one explant were mounted on slides, deparaffinized, rehydrated, and stained with hematoxylin-eosin (Fig. 1, A and B). The gonocytes were identified by their large, spherical, lightly stained nuclei containing fine chromatin granules and two or more globular nucleoli, and by a clearly visible cytoplasmic membrane. All the gonocytes in every 3, 10, or 20 sections (14.5 dpc, 18.5 dpc, or 3 dpp, respectively) were counted, and the total number of gonocytes was taken as the total count (TC). This number was multiplied by 3 (14.5 dpc), 10 (18.5 dpc), or 20 (3 dpp) to obtain the crude count (CC) of gonocytes per testis. The Abercrombie formula was used to correct for any double counting resulting from the appearance of a single cell in two successive sections: TC=CCxS/(S+D) where "TC" is the true count, "S" is the section thickness (5 µm) and "D" is the mean diameter of the gonocytes nuclei [26]. "D" equals the average of the nuclear diameters measured on the section (DM) divided by 
/4 to correct for the over-representation of smaller profiles in sections through spherical particles. DM was measured in each testis studied, by at least 100 random determinations using a micrometer eye piece and calibrated with an object micrometer on the microscope plate. All counts and measurements were done blind.
|
Identification and Numeration of the Sertoli Cells
All the Sertoli cells in 14.5-dpc explants could not be identified with certitude by simple histological observation because of the disorganization of the seminiferous cords induced by RA treatment at this age. Therefore, Sertoli cells were identified by immunostaining for AMH as previously described, with minor modifications [27]. Deparaffinized sections were rehydrated, and endogenous peroxidases and nonspecific protein binding were blocked by incubations in hydrogen peroxide and then normal goat serum. The sections were subsequently incubated overnight in the anti-AMH antibody (16 µg/ml) in a humidified chamber at 4°C, and the distribution of this primary antibody was revealed with a biotinylated goat anti-rabbit secondary antibody and an avidin-biotin-peroxidase complex (Vector Laboratories, Burlingame, CA). Peroxidase was visualized with vector SG. Sections were rinsed in PBS between each step. All the AMH-positive cells in every 3rd section were counted, and the Abercrombie formula was applied.
For 18.5-dpc and 3-dpp testes, a simple histological observation of hematoxylin-eosine-stained slides was sufficient to identify the Sertoli cells since there was no RA-induced disorganization at these ages and Sertoli cells are easily characterized by their basal position in the seminiferous cords and their cytological features (irregular nucleus and no clear cytoplasmic membrane). For numeration of the Sertoli cells, the areas of one out of 10 (18.5 dpc) or 20 (3 dpp) consecutive sections were measured with a computerized video densitometer (Biocom, Les Ulis, France) and added together. The Sertoli cells in 3 sections of each testis were counted, and the number was divided by the corresponding area to determine the average of Sertoli cell density per surface unit. This density was then multiplied by the cumulative areas of the sections to obtain the total Sertoli cell number for the whole testis. The Abercrombie formula was also used. The diameter of Sertoli cell nuclei was constant (6.66 ± 0.05 µm) regardless of treatment or explantation age.
Identification and Numeration of the Leydig Cells
Leydig cells were identified in 14.5-dpc testes by immunocytochemical detection of 3ßHSD activity. Immunostaining was performed with the Vectastain Elite ABC kit (Vector Laboratories). Deparaffinized sections were rehydrated, placed in citrate buffer, and heated in a microwave oven for 3 x 5 min at 650 W. Endogenous peroxidases and nonspecific protein binding were blocked by incubations in hydrogen peroxide and then normal goat serum. The sections were then incubated overnight in anti-3ßHSD antibody (1:200) in a humidified chamber at 4°C. The distribution of this primary antibody was revealed with a biotinylated goat anti-rabbit secondary antibody and the avidin-biotin-peroxidase complex. Peroxidase was visualized with 3,3'-diaminobenzidine. Sections were rinsed in PBS between each step. The specificity of 3ßHSD staining was checked by replacing the antibody with nonimmune IgG.
All the Leydig cells in every 3rd section of whole testis were counted, and the Abercrombie formula was applied. The diameter of Leydig cell nuclei was unchanged (7.73 ± 0.03 µm) after incubation with RA for 3 days.
Measurement of BrdU Incorporation Index
The procedure was as previously described [23]. Cultured 14.5-dpc, 18.5-dpc, and 3-dpp testes were labeled with BrdU (labeling reagent diluted 1:100 according to the instructions of the cell proliferation kit; Amersham, Bucks, UK) during the last 3 h of culture. BrdU incorporation into proliferating cells was detected by immunocytochemistry according to the manufacturer's recommendations. Briefly, sections randomly chosen were mounted and incubated successively with 0.3% H2O2 in methanol at 20°C for 30 min to inactivate endogenous peroxidases, and in a mouse anti-BrdU monoclonal antibody (cell proliferation kit, Amersham) at 20°C for 1 h. The antibody bound to the nuclei was detected by a peroxidase-linked anti-mouse IgG. Finally, slides were stained with DAB (3,3'-diaminobenzidine (Sigma; Fig. 1, C and D). The BrdU incorporation index (percentage of cells showing a clear positive immunoreaction to BrdU) was obtained by a blind counting of all the gonocyte or Sertoli cell nuclei on the sections. For evaluating the BrdU incorporation index in the Sertoli cells of 14.5-dpc explants, a double immunostaining for BrdU and AMH was performed.
Measurement of DNA Fragmentation Index
Apoptotic cells were detected in situ using a modified version of the TUNEL method as previously described [23]. Sections randomly chosen were incubated successively in hydrogen peroxide, permeabilizing solution (0.1% Triton X-100 in 0.1% sodium citrate) and TdT buffer containing TdT (0.6 U/µl; Boehringer Mannheim, Indianapolis, IN) and fluorescein-dUTP (Boehringer Mannheim). The sections were incubated with anti-fluorescein antibody conjugated to peroxidase, stained with DAB, and counterstained by brief immersion in hematoxylin (Fig. 1, E and F). Positive controls were incubated with deoxyribonuclease I (100 µg/ml) for 10 min at 20°C to induce DNA strand breaks. Negative controls were incubated without TdT. Sections were rinsed in PBS between each step. The DNA fragmentation index (percentage of cells with a clear positive TUNEL staining) was obtained from blind counting of all the gonocyte nuclei on the sections.
Testosterone RIA
The testosterone secreted into the medium was measured in duplicate by an RIA previously described [9]. No extraction or chromatography was performed because 17ß-hydroxy-5-androstane-3-one, the only steroid that cross-reacts (64%) significantly with testosterone, is secreted in minute amounts by the fetal rat testis [9].
Cyclic AMP RIA
Cyclic AMP production was estimated after 27 h or 72 h of culture by incubating the testes for 3 h with fresh medium containing 1 mM isobutylmethylxanthine (IBMX), in the presence or absence of 200 mIU/ml recombinant hFSH. The media were collected, acetylated, and assayed for cAMP by RIA [28].
Statistical Analysis
All values are the means ± SEM. The significance of the differences between the mean values for the treated and untreated controls were evaluated using Student's paired t-test. Other means were compared using one-way ANOVA (Fisher's test).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Testes from 14.5- and 18.5-day-old fetuses and from 3-day-old neonates were cultured for 3 days with or without RA, and histological analyses were performed.
At the time of explantation, Sertoli cells in 14.5-dpc testes were almost all aggregated in cords (Fig. 2A). The seminiferous cords cultured in control medium correctly ended their formation (Fig. 2B), whereas those cultured in RA (10-6 M) had altered organization, and most were disrupted. Many Sertoli cells were isolated or packed in clusters (Fig. 2C). The seminiferous cords were also disrupted after culture in 3.10-8 M RA, but over smaller areas (data not shown).
|
The seminiferous cords of explants of 18.5-dpc and 3-dpp testes were not disrupted by RA (10-6 M). However, RA increased the mean diameter of the cords from 39.0 ± 0.8 (control) to 42.7 ± 0.5 µm of 18.5-dpc testes (n = 4, P < 0.05) and from 42.0 ± 0.5 (control) to 48.7 ± 0.6 µm for 3-dpp testes (n = 3, P < 0.05). The number of Sertoli cells in 18.5-dpc testes was unaffected by RA (see below), and the sex cords contained almost the same number of gonocytes (see below). Consequently, the enlargement of the cord diameter in response to RA was due to an increase in Sertoli cell size. For 3-dpp testes, simple observation of the sections was sufficient to show that RA also caused Sertoli cell hypertrophy (Fig. 3).
|
Effect of RA on the Proliferation and Function of Sertoli Cells
The number of Sertoli cells and their mitotic and apoptotic indexes were evaluated after 3 days in culture. RA did not change the total number of Sertoli cells in 14.5- and 18.5-dpc testes, but it increased this number in 3-dpp testes (Fig. 4A). The mitotic index was maximal for 18.5-dpc control testes as compared with 14.5-dpc and 3-dpp ones, in accordance with earlier results [29] (Fig. 4B). RA (10-6 M) did not modify the mitotic indexes for 14.5- and 18.5-dpc testes but increased this index in 3-dpp testes. No TUNEL-positive Sertoli cells were found in control or RA-treated testes (data not shown).
|
The influence of RA on the differentiated functions of the Sertoli cells was assessed by measuring the cAMP produced in response to acute stimulation with FSH (3 h) at the end of the culture (Fig. 5). The FSH-induced cAMP production gradually increased as a function of the age at explantation in controls and was strikingly reduced by 10-6 M RA at all the ages studied.
|
Since FSH, which is produced in vivo from late fetal life onwards [30], also acts on mitosis and differentiation of the Sertoli cells [31], we determined the relationship between the effects of RA and FSH on cAMP production and mitosis (Table 1). The first experiments were done on testes cultured for 3 days. In the presence of FSH, RA decreased cAMP production, as did RA alone, but RA had no detectable effect on mitosis. However, the mitogenic effect of FSH was very slight after 3 days. We therefore cultured testes for only one day, which allowed the exhibition of the strong mitogenic effect of FSH. RA inhibited the acute FSH-stimulated production of cAMP and decreased Sertoli cell mitosis in the presence and absence of FSH.
|
Effect of RA on the Proliferation of the Gonocytes
The total number of gonocytes in explanted 14.5-dpc testes was counted at the time of explantation, and after 2 and 3 days of culture with or without RA (Fig. 6A). The number of gonocytes in control testes increased 2.1-fold after 2 days and 2.3-fold after 3 days in culture. RA (10-6 M) reduced the number of gonocytes 3-fold after 2 days and 6-fold after 3 days in culture. This effect was dose-dependent and statistically significant even with 3.10-8 M RA (P < 0.05). Curiously, RA slightly increased the diameter of the gonocyte nuclei (10.01 ± 0.04 µm for control cells vs. 10.54 ± 0.19 µm for RA-treated cells, n = 3, P < 0.05).
|
The mitotic index of gonocytes in 14.5-dpc testes was measured at the time of explantation after 2 and 3 days in culture (Fig. 6B). It decreased as a function of the time in culture in control testes, similar to the change in vivo [6, 32]. Although there were considerably fewer gonocytes in RA-treated testes, more BrdU was incorporated into the remaining gonocytes after 2 and 3 days in culture. The percentage of BrdU-labeled gonocytes was increased to reach 125% and 247% of the control values after 2 and 3 days with 10-6 M RA. This positive effect was also dose-dependent on Day 3.
These results thus suggest that RA causes apoptosis in the gonocytes in 14.5-dpc cultured testes. We checked this by measuring the rate of apoptosis in the gonocytes by TUNEL method. Since the TUNEL method needs fixation with formaldehyde, which gives poor-quality histological pictures, the identification of gonocytes in 3-day-treated testes was not certain because of the great disorganization. We therefore examined explants cultured for 2 days, in which gonocytes could be identified (Fig. 6C). The number of TUNEL-positive gonocytes in the RA-treated testes was 2-fold greater than in the paired controls. It should be noted that mitotic and apoptotic indexes cannot be compared since there is a great difference in the duration of the cell cycle (many hours) and that of apoptosis (from a few minutes to under 2 h).
Explants of 18.5-dpc testes (beginning of the quiescent period in gonocytes in vivo) showed no change in the number of gonocytes and no induction of mitotic or apoptotic activities after 3 days in culture in control or RA-supplemented medium (Fig. 7A).
|
Explants of 3-dpp testes had a constant number of gonocytes throughout the control culture, although mitotic and apoptotic activities had resumed (Fig. 7, A–C). RA (10-6 M) slightly but significantly increased the number and percentage of BrdU-positive gonocytes without any change in the percentage of TUNEL-positive gonocytes. The mean diameter of the gonocyte nuclei was not altered by RA in 18.5-dpc testes (10.18 ± 0.04 [control] and 10.16 ± 0.04 µm [RA], n = 3) or in 3-dpp explants (10.00 ± 0.04 [control] and 9.89 ± 0.10 µm [RA], n = 3).
Effect of RE and RA on Fetal Leydig Cell Number and Function
Explanted 14.5-dpc testes were cultured for 120 h with or without various concentrations of RA, and daily testosterone production (basal secretion) was measured by RIA. LH (100 ng/ml) was added to all media for 3 h, and the LH-stimulated production of testosterone was assayed. As previously shown [24], testosterone secretion increased during the first 3 or 4 days in control culture, with a pattern which mimicked that in vivo [9] (Fig. 8, A and B). RA decreased both basal and stimulated testosterone secretion in a dose-dependent manner to reach 8% of the control value with 10-5 M RA (Fig. 8A). The lowest effective dose was 3.10-8 M RA.
|
We compared the effect of RE on in vitro testosterone production with that of RA (Fig. 8B). Like RA, RE decreased both basal and LH-stimulated testosterone secretion. However, RE was less potent than RA. The negative effect of 10-5 M RE appeared on the second day in culture, whereas RA was active on Day 1. The lowest effective concentration of RE was 10-6 M while it was 3.10-8 M for RA. Lastly, 10-6 M RE inhibited basal and LH-stimulated testosterone production on Day 5 significantly less than did 10-6 M RA. Hence, fetal testis may be able to convert RE into RA.
We determined whether or not RA inhibits testosterone production by decreasing the steroidogenic activity of the fetal Leydig cell and/or reducing the number of Leydig cells per testis; to do this, we counted the total number of 3ßHSD-positive cells per testis at explantation (D0) and after 3 days in culture (D3) (Fig. 9). Very few 3ßHSD-positive cells were detected at explantation. Their number increased more than 50-fold after 3 days in culture in control medium, and RA (10-6 M) had no effect on this spontaneous differentiation. Thus, RA inhibited testicular steroidogenesis in 14.5-dpc testes exclusively by decreasing the steroidogenic activity of each fetal Leydig cell.
|
With 18.5-dpc or 3-dpp explants, the basal testosterone production decreased during the 3 days of culture in controls (Fig. 10), as previously described in vitro [24] and as in vivo [33]. RA had no effect on either basal or LH-stimulated testosterone production.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We first studied the in vitro effect of RA on Sertoli cell organization, growth, proliferation and differentiation. Although RA stimulates the growth and differentiation of epithelial structures in fetal organs such as kidney and lung [12, 34], we found that culture with RA for 3 days causes disorganization of the seminiferous cords of the 14.5-dpc testis. This agrees with a previous report showing that RA treatment for 3 h at the beginning of the culture is sufficient to prevent morphogenesis of the seminiferous cords in the developing rat testis [21]. This effect was recently confirmed and extended by Cupp et al. [19], who showed that RAR-selective agonist and all-trans RA completely inhibited seminiferous cord formation in 13.5-dpc cultured testis. This effect is probably linked with the formation of the basal lamina for the following reasons. First, RA inhibits the synthesis and modifies the deposition of extracellular matrix in many cell types [35, 36]. Second, as Marinos et al. [21] have shown, RA perturbs the deposition of some components of the basement membrane of the sex cords, such as laminin and collagen IV. Third, the proline competitor L-azetidine-2 carboxylic acid (LACA), which inhibits collagen synthesis, prevents the in vitro formation of seminiferous cords in the 13.5-dpc rat testis [37]. Finally, the older stages here studied (18.5-dpc and 3-dpp), when the basement membrane is fully established [38], showed no disorganization caused by RA. The same critical period has been found for other treatments that disorganize the cords, such as adding fetal calf serum, cAMP, or LACA in vitro [37, 39, 40]. Interestingly, we showed that the proliferation of the Sertoli cells was not affected by the disruption of the seminiferous cords by identifying the Sertoli cells with AMH immunostaining. This suggests that the proliferation and organization of the Sertoli cells are two independently controlled processes.
RA had other effects on the development of the Sertoli cells in 18.5-dpc and 3-dpp testes. At both ages, RA increased the size of the Sertoli cells. RA alone also enhanced the mitotic index and the number of Sertoli cells in 3-dpp testes cultured for 3 days, but not in 18.5-dpc testes. Thus, RA may have 3 actions depending on the developmental stage: 1) RA reverses testis morphogenesis at the beginning of the seminiferous cord formation; 2) RA increases the volume of the Sertoli cells when the sex cords are totally differentiated; and 3) RA enhances the mitotic activity of these cells in the neonate testis. The capacity of RA to increase mitosis probably continues in older stages: Jaillard et al. [41] observed an increase in DNA synthesis in immature pig Sertoli cells in response to RA in vitro. However, the effect of RA on Sertoli cell mitosis in 3-dpp testes appeared to be complex, since RA alone had a short-term (27 h) negative effect and a long-term (72 h) positive effect. RA and FSH also had opposite effects: RA impaired the FSH-induced mitosis observed after 27 h in culture, and FSH inhibited the RA-induced mitosis observed after 72 h in culture. These results agree and extend the recent finding by Cupp et al. [19] showing that RA inhibits the FSH-induced incorporation of tritiated thymidine into neonatal testicular cells cultured for 24 h. However, Cupp et al. saw no negative effect of RA alone after 24 h, and this difference from our results is probably due to the fact that our organotypic cultures produced a higher concentration of testicular growth factors.
The differentiated functions of Sertoli cells were estimated by the cAMP produced in response to acute stimulation by FSH. The 14.5-dpc control testes cultured for 3 days responded to FSH, which agrees with previous data showing that FSH receptors are present and functional in Sertoli cells as soon as fetal day 15.5 [42, 43]. The FSH-induced cAMP production gradually increased with older explants, which agrees with previous findings [44]. RA decreased the acute FSH-induced cAMP production, with a greater intensity in older explants, where the response to FSH was greater. The effect of RA on the acute cAMP response of older testes to FSH is controversial, since RA has a negative effect in cultures of Sertoli cells from 20-dpp rat testes [45] and a positive effect in immature pig Sertoli cells [41].
The negative effect of RA on mitosis may be linked to its negative effect on cAMP production, while the mitogenic effect of RA in 3-dpp testes cultured for 3 days is independent of the cAMP pathway since RA decreased cAMP production under these conditions.
The second testicular cell type affected by RA in our organ culture system was the gonocyte. Here also, RA effect depended on the age at explantation. Culturing the 14.5-dpc control testes for 3 days resulted in an increase in the number of gonocytes, which mimics the in vivo development. RA strongly reduced the number of gonocytes; this resulted from a high increase in apoptosis that greatly exceeded the slight RA-induced increase of mitosis. This is, we believe, the first report that RA causes DNA fragmentation in germ cells. It is consistent with the action of retinoids to cause apoptosis in many other cell types or lines [46, 47]. The observed RA-induced decrease in gonocyte number is probably not linked to the disruption of the seminiferous cords. Indeed, in the RA-treated testes, we often observed gonocytes outside the cords and a lack of gonocytes inside cords that were otherwise correctly organized. This independence of seminiferous cord morphogenesis and proliferation of germ cells has been seen in other systems. For instance, culture with dibutyryl cAMP causes a disorganization of the seminiferous cords without affecting the number of gonocytes [40]. Furthermore, the organization of the seminiferous cords is unaffected in rat testis treated by busulfan or in W/W mouse testis in spite of the great depletion in germ cells [38, 48].
RA had no effect on mitosis and apoptosis of the gonocytes in testes from 18.5-dpc fetuses, in which these processes are arrested [23]. RA increased mitosis but not apoptosis in 3-dpp neonates rat testes, in which the mitosis and apoptosis have resumed in the gonocytes, resulting in more germ cells. This positive effect on neonatal gametogenesis is comparable to the increase in BrdU incorporation into spermatogonia of vitamin A-deficient adults in response to RA [15]. Therefore, RA can enhance both apoptosis and mitosis of the gonocytes. In our system, its effect on apoptosis is predominant during the first fetal period, whereas it is mitosis that is most influenced during the neonatal period. It must be noted that the relative importance of the effect on mitosis and apoptosis may also depend on the experimental conditions. Indeed, Koshimizu et al. showed that RA increased the number of mouse primordial germ cells cocultured with somatic cells [49]. In the same way, data from our laboratory indicate that with 3-dpp testes, RA decreased the number of gonocytes when they were cultured as dispersed cells [20].
The third cell type affected by RA was the fetal Leydig cell, and one major finding is that retinoids have a negative effect on the onset of steroidogenesis in the 14.5-dpc fetal testis. This is due to inhibition of the activity of each new Leydig cell, since the differentiation of 3ßHSD-positive cells was not altered. This negative effect of RA could be the result of the disorganization of the seminiferous cords. This seemed to be supported by data showing that LACA, which alters the formation of sex cords, also decreases testosterone production [37] and that stimulation of Sertoli cell activity by FSH increases the steroidogenic activity of the fetal Leydig cells [42]. In the 18.5-dpc testis, the number of Leydig cells continues to increase [50], whereas the production of testosterone by each cell decreases [10], and RA had no more effect. Hence, RA is able to reduce the differentiation of steroidogenesis but seems unable to increase its decline. The effect of RA on the initial differentiation of fetal Leydig cell seems to be specific to the fetus. Indeed, Vitamin A deficiency leads to a decrease in plasma testosterone in adult rats, and this can be reversed with RE [51, 52]. Furthermore, retinoids increase steroid production and the P-450 17 mRNA level in adult Leydig cells and in derived cell lines in vitro [53–55].
Finally, RE was active but less potent than RA in inhibiting fetal testicular steroidogenesis. This suggests that the fetal testis can convert RE into RA. Interestingly, the effective concentration of RE in vitro (10-6 M) is also the RE concentration in the plasma of the rat fetuses in vivo [13], suggesting that the concentrations of retinoids used here are physiological.
Some of the observed effects of RA may be linked. RA had a negative effect on testicular morphogenesis, steroidogenesis, and gametogenesis in 14.5-dpc explants and a positive effect on the proliferation of both Sertoli cells and gonocytes in 3-dpp explants cultured for 3 days. However, we cannot establish any correlation between the age-related change in the effect of RA on steroidogenesis, gametogenesis, and the Sertoli cell cAMP response to FSH. Each of these effects requires further investigation to determine whether RA acts directly on the 3 cell types, or whether there are relationships between these effects. Recent knowledge of the distribution of RA receptors in the various cell types throughout fetal testicular development are not very informative, since most of the three RAR receptors were detected in all cell types in the fetal and neonatal testis [19]. In the same way, data from our laboratory show that most of the three RXRs are also largely expressed in the fetal and neonatal testes [20].
Whether RA acts directly or via autocrine/paracrine factors also remains to be investigated. Curiously, transforming growth factor ß (TGFß) has almost the same effects as RA on the developing testis. It decreases the immature Sertoli cell response to FSH [56] and affects fetal Sertoli cell morphology [57]. It reduces testosterone secretion by the fetal Leydig cells [57] and enhances the apoptosis of gonocytes [23]. Moreover, RA stimulates the TGFß1, TGFß2, and TGFß3 mRNA expression in testicular cells from 0-dpp neonates [19]. Taken together, these data suggest that the effects of RA could be mediated, at least in part, by TGFß.
In conclusion, we have shown that RA is a powerful potential regulator of the development of the endocrine and exocrine functions of the fetal and neonatal testis. RA has negative effects on the onset of steroidogenesis and spermatogenesis in the fetal testis. These negative effects seem specific to early development and are the opposite of those observed in the neonatal and adult testes. Whether RA has any physiological role in vivo remains to be established.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported by INSERM, INRA, and Université Paris 7. G.L. holds a fellowship from the Ministère de l'Education Nationale de la Recherche et de la Technologie. Part of this study was presented in XVe Congrès de la Société d'Andrologie de Langue Française in Lyon on December 10–12th 1998. 
2 Correspondence: René Habert, INSERM U 418—Université Paris 7, Tour 33/43, 2 Place Jussieu, 75251 Paris Cedex 05, France. FAX: 33 1 44 27 56 11; habert{at}paris7.jussieu.fr
Accepted: December 20, 1999.
Received: August 3, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
mutant mice. Dev Biol 1993; 90:7225–7229. and RXR induces differentiation and apoptosis, respectively, in HL-60 cells. Cell Growth Differ 1996; 7:179–186.[Abstract]-selective compounds in mouse thymocytes through a novel apoptosis pathway. Mol Pharmacol 1997; 51:972–982.
5–4-isomerase and 17-hydroxylase/C17–20 lyase cytochrome P-450 messenger ribonucleic acid levels in the K9 mouse Leydig cell line. Mol Cell Endocrinol 1994; 106:31–39.[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  S. Mhaouty-Kodja, A. Lozach, R. Habert, M. Tanneux, C. Guigon, S. Brailly-Tabard, J.-P. Maltier, and C. Legrand-Maltier Fertility and spermatogenesis are altered in {alpha}1b-adrenergic receptor knockout male mice J. Endocrinol., November 1, 2007; 195(2): 281 - 292. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Delbes, C. Duquenne, J. Szenker, J. Taccoen, R. Habert, and C. Levacher Developmental Changes in Testicular Sensitivity to Estrogens throughout Fetal and Neonatal Life Toxicol. Sci., September 1, 2007; 99(1): 234 - 243. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Wang and M. Culty Identification and Distribution of a Novel Platelet-Derived Growth Factor Receptor {beta} Variant: Effect of Retinoic Acid and Involvement in Cell Differentiation Endocrinology, May 1, 2007; 148(5): 2233 - 2250. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Lambrot, H. Coffigny, C. Pairault, A.-C. Donnadieu, R. Frydman, R. Habert, and V. Rouiller-Fabre Use of Organ Culture to Study the Human Fetal Testis Development: Effect of Retinoic Acid J. Clin. Endocrinol. Metab., July 1, 2006; 91(7): 2696 - 2703. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Delbes, C. Levacher, C. Duquenne, C. Racine, P. Pakarinen, and R. Habert Endogenous Estrogens Inhibit Mouse Fetal Leydig Cell Development via Estrogen Receptor {alpha} Endocrinology, May 1, 2005; 146(5): 2454 - 2461. [Abstract] |
