当前位置: 首页 > 期刊 > 《内分泌学杂志》 > 2006年第1期 > 正文
编号:11416160
Retinoic Acid Metabolism and Signaling Pathways in the Adult and Developing Mouse Testis
http://www.100md.com 《内分泌学杂志》
     Institut de Genetique et de Biologie Moleculaire et Cellulaire (IGBMC) (N.V., C.D., C.R.-E., M.O.-A., P.C., N.B.G., M.M.)

    Institut Clinique de la Souris (ICS) (P.C., M.M.), Centre National de la Recherche Scientifique (CNRS)/Institut National de la Sante et de la Recherche Medicale (INSERM)/Universite Louis Pasteur de Strasbourg (ULP)/College de France, 67404 Illkirch Cedex, Communaute Urbaine de Strasbourg, France

    Abstract

    As a first step in investigating the role of retinoic acid (RA) in mouse testis, we analyzed the distribution pattern of the enzymes involved in vitamin A storage (lecithin:retinol acyltransferase), RA synthesis (-carotene 15,15'-monoxygenase and retinaldehyde dehydrogenases) and RA degradation (cytochrome P450 hydroxylases) as well as those of all isotypes of receptors transducing the RA signal [RA receptors (RARs) and rexinoid receptors (RXRs)]. Our data indicate that in adult testis 1) cytochrome P450 hydroxylase enzymes may generate in peritubular myoid cells a catabolic barrier that prevents circulating RA and RA synthesized by Leydig cells to enter the seminiferous epithelium; 2) the compartmentalization of RA synthesis within this epithelium may modulate, through paracrine mechanisms, the coupling between spermatogonia proliferation and spermatogenesis; 3) retinyl esters synthesized in round spermatids by lecithin:retinol acyltransferase may be transferred and stored in Sertoli cells, in the form of adipose differentiation-related protein-coated lipid droplets. We also show that RAR and RXR are confined to Sertoli cells, whereas RAR is expressed in spermatogonia and RAR, RXR, and RXR are colocalized in step 7–8 spermatids. Correlating these expression patterns with the pathological phenotypes generated in response to RAR and RXR mutations and to postnatal vitamin A deficiency suggests that spermiation requires RXR/RAR heterodimers in Sertoli cells, whereas spermatogonia proliferation involves, independently of RXR, two distinct RAR-mediated signaling pathways in both Sertoli cells and spermatogonia. Our data also suggest that the involvement of RA in testis development starts when primary spermatogonia first appear.

    Introduction

    VITAMIN A (RETINOL) is essential for male reproduction. In the adult genital tract, it is necessary for the maintenance of the normal differentiated state of genital ducts, prostate, and seminal vesicle epithelia (1, 2). In the testis, it enhances testosterone production (3, 4), permits the maintenance of Sertoli cell tight junctions that contribute to the blood-testis barrier (Ref.5 and references therein), and plays indispensable roles in spermatogenesis by promoting spermatogonia differentiation, adhesion of germ cells to Sertoli cells, and release of mature spermatids into the lumen of seminiferous tubules (Refs.6 and 7 and references therein).

    Systemic administration of retinoic acid (RA), the active metabolite of vitamin A, can restore spermatogenesis from growth-arrested A spermatogonia in vitamin A-deficient (VAD) testis (8). However, under physiological conditions, less than 1% of the RA present in the testis originates from the circulation (9). Instead, testicular RA is synthesized in situ (10). It is predominantly produced from retinol (ROL) through a two-step metabolic pathway (reviewed in Ref.11), ROL being provided either by the blood where it circulates bound to the ROL-binding protein (RBP) (12) or by the local stores of retinyl esters (reviewed in Ref.13). The first step, oxidation of ROL into retinaldehyde, involves alcohol dehydrogenases, of which three are expressed in the mouse testis: the ubiquitous ADH3, ADH1 in Sertoli cells, and ADH4 in late spermatids (14, 15). The second step, oxidation of retinaldehyde into RA, is catalyzed by four retinaldehyde dehydrogenases (RALDH1–4, encoded by the Aldh1a1 to Aldh1a3 and Aldh8a1 genes, respectively) (11, 16), of which Aldh1a1 and Aldh1a2 are expressed in the rodent testis (17, 18). Alternatively, RA can be produced from -carotenes carried along into chylomicron remnants (19), through their conversion into retinaldehyde by the -carotene 15,15'-monoxygenase, which is expressed in testis (reviewed in Ref.20). Aside from synthesis, degradation of RA is also an important balancing mechanism that protects cells from inadequate RA stimulation (21). It is catalyzed by at least three cytochrome P450 hydroxylases (CYP26A1, CYP26B1, and CYP26C1), which repeatedly hydroxylate RA and its metabolites into increasingly water-soluble products that are less active and readily excretable (Ref.22 and references therein). Whether they are expressed in testis is not known.

    Inside cells, ROL and RA are bound to cellular binding proteins (23). The ROL-binding proteins, CRBP1 and CRBP2, are both involved in ROL storage in vivo (2, 24). In the adult rat testis, CRBP1 is localized mainly in Sertoli cells (25, 26), whereas CRBP2 is not expressed (27, 28). As for RA-binding proteins, CRABP1 and CRABP2, their actual functions remain controversial (23, 29, 30). In the rat testis, CRABP1 is detected in spermatogonia (31) and CRABP2 in developing Sertoli and Leydig cells (32).

    In the nucleus, RA binds to two types of nuclear receptors, the RA receptors (RAR, -, and - isotypes that bind both all-trans and 9-cis RA stereoisomers) and the rexinoid receptors (RXR, -, and - isotypes that bind 9-cis RA only). Although RARs and RXRs may function as homodimers in vitro (33), RXR/RAR heterodimers have been clearly shown to be the functional units transducing the RA signal in the mouse, both during development and in adult tissues (34, 35), where they act as regulators of transcription by binding to RA response elements located in the regulatory regions of target genes (33). Unliganded RAR/RXR heterodimers, bound to RA response elements, recruit transcriptional corepressor complexes associated with histone deacetylase activity, resulting in chromatin condensation and transcriptional silencing. Upon binding of RA, corepressors are released, and the recruitment of coactivator complexes exhibiting histone transacetylase activity results in chromatin decondensation and activation of target gene transcription (36). In addition, transcriptional activation by retinoids might extend beyond RXR/RAR heterodimers, because RXR is also a heterodimeric partner of a number of other nuclear receptors (e.g. peroxisome proliferator-activated receptors, liver X receptors, farnesoid X receptor, and orphan receptors) (reviewed in Ref.37). It is finally worth noting that liganded RAR can alternatively modulate, independently from RXR, the expression of genes through repressing activity of activator protein 1 (AP-1) transcription complexes (heterodimers between fos- and jun-related proteins) (Ref.38 and references therein).

    Previous immunohistochemical (IHC) and in situ hybridization (ISH) studies have shown the distributions of RARs, RXRs, ROL- and RA-metabolizing enzymes, and ROL- and RA-binding proteins in testis (25, 26, 31, 32, 39, 40, 41, 42, 43, 44) as well as in Sertoli cell lines (45). However, most of these studies were performed in rats, and many IHC experiments lacked the proper controls (46). Because targeted mutagenesis makes the mouse a unique model to study gene function, we decided to thoroughly investigate the cellular localization of RA-metabolizing enzymes, ROL- and RA-binding proteins, RARs, and RXRs in the mouse testis, at different stages of postnatal development and in adults. Our results show that the previous IHC data on RAR and RXR distribution need to be revised. They further reveal that RA is likely to act, in the mouse seminiferous epithelium, as a paracrine signal whose level is tightly controlled and varies according to the stage of the seminiferous epithelium cycle. This study provides the solid basis that is required to undertake the somatic mutagenesis experiments aimed at determining the mechanisms of action and the cellular and molecular targets of RA in male reproduction.

    Materials and Methods

    Mice

    Mice on a mixed C57BL/6-129/Sv genetic background were housed in a facility licensed by the French Ministry of Agriculture (agreement B67-218-5). Animal experiments were supervised by M.M. and N.B.G. (agreements 67-62 and 67-205), in compliance with the European legislation on care and use of laboratory animals.

    Staging of the seminiferous epithelium

    The cycle of the seminiferous epithelium is divided into 12 stages, each defined by a specific association of germ cells (47). These epithelial stages are precisely delineated on periodic acid-Schiff-stained histological sections from Bouin-fixed testes, but not necessarily on sections destined for IHC and ISH analyses. In these instances, we identified an epithelial stage or a cohort of consecutive stages using one or two hallmark(s) such as numerous mitotic figures at stage XII; condensed and elongated spermatids, scattered throughout the thickness of the epithelium, at stages II–VI; and preleptotene spermatocytes at stages VII–VIII. Stage I was defined as a segment containing small round spermatids continuous with a segment at stage XII. Stages IX, X, and XI were distinguished from one another by the degree of elongation of spermatid nuclei.

    Histology

    Tissues were fixed in Bouin’s fluid and embedded in paraffin. Five-micrometer-thick sections were stained with periodic acid-Schiff or Mallory’s trichrome (see supplemental data published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org).

    IHC

    Antibodies used in IHC experiments are listed in supplemental Table 1. Five different fixation protocols were tested (supplemental Table 1, and see supplemental data for additional details on fixations), and each experiment was repeated twice on at least three different mice per age group. After rinsing in PBS containing 0.1% (vol/vol) Tween 20 (three times for 5 min each at room temperature), sections were incubated with the affinity-purified primary antibodies for 16 h at 4 C in a humidified chamber. Detection of the bound primary antibodies was achieved for 45 min at room temperature in a humidified chamber, using either a Cy3-conjugated goat antirabbit IgG (Biomol Immuno Research Laboratories, Exeter, UK), a Cy3-conjugated goat antiguinea pig IgG (Euromedex, Souffelweyersheim, France), or a biotinylated antimouse antibody (Vectastain; Vector Laboratories, Burlingame, CA), depending upon the origin of the primary antibody (supplemental Table 1). Nuclei were counterstained with 4',6-diamidino-2-phenyl-indole (DAPI) (Roche Diagnostics, Meylan, France) diluted in the mounting medium at 10 μg/ml (Vectashield; Vector).

    ISH

    Ten-micrometer-thick cryosections were used. ISH with 35S-labeled probes was performed as described (48). The sections were counterstained with toluidine blue. ISH with digoxigenin-labeled probes was performed as described (see supplemental data). Each experiment was repeated on at least three mice per age group. Testes were fixed in 4% (wt/vol) phosphate-buffered paraformaldehyde for 16 h at 4 C and cryoprotected in series of 5, 10, and 20% (wt/vol) sucrose in PBS for 0.5, 2, and 16 h, respectively. We found that intracardiac perfusion of the fixative strikingly increased the ISH signal compared with testes fixed only by immersion. Posthybridization washes were done in MABT [100 mM maleic acid (pH 7.5), 150 mM NaCl, 0.1% (vol/vol) Tween 20). For genes expressed at low levels, 5% (vol/vol) polyvinyl alcohol (Sigma-Aldrich Co., Lyon, France) was added to the staining solution to increase the sensitivity (49). Nuclei were counterstained with DAPI diluted in the mounting medium at 10 μg/ml (Vectashield; Vector).

    The plasmids containing full-length Rara (1.6 kb), Rarb (1.7 kb), Rarg (1.9 kb), Rxra (1.4 kb), Rxrb (1.3 kb), Rxrg (1.4 kb), Cyp26a1 (1,7 kb), Rbp1 (650 bp), Crabp1 (760 bp), and Crabp2 (830 bp) cDNAs or parts of Cyp26b1 (680 bp; exons 2–5), Cyp26c1 (470 bp; exons 4–5), Adfp (800 bp; exons 3–7), Aldh1a1 (1 kb; exons 8–13), Aldh1a2 (468 bp, exons 2–6), Aldh1a3 (500 bp; exons 6–9), Bcmo1 (420 bp; exons 9–11), and Lrat (406 bp; exons 1–2) cDNA were linearized and used as templates for the synthesis of the sense or antisense riboprobes.

    Analysis of -galactosidase activity

    Unfixed testes were frozen in embedding medium. Ten-micrometer-thick sections were air dried, hydrated in PBS, fixed in 1% (wt/vol) paraformaldehyde, 0.1% (wt/vol) glutaraldehyde in PBS containing 1% (vol/vol) Tween 20 for 1 min at room temperature, washed in PBS, and then incubated in PBS containing 2 mM MgCl2, 5 mM K3Fe(CN)6, 5 mM K4Fe(CN)6, and 1 mg/ml 5-bromo-4-chloro-3-indolyl--D-galactoside (X-gal) at 30 C for 20 h. Sections were counterstained with DAPI diluted in the mounting medium at 10 μg/ml (Vectashield; Vector).

    Gold chloride method for detection of retinyl esters

    Mice were perfusion fixed (see supplemental data) with 0.05% (wt/vol) chromic acid in water, under dim light. After dissection, organs were cut in 5 x 10 x 10-mm slices, fixed for 30 min in the perfusion solution, and immediately embedded in 0.8–4% (wt/vol) melting agar according to the stiffness of the sample. Fifty-micrometer-thick vibratome sections were immersed for 10 min in 0.05% chromic acid before staining in gold chloride solution overnight at room temperature. Sections were washed three times in water and mounted in 70% (vol/vol) glycerol in PBS.

    Results

    Expression of RARs in the adult mouse testis

    Using radioactive ISH, Rara and Rxrb transcripts were detected in all tubule sections and spanned the whole thickness of the seminiferous epithelium (Fig. 1, A and E) (see also Ref.39). Rarb and Rxra transcripts were colocalized in the round spermatid layer of tubule sections from stages VII and VIII of the seminiferous epithelium cycle (asterisks and double asterisks in Fig. 1, B and D). Low levels of Rarg transcripts were unevenly distributed at the periphery of all tubule sections (Fig. 1C). Rxrg transcripts were detected in some tubule sections from epithelial stages VII and VIII, but only in part of the circumference of the round spermatid layer (double asterisks in Fig. 1F, compare with B and D). Nonradioactive ISH combined with a DAPI nuclear counterstain allowed accurate identification of Rxra- and Rxrg-expressing cells: all steps 7 and 8 round spermatids expressed Rxra (R, Fig. 1G), whereas only a subset of these spermatids expressed Rxrg transcripts, at the time of sperm release (R, Fig. 1H). Low levels of Rxra transcripts were also detected in steps 2, 3, 4, 5, and/or 6 spermatids (which can hardly be distinguished from one another on histological sections processed for ISH; see Materials and Methods), as well as in step 9 spermatids (not shown). Therefore, Rara, Rarg, and Rxrb are expressed in all tubules irrespective of the epithelial stages, in contrast to Rarb, Rxra, and Rxrg, which are expressed mainly or exclusively in steps 7 and 8 spermatids (Table 1).

    We have used three antibodies to detect RAR by IHC: RP(F), Ab9(F), and sc-551 (supplemental Table 1). In wild-type (WT) males, the RP(F) antibody yielded a strong signal in the nuclei of all Sertoli cells (S, Fig. 2, A and B), in agreement with previous observations using a different protocol (44). RAR was undetectable in germ cells (E, P, and R, Fig. 2B) as well as in peritubular cells (M, Fig. 2B). Importantly, the RP(F) antibody yielded no nuclear signal in Rara-null testis (50), demonstrating that it was actually specific for RAR (Fig. 2, C and D). In contrast, the weak cytoplasmic signal exhibited by Leydig cells (LY, Fig. 2, A–D) was unrelated to RAR, because it was detected in WT (Fig. 2, A and B) and Rara-null testis (Fig. 2, C and D). In WT testis, sc-551 yielded a strong cytoplasmic signal in Sertoli cells (S, Fig. 2, E and F), whereas Ab9(F) gave a strong nuclear signal in Sertoli and germ cells (supplemental Fig. 1 and supplemental Table 1). However, irrespective of the IHC protocol that was employed, identical signals were observed in Rara-null testis (S, Fig. 2, G and H, and supplemental Table 1), indicating that neither sc-551 nor Ab9(F) did specifically recognize RAR. Altogether, our results show that RAR is essentially, if not only, present in the nuclei of Sertoli cell in the adult mouse testis (Table 1).

    In WT testis, the RP(mF) antibody, recognizing the RAR isotype (supplemental Table 1), yielded a nuclear signal within a few cells that were identified as A spermatogonia based on their distribution at the periphery of all seminiferous tubules (A, Fig. 2, K and L) and their cytological features (oval nuclei lying in parallel with the basement membrane and containing little heterochromatin; A, Fig. 2, M–P). The RP(mF) antibody also yielded a faint positive nuclear signal in pachytene spermatocytes (white arrowheads, Fig. 2, M and O). In RargZL–/ZL– mutants, carrying null alleles of Rarg knocked-in with LacZ reporter genes (51), no immunostaining could be detected in spermatogonia, but spermatocytes exhibited a faint signal identical to that of WT spermatocytes (Fig. 2, Q and R, and not shown). Importantly in RargZL–/ZL– testis, A spermatogonia were the only cells to display the -galactosidase activity reporting the cells where Rarg was expressed before site-directed mutagenesis (A, Fig. 2, S and T). We conclude that RAR is essentially, if not only, present in the nuclei of A spermatogonia in the adult mouse testis (Table 1).

    RXR was detected using the RPRX- and 4RX3A2 antibodies (supplemental Table 1). Because of the embryonic lethality of Rxra-null mutants (34), testes lacking RXR are not available. However, skin of mice lacking Rxra in all epidermal keratinocytes (Rxraep–/–) was used as a negative control of IHC (52). Importantly, the immunostaining generated by RPRX- was indistinguishable in WT and Rxraep–/– skin samples, indicating that RPRX- did not specifically recognize RXR in IHC experiments. Therefore, the antigens revealed by this antibody in the nuclei of several somatic and germ cell types are likely to be unrelated to RXR (not shown). In contrast, the 4RX3A2 antibody yielded no nuclear signal in the keratinocytes of Rxraep–/– mice, demonstrating its specificity for RXR (not shown; see Ref.52). In the adult testis, 4RX3A2 specifically labeled steps 7 and 8 round spermatid nuclei (R, Fig. 2I, compare with J), in accordance with our ISH data (Fig. 1, D and G). Therefore, RXR is essentially, if not only, present in the nuclei of steps 7 and 8 round spermatids in the adult mouse testis (Table 1).

    We have reported elsewhere the immunolocalization of RXR in Sertoli cells, using testis from Rxrb-null mice as negative controls (39). Here, we found that the sc-831 antibody (supplemental Table 1) always gave identical, nonspecific, signals on testis sections from WT and Rxrb-null mice, even though five different IHC protocols were used (supplemental Table 1).

    Finally, RAR and RXR could not be detected in adult testis using the antibodies RP(F)2 and sc-555 (supplemental Table 1), respectively. These antibodies efficiently and specifically recognize RAR and RXR in mouse tissues (53, 54). Therefore, RAR and RXR are either absent from testicular cells or expressed at levels below the detection limits of IHC (Table 1). Altogether, the present data emphasize the importance of tissues from null mutants as the only valid controls to assess the specificity of IHC experiments (46).

    Expression of retinoid receptors in the developing mouse testis

    Expression of the RARs and RXRs was studied at postnatal d 1 (P1), at P5 (when gonocytes differentiate into primitive spermatogonia), at P10 (i.e. at the onset of meiosis), and at P20 (when postmeiotic cells first appear) (55). RAR was confined to the nuclei of Sertoli cell precursors at P1 and P5 (S, Fig. 3, A–D) and subsequently to those of immature Sertoli cells at P10 and P20 (S, Fig. 3, E–H). Germ cells (G, A, PR, P), as well as precursors of peritubular myoid (M) and Leydig (LY) cells, were not immunostained by the RP(F) antibody (Fig. 3, A–H). RAR was first detected at P5 in primitive spermatogonia (A, Fig. 3, K and L) and in A spermatogonia at P10 and P20 (A, Fig. 3, M–P). Earlier spermatogonia precursors, namely gonocytes, did not express RAR (G, Fig. 3, I and J).

    Rxra transcripts were undetectable in the seminiferous tubules at P5, P10, and P20. However, at these developmental stages, Rxra was expressed in the interstitial compartment of the testis, which contains the Leydig cell precursors (LY, Fig. 3Q and not shown). In contrast, Rxrb transcripts were detected within the developing seminiferous tubules at P5, P10, and P20. As exemplified at P5, Rxrb transcripts were confined to the central portion of the tubule sections, which corresponds to the cytoplasm of Sertoli precursor cells (T, Fig. 3R). As expected from their localization in late round spermatids in the adult testis, Rarb and Rxrg transcripts were undetectable in the P1–20 developing testis (data not shown).

    Altogether, these data indicate that Rara, Rarg, and Rxrb are expressed in given testicular cell types throughout postnatal development and adult life. They also suggest that RXR can perform distinct functions in prepubertal vs. adult testis.

    Expression of RA-synthesizing and RA-degrading enzymes in the adult and developing testis

    In adult testis, Leydig and Sertoli cells expressed high and low levels of Aldh1a1 transcripts, respectively (Fig. 4, A and B). Aldh1a2 transcripts were detected in germ cells. They were most abundant in late pachytene spermatocytes and at epithelial stages VII–X (P, Fig. 4, E–H) and in diplotene spermatocytes at stage XI (not shown). They were also present in round spermatids (R2–6) at epithelial stages I (not shown) and II–VI (Fig. 4, E and F). In contrast, they were undetectable in spermatogonia, preleptotene (PR), leptotene, zygotene, and early pachytene spermatocytes as well as in late, steps 7 and 8, round spermatids (R7–8) and in elongating spermatids (e.g. E16) (Fig. 4, E–H, and not shown). Low levels of Aldh1a3 transcripts were detected in adult Leydig cells (data not shown).

    In the developing testis, Aldha1 expression was undetectable at P1 (not shown), very robust in Sertoli cell precursors at P5 (S, Fig. 4, C and D), but weak at P10 and P20, and started in Leydig cells around P20 (not shown). As in the adult testis, Aldh1a1 transcripts were absent from germ cells (not shown). Low levels of Aldh1a2 transcripts were detected at P5 in Sertoli cells (not shown) and strong expression was observed at P20, concomitantly with the appearance of late pachytene spermatocytes (P, Fig. 4, I and J). Aldh1a3 transcripts were undetectable in the developing testis (not shown).

    The Bcmo1 transcripts, encoding the -carotene cleaving enzyme, are present in whole-testis extracts (20). They were restricted to germ cells during the late stages of spermiogenesis (Fig. 4, K and L), namely round spermatids at steps 7 and 8 (R7–8), elongating spermatids at steps 9–12 (E10, Fig. 4, K and L), and to a lesser extent, elongated and condensed spermatids at steps 13–15 (E13–15). Bcmo1 transcripts were absent from developing testes analyzed from P1–20 (not shown).

    The Cyp26a1, Cyp26b1, and Cyp26c1 transcripts, encoding RA-hydroxylating enzymes, were all restricted to the peritubular myoepithelial cells (M) both in the adult and the developing testis (Fig. 5, A–H, and not shown).

    Altogether, these results indicate that 1) RALDH1 may perform distinct functions in the developing and adult testes, 2) RALDH2 appears to be responsible for essentially all RA synthesis within the seminiferous epithelium, 3) RALDH2 and -carotene 15,15'-monoxygenase activities are tightly regulated during the seminiferous epithelium cycle and do not overlap in the postmeiotic populations of this epithelium, and 4) all the known RA-degrading CYP26 activities are confined to a single somatic cell-type both in the adult and the developing testis (Table 1).

    Expression of the cellular RA-binding protein in the adult and developing testis

    In the adult mouse testis, Crabp1 transcripts were detected exclusively in A and B spermatogonia. During development, Crabp1 expression was absent at P1 and was very robust in apparently all the spermatogonia at P5, P10, and P20. Other germ cell types as well as the testicular somatic cells did not express Crabp1 (Table 1, supplemental Fig. 2, and not shown). These results indicate that the pattern of Crabp1 expression has been conserved between the rat (31, 32) and the mouse. Contrary to the rat testis (32), the mouse testis did not express Crabp2 during development, and Crabp2 transcripts were also absent from the adult organ. Note that we checked that the probe used in our ISH assays gave a strong and specific signal in mouse embryonic tissues known to express Crabp2 (not shown).

    Vitamin A storage in the adult mouse testis

    Using the gold chloride method (56), retinyl esters were detected in the lipid droplets located in Sertoli cells (L1, Fig. 6A). Minute positive droplets were also occasionally detected at the luminal side of the seminiferous epithelium (L2, Fig. 6A). Because the technique preserves germ cells only poorly, it is unclear whether these small droplets belonged to spermatids or to Sertoli cells. In contrast, the lipid-containing Leydig cells did not react with the gold chloride, indicating that the staining was specific (LY, Fig. 6A). Along these lines, the vitamin A-storing liver stellate cells (56) and the lipid- but not vitamin A-storing brown adipocytes served, respectively, as positive and negative controls of the histochemical reaction (supplemental Fig. 3).

    Under physiological conditions, lecithin:retinol acyltransferase (LRAT) is the most potent esterifying enzyme for ROL storage (57). Its efficiency is enhanced by CRBP1 (58, 59), whose gene (i.e. Rbp1) inactivation dramatically reduces the retinyl ester stores (2). Both of these proteins are expressed in the rat testis (25, 26, 60). Thus, we used ISH to identify the cell types expressing them in the adult mouse testis. The Lrat transcripts were detected only at epithelial stages II–VI, in step 2–6 round spermatids (Fig. 6, B and C), whereas the Rbp1 transcripts were detected mostly in Leydig cells (LY, Fig. 6, D–F) and at much lower levels in Sertoli cells, at epithelial stages X–XI (Fig. 6, E and F). Note that we did not check the expression of Rbp2 (encoding CRBP2) because this gene is expressed only in the adult intestine in the rat (27, 28).

    The adipose differentiation-related protein (ADFP), which is localized at the surface of lipid droplets in many tissues (reviewed in Ref.61), including Sertoli cells (62), stimulates lipid droplet formation (63). To investigate the formation of retinyl ester-containing lipid droplets in Sertoli cells, we analyzed the distribution of ADFP and Adfp transcripts in the adult mouse seminiferous epithelium. ADFP was associated with 1) minute lipid droplets in elongated spermatids (L2, Fig. 6, G and H), 2) residual bodies that detach from these mature spermatids and are phagocytosed by Sertoli cells at spermiation (not shown), and 3) lipid droplets at the base of Sertoli cells (L1, Fig. 6, G and H). In contrast, Adfp transcripts were absent from Sertoli cells but expressed in round spermatids at steps 2–6 and at the beginning of step 7 (Fig. 6, I and J, and not shown).

    Altogether, these data suggest that 1) retinyl ester synthesis by LRAT may not involve CRBP1 in the mouse testis, because the latter is expressed in Sertoli cells, whereas the former is expressed in early spermatids (Table 1), and 2) the vitamin A-containing droplets located in Sertoli cells arise, at least in part, from those present in spermatids.

    Spermatogenesis is not altered in Rarb-, Rarg-, Rrxg-, Aldh1a1-, Aldh1a3-, Crabp1-, and Crabp2-null mutants

    The Rara- and Rxrb-null males are sterile because of testicular defects, namely degeneration and/or abnormal spermiation (39, 50, 62). In contrast, young Rarb-, Rrxg-, and Aldh1a1-null males are fertile (53, 64, 65). At 12 months of age, Rarb-, Rxrg-, and Aldh1a1-null males (n = 4 for each genotype) were also fertile but occasionally displayed one to three degenerating seminiferous tubules. This abnormality was, however, not related to the mutations, because it was observed with the same penetrance in age-matched WT littermates (not shown). Moreover, failure of spermiation was never observed in these mutant mice. Likewise, the testis from 12-month-old Aldh1a3-null mutants (n = 2), whose lethal malformations were rescued by maternal administration of RA (66), were histologically normal (not shown). Accordingly, these Aldh1a3-null males were fertile. Crabp1- and Crabp2-null mutants as well as Crabp1/Crabp2 double-null mutants (Ref.29 and references therein) have been bred for years in the homozygous state in our animal facility. Histological analysis of the testes from two of each of these mutants at 6 months of age did not reveal any abnormality (not shown).

    The sterility of Rarg-null males has been ascribed to the squamous metaplasia of their prostate and seminal vesicle epithelia (67). Accordingly, the testes and epididymides of young (i.e. 2- to 6-month-old) mutants were histologically normal (Fig. 7, A and B). In contrast, the testes from 12-month-old Rarg-null mutants (n = 5) consistently displayed signs of degeneration manifested by a thinning of the seminiferous epithelium and by a loss of germ cells occasionally yielding tubules with only Sertoli cells (D, Fig. 7C). These lesions were associated with accumulations of spermatozoa within the lumen of some seminiferous tubules (SZ, Fig. 7C) and with an extensive keratinization of the caudal epididymis epithelium (bracket, Fig. 7D). These testicular lesions were mimicked in WT males upon ligature of the efferent ductules (Fig. 7E) and resembled those of Esr1-null (formerly estrogen receptor-, ER-null) males in which testicular fluid absorption by the epithelial cells lining the efferent ductules is inhibited (68). In contrast, the epididymal epithelium did not keratinize in efferectomized males (Fig. 7F), indicating that squamous metaplasia in Rarg-null males is actually related to the loss of RAR.

    Altogether, these observations indicate that Rarb-, Rxrg- Aldh1a1-, and Aldh1a3-null males are not prone to age-related testicular defects, whereas the testis degeneration exhibited by old Rarg-null males is most probably caused by an obstruction of the epididymal lumen by desquamated corneocyte-like cells generated upon the squamous metaplasia of the epithelium (SQ, Fig. 7D). It should finally be noted that spermatogenesis in Rxra-, Aldh1a2-, Cyp26a1-, and Cyp26b1-null mutants could not be analyzed, because all of these mutants die in utero or at birth (Refs.21 , 34 , and 69 and references therein).

    Discussion

    During the last decade, the localization of retinoid receptors has been extensively analyzed in rodent testes. However, we do not confirm the results from previous studies (40, 41, 42, 43, 45), which in fact were performed with antibodies that do not recognize specifically RAR (i.e. sc-551), RXR (i.e. RPRX-), and RXR (i.e. sc-831) (supplemental Table 1). Likewise, our data demonstrating a nuclear-restricted localization of RAR do not support the proposal that shifts between the cytoplasm and nucleus may regulate the functions of this receptor in Sertoli cells (43, 45). As for RA-synthesizing and RA-degrading enzymes, most of their expression patterns were not studied.

    The present work indicates that in the mouse testis during postnatal development and adulthood, RAR and RXR are expressed in well defined cell populations (Table 1 and Fig. 8): Rara and Rxrb in Sertoli cells, Rarb, Rxra, and Rxrg in steps 7 and 8 spermatids, and Rarg in spermatogonia. Spermatocytes do not express RARs. As for RA-synthesizing enzymes, Aldh1a1 is mainly expressed in immature Sertoli cells and in adult Leydig cells and Aldh1a2 in pachytene spermatocytes and round spermatids, Aldh1a3 is faintly expressed in Leydig cells of adults only, and Bcmo1 is restricted to late round and elongating spermatids. As for RA-degrading enzymes, all three Cyp26 genes are confined to peritubular myoid cells, both in the adult and the developing testis. Lastly, our data indicate that Rbp1 is expressed mainly in Leydig cells and at much lower levels in Sertoli cells, and Lrat is expressed in early spermatids, whereas retinyl ester droplets are located in Sertoli cells.

    Actions of RA on spermatogenesis may not involve RXR/RAR heterodimers

    The coexpression of Rarb, Rxra, and Rxrg in step 7 and 8 spermatids suggests that RXR/RAR and RXR/RAR heterodimers could be functional in these cells. However, Rarb-, Rxrg-, and Rarb/Rxrg-null mice do not display reproductive defects, indicating that both RAR and RXR are dispensable for spermatogenesis in mice fed a standard breeding diet (Refs.53 and 64 and present report). On the other hand, the coexpression of Rara and Rxrb in Sertoli cells and the phenotypes resulting from Rara and Rxrb ablations suggest that RXR/RAR heterodimers may control spermiation (39, 70).

    With exception of spermiation defects, Rara- and Rxrb-null testes exhibit distinct sets of abnormalities. Those observed in Rxrb-null testes can be ascribed to a loss of the RXR/liver X receptor--mediated signal transduction (62). On the other hand, because no RXR other than RXR is apparently expressed in Sertoli cells, the early testicular degeneration observed in Rara-null (50) but not in Rxrb-null mutants (39, 62) may highlight functions of RAR independent of RXR, for example through a RAR-mediated AP-1 transrepression. Likewise, because no RXR is detectable in spermatogonia, a RAR-mediated AP-1 transrepression may play a role in the mitotic phase of spermatogenesis (71).

    Some RA effects on the seminiferous epithelium are apparently not dependent on RAR

    The testicular degenerations resulting from Rara ablation and induced by VAD share similar features that strongly support the view that a liganded RAR is instrumental for the adhesion of germ cells to Sertoli cells (Ref.50 and our unpublished results). Another hallmark of VAD, namely the arrest of spermatogonia differentiation (7, 8), is not seen in the Rara-null testis (50). As shown here, RAR is a suitable candidate for mediating RA actions in spermatogonia. However, Rarg ablation has only an indirect impact on testicular histology. This apparent dispensability of RAR for spermatogonia differentiation may be accounted for by a functional redundancy between RARs, because low levels of RAR and/or RAR in spermatogonia, below the detection limits of IHC and ISH, could possibly compensate for the loss of RAR. Alternatively, the arrest of spermatogonia differentiation observed upon dietary VAD could result from a block of RA signal transduction occurring simultaneously in spermatogonia and Sertoli cells. This hypothesis will be tested through somatic mutagenesis of the RAR genes (72), using available transgenic lines expressing the Cre recombinase either in Sertoli cells (73) or in spermatogonia (74).

    The peritubular cell layer functions as a catabolic barrier preventing RA from entering the seminiferous tubules

    Because the testicular RA is not derived from the circulation (9), it has to be synthesized locally. We found two potential sites of RA synthesis in the adult mouse testis, one extratubular and the other intratubular. The former source relies on the expression of Aldh1a1 in Leydig cells. Although RALDH1 has only a weak retinaldehyde oxidizing activity in vitro, it actually synthesizes RA in vivo (65, 75). Thus, its high level of expression may permit conversion of retinaldehyde into RA in Leydig cells (Ref.18 and present report). That Aldh1a1-null mutants do not display reproductive defects possibly reflects the existence of a functional compensation by RALDH3, which is detected at low levels in Leydig cells.

    Any RA made outside of the seminiferous tubules is nevertheless unlikely to reach Sertoli or germ cells, because of the expression, in peritubular myoid cells, of CYP26 (A1, B1, and C1) RA-degrading enzymes. The existence of this catabolic barrier provides the explanation for the inability of low doses of RA to restore spermatogenesis in VAD rodents (8, 9). In this context, it is noteworthy that although pharmacological doses of 4-oxo-RA are able to exert biological activities in the mouse testis (76), most of the oxidized RA metabolites produced by CYP26 enzymes (including 4-oxo-RA) cannot be detected in vivo (10) and are unlikely to exert physiological functions in the mouse (21).

    The sources of RA synthesis and the loci of RA signal transduction are distinct in the mouse germ cells

    In the testis, both ROL and -carotene represent potential substrates for RA synthesis (19, 77). Late round, elongating, and mature spermatids (steps 7–15) have the capacity to generate retinaldehyde from ROL and -carotenes through their expression of Adh4 (14) and Bcmo1 (present report), respectively. However, they cannot synthesize RA because they are devoid of RALDH (Fig. 8). Retinaldehyde originating from these spermatids may diffuse toward the closest germ cell population present within the same cell association, namely late pachytene spermatocytes, diplotene spermatocytes, and early (steps 1–6) spermatids, where it can be converted into RA by RALDH2 (Fig. 8). These three germ cell populations represent therefore essential sites of RA synthesis in the seminiferous epithelium. Variations in the expressions levels of Aldh1a2 and Bcmo1 during the seminiferous epithelium cycle may provide a means to modulate local RA concentrations and thus to control specific events of spermatogenesis.

    No RAR is expressed in the RA-synthesizing spermatocytes (Fig. 8). Conversely, no RALDH is detected in germ cells that express RAR, namely spermatogonia and steps 7 and 8 spermatids. Despite expression of the alcohol dehydrogenase Adh1 (14) and of Aldh1a1 (present report), the capacity of Sertoli cells to convert ROL into RA is probably limited, because RALDH1 is present at low levels (Table 1) and has a weak retinaldehyde-oxidizing activity (see above). It appears therefore likely that steps 7 and 8 spermatids use RA synthesized by RALDH2 in their immediate precursors (i.e. step 6 spermatids). As for spermatogonia, they may receive most of their RA from the nearest Aldh1a2-expressing cells, i.e. the late pachytene and diplotene spermatocytes (from which they are separated only by slender Sertoli cell processes; Fig. 8), at least under physiological conditions. Interestingly, this compartmentalization of RA synthesis within the seminiferous epithelium may be used to modulate, through a paracrine mechanism, the synchronization between the mitotic and meiotic phases of spermatogenesis. On the other hand, it is well established that spermatogonial differentiation can resume normally under various conditions where more advanced germ cells are absent, notably during recovery from irradiation, during repopulation after transplantation, after replacement of cryptorchid testes in the scrotum, and once VAD mice are put again on a vitamin A-containing diet. Therefore, spermatogonia should also receive some RA synthesized by the neighboring Sertoli cells.

    RA signaling appears dispensable for prenatal but not for postnatal testis development

    At P1, there is no intratubular source of RA, and RARs are not expressed in gonocytes. Accordingly, fetuses lacking Rara and Rarb, Rara and Rarg, or Rarb and Rarg do not display testicular defects (Ref.53 and references therein). In addition, the CYP26-mediated catabolic barrier that prevents RA from reaching the seminiferous epithelium is present in peritubular myoid cell precursors at P1 (present report) and most probably also at fetal stages (78). Additionally, because the survival of cultured mouse gonocytes and Sertoli cell precursors is dramatically reduced by RA (43), it is possible that RA must be excluded from the fetal testis.

    At P5, the seminiferous epithelium is still insulated from exogenous sources of RA through the Cyp26-expressing peritubular myoid cells. However, at this time, the Sertoli cell precursors display a robust expression of Aldh1a1 that may provide a ligand to activate RAR in these cells and RAR in spermatogonia. These observations suggest that, at P5, a RA signal may play important functions in the differentiation of Sertoli cell precursors and primitive spermatogonia.

    Retinyl esters are stored in Sertoli cells and may originate from spermatids

    We have shown here that mouse Sertoli cells and spermatids contain retinyl esters similarly to their rat counterparts (79). LRAT is the most potent ROL-esterifying enzyme, whose efficiency is enhanced by CRBP1 (58, 59). Rat Sertoli cells accordingly coexpress Lrat and Rbp1, whereas spermatids express Lrat but not Rbp1 (26, 60). In the mouse, we show here that Sertoli cells also express Rbp1 but not Lrat, whereas spermatids express Lrat but not Rbp1. Altogether, these observations suggest that 1) in the testis of both species the main site of Lrat expression is the spermatid population, where CRBPI is absent, and 2) as previously observed in cultured cells (80), ROL esterification by LRAT may not necessarily require CRBP. Importantly, many of the Lrat-null mouse males are sterile (81), suggesting that expression of Lrat in spermatids plays an important role in fertility. On the other hand, our data do not allow us to rule out the possibility that an acyl CoA:retinol acyltransferase activity may catalyze the esterification of ROL in Sertoli cells (82).

    ADFP, which is localized at the surface of lipid droplets (reviewed in Ref.61), stimulates their formation (63). Interestingly, the Lrat-expressing spermatids are the only cells of the seminiferous epithelium that contain Adfp transcripts, which are translated into proteins in elongating spermatids. We therefore propose that coexpression of Lrat and Adfp couples retinyl ester synthesis with their packaging in lipid droplets. Our IHC observations also support the possibility that Sertoli cell retinyl ester stores are derived from spermatid lipid droplets that are delivered to the Sertoli cells after spermiation.

    Acknowledgments

    We thank B. Feret, B. Weber, and the staff of the Institut de Genetique et de Biologie Moleculaire et Cellulaire and Institut Clinique de la Souris common services for their technical assistance.

    Footnotes

    This work was supported by funds from the Centre National de la Recherche Scientifique (CNRS), the Institut National de la Sante et de la Recherche Medicale (INSERM), the Hpital Universitaire de Strasbourg, the College de France, and the Institut Universitaire de France (IUF). N.V. was supported by an IUF fellowship.

    First Published Online October 6, 2005

    Abbreviations: ADFP, Adipose differentiation-related protein; AP-1, activator protein 1; CRABP, cellular retinoic acid-binding protein; CRBP, cellular retinol-binding protein; CYP, cytochrome P450 hydroxylase; DAPI, 4',6-diamidino-2-phenyl-indole; IHC, immunohistochemistry; ISH, in situ hybridization; LRAT, lecithin:retinol acyltransferase; P1, postnatal d 1; RA, retinoic acid; RALDH, retinaldehyde dehydrogenase; RAR, RA receptor; RBP, retinol-binding protein; ROL, retinol; RXR, rexinoid receptor; VAD, vitamin A-deficient; WT, wild type.

    Accepted for publication September 22, 2005.

    References

    Wolbach SB, Howe PR 1925 Tissue changes following deprivation of fat-soluble A vitamin. J Exp Med 42:753–777

    Ghyselinck NB, Bavik C, Sapin V, Mark M, Bonnier D, Hindelang C, Dierich A, Nilsson CB, Hakansson H, Sauvant P, Azais-Braesco V, Frasson M, Picaud S, Chambon P 1999 Cellular retinol-binding protein I is essential for vitamin A homeostasis. EMBO J 18:4903–4914

    Appling DR, Chytil F 1981 Evidence of a role for retinoic acid (vitamin A-acid) in the maintenance of testosterone production in male rats. Endocrinology 108:2120–2124

    Huang HF, Dyrenfurth I, Hembree WC 1983 Endocrine changes associated with germ cell loss during vitamin A-induced recovery of spermatogenesis. Endocrinology 112:1163–1171

    Morales A, Cavicchia JC 2002 Spermatogenesis and blood-testis barrier in rats after long-term vitamin A deprivation. Tissue Cell 34:349–355

    Huang HF, Marshall GR 1983 Failure of spermatid release under various vitamin A states: an indication of delayed spermiation. Biol Reprod 28:1163–1172

    Griswold MD, Bishop PD, Kim KH, Ping R, Siiteri JE, Morales C 1989 Function of vitamin A in normal and synchronized seminiferous tubules. Ann NY Acad Sci 564:154–172

    van Pelt AM, de Rooij DG 1991 Retinoic acid is able to reinitiate spermatogenesis in vitamin A-deficient rats and high replicate doses support the full development of spermatogenic cells. Endocrinology 128:697–704

    Kurlandsky SB, Gamble MV, Ramakrishnan R, Blaner WS 1995 Plasma delivery of retinoic acid to tissues in the rat. J Biol Chem 270:17850–17857

    Schmidt CK, Brouwer A, Nau H 2003 Chromatographic analysis of endogenous retinoids in tissues and serum. Anal Biochem 315:36–48

    Duester G 2001 Genetic dissection of retinoid dehydrogenases. Chem Biol Interact 130–132:469–480

    Soprano DR, Blaner WS 1994 Plasma retinol-binding protein. In: Sporn MB, Roberts AB, Goodman DS, eds. The retinoids: biology, chemistry and medicine. 2nd ed. New York: Raven Press; 257–281

    Ross AC 2003 Retinoid production and catabolism: role of diet in regulating retinol esterification and retinoic acid oxidation. J Nutr 133:291S–296S

    Deltour L, Haselbeck RJ, Ang HL, Duester G 1997 Localization of class I and class IV alcohol dehydrogenases in mouse testis and epididymis: potential retinol dehydrogenases for endogenous retinoic acid synthesis. Biol Reprod 56:102–109

    Molotkov A, Fan X, Deltour L, Foglio MH, Martras S, Farres J, Pares X, Duester G 2002 Stimulation of retinoic acid production and growth by ubiquitously expressed alcohol dehydrogenase Adh3. Proc Natl Acad Sci USA 99:5337–5342

    Lin M, Zhang M, Abraham M, Smith SM, Napoli JL 2003 Mouse retinal dehydrogenase 4 (RALDH4), molecular cloning, cellular expression, and activity in 9-cis-retinoic acid biosynthesis in intact cells. J Biol Chem 278:9856–9861

    Zhai Y, Sperkova Z, Napoli JL 2001 Cellular expression of retinal dehydrogenase types 1 and 2: effects of vitamin A status on testis mRNA. J Cell Physiol 186:220–232

    Lopez-Fernandez LA, del Mazo J 1997 The cytosolic aldehyde dehydrogenase gene (Aldh1) is developmentally expressed in Leydig cells. FEBS Lett 407:225–229

    Napoli JL, Race KR 1988 Biogenesis of retinoic acid from -carotene. Differences between the metabolism of -carotene and retinal. J Biol Chem 263:17372–17377

    Paik J, During A, Harrison EH, Mendelsohn CL, Lai K, Blaner WS 2001 Expression and characterization of a murine enzyme able to cleave -carotene. The formation of retinoids. J Biol Chem 276:32160–32168

    Niederreither K, Abu-Abed S, Schuhbaur B, Petkovich M, Chambon P, Dolle P 2002 Genetic evidence that oxidative derivatives of retinoic acid are not involved in retinoid signalling during mouse development. Nat Genet 31:84–88

    Taimi M, Helvig C, Wisniewski J, Ramshaw H, White J, Amad M, Korczak B, Petkovich M 2004 A novel human cytochrome P450, CYP26C1, involved in metabolism of 9-cis and all-trans isomers of retinoic acid. J Biol Chem 279:77–85

    Napoli JL 1999 Retinoic acid: its biosynthesis and metabolism. Prog Nucleic Acids Res Mol Biol 63:139–188

    E X, Zhang L, Lu J, Tso P, Blaner WS, Levin MS, Li E 2002 Increased neonatal mortality in mice lacking cellular retinol-binding protein II. J Biol Chem 277:36617–36623

    Rajan N, Sung WK, Goodman DS 1990 Localization of cellular retinol-binding protein mRNA in rat testis and epididymis and its stage-dependent expression during the cycle of the seminiferous epithelium. Biol Reprod 43:835–842

    Zhai Y, Higgins D, Napoli JL 1997 Coexpression of the mRNAs encoding retinol dehydrogenase isozymes and cellular retinol-binding protein. J Cell Physiol 173:36–43

    Li E, Demmer LA, Sweetser DA, Ong DE, Gordon JI 1986 Rat cellular retinol-binding protein II: use of a cloned cDNA to define its primary structure, tissue-specific expression, and developmental regulation. Proc Natl Acad Sci USA 83:5779–5783

    Levin MS, Li E, Ong DE, Gordon JI 1987 Comparison of the tissue-specific expression and developmental regulation of two closely linked rodent genes encoding cytosolic retinol-binding proteins. J Biol Chem 262:7118–7124

    Lampron C, Rochette-Egly C, Gorry P, Dolle P, Mark M, Lufkin T, LeMeur M, Chambon P 1995 Mice deficient in cellular retinoic acid binding protein II (CRABPII) or in both CRABPI and CRABPII are essentially normal. Development 121:539–548

    Noy N 2000 Retinoid-binding proteins: mediators of retinoid action. Biochem J 348:481–495

    Rajan N, Kidd GL, Talmage DA, Blaner WS, Suhara A, Goodman DS 1991 Cellular retinoic acid-binding protein messenger RNA: levels in rat tissues and localization in rat testis. J Lipid Res 32:1195–1204

    Zheng WL, Bucco RA, Schmitt MC, Wardlaw SA, Ong DE 1996 Localization of cellular retinoic acid-binding protein (CRABP) II and CRABP in developing rat testis. Endocrinology 137:5028–5035

    Chambon P 1996 A decade of molecular biology of retinoic acid receptors. FASEB J 10:940–954

    Kastner P, Mark M, Ghyselinck N, Krezel W, Dupe V, Grondona JM, Chambon P 1997 Genetic evidence that the retinoid signal is transduced by heterodimeric RXR/RAR functional units during mouse development. Development 124:313–326

    Mark M, Chambon P 2003 Functions of RARs and RXRs in vivo: genetic dissection of the retinoid signalling pathway. Pure Appl Chem 75:1709–1732

    Glass CK, Rosenfeld MG 2000 The coregulator exchange in transcriptional functions of nuclear receptors. Genes Dev 14:121–141

    Laudet V, Gronemeyer H 2002 The nuclear receptor facts book. London: Academic Press

    Benkoussa M, Brand C, Delmotte MH, Formstecher P, Lefebvre P 2002 Retinoic acid receptors inhibit AP1 activation by regulating extracellular signal-regulated kinase and CBP recruitment to an AP1-responsive promoter. Mol Cell Biol 22:4522–4534

    Kastner P, Mark M, Leid M, Gansmuller A, Chin W, Grondona JM, Decimo D, Krezel W, Dierich A, Chambon P 1996 Abnormal spermatogenesis in RXR mutant mice. Genes Dev 10:80–92

    Akmal KM, Dufour JM, Kim KH 1997 Retinoic acid receptor- gene expression in the rat testis: potential role during the prophase of meiosis and in the transition from round to elongating spermatids. Biol Reprod 56:549–556

    Gaemers IC, van Pelt AM, van der Saag PT, Hoogerbrugge JW, Themmen AP, de Rooij DG 1998 Differential expression pattern of retinoid X receptors in adult murine testicular cells implies varying roles for these receptors in spermatogenesis. Biol Reprod 58:1351–1356

    Dufour JM, Kim KH 1999 Cellular and subcellular localization of six retinoid receptors in rat testis during postnatal development: identification of potential heterodimeric receptors. Biol Reprod 61:1300–1308

    Boulogne B, Levacher C, Durand P, Habert R 1999 Retinoic acid receptors and retinoid X receptors in the rat testis during fetal and postnatal development: immunolocalization and implication in the control of the number of gonocytes. Biol Reprod 61:1548–1557

    Chapellier B, Mark M, Garnier JM, LeMeur M, Chambon P, Ghyselinck NB 2002 A conditional floxed (loxP-flanked) allele for the retinoic acid receptor- (RAR) gene. Genesis 32:87–90

    Dufour JM, Vo MN, Bhattacharya N, Okita J, Okita R, Kim KH 2003 Peroxisome proliferators disrupt retinoic acid receptor- signalling in the testis. Biol Reprod 68:1215–1224

    Saper CB, Sawchenko PE 2003 Magic peptides, magic antibodies: guidelines for appropriate controls for immunohistochemistry. J Comp Neurol 465:161–163

    Russell LD, Ettlin RA, Sinha Hikim AP, Clegg ED 1990 Histological and histopathological evaluation of the testis. Clearwater, FL: Cache River Press

    Decimo D, Georges-Labouesse E, Dolle P 1995 In situ hybridization of nucleic acid probes to cellular RNA. In: Hames BD, Higgins S, eds. Gene probes 2: a practical approach. Oxford, UK: Oxford University Press; 183–210

    De Block M, Debrouwer D 1993 RNA-RNA in situ hybridization using digoxigenin-labeled probes: the use of high-molecular-weight polyvinyl alcohol in the alkaline phosphatase indoxyl-nitroblue tetrazolium reaction. Anal Biochem 215:86–89

    Lufkin T, Lohnes D, Mark M, Dierich A, Gorry P, Gaub MP, LeMeur M, Chambon P 1993 High postnatal lethality and testis degeneration in retinoic acid receptor- mutant mice. Proc Natl Acad Sci USA 90:7225–7229

    Schnütgen F, Doerflinger N, Calleja C, Wendling O, Chambon P, Ghyselinck NB 2003 A directional strategy for monitoring Cre-mediated recombination at the cellular level in the mouse. Nat Biotechnol 21:562–565

    Chapellier B, Mark M, Messaddeq N, Calleja C, Warot X, Brocard J, Gerard C, Li M, Metzger D, Ghyselinck NB, Chambon P 2002 Physiological and retinoid-induced proliferations of epidermis basal keratinocytes are differently controlled. EMBO J 21:3402–3413

    Ghyselinck NB, Dupe V, Dierich A, Messaddeq N, Garnier JM, Rochette-Egly C, Chambon P, Mark M 1997 Role of the retinoic acid receptor- (RAR) during mouse development. Int J Dev Biol 41:425–447

    Krezel W, Kastner P, Chambon P 1999 Differential expression of retinoid receptors in the adult mouse central nervous system. Neuroscience 89:1291–1300

    Bellve AR, Cavicchia JC, Millette CF, O’Brien DA, Bhatnagar YM, Dym M 1977 Spermatogenic cells of the prepubertal mouse. Isolation and morphological characterization. J Cell Biol 74:68–85

    Wake K, Motomatsu K, Senoo H, Masuda A, Adachi E 1986 Improved Kupffer’s gold chloride method for demonstrating the stellate cells storing retinol (vitamin A) in the liver and extrahepatic organs of vertebrates. Stain Technol 61:193–200

    Ruiz A, Winston A, Lim YH, Gilbert BA, Rando RR, Bok D 1999 Molecular and biochemical characterization of lecithin retinol acyltransferase. J Biol Chem 274:3834–3841

    Yost RW, Harrison EH, Ross AC 1988 Esterification by rat liver microsomes of retinol bound to cellular retinol-binding protein. J Biol Chem 263:18693–18701

    Herr FM, Ong DE 1992 Differential interaction of lecithin-retinol acyltransferase with cellular retinol binding proteins. Biochemistry 31:6748–6755

    Schmitt MC, Ong DE 1993 Expression of cellular retinol-binding protein and lecithin-retinol acyltransferase in developing rat testis. Biol Reprod 49:972–979

    Londos C, Brasaemle DL, Schultz CJ, Segrest JP, Kimmel AR 1999 Perilipins, ADRP, and other proteins that associate with intracellular neutral lipid droplets in animal cells. Semin Cell Dev Biol 10:51–58

    Mascrez B, Ghyselinck NB, Watanabe M, Annicotte JS, Chambon P, Auwerx J, Mark M 2004 Ligand-dependent contribution of RXR to cholesterol homeostasis in Sertoli cells. EMBO Rep 5:285–290

    Imamura M, Inoguchi T, Ikuyama S, Taniguchi S, Kobayashi K, Nakashima N, Nawata H 2002 ADRP stimulates lipid accumulation and lipid droplet formation in murine fibroblasts. Am J Physiol Endocrinol Metab 283:775–783

    Krezel W, Dupe V, Mark M, Dierich A, Kastner P, Chambon P 1996 RXR null mice are apparently normal and compound RX+/–/RXR–/–/RXR–/– mutant mice are viable. Proc Natl Acad Sci USA 93:9010–9014

    Fan X, Molotkov A, Manabe S, Donmoyer CM, Deltour L, Foglio MH, Cuenca AE, Blaner WS, Lipton SA, Duester G 2003 Targeted disruption of Aldh1a1 (Raldh1) provides evidence for a complex mechanism of retinoic acid synthesis in the developing retina. Mol Cell Biol 23:4637–4648

    Dupe V, Matt N, Garnier JM, Chambon P, Mark M, Ghyselinck NB 2003 A newborn lethal defect due to inactivation of retinaldehyde dehydrogenase type 3 is prevented by maternal retinoic acid treatment. Proc Natl Acad Sci USA 100:14036–14041

    Lohnes D, Kastner P, Dierich A, Mark M, LeMeur M, Chambon P 1993 Function of retinoic acid receptor- in the mouse. Cell 73:643–658

    Eddy EM, Washburn TF, Bunch DO, Goulding EH, Gladen BC, Lubahn DB, Korach KS 1996 Targeted disruption of the estrogen receptor gene in male mice causes alteration of spermatogenesis and infertility. Endocrinology 137:4796–4805

    Yashiro K, Zhao X, Uehara M, Yamashita K, Nishijima M, Nishino J, Saijoh Y, Sakai Y, Hamada H 2004 Regulation of retinoic acid distribution is required for proximodistal patterning and outgrowth of the developing mouse limb. Dev Cell 6:411–422

    Chung SS, Wang X, Wolgemuth DJ 2005 Male sterility in mice lacking retinoic acid receptor- involves specific abnormalities in spermiogenesis. Differentiation 73:188–198

    Schultz R, Penttila TL, Parvinen M, Persson H, Hokfelt T, Pelto-Huikko M 1995 Expression of immediate early genes in tubular cells of rat testis. Biol Reprod 52:1215–1226

    Metzger D, Chambon P 2001 Site- and time-specific gene targeting in the mouse. Methods 24:71–80

    Lecureuil C, Fontaine I, Crepieux P, Guillou F 2002 Sertoli and granulosa cell-specific Cre recombinase activity in transgenic mice. Genesis 33:114–118

    Weber P, Cammas F, Gerard C, Metzger D, Chambon P, Losson R, Mark M 2002 Germ cell expression of the transcriptional co-repressor TIF1 is required for the maintenance of spermatogenesis in the mouse. Development 129:2329–2337

    Haselbeck RJ, Hoffmann I, Duester G 1999 Distinct functions for Aldh1 and Raldh2 in the control of ligand production for embryonic retinoid signalling pathways. Dev Genet 25:353–364

    Gaemers IC, van Pelt AM, van der Saag PT, de Rooij DG 1996 All-trans-4-oxo-retinoic acid: a potent inducer of in vivo proliferation of growth-arrested A spermatogonia in the vitamin A-deficient mouse testis. Endocrinology 137:479–485

    Yamanushi T, Igarashi O 1995 The mobilization and tissue distribution of -carotene in the rat by the venous injection method. J Nutr Sci Vitaminol (Tokyo) 41:169–177

    Abu-Abed S, MacLean G, Fraulob V, Chambon P, Petkovich M, Dolle P 2002 Differential expression of the retinoic acid-metabolizing enzymes CYP26A1 and CYP26B1 during murine organogenesis. Mech Dev 110:173–177

    Rajguru SU, Kang YH, Ahluwalia BS 1982 Localization of retinol (vitamin A) in rat testes. J Nutr 112:1881–1891

    Shingleton JL, Skinner MK, Ong DE 1989 Retinol esterification in Sertoli cells by lecithin-retinol acyltransferase. Biochemistry 28:9647–9653

    Batten ML, Imanishi Y, Maeda T, Tu DC, Moise AR, Bronson D, Possin D, Van Gelder RN, Baehr W, Palczewski K 2004 Lecithin-retinol acyltransferase is essential for accumulation of all-trans-retinyl esters in the eye and in the liver. J Biol Chem 279:10422–10432

    Chaudhary LR, Nelson EC 1987 Some properties and subcellular distribution of acyl-coenzyme A: retinol acyltransferase activity in rat testes. Biochim Biophys Acta 917:24–32(Nadege Vernet, Christine Dennefeld, Ceci)