Modulation of the mouse testis transcriptome during postnatal development and in selected models of male infertility.

The aim of this study is to develop an overview of genetic events during spermatogenesis using a novel, specifically targeted gonadal gene set. Two subtracted cDNA libraries enriched for testis specific and germ cell specific genes were constructed, characterized and sequenced. The combined libraries contain >1905 different genes, the vast majority previously uncharacterized in testis. cDNA microarray analysis of the first wave of murine spermatogenesis and of selected germ cell-deficient models was used to correlate the expression of groups of genes with the appearance of defined germ cell types, suggesting their cellular expression patterns within the testis. Real-time RT-PCR and comparison to previously known expression patterns confirmed the array-derived transcription profiles of 65 different genes, thus establishing high confidence in the profiles of the uncharacterized genes investigated in this study. A total of 1748 out of 1905 genes showed significant change during the first spermatogenic wave, demonstrating the successful targeting of the libraries to this process. These findings highlight unknown genes likely to be important in germ cell production, and demonstrate the utility of these libraries in further studies. Transcriptional analysis of well-characterized mouse models of infertility will allow us to address the causes and progression of the pathology in related human infertility phenotypes.

Absence of mDazl produces a final block on germ cell development at meiosis.

The autosomal gene DAZL is a member of a family of genes (DAZL, DAZ, BOULE), all of which contain a consensus RNA binding domain and are expressed in germ cells. Adult male and female mice null for Dazl lack gametes. In order to define more precisely the developmental stages in germ cells that require Dazl expression, the patterns of germ cell loss in immature male and female wild-type (+/+, WT) and Dazl -/- (DazlKO) mice were analysed. In females, loss of germ cells occurred during fetal life and was coincident with progression of cells through meiotic prophase. In males, testes were recovered from WT and DazlKO males obtained before and during the first wave of spermatogenesis (days 2-19). Mitotically active germ cells were present up to and including day 19. Functional differentiation of spermatogonia associated with detection of c-kit positive cells did not depend upon expression of Dazl. RBMY-positive cells (A, intermediate, B spermatogonia, zygotene and preleptotene spermatocytes) were reduced in DazlKO compared with WT testes. Staining of cell squashes from day 19 testes with anti-gamma-H2AX and anti-SCP3 antibodies showed that germ cells from DazlKO males were unable to progress beyond the leptotene stage of meiotic prophase I. It was concluded that in the absence of Dazl, germ cells can complete mitosis, and embark on functional differentiation but that, in both sexes, progression through meiotic prophase requires this RNA binding protein.

Does Rbmy have a role in sperm development in mice?

The Y(d1) deletion in mice removes most of the multi-copy Rbmy gene cluster that is located adjacent to the centromere on the Y short arm (Yp). XY(d1) mice develop as females because Sry is inactivated, probably because it is now juxtaposed to centromeric heterochromatin. We have previously produced XY(d1)Sry transgenic males and found that they have a substantially increased frequency of abnormal sperm. Staining of testis sections with a polyclonal anti-RBMY antibody appeared to show a marked decrease of RBMY protein in the spermatids of XY(d1)Sry males compared to control males, which led us to suggest that this may be responsible for the increase in sperm anomalies. In the current study we sought to determine whether augmenting Rbmy expression specifically in the spermatids of XY(d1)Sry males would ameliorate the sperm defects. An expressing Rbmy transgene driven by the spermatid-specific mouse protamine 1 promotor (mP1Rbmy) was therefore introduced into XY(d1)Sry males. This failed to reduce the frequency of abnormal sperm. In the course of this study, a new RBMY antibody was generated that, in contrast to the original antibody, failed to detect RBMY in spermatid stages by immunostaining. The lack of RBMY was confirmed by western blotting of lysates from purified round spermatids and elongating spermatids. The implications of these results for the proposed role for RBMY in sperm development are discussed.

Evidence that postnatal growth retardation in XO mice is due to haploinsufficiency for a non-PAR X gene.

XO Turner women, irrespective of the parental source of the X chromosome, are of short stature, and this is now thought to be largely a consequence of haploinsufficiency for the pseudoautosomal region (PAR) gene SHOX. X(p)O mice (with a paternal X) are developmentally retarded in fetal life, are underweight at birth, and show reduced weight gain in the first few weeks after birth. X(m)O mice, on the other hand, are more developmentally advanced than their XX siblings in fetal life; their postnatal growth has not hitherto been assessed. Here we show that X(m)O mice are not underweight at birth, but they nevertheless show reduced weight gain postnatally. The fact that postnatal growth is affected in X(p)O and X(m)O mice, means that this must be due to X dosage deficiency. In order to see if haploinsufficiency for a PAR gene was responsible for this growth deficit (cf SHOX deficiency in Turner women), X(m)Y*(X) females, in which the Y*(X) chromosome provides a second copy of the PAR, were compared with XX females. These X(m)Y*(X) females were also growth-retarded relative to their XX sibs, suggesting that it may be haploinsufficiency for a non-dosage-compensated X gene or genes outside the PAR that is responsible for the postnatal growth deficit in XO mice. The X genes known to escape X inactivation in the mouse have closely similar Y homologues. X(m)YSRY-negative females were therefore compared with XX females to see if the presence of the SRY-negative Y chromosome corrected the growth deficit; this proved to be the case. The postnatal growth deficit of XO mice is therefore probably due to haploinsufficiency for a non-dosage-compensated X gene that has a Y homologue that provides an equivalent function in the somatic tissues of males.

M31 and macroH2A1.2 colocalise at the pseudoautosomal region during mouse meiosis.

Progression through meiotic prophase is associated with dramatic changes in chromosome condensation. Two proteins that have been implicated in effecting these changes are the mammalian HP1-like protein M31 (HP1beta or MOD1) and the unusual core histone macroH2A1.2. Previous analyses of M31 and macroH2A1.2 localisation in mouse testis sections have indicated that both proteins are components of meiotic centromeric heterochromatin and of the sex body, the transcriptionally inactive domain of the X and Y chromosomes. This second observation has raised the possibility that these proteins co-operate in meiotic sex chromosome inactivation. In order to investigate the roles of M31 and macroH2A1.2 in meiosis in greater detail, we have examined their localisation patterns in surface-spread meiocytes from male and female mice. Using this approach, we report that, in addition to their previous described staining patterns, both proteins localise to a focus within the portion of the pseudoautosomal region (PAR) that contains the steroid sulphatase (Sts) gene. In light of the timing of its appearance and of its behaviour in sex-chromosomally variant mice, we suggest a role for this heterochromatin focus in preventing complete desynapsis of the terminally associated X and Y chromosomes prior to anaphase I.

A Y-encoded subunit of the translation initiation factor Eif2 is essential for mouse spermatogenesis.

In mouse and man, deletions of specific regions of the Y chromosome have been linked to early failure of spermatogenesis and consequent sterility; the Y chromosomal gene(s) with this essential early role in spermatogenesis have not been identified. The partial deletion of the mouse Y short arm (the Sxrb deletion) that occurred when Tp(Y)1CtSxr-b (hereafter Sxrb) arose from Tp(Y)1CTSxr-b (hereafter Sxra) defines Spy, a Y chromosomal factor essential for normal spermatogonial proliferation. Molecular analysis has identified six genes that lie within the deletion: Ube1y1 (refs. 4,5), Smcy, Uty, Usp9y (also known as Dffry), Eif2s3y (also known as Eif-2gammay) and Dby10; all have closely similar X-encoded homologs. Of the Y-encoded genes, Ube1y1 and Dby have been considered strong candidates for mouse Spy function, whereas Smcy has been effectively ruled out as a candidate. There is no Ube1y1 homolog in man, and DBY, either alone or in conjunction with USP9Y, is the favored candidate for an early spermatogenic role. Here we show that introduction of Ube1y1 and Dby as transgenes into Sxrb-deletion mice fails to overcome the spermatogenic block. However, the introduction of Eif2s3y restores normal spermatogonial proliferation and progression through meiotic prophase. Therefore, Eif2s3y, which encodes a subunit of the eukaryotic translation initiation factor Eif2, is Spy.

Spermatogenic failure in male mice with four sex chromosomes.

There is accumulating evidence that meiosis, like mitosis, is monitored by a number of checkpoints. In mammals, the presence of asynapsed chromosomes at pachytene triggers a checkpoint (the pachytene or synapsis checkpoint) that removes cells via a p53-independent apoptotic pathway. In the special case of the sex bivalent in males, it is pseudoautosomal region (PAR) asynapsis that triggers the checkpoint. In male mice with three sex chromosomes (XYY or XYY(*X)) some pachytene spermatocytes achieve full (trivalent) PAR synapsis, but in many cells one sex chromosome remains as a univalent, thus triggering the checkpoint. Sperm counts in these males have been shown to be positively correlated with trivalent frequencies. In the present study sperm production and levels of sex chromosome synapsis were studied in mice with four sex chromosomes (XYYY(*X)) and XYY(*X)Y(*X)). These mice proved to be more severely affected than XYY or XYY(*X) mice. Nevertheless, pachytene synaptonemal complex analysis revealed that full PAR synapsis was achieved through the formation of radial quadrivalents or through the formation of two sex bivalents in 21%-49% of cells analysed. Given these levels of full PAR synapsis, the sperm counts were consistently lower than would have been predicted from the relationship between levels of PAR synapsis and sperm counts in mice with three sex chromosomes. It has been suggested that the inactivation of the asynapsed non-PAR X and Y axes of the XY bivalent of normal males (MSCI), which occurs during meiotic prophase, may be driven by Xist transcripts originating from the X. If this is the case, the non-PAR Y axes of YY and YY(*X) bivalents would fail to undergo MSCI. This could be cell lethal, either because of 'inappropriate' Y gene expression, or because the non-PAR Y axis may now trigger the synapsis checkpoint.

Evidence that the testis determination pathway interacts with a non-dosage compensated, X-linked gene.

In a number of mammals, including mouse and man, it has been shown that at equivalent gestational ages, males are developmentally more advanced than females, even before the gonads form. In mice, although some strains of Y chromosome exert a minor accelerating effect in pre-implantation development, it is a post-implantation effect of the difference in X chromosome constitution that is the major cause of the male/female developmental difference. Thus XX females are retarded in their development by about 1.5 h relative to X(M)O females or XY males; however, they are more advanced than X(P)O females by about 4 h. It has been suggested that this early developmental difference between XX and XY embryos may "weight the dice" in favour of ovarian and testicular development, respectively, although expression of Sry will normally overcome any such bias. Here we test this proposal by comparing the relative frequencies of female, hermaphrodite and male development in X(P)O, XX and X(M)O mice that carry an incompletely penetrant Sry transgene. The results show that testicular tissue develops more frequently in XX,Sry transgenics than in either of the two types of XO transgenics. Thus the incidence of testicular development is affected by X dosage rather than by the developmental hierarchy. This implies there is a non-dosage compensated gene (or genes) on the X chromosome, which interacts with the testis-determining pathway. Since the pseudoautosomal region (PAR) is known to escape X-inactivation, penetrance of the Sry transgene was also assessed in X(M)Y(*X) mice that have two doses of the PAR but have a single dose of all genes proximal to the distal X marker Amel. These mice showed similar levels of testicular development to X(M)O mice with the transgene; thus the non-dosage compensated X gene maps outside the PAR.

Recombinational DNA double-strand breaks in mice precede synapsis.

In Saccharomyces cerevisiae, meiotic recombination is initiated by Spo11-dependent double-strand breaks (DSBs), a process that precedes homologous synapsis. Here we use an antibody specific for a phosphorylated histone (gamma-H2AX, which marks the sites of DSBs) to investigate the timing, distribution and Spo11-dependence of meiotic DSBs in the mouse. We show that, as in yeast, recombination in the mouse is initiated by Spo11-dependent DSBs that form during leptotene. Loss of gamma-H2AX staining (which in irradiated somatic cells is temporally linked with DSB repair) is temporally and spatially correlated with synapsis, even when this synapsis is 'non-homologous'.

Analysis of male meiotic "sex body" proteins during XY female meiosis provides new insights into their functions.

During male meiosis in mammals the X and Y chromosomes become condensed to form the sex body (XY body), which is the morphological manifestation of the process of meiotic sex chromosome inactivation (MSCI). An increasing number of sex body located proteins are being identified, but their functions in relation to MSCI are unclear. Here we demonstrate that assaying male sex body located proteins during XY female mouse meiosis, where MSCI does not take place, is one way in which to begin to discriminate between potential functions. We show that a newly identified protein, "Asynaptin" (ASY), detected in male meiosis exclusively in association with the X and Y chromatin of the sex body, is also expressed in pachytene oocytes of XY females where it coats the chromatin of the asynapsed X in the absence of MSCI. Furthermore, in pachytene oocytes of females carrying a reciprocal autosomal translocation, ASY associates with asynapsed autosomal chromatin. Thus the location of ASY to the sex body during male meiosis is likely to be a response to the asynapsis of the non-homologous regions [outside the pseudoautosomal region (PAR)] of the heteromorphic X-Y bivalent, rather than being related to MSCI. In contrast to ASY, the previously described sex body protein XY77 proved to be male sex body specific. Potential functions for MSCI and the sex body are discussed together with the possible roles of these two proteins.

An analysis of meiotic impairment and of sex chromosome associations throughout meiosis in XYY mice.

The existing XYY meiotic data for mice present a very heterogeneous picture with respect to the relative frequencies of different sex chromosome associations, both at pachytene and diakinesis/metaphase I. Furthermore, where both pachytene and diakinesis/MI data are available for the same males, the frequencies of the different configurations at the two stages are very different. In the present paper we utilise "XYY" and "XY/XYY" mosaic mice with cytologically distinguishable Y chromosomes to investigate the factors responsible for this heterogeneity between different males and between the two meiotic stages. It is concluded (1) that the initial pattern of synapsis is driven by the relatedness of the three pseudoautosomal regions (PARs); (2) that the order and extent of PAR synapsis within radial trivalents are also affected by PAR relatedness and that this leads to chiasmata being preferentially formed between closely related PARs; (3) that trivalents with a single chiasma resolve into a bivalent + univalent by the diakinesis stage; (4) that although many spermatocytes with asynapsed sex chromosomes are eliminated between pachytene and diakinesis, those that survive this phase of elimination progress to the first meiotic metaphase (MI) and accumulate in large numbers, leading to an over-representation of those with univalents as compared to radial trivalents; and (5) that the arrested MI cells are eventually eliminated, so that very few "XYY" cells contribute products to MII.

Evidence that sex chromosome asynapsis, rather than excess Y gene dosage, is responsible for the meiotic impairment of XYY mice.

There is extensive evidence for the existence of a meiotic checkpoint that acts to eliminate spermatocytes that fail to achieve full sex chromosome synapsis at the pachytene stage of the first meiotic prophase. XYY mice are nearly always sterile, with clear signs of meiotic impairment, and sex chromosome asynapsis has been proposed to underlie this impairment. However, a study of XYY*(X) mice (mice having three sex chromosomes but only a single dose of Y genes) revealed that these mice are fertile, and thus implicated Y gene dosage as a major factor in the sterility of XYY mice. To address this question further, sex chromosome synapsis and spermatogenic proficiency were compared between XYY*(X) and XYY mice generated in the same litters. This established that differences in spermatogenic proficiency within and between the two genotypes correlated with the frequency of radial trivalent formation (full sex chromosome synapsis); XYY*(X) males, as a group, had double the radial trivalent frequency of XYY males. This observation provides strong support for the view that sex chromosome asynapsis (or some consequence thereof), rather than Y gene dosage, is the major factor leading to the meiotic impairment of XYY mice.

A high frequency of XO offspring from X(Paf)Y* male mice: evidence that the Paf mutation involves an inversion spanning the X PAR boundary.

It has previously been reported that 19% of the daughters of males carrying the X-linked mutation patchy fur (Paf) are XO with a maternally derived X chromosome. We now report that hemizygous Paf males that also carry the variant Y chromosome Y*, show a much increased XO production ( approximately 40% of daughters). We hypothesize that the Paf mutation is associated with an inversion spanning the pseudoautosomal region (PAR) boundary, and that this leads to preferential crossing over between the resulting inverted region of PAR and an equivalent inverted PAR region within the compound Y* PAR. This would lead to the production of dicentric X and acentric Y products and consequent sex chromosome loss. This interpretation is supported by analysis of the sex chromosome complements at the second meiotic metaphase, which revealed a high incidence of dicentrics. Another curious feature of the Paf mutation is that mice that are homozygous Paf have more hair than mice that are hemizygous Paf. This can be explained if the Paf mutation is a hypomorphic mutation that escapes X inactivation because, unlike the wild type allele, it is now located within the PAR.

The role of Y-encoded genes in mammalian spermatogenesis.

As long ago as 1931 Fisher outlined the reasons for the accumulation of male 'benefit genes' (e.g. male fertility factors) on the Y chromosome, but it was over four decades later that a study of men with partial Y chromosome deletions revealed that a factor essential for male fertility was present on the human Y. Today, the Y deletion interval containing this 'Azoospermia Factor' (AZF) has been subdivided into three subintervals associated with different degrees of spermatogenic impairment. Furthermore, three deletion intervals have been identified on the mouse Y that impact on the spermatogenic process. This review examines these deletion intervals in mouse and man and summarises progress towards identifying candidate genes for their respective spermatogenic functions.

The Y* rearrangement in mice: new insights into a perplexing PAR.

In essence, the Y* rearrangement in the mouse is a Y chromosome that has been hijacked by a non-Y centromere attached distal to the pseudoautosomal region (PAR). All the Y-unique material is thought to be unaltered, but the recombinatory behaviour of the Y* with the X during male meioisis led to the conclusion that part of the PAR is inverted. In the course of a cross set up to introduce the X-linked mutation Patchy fur (Paf) into XY* males, the Y* chromosome was found to carry the wild type allele of Paf. Paf maps close to the X PAR boundary, so we hypothesised that the inverted region of the Y* PAR originated from an X chromosome that provided not only an inverted copy of proximal PAR, but also an X PAR boundary together with some adjacent X-unique material that included the Paf locus. This hypothesis was validated by Southern analysis using an X PAR boundary probe to show that Y* has an X PAR boundary. Thus the Y* PAR has resulted from an end to end fusion of an X and a Y PAR. Furthermore, it was shown that in conjunction with this PAR-PAR fusion, there has been deletion of both copies of the distally located pseudoautosomal gene Steroid sulfatase (Sts).

The mammalian Y chromosome: a new perspective.

For several decades, the mammalian Y chromosome was considered a genetic "desert," with the testis determinant being the sole survivor of the attrition that followed the chromosome's inception. Aside from the addition of a genetic factor required for spermatogenesis to the human Y chromosome in 1976, this view held sway until the mid-1980s. The ensuing molecular genetic analysis, culminating in the recent paper in Science by Lahn and Page, has identified more than 20 genes or gene families on the human Y. This has led to a reappraisal of the evolution and functions of this unique chromosome.

Mouse homologues of the human AZF candidate gene RBM are expressed in spermatogonia and spermatids, and map to a Y chromosome deletion interval associated with a high incidence of sperm abnormalities.

An RNA-binding motif (RBM) gene family has been identified on the human Y chromosome that maps to the same deletion interval as the 'azoospermia factor' (AZF). We have identified the homologous gene family (Rbm) on the mouse Y with a view to investigating the proposal that this gene family plays a role in spermatogenesis. At least 25 and probably >50 copies of Rbm are present on the mouse Y chromosome short arm located between Sry and the centromere. As in the human, a role in spermatogenesis is indicated by a germ cell-specific pattern of expression in the testis, but there are distinct differences in the pattern of expression between the two species. Mice carrying the deletion Yd1, that maps to the proximal Y short arm, are female due to a position effect resulting in non-expression of Sry ; sex-reversing such mice with an Sry transgene produces males with a high incidence of abnormal sperm, making this the third deletion interval on the mouse Y that affects some aspect of spermatogenesis. Most of the copies of Rbm map to this deletion interval, and the Yd1males have markedly reduced Rbm expression, suggesting that RBM deficiency may be responsible for, or contribute to, the abnormal sperm development. In man, deletion of the functional copies of RBM is associated with meiotic arrest rather than sperm anomalies; however, the different effects of deletion are consistent with the differences in expression between the two species.

The meiotic checkpoint monitoring synapsis eliminates spermatocytes via p53-independent apoptosis.

Evidence is accumulating that meiosis is subject to 'checkpoints' that monitor the quality of this complex process. In yeast, unresolved double strand breaks (DSBs) in DNA are thought to trigger a 'recombination checkpoint' that leads to pachytene arrest. In higher eukaryotes, there is evidence for a checkpoint that monitors chromosome synapsis and in mammals the most compelling evidence relates to the sex chromosomes. In normal male mice, there is synapsis between the X and Y pseudoautosomal regions; in XSxr(a)O mice, with a single asynaptic sex chromosome, there is arrest at the first meiotic metaphase, the arrested cells being eliminated by apoptosis (our unpublished data). Satisfying the requirement for pseudoautosomal synapsis by providing a pairing partner for the XSxr(a) chromosome avoids this arrest. We have considered that this 'synapsis checkpoint' may be a modification of the yeast 'recombination checkpoint' with unresolved DSBs (a corollary of asynapsis) providing the trigger for apoptosis. DSBs induced by irradiation are known to trigger apoptosis in a number of cell types via a p53-dependent pathway, and we now show that irradiation-induced spermatogonial apoptosis is also p53-dependent. In contrast, the apoptotic elimination of spermatocytes with synaptic errors proved to be p53-independent.

Timing of mating, developmental asynchrony and the sex ratio in mice.

According to the developmental asynchrony hypothesis, changing the time of mating within the estrous cycle could alter the interval between completion of blastocyst development and uterine responsiveness for implantation. This may then lead to sex ratio skews in animals that exhibit sex-differential blastocyst development, because uterine stage may now benefit either slow (female) or fast (male) developing blastocysts. To test this hypothesis, the responses of two strains of mice to altered mating dynamics were compared. In a strain that exhibits higher male than female blastocyst developmental rates, sex ratios became significantly female-biased when mated late during the estrous cycle as opposed to early mating. However, timing of mating did not affect sex ratios in a strain with synchronous development of male and female preimplantation embryos. Hence, it is concluded that developmental asynchrony between male and female blastocysts on the one hand, and blastocysts and uterus on the other, are indeed responsible for the effect of timing of mating on litter sex ratios in mice.

Hormonal and cellular regulation of Sertoli cell anti-Müllerian hormone production in the postnatal mouse.

Anti-Müllerian hormone (AMH) is secreted by immature testicular Sertoli cells. Clinical studies have demonstrated a negative correlation between serum AMH and testosterone in puberty but not in the neonatal period. We investigated AMH regulation using mouse models mimicking physiopathological situations observed in humans. In normal mice, intratesticular, not serum, testosterone repressed AMH synthesis, explaining why AMH is downregulated in early puberty when serum testosterone is still low. In neonatal mice, AMH was not inhibited by intratesticular testosterone, due to the lack of expression of the androgen receptor in Sertoli cells. We had shown previously that androgen-insensitive patients exhibit elevated AMH in coincidence with gonadotropin activation. In immature normal and in androgen-insensitive Tfm mice, follicle stimulating hormone (FSH) administration resulted in elevation of AMH levels, indicating that AMH secretion is stimulated by FSH in the absence of the negative effect of androgens. The role of meiosis on AMH expression was investigated in Tfm and in pubertal XXSxrb mice, in which germ cells degenerate before meiosis. We show that meiotic entry acts in synergy with androgens to inhibit AMH. We conclude that AMH represents a useful marker of androgen and FSH action within the testis, as well as of the onset of meiosis.