Sequencing of a 'mouse azoospermia' gene panel in azoospermic men: identification of RNF212 and STAG3 mutations as novel genetic causes of meiotic arrest.

STUDY QUESTION: What is the diagnostic potential of next generation sequencing (NGS) based on a 'mouse azoospermia' gene panel in human non-obstructive azoospermia (NOA)? SUMMARY ANSWER: The diagnostic performance of sequencing a gene panel based on genes associated with mouse azoospermia was relatively successful in idiopathic NOA patients and allowed the discovery of two novel genes involved in NOA due to meiotic arrest. WHAT IS KNOWN ALREADY: NOA is a largely heterogeneous clinical entity, which includes different histological pictures. In a large proportion of NOA, the aetiology remains unknown (idiopathic NOA) and yet, unknown genetic factors are likely to play be involved. The mouse is the most broadly used mammalian model for studying human disease because of its usefulness for genetic manipulation and its genetic and physiological similarities to man. Mouse azoospermia models are available in the Mouse Genome Informatics database (MGI: http://www.informatics.jax.org/). STUDY DESIGN, SIZE, DURATION: The first step was to design of a 'mouse azoospermia' gene panel through the consultation of MGI. The second step was NGS analysis of 175 genes in a group of highly selected NOA patients (n = 33). The third step was characterization of the discovered gene defects in human testis tissue, through meiotic studies using surplus testicular biopsy material from the carriers of the RNF212 and STAG3 pathogenic variants. The final step was RNF212 and STAG3 expression analysis in a collection of testis biopsies. PARTICIPANTS/MATERIALS, SETTING, METHODS: From a total of 1300 infertile patients, 33 idiopathic NOA patients were analysed in this study, including 31 unrelated men and 2 brothers from a consanguineous family. The testis histology of the 31 unrelated NOA patients was as follows: 20 Sertoli cell-only syndrome (SCOS), 11 spermatogenic arrest (6 spermatogonial arrest and 5 spermatocytic arrest). The two brothers were affected by spermatocytic arrest. DNA extracted from blood was used for NGS on Illumina NextSeq500 platform. Generated sequence data was filtered for rare and potentially pathogenic variants. Functional studies in surplus testicular tissue from the carriers included the investigation of meiotic entry, XY body formation and metaphases by performing fluorescent immunohistochemical staining and immunocytochemistry. mRNA expression analysis through RT-qPCR of RNF212 and STAG3 was carried out in a collection of testis biopsies with different histology. MAIN RESULTS AND THE ROLE OF CHANCE: Our approach was relatively successful, leading to the genetic diagnosis of one sporadic NOA patient and two NOA brothers. This relatively high diagnostic performance is likely to be related to the stringent patient selection criteria i.e. all known causes of azoospermia were excluded and to the relatively high number of patients with rare testis histology (spermatocytic arrest). All three mutation carriers presented meiotic arrest, leading to the genetic diagnosis of three out of seven cases with this specific testicular phenotype. For the first time, we report biallelic variants in STAG3, in one sporadic patient, and a homozygous RNF212 variant, in the two brothers, as the genetic cause of NOA. Meiotic studies allowed the detection of the functional consequences of the mutations and provided information on the role of STAG3 and RNF212 in human male meiosis. LIMITATIONS, REASONS FOR CAUTION: All genes, with the exception of 5 out of 175, included in the panel cause azoospermia in mice only in the homozygous or hemizygous state. Consequently, apart from the five known dominant genes, heterozygous variants (except compound heterozygosity) in the remaining genes were not taken into consideration as causes of NOA. We identified the genetic cause in approximately half of the patients with spermatocytic arrest. The low number of analysed patients can be considered as a limitation, but it is a very rare testis phenotype. Due to the low frequency of this specific phenotype among infertile men, our finding may be considered of low clinical impact. However, at an individual level, it does have relevance for prognostic purposes prior testicular sperm extraction. WIDER IMPLICATIONS OF THE FINDINGS: Our study represents an additional step towards elucidating the genetic bases of early spermatogenic failure, since we discovered two new genes involved in human male meiotic arrest. We propose the inclusion of RNF212 and STAG3 in a future male infertility diagnostic gene panel. Based on the associated testis phenotype, the identification of pathogenic mutations in these genes also confers a negative predictive value for testicular sperm retrieval. Our meiotic studies provide novel insights into the role of these proteins in human male meiosis. Mutations in STAG3 were first described as a cause of female infertility and ovarian cancer, and Rnf212 knock out in mice leads to male and female infertility. Hence, our results stimulate further research on shared genetic factors causing infertility in both sexes and indicate that genetic counselling should involve not only male but also female relatives of NOA patients. STUDY FUNDING/COMPETING INTEREST(S): This work was funded by the Spanish Ministry of Health Instituto Carlos III-FIS (grant number: FIS/FEDER-PI14/01250; PI17/01822) awarded to CK and AR-E, and by the European Commission, Reproductive Biology Early Research Training (REPROTRAIN, EU-FP7-PEOPLE-2011-ITN289880), awarded to CK, WB, and AE-M. The authors have no conflict of interest.

Evolution of testis-specific kinases TSSK1B and TSSK2 in primates.

The testis-specific serine/threonine protein kinases TSSK1 and TSSK2 are known to be essential for male fertility, in mice. The enzymes are present in elongating spermatids, and targeted deletion of the two genes Tssk1 and Tssk2 results in dysregulation of spermiogenesis. The mouse genes are genetically closely linked, forming a Tssk1-Tssk2 tandem. In human, TSSK1 is present in the form of a pseudogene, TSSK1A, which is linked to an intact TSSK2 gene, and in the form of an intact gene, TSSK1B, which is not genetically linked to TSSK2. Studies on conservation of genes and gene function between mouse and human are relevant, to be able to use mouse models for studies on human infertility, and to evaluate possible targets for non-hormonal contraception targeting the male. Therefore, we have performed a detailed analysis of the evolution of genes encoding TSSK1 and TSSK2 among mammals, in particular among primates. This study includes functional analysis of replacement mutation K27R in TSSK2, which is frequently observed among humans. In primates, the kinase domains of TSSK1B and TSSK2 have evolved under negative selection, reflecting the importance to maintain their kinase activity. Positive selection was observed for the C-terminal domain of TSSK1B, which indicates that TSSK1B and TSSK2 may perform at least partly differential functions.

Mutations that affect meiosis in male mice influence the dynamics of the mid-preleptotene and bouquet stages.

Meiosis pairs and segregates homologous chromosomes and thereby forms haploid germ cells to compensate the genome doubling at fertilization. Homologue pairing in many eukaryotic species depends on formation of DNA double strand breaks (DSBs) during early prophase I when telomeres begin to cluster at the nuclear periphery (bouquet stage). By fluorescence in situ hybridization criteria, we observe that mid-preleptotene and bouquet stage frequencies are altered in male mice deficient for proteins required for recombination, ubiquitin conjugation and telomere length control. The generally low frequencies of mid-preleptotene spermatocytes were significantly increased in male mice lacking recombination proteins SPO11, MEI1, MLH1, KU80, ubiquitin conjugating enzyme HR6B, and in mice with only one copy of the telomere length regulator Terf1. The bouquet stage was significantly enriched in Atm(-/-), Spo11(-/-), Mei1(m1Jcs/m1Jcs), Mlh1(-/-), Terf1(+/-) and Hr6b(-/-) spermatogenesis, but not in mice lacking recombination proteins DMC1 and HOP2, the non-homologous end-joining DNA repair factor KU80 and the ATM downstream effector GADD45a. Mice defective in spermiogenesis (Tnp1(-/-), Gmcl1(-/-), Asm(-/-)) showed wild-type mid-preleptotene and bouquet frequencies. A low frequency of bouquet spermatocytes in Spo11(-/-)Atm(-/-) spermatogenesis suggests that DSBs contribute to the Atm(-/-)-correlated bouquet stage exit defect. Insignificant changes of bouquet frequencies in mice with defects in early stages of DSB repair (Dmc1(-/-), Hop2(-/-)) suggest that there is an ATM-specific influence on bouquet stage duration. Altogether, it appears that several pathways influence telomere dynamics in mammalian meiosis.

Chromatin dynamics in the male meiotic prophase.

During the male meiotic prophase in mouse and man, pairing and recombination of homologous chromosomes is accompanied by changes in chromatin structure. In this review, the dynamics of assembly and disassembly of the chromatin-associated complexes that mediate sister chromatid cohesion (cohesin) and maintain chromosome pairing (the synaptonemal complex) are described. Special features of the meiotic S phase are discussed, and also the dynamics of several key players that act together after the S phase at sites of meiotic double-strand break DNA repair. Current knowledge on histone modifications that occur during the male meiotic prophase is discussed, with special attention for the inactive chromatin of the X and Y chromosomes that constitutes the sex body. Finally, it is discussed that in the future, it will be possible to view the true chromatin dynamics during male meiosis in time, in living cells, through analysis of fluorescent-tagged proteins expressed in transgenic mice, using advanced fluorescent microscopy techniques.

DNA repair mechanisms and gametogenesis.

In mammals, there is a complex and intriguing relationship between DNA repair and gametogenesis. DNA repair mechanisms are involved not only in the repair of different types of DNA damage in developing germline cells, but also take part in the meiotic recombination process. Furthermore, the DNA repair mechanisms should tolerate mutations occurring during gametogenesis, to a limited extent. In the present review, several gametogenic aspects of DNA mismatch repair, homologous recombination repair and postreplication repair are discussed. In addition, the role of DNA damage-induced cell cycle checkpoint control is considered briefly. It appears that many genes encoding proteins that take part in DNA repair mechanisms show enhanced or specialized expression during mammalian gametogenesis, and several gene knockout mouse models show male or female infertility. On the basis of such knowledge and models, future experiments may provide more information about the precise relationship between DNA repair, chromatin dynamics, and genomic stability versus instability during gametogenesis.

Molecular and cellular mechanisms in spermatogenesis.

Mammalian spermatogenesis shows a strict control of many specific molecular and cellular events. This control involves Sertoli cell-germ cell interaction, as well as a programmed performance of changes in chromatin structure and gene expression in the developing germ cells. In recent years, much knowledge about the functions of defined genes in spermatogenesis has been gained by making use of mouse transgenic and gene knockout models. Several of these models are discussed in this brief overview, with an emphasis on genes encoding proteins involved in the control of gene transcription, mRNA translation, DNA repair and protein ubiquitination. A better understanding of the molecular and cellular biology of spermatogenesis in the mouse may provide concepts that can improve our understanding of human male infertility and may also lead to the identification of novel targets for contraceptive intervention.

Specific aspects of the ubiquitin system in spermatogenesis.

The ubiquitin system is involved in numerous cellular processes, regulating the amounts and/or activities of specific proteins through posttranslational coupling with ubiquitin or ubiquitin-like proteins. In spermatogenesis, there appears to be a special requirement for certain components of the ubiquitin system, as exemplified in human and mouse by mutation of USP9Y and HR6B, respectively. Both genes encode proteins which take part in the ubiquitin system and are ubiquitously expressed, but their mutation generates no apparent phenotype other than male infertility. Different phases of mammalian spermatogenesis probably require different specialized activities of the ubiquitin system. It is anticipated that ubiquitination activities similar to those required during mitotic cell cycle regulation will play some role in control of the meiotic divisions. In spermatocytes, there is an intricate link among DNA repair, the ubiquitin system, and regulation of meiotic chromatin structure, as indicated by the co-localization of proteins involved in these processes on meiotic recombination complexes. HR6B and its nearly identical homolog HR6A are multiple function proteins, with ubiquitin-conjugating activity and essential roles in post-replication DNA repair. HR6B, possibly together with the ubiquitin-ligating enzyme mRAD1 8Sc, is most likely involved in chromatin re-organization during the meiotic and post-meiotic phases of spermatogenesis. Biochemical data indicate that, in particular during spermiogenesis, the general activity of the ubiquitin system is high, which most likely is related to the high requirement for massive breakdown of cytoplasmatic and nuclear proteins during this last phase of spermatogenesis.

Characterization of mRAD18Sc, a mouse homolog of the yeast postreplication repair gene RAD18.

The RAD18 gene of the yeast Saccharomyces cerevisiae encodes a protein with ssDNA binding activity that interacts with the ubiquitin-conjugating enzyme RAD6 and plays an important role in postreplication repair. We identified and characterized the putative mouse homolog of RAD18, designated mRAD18Sc. The mRAD18Sc open reading frame encodes a 509-amino-acid polypeptide that is strongly conserved in size and sequence between yeast and mammals, with specific conservation of the RING-zinc-finger and the classic zinc-finger domain. The degree of sequence conservation between mRAD18Sc, RAD18, and homologous sequences identified in other species (NuvA from Aspergillus nidulans and Uvs-2 from Neurospora crassa) is entirely consistent with the evolutionary relationship of these organisms, strongly arguing that these genes are one another's homologs. Consistent with the presence of a nuclear translocation signal in the amino acid sequence, we observed the nuclear localization of GFP-tagged mRAD18Sc after stable transfection to HeLa cells. mRNA expression of mRAD18Sc in the mouse was observed in thymus, spleen, brain, and ovary, but was most pronounced in testis, with the highest level of expression in pachytene-stage primary spermatocytes, suggesting that mRAD18Sc plays a role in meiosis of spermatogenesis. Finally, we mapped the mRAD18Sc gene on mouse chromosome 6F.

The ubiquitin system in gametogenesis.

Ubiquitin is a ubiquitous and highly conserved protein of 76 amino acid residues, that can be covalently attached to cellular acceptor proteins. The attachment of ubiquitin to target proteins is achieved through a multi-step enzymatic pathway, which involves activities of ubiquitin-activating E1 enzymes, ubiquitin-conjugating E2 enzymes, and ligating E3 enzymes. Mono- or poly-ubiquitination of proteins can lead to protein degradation or modification of protein activity. Many components of the complex ubiquitin system show remarkable evolutionary conservation, from yeast to mammalian species. The ubiquitin system is essential to all eukaryotic cells. Among others, several signal transduction cascades show involvement of the ubiquitin system, but there are currently little data supporting a specific role of the ubiquitin system in hormonal control of reproduction. Interestingly, during gametogenesis, many specialized and important aspects of the ubiquitin system become apparent. Components of the ubiquitin system appear to be involved in different steps and processes during gametogenesis, including control of meiosis, and reorganization of chromatin structure.

Follicle-stimulating hormone and testosterone stimulation of immature and mature Sertoli cells in vitro: inhibin and N-cadherin levels and round spermatid binding.

The in vitro response of Sertoli cells isolated from adult rat testes to testosterone (T) and follicle-stimulating hormone (FSH) treatment was investigated. Sertoli cells from >70-day-old Sprague-Dawley rats were isolated by a combined enzymatic treatment followed by the removal of the majority of contaminating germ cells with immobilized peanut agglutinin lectin. Sertoli cells were then cultured for 6-10 days, forming a confluent layer with a cell viability of >83% and 74-77% purity. The contaminating cells were peritubular cells (4-6%), pachytene spermatocytes (4-5%), round spermatids (<2%), elongated spermatids (<1%), and degenerating germ cells (14.8%). The proportion of degenerating germ cells decreased from 14.8% to 8.6% between days 6 and 10 in culture. After a prestimulation culture period of 4 days, FSH treatment over a 2-day period resulted in a dose-related increase of inhibin with a median effective dose (ED50) value of 36.7+/-20.4 ng/ml in comparison with an ED50 value of 4.4+/-0.9 ng/ml obtained with immature Sertoli cell cultures from 20-day-old rats. Mature Sertoli cells, in contrast to immature Sertoli cells, were unresponsive to combined FSH + T treatment for the production of the cell adhesion protein N-cadherin. FSH treatment promoted the in vitro binding of round spermatids isolated from adult testis to adult Sertoli cells in coculture. It is concluded that mature Sertoli cells in culture are responsive to FSH in terms of inhibin production and round-spermatid binding. The lack of an FSH + T-induced increase in N-cadherin or round spermatid binding is attributed to either a reduced sensitivity, or an alteration in the regulation of mature Sertoli cells by FSH + T.

Histone ubiquitination and chromatin remodeling in mouse spermatogenesis.

Male infertility in HR6B knockout mice is associated with impairment of spermatogenesis. The HR6B gene is a mammalian, autosomal homolog of the Saccharomyces cerevisiae gene Rad6 encoding a ubiquitin-conjugating enzyme. In addition, X-chromosomal HR6A has been identified, in human and mouse. RAD6 in yeast is required for a variety of cellular functions, including sporulation, DNA repair, and mutagenesis. Since RAD6 and its mammalian homologs can ubiquitinate histones in vitro, we have investigated the pattern of histone ubiquitination in mouse testis. By immunoblot and immunohistochemical analysis of wild-type mouse testis, a high amount of ubiquitinated H2A (uH2A) was detected in pachytene spermatocytes. This signal became undetectable in round spermatids, but then increased again during a relatively short developmental period, in elongating spermatids. No other ubiquitinated histones were observed. In the HR6B knockout mice, we failed to detect an overt defect in the overall pattern of histone ubiquitination. For somatic cell types, it has been shown that histone ubiquitination is associated with destabilization of nucleosomes, in relation to active gene transcription. Unexpectedly, the most intense uH2A signal in pachytene spermatocytes was detected in the sex body, an inactive nuclear structure that contains the heterochromatic X and Y chromosomes. The postmeiotic uH2A immunoexpression in elongating spermatids indicates that nucleosome destabilization induced by histone ubiquitination may play a facilitating role during histone-to-protamine replacement.

Targeted inactivation of mouse RAD52 reduces homologous recombination but not resistance to ionizing radiation.

The RAD52 epistasis group is required for recombinational repair of double-strand breaks (DSBs) and shows strong evolutionary conservation. In Saccharomyces cerevisiae, RAD52 is one of the key members in this pathway. Strains with mutations in this gene show strong hypersensitivity to DNA-damaging agents and defects in recombination. Inactivation of the mouse homologue of RAD52 in embryonic stem (ES) cells resulted in a reduced frequency of homologous recombination. Unlike the yeast Scrad52 mutant, MmRAD52(-/-) ES cells were not hypersensitive to agents that induce DSBs. MmRAD52 null mutant mice showed no abnormalities in viability, fertility, and the immune system. These results show that, as in S. cerevisiae, MmRAD52 is involved in recombination, although the repair of DNA damage is not affected upon inactivation, indicating that MmRAD52 may be involved in certain types of DSB repair processes and not in others. The effect of inactivating MmRAD52 suggests the presence of genes functionally related to MmRAD52, which can partly compensate for the absence of MmRad52 protein.

Immunoexpression of testis-specific histone 2B in human spermatozoa and testis tissue.

During mammalian spermatogenesis, the chromatin of the spermatogenic cells is profoundly reorganized. Somatic histones are partly replaced by testis-specific histones. These histones are then replaced by transition proteins and finally by protamines. This series of nucleoprotein rearrangements results in a highly condensed sperm cell nucleus. In contrast to spermatozoa from other species, human spermatozoa still contain a significant amount of histones, including testis-specific histone 2B (TH2B). In the present study it is shown that an antibody targeting tyrosine hydroxylase, which has been found previously to cross-react with rat TH2B, also specifically immunoreacts with human TH2B on Western blots, in immunohistochemistry of human testis tissue, and in immunocytochemistry of decondensed human spermatozoa. In human testis tissue, TH2B immunostaining first apparent in spermatogonia, shows marked variation, especially at the pachytene spermatocyte stage, and then reaches an intense signal in round spermatids. Shortly before spermatid elongation, a portion of the spermatid nucleus, corresponding to the acrosomal region, loses its immunoreactivity. During condensation of the spermatid nucleus, the immunodetectability of TH2B disappears gradually, from the anterior region of the nucleus onwards. At the final stages of spermiogenesis, the immunostaining is completely absent. Immunocytochemical staining of spermatozoa revealed no TH2B immunosignal, but immunostaining was observed when spermatozoa obtained from semen were decondensed to make nuclear proteins accessible to the antibody. There was, however, a striking intercellular variability in the intensity of staining of spermatozoa within an ejaculate. In a population of 35 men attending our Andrology Clinic, we observed interindividual differences in total sperm TH2B content, which showed a significant, although not very pronounced, negative correlation with normal morphology (P = 0.05).

Knockout mouse model and gametogenic failure.

To evaluate the function of a defined gene in gametogenesis, exciting opportunities are offered by the introduction of techniques to generate knockout mice. In this short article, we briefly describe a few gene knockout mouse models, which show a phenotype that involves impairment of gametogenesis and/or fertility. The focus will be on the mHR6B gene knockout mouse, which shows male infertility. The mHR6B gene encodes an ubiquitin-conjugating enzyme, and the data point to an important role of the ubiquitin pathway in gametogenesis.

Testis-specific expression of a functional retroposon encoding glucose-6-phosphate dehydrogenase in the mouse.

The X-chromosomal gene glucose-6-phosphate dehydrogenase (G6pd) is known to be expressed in most cell types of mammalian species. In the mouse, we have detected a novel gene, designated G6pd-2, encoding a G6PD isoenzyme. G6pd-2 does not contain introns and appears to represent a retroposed gene. This gene is uniquely transcribed in postmeiotic spermatogenic cells in which the X-encoded G6pd gene is not transcribed. Expression of the G6pd-2 sequence in a bacterial system showed that the encoded product is an active enzyme. Zymogramic analysis demonstrated that recombinant G6PD-2, but not recombinant G6PD-1 (the X-chromosome-encoded G6PD), formed tetramers under reducing conditions. Under the same conditions, G6PD tetramers were also found in extracts of spermatids and spermatozoa, indicating the presence of G6pd-2-encoded isoenzyme in these cell types. G6pd-2 is one of the very few known expressed retroposons encoding a functional protein, and the presence of this gene is probably related to X chromosome inactivation during spermatogenesis.

Inactivation of the HR6B ubiquitin-conjugating DNA repair enzyme in mice causes male sterility associated with chromatin modification.

The ubiquitin-conjugating yeast enzyme RAD6 and its human homologs hHR6A and hHR6B are implicated in postreplication repair and damage-induced mutagenesis. The yeast protein is also required for sporulation and may modulate chromatin structure via histone ubiquitination. We report the phenotype of the first animal mutant in the ubiquitin pathway: inactivation of the hHR6B-homologous gene in mice causes male infertility. Derailment of spermatogenesis becomes overt during the postmeiotic condensation of chromatin in spermatids. These findings provide a parallel between yeast sporulation and mammalian spermatogenesis and strongly implicate hHR6-dependent ubiquitination in chromatin remodeling. Since heterozygous male mice and even knockout female mice are completely normal and fertile and thus able to transmit the defect, similar hHR6B mutations may cause male infertility in man.

Anti-müllerian hormone and anti-müllerian hormone type II receptor messenger ribonucleic acid expression during postnatal testis development and in the adult testis of the rat.

Anti-müllerian hormone (AMH) induces degeneration of the müllerian ducts during male sex differentiation and may have additional functions concerning gonadal development. In the immature rat testis, there is a marked developmental increase in AMH type II receptor (AMHRII) messenger RNA (mRNA) expression in Sertoli cells, concomitant with the initiation of spermatogenesis. AMHRII mRNA is also expressed at a high level in Sertoli cells in adult rats. To obtain information about the possible functions of AMH in the testis, we investigated the postnatal expression patterns of the genes encoding AMH and AMHRII in the rat testis in more detail. Using RNase protection assays, AMH and AMHRII mRNA expression was measured in total RNA preparations from testes or testicular tubule segments isolated from control rats and from rats that had received various treatments. The testicular level of AMHRII mRNA was found to be much higher than that of AMH mRNA in adult rats. AMH mRNA was detected at a maximal level at stage VII of the spermatogenic cycle and at a low level at the other stages. AMHRII mRNA increases from stage XIII, is highest at stages VI and VII, and then rapidly declines at stage VIII to almost undetectable levels at stages IX-XII. It was found that the increase in testicular AMHRII mRNA expression during the first 3 weeks of postnatal development also occurs in sterile rats (prenatally irradiated), and hence, is independent of the presence or absence of germ cells. Yet, the total testicular level of AMHRII mRNA was decreased in sterile adult rats (prenatally irradiated or experimental cryptorchidism), as compared with intact control rats. However, treatment of adult rats with methoxyacetic acid or hydroxyurea, which resulted in partial germ cell depletion, had no effect on total testicular AMHRII mRNA expression. We conclude that a combination of multiple spermatogenic cycle events, possibly involving changes of Sertoli cell structure and/or Sertoli cell-basal membrane interactions, regulate autocrine AMH action on Sertoli cells, in particular at stage VII of the spermatogenic cycle.

Anti-müllerian hormone and anti-müllerian hormone type II receptor messenger ribonucleic acid expression in rat ovaries during postnatal development, the estrous cycle, and gonadotropin-induced follicle growth.

During fetal development, anti-müllerian hormone (AMH) is produced only by Sertoli cells, but postnatally, granulosa cells also produce this peptide growth/differentiation factor. We recently identified a candidate AMH type II receptor (AMHRII). In the present study, postnatal ovarian AMH and AMHRII messenger RNA (mRNA) expression was studied by in situ hybridization and ribonuclease protection. In ovaries from adult rats, AMH and AMHRII mRNAs were found to be mainly expressed in granulosa cells from preantral and small antral follicles. Corpora lutea and large antral follicles express little or no AMH and AMHRII mRNA, and primordial follicles and oocytes appeared to be AMH and AMHRII mRNA negative. Thecal and interstitial cells express no detectable AMH mRNA and little or no AMHRII mRNA. The colocalization of AMH and AMHRII mRNAs in granulosa cells of specific follicle types suggests that actions of AMH via AMHRII are autocrine in nature. There is a decreased level of AMH and AMHRII mRNA expression when follicles become atretic. Both mRNA species are eventually lost from atretic follicles, although AMHRII mRNA expression seems to persist somewhat longer than AMH mRNA. During the estrous cycle, no marked changes in the patterns of AMH and AMHRII mRNA expression were detected, except at estrus, when expression of both mRNA species in preantral follicles was decreased compared to that on the other days of the cycle. On postnatal day 5, total ovarian AMH mRNA expression is low and is located in small preantral follicles. During the first weeks of postnatal development, AMH mRNA expression in preantral follicles increases, and the later formed small antral follicles also express AMH mRNA. In contrast, AMHRII mRNA is expressed on postnatal day 5 at a higher level than AMH mRNA, but cannot be localized to specific cell types. From postnatal day 15 onward, AMHRII mRNA expression becomes more restricted to the preantral and small antral follicles. Treatment of prepubertal rats with GnRH antagonist (Org 30276) and human recombinant FSH (Org 32489) or with GnRH antagonist and estradiol benzoate resulted in follicle growth and inhibition of AMH and AMHRII mRNA expression in some, but not all, preantral and small antral follicles. These results indicate that FSH and estrogens may play a role in the down-regulation of AMH and AMHRII mRNA expression in vivo when small antral follicles differentiate into large antral follicles. Furthermore, the FSH surge on the morning of estrus may inhibit AMH and AMHRII mRNA expression in preantral follicles.(ABSTRACT TRUNCATED AT 400 WORDS)