| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||
REVIEW |
1 Monash Institute of Medical Research, Monash University, Melbourne, Victoria 3168, Australia2 Australian Research Centre of Excellence in Biotechnology and Development, Monash University, Melbourne, Victoria 3168, Australia3 Queensland Institute of Medical Research, Brisbane, Queensland 4072, Australia
Correspondence should be addressed to M K O'Bryan; Email: moira.obryan{at}med.monash.edu.au
| Abstract |
|---|
| Introduction |
|---|
Spermatogenesis is a highly regulated process; it is not surprising that epigenetic mechanisms are involved and that perturbations in these mechanisms can result in male infertility. There have been various links drawn between epigenetic mechanisms and important events during spermatogenesis, including chromosome condensation, XY body formation, retrotransposon silencing and packaging of DNA into the spermatid nucleus; however, the underlying molecular mechanisms are only now becoming clear. While epigenetic regulation is crucial in many systems within the body, this review will concentrate on spermatogenesis, male fertility and the consequences of aberrant regulation for offspring.
| Epigenetic modifications |
|---|
Chromatin is made-up of histones and DNA, both of which may be modified to affect the way in which they interact. Histones are composed of predominantly positively charged amino acids, which allow them to bind with a high affinity to negatively charged DNA (Alberts et al. 1989). There are four types of nucleosomal histones: H2A, H2B, H3 and H4. Histones are responsible for the coiling of DNA into nucleosomes. Histone variants exist for histones H3 (CENP-A, H3.3), H2A (H2AZ, H2AX, macroH2A) and H2B (TH2B). Each has a different role in maintaining chromatin structure. Two copies of each nucleosomal histone (H2A, H2B, H3 and H4) form a histone octamer core around which 147 bp of double-stranded DNA is wound creating a nucleosome (reviewed in Zhang & Reinberg 2001; Fig. 1). Nucleosomes are the basic subunit of chromatin. Linker DNA as well as histone H1 links all the nucleosomes together to create a 30 nm chromatin fibre. Epigenetic mechanisms regulate the association between the DNA and histones in such a way that genes are made transcriptionally active or inactive.
|
DNA methylation
DNA methylation is one of the most well-studied epigenetic genomic modifications. DNA methylation involves the transfer of a methyl group to the five-positioned carbon of a cytosine within a CpG dinucleotide (Nakao 2001, Aapola et al. 2004; Table 1). Approximately 60–90% of all CpG sequences in the genome are methylated, while unmethylated CpG dinucleotides are mainly clustered in the CpG-rich sequence, termed CpG island, of the gene promoter region (Ng & Bird 1999). DNA methylation at promoter regions is involved with gene silencing. DNA methylation is also used to silence retrotransposons and imprinted genes (see below).
|
|
|
Histone acetyltransferases (HATs) can acetylate specific lysine residues in the N-termini of all core histones and are associated with gene activation (An 2007; Table 1, Fig. 2). Histone deacetyltransferases reverse the action of HATs and are associated with gene repression (Table 1). The acetylation and deacetylation of several lysine residues on histones H3 and H4 are important during male meiosis and will be discussed later in this review.
Histone phosphorylation occurs on several serine residues of all the histones and is generally associated with transcriptional activation (reviewed in Berger 2002; Table 1, Fig. 2). By contrast, phosphorylation of the histone variant H2AX (forming
H2AX) correlates with chromosome condensation and gene silencing during male meiosis (Fernandez-Capetillo et al. 2003; Table 1).
Ubiquitylation on specific lysine residues can either enhance or repress transcription (reviewed in Conaway et al. 2002; Table 1). Ubiquitylation is the modification of a protein by the covalent attachment of an ubiquitin monomer to lysine residues. Histone H2A can be mono-ubiquitylated and its presence is linked to transcriptional repression during male meiosis (Baarends et al. 2005; Fig. 2). Conversely, H2B mono-ubiquitylation is linked to transcriptional activation (Zhu et al. 2005; Fig. 2).
Histone lysine residues may also be sumoylated and have been linked to gene repression (Shiio & Eisenman 2003; Table 1). Sumoylation involves the covalent attachment of small ubiquitin-related modifier (SUMO) proteins to lysine residues (Shiio & Eisenman 2003). Sumoylation generally prevents the activation of PTMs such as acetylation. There are four known SUMO proteins (SUMO1–4). To date, SUMO1 is the only one associated with spermatogenesis specifically, in silencing the XY body during meiosis (Rogers et al. 2004).
ATP-dependent chromatin remodelling
Unlike DNA methylation and histone PTMs, ATP-dependent chromatin remodelling complexes are correlated with non-convalent chromatin remodelling. ATP-dependent complexes use the energy of ATP hydrolysis to change the position or structure of the nucleosomes and therefore affect the accessibility of transcription factors and other regulatory elements to the DNA (reviewed in Narlikar et al. 2002). ATP-dependent chromatin remodelling has been implicated in both transcriptional activation and repression. The predominant mechanism of ATP-dependent chromatin remodelling involves sliding the histones along the DNA. In addition, these complexes may twist, spool and bulge DNA (Fan et al. 2003). There are currently three known families of ATP-dependent chromatin remodelling complexes, the SWI/SNF (switch/sucrose nonfermenter) family, the ISWI (imitation switch) family and the MI-2 (myositis specific autoantigen 2) family, all of which are defined by the presence of a conserved catalytic ATPase subunit (Wang & Zhang 2001, Fan et al. 2005).
It has previously been shown that Smarca5 (encoding SNF2H) from the ISWI family is expressed in the testis where it may have a role in DNA repair and recombination (Lazzaro & Picketts 2001, Chong et al. 2007). As discussed later in this review, a hypomorphic allele of Smarca5 has been generated through an N-ethylnitrosourea (ENU) screen for epigenetic modifiers and has revealed that haploinsufficiency for Smarca5 shows paternal-specific effects on the phenotype of offspring (Chong et al. 2007; Table 2). The ablation of Smarca5 results in embryonic lethality at the peri-implantation stage (Stopka & Skoultchi 2003). RAD54 from the SWI/SNF family has also been implicated in DNA repair during spermatogenesis. Disruption of Rad54 results in aberrant homologous recombination and DNA repair defects (Essers et al. 1997; Table 2).
| The epigenetic state during spermatogenesis |
|---|
Transposable elements
In the prenatal testis, the only germ cells present are gonocytes. The gonocytes proliferate for a few days and then become arrested (Huckins & Clermont 1968). Shortly after birth, the gonocytes move to the basement membrane to become spermatogonia and resume proliferation (Huckins & Clermont 1968). Spermatogonia divide by mitosis and constitute the pool of stem cells from which meiosis and spermatogenesis proceed.
Prior to spermatogenesis the silencing of TEs takes place in the gonocytes and prospermatogonia. TEs are pieces of mobile DNA and include DNA transposons, long terminal repeat (LTR) retrotransposons, long interspersed nuclear elements (LINEs) and short interspersed nuclear elements (reviewed in Deininger & Batzer 2002). TEs comprise 45% of human and 37% of mouse DNA (Deininger et al. 2003). Their movement can be mutagenic and can cause chromosome breakage, improper recombination and genome rearrangement if not silenced via methylation (Deininger et al. 2003). TEs can amplify and re-insert into the genome causing ectopic activation or inactivation of genes. Normally, DNA methylation is established in the gonocytes to silence TEs thereby preventing propagation.
DNMT3L facilitated methylation is involved in the silencing of TEs in the testis as evidenced by an absence of DNMT3L causing a loss of methylation at LINE-1 and the intracisternal A-particle (IAP) transposons (Bourc'his & Bestor 2004, Webster et al. 2005, Hata et al. 2006; Table 2). Consistent with such a role, DNMT3L is expressed in the gonocytes at 14–18 days post-coitum (dpc) when global DNA methylation is occurring (Bourc'his & Bestor 2004). Indeed, it has been proposed that the absence of TE silencing in this model contributes directly to meiotic arrest and infertility in the adult mouse.
Paternal imprinting
Paternally imprinted genes are genes that are subject to monoallelic expression. That is, expression occurs only from the allele inherited from the maternal parent. Paternally imprinted genes are methylated, and silenced, in male germ cells. Paternal imprinting of genes starts to be established in the gonocytes at about 15.5 dpc, and continues on into the spermatogonia (Fig. 3). This methylation and gene silencing are carried through the remainder of spermatogenesis and into any resultant sperm and offspring. Only three paternally imprinted regions, H19-Igf2, Rasgrf and Dlk1-Gtl2 have been identified to date (Davis et al. 1999, Li et al. 2004). Many more maternally imprinted genes have been identified.
|
While the critical importance of appropriate paternal imprinting has been established for offspring inheriting such alleles naturally (see below), a recent paper showing an increased prevalence of abnormal methylation in a set of seven imprinted genes in infertile men with oligospermia (decreased sperm count) compared with normospermic men from infertile relationships (Marques et al. 2004, Kobayashi et al. 2007) raises the possibility that a failure of imprinting may actually cause infertility. Alternatively, the observed hypomethylation of paternally imprinted alleles may be indicative of inappropriate global methylation, and as such inappropriate gene expression. Comfortingly, the study by Kobayashi et al. (2007) suggests that at least in their study, global methylation defects had not occurred. Normally, of course, such situations would be self-limiting because of the resultant infertility; however, with increasing access to intracytoplasmic sperm injection (ICSI) there is a possibility that such hypomethylated sperm are used to conceive children wherein the affects may be much more debilitating. It should be noted, however, that at the time of writing this review this possibility was only theoretical as there is no published evidence of an increase in paternally imprinted diseases in children conceived via ICSI compared with standard in vitro fertilisation where embryos have been cultured under the same conditions.
Meiosis
Spermatogonia undergo differentiation and continuous rounds of mitosis to generate primary spermatocytes, which enter meiosis (de Kretser & Kerr 1994). Meiosis is characterised by two cell divisions, during which chromosome numbers are halved. In the first division, primary spermatocytes duplicate DNA content then divide to become secondary spermatocytes (Clermont 1972). In the second division, secondary spermatocytes divide to become round spermatids (Clermont 1972). A large part of meiosis is spent in prophase of the first division. Prophase accounts for over 90% of the duration of meiosis and lasts about 3 weeks in the mouse (Cobb & Handel 1998). Prophase 1 is further broken up into four stages: leptotene, zygotene, pachytene and diplotene. During leptotene of meiosis, the chromosomes start condensing and double-strand breaks (DSBs) begin forming. DSBs are essential for the recombination between all homologous chromosomes that occur later in meiosis (Roeder 1997, Cobb & Handel 1998). During zygotene, sister chromatids begin to pair and form synaptonemal complexes (SCs). SCs are proteinaceous meiosis-specific structures composed of two axial elements, which form along the length of the sister chromatids and are joined by a central element which zips up the two axial elements (now called lateral elements; Morelli & Cohen 2005). Synapsis is completed during pachytene. Homologous recombination also occurs in pachytene spermatocytes. Prophase ends at diplotene when chromosome separation is initiated and SCs start disappearing (reviewed in Baarends & Grootegoed 2003). Following prophase, the spermatocytes undergo metaphase, anaphase and telophase of the first division to produce two secondary spermatocytes per primary spermatocyte. During metaphase, the homologous chromosomes condense further and become attached to spindle fibres. The chromosomes separate, but sister chromatids remain attached and move towards opposite poles during anaphase. The spindle is squeezed and the nuclei and cytoplasm start dividing during telophase.
The secondary spermatocytes contain a single set of chromosomes but have a 2n DNA content. During the second meiotic division, the chromosomes condense and the spermatocytes undergo prophase, metaphase, anaphase and telophase of meiosis II. During meiosis II, sister chromatids separate and divide to produce round spermatids which have a 1n content (de Kretser & Kerr 1994). The end products of meiosis are four haploid round spermatids per spermatocyte. Several aspects of meiosis are critically regulated by epigenetic modifications.
Chromatin remodelling and methylation during meiosis
As evidenced in several studies, epigenetic modifications are essential for several aspects of meiotic chromosome packing, pairing and recombination (Fig. 3). DNMT3L is involved in facilitating the methylation pattern required during male meiosis. In the absence of DNMT3L, meiotic chromosomes fail to form heterochromatin appropriately and fail to pair at the zygotene stage (Bourc'his & Bestor 2004, Webster et al. 2005; Table 3). Leptotene and zygotene spermatocytes displayed non-homologous synapsis and spermatocytes were unable to reach the full pachytene stage (Bourc'his & Bestor 2004). As a consequence, all germ cells eventually die by apoptosis or sloughing at the zygotene–pachytene stage (Bourc'his & Bestor 2004, Webster et al. 2005). It is believed that in Dnmt3l knockout males the loss of methylation resulted in an inappropriate euchromatic state during meiosis and genes that should be repressed, including imprinted regions, retrotransposons and repeat elements, remained active resulting in what the Bestor lab described as a meiotic catastrophe (Bourc'his & Bestor 2004, Webster et al. 2005). It is clear that DNMT3L-mediated methylation is needed for meiotic cells to progress through spermatogenesis.
|
, PPβ and HP1
(Lachner et al. 2001), and the subsequent binding of the HMTase SUV420H and the methyltransferases DNMT3A/DNMT3B and the di- and tri-methylation of H4K20. Cumulatively, this leads to the establishment of a transcriptionally repressed state (Lehnertz et al. 2003, Schotta et al. 2004). Suv39h-deficient mice displayed impaired spermatogenesis. Spermatocytes undergo apoptosis in the pachytene stage as a consequence of incomplete homologue pairing and synapsis (Peters et al. 2001). More recently, the HMTase G9A, which in somatic cells is involved in H3K9 mono- and di-methylation at euchromatic regions, has been shown to be important in male meiosis. Deletion of G9a resulted in embryonic lethality at mid-gestation, for this reason germ lineage-specific conditional G9a knockout mice were generated. These mice were sterile as spermatocytes failed to progress past the pachytene stage (Tachibana et al. 2007). Many spermatocytes were unable to form SCs. It was proposed that this perturbation in SC formation may be due to the chromosomes failing to find and pair with a homologue or to the overexpression of G9A-regulated genes (Tachibana et al. 2007).
In addition to dynamic changes in histone modifications, successful male meiosis also requires the use of specialised histone variants. The histone variant H2AX is an important part of the nucleosome of meiotic cells. In response to DNA DSBs, H2AX is phosphorylated at serine residue Ser139 to form phosphorylated histone H2AX (
H2AX; Mahadevaiah et al. 2001). On autosomes during meiosis,
H2AX localises to sites of DSBs and is involved in the chromatin condensation which occurs during meiotic recombination in leptotene and zygotene spermatocytes (Redon et al. 2002). It has been proposed that
H2AX may act to recruit members of the DNA repair machinery to DSB sites (Baarends & Grootegoed 2003). Once DSBs disappear and synapsis is achieved,
H2AX is no longer localised to DSB sites.
Other DNA repair proteins that are present during meiosis are RAD50 and RAD51. RAD50 is the first to be recruited to the site of DNA DSBs on the autosomes and is present from preleptotene until the end of zygotene (Eijpe et al. 2000). RAD51 is associated with the axial elements formed during leptotene (Tarsounas & Moens 2001). RAD54, as outlined above, is a DNA-dependent ATPase needed for recombination and DNA repair (Table 2). RAD54 interacts with RAD51 during DNA repair, together these proteins are believed to promote chromatin remodelling by displacing nucleosomes during homologous recombination (Alexeev et al. 2003).
The ubiquitin-conjugating enzyme HR6B is also involved in DNA damage repair and has been implicated in the regulation of chromatin structure during meiotic recombination. In Hr6b knockout mice, spermatogenesis was impaired due to major changes found in the structure of the SCs. These mice also displayed an increased frequency of meiotic recombination (Baarends et al. 2003; Table 2).
Meiotic sex chromosome inactivation (MSCI)
In the pachytene stage of meiosis, the X and Y chromosomes form an XY (or sex) body and become transcriptionally silent in a process known as MSCI (Handel 2004, Turner 2007; Fig. 3). Unlike the autosomes, in males sex chromosomes are heterologous and only synapse at their pseudoautosomal regions. The development of the XY body and the maintenance of its heterochromatic state have been the focus of several recent studies (Baarends et al. 2007, van der Heijden et al. 2007). There are now a large number of known epigenetic modifications that are associated with MSCI, although the exact order of events are unclear (Table 3). Several of these modifications and modifiers are detailed below.
During pachytene when the autosomes synapse,
H2AX localises to the XY body independently of DSBs occurring (Fernandez-Capetillo et al. 2003). The importance of H2AX in this process is indicated in H2ax-deficient spermatocytes which are unable to form an XY body and arrest at the pachytene stage (Fernandez-Capetillo et al. 2003; Tables 2 and 3). H2AX phosphorylation is dependent on the DNA repair protein ataxia telangiectasia and Rad3 related (ATR). ATR localises to the XY body at the onset of MSCI (Turner et al. 2004). The recruitment of ATR to the XY body depends on the tumour suppressor protein breast cancer 1, early onset (BRCA1; Turner et al. 2004). Meiotic cells of mice hypomorphic for Brca1 do not undergo H2AX phosphorylation and the pachytene cells do not undergo MSCI (Turner et al. 2004). H2AX phosphorylation, ATR and BRCA1 all act together to initiate MSCI. This is one of the earliest events which takes place on the XY body during MSCI.
In order for MSCI to be maintained throughout pachytene, there are many other epigenetic modifications that occur on the XY body (reviewed in Hoyer-Fender 2003, Handel 2004, Turner 2007). They include ubiquitylation of H2A (forming uH2A), sumoylation (involving SUMO1), methylation of H3K27, di-methylation of H3K9, H4K20, H3K79 and H3K27, tri-methylation of H3K9 and H4K20, and deacetylation of H3K9, H4K12 and H4K16 (Table 3). Modifications known to be involved with gene activation such as H4K5 and H4K8 hyperacetylation are also found on the XY body. The exact function each performs, however, are unknown.
In addition, it is hypothesised that meiosis-induced factor containing a PR/SET domain and zinc-finger motif (MEISETZ), which is involved in H3K4 methylation, may have a role in activating genes essential for meiosis via XY body formation. Meisetz is expressed in the germ cells of the foetal female gonad and in the post-natal testis (Hayashi et al. 2005). Meisetz knockout males and females have an impaired DSB repair pathway and homologous chromosomes failed to pair (Hayashi et al. 2005; Table 2). Of relevance here, in the males, the sex body does not form and the spermatocytes displayed a meiotic arrest.
The appearance of the H2A variant H2AZ is also associated with MSCI in very late pachytene spermatocytes when many other heterochromatic marks are disappearing from the XY body (Greaves et al. 2006). At the completion of MSCI, H2AZ is incorporated into the XY body where it is thought to at least partially maintain the inactive state of the XY body in round and elongating spermatids (Greaves et al. 2006).
The H3 histone variant H3.3 has been implicated in MSCI. H3.3 is incorporated into the XY body during MSCI and coincides with the loss of not only the histone variants H3.1 and H3.2 but also most of the histone PTMs (van der Heijden et al. 2007). PTMs, such as mono-, di- and tri-methylated H3K4; di- and tri-methylated H3K9; mono- and di-methylated H3K27; di-methylated H3K79; and di- and tri-methylated H4K20, were all lost after H3.3 incorporation into the XY body (van der Heijden et al. 2007). In late meiotic and post-meiotic (i.e. round spermatid) stages, a select number of these histone modifications, predominantly those involved in gene silencing, are re-acquired (van der Heijden et al. 2007). H3.3 is known to be associated with modifications involved in gene activation including acetylation of K9, K14 and K18 (McKittrick et al. 2004). The exact role H3.3 performs in MSCI is unclear, although it is evident that H3.3 incorporation into the XY body promotes extensive chromatin remodelling and is essential for gene silencing on the XY body during the later stages of MSCI and the post-meiotic stages of spermatogenesis (van der Heijden et al. 2007).
In addition, SCMH1 (the mammalian homologue of Drosophila sex comb on midleg) has recently been shown to be important in regulating some of the chromatin modifications on the XY body. SCMH1 is part of the mammalian Polycomb repressive complex 1 (PRC1) and is involved in the repression of numerous genes (Levine et al. 2002). Within the testis, SCMH1 is expressed in the zygotene–pachytene spermatocytes and round spermatids where it is involved in the exclusion of the PRC1 complex from the XY body. As illustrated by a recent Scmh1 knockout mouse line (missing the SPM domain), a loss of Scmh1 leads to male sub-fertility characterised by a loss of spermatocytes by apoptosis (Takada et al. 2007). Scmh1-deficient pachytene spermatocytes failed to maintain exclusion of the PRC1 complex from the XY body leading to the XY body H3K27me3. Interestingly, a double knockout of Scmh1 and Phc2 (another polycomb protein) corrected this phenotype. SCMH1 binds to PHC2, via its SPM domain (Isono et al. 2005). These data indicate the SCMH1 is a key regulatory component of the PRC1 in maintaining exclusion from the XY body. PHC2 on the other hand may promote such an association (Takada et al. 2007). Gene expression data also indicated that inappropriate PRC1 exclusion from the XY body did not affect MSCI (Takada et al. 2007).
Post-meiotic sex chromosome repression (PMSR)
It is now evident that at least partial XY body silencing persists after meiosis in a process known as PMSR (Namekawa et al. 2006). Through microarray studies focusing on X chromosome-linked genes, Namekawa et al. (2006) found that 87% of 676 genes on the X chromosome remain suppressed post-meiotically. This was comparable with pachytene spermatocytes where 92% of genes were repressed. Many studies have found that modifications associated with a heterochromatic state such as tri-methylation of H3K9 and deacetylation of H3K9 and H4K12 remain on the sex chromosomes post-meiotically (Khalil et al. 2004, van der Heijden et al. 2007). It has been suggested that this silencing persists into the zygote; that is, the paternally inherited X chromosome (Xp) is pre-inactivated in female pre-implantation embryos (Huynh & Lee 2003). This theory is, however, hotly debated by others who propose that the Xp is not pre-inactivated, but rather is inactivated solely within the zygote by expression of the Xist gene (Okamoto et al. 2005).
Spermiogenesis
In addition to extensive cytoplasmic metamorphoses, during this post-meiotic phase the haploid round spermatids undergo extensive chromatin remodelling to develop into mature spermatozoa (Fig. 3). This involves re-shaping and condensation of the nucleus to transcriptionally inactivate and protect the DNA. This is assisted by DSBs and the histone-to-protamine transition (reviewed in Doenecke et al. 1997, Govin et al. 2004).
The histone-to-protamine transition involves the replacement of the histones with transition nuclear proteins (TP1 and TP2), then subsequently the replacement of the TPs with protamines (PRM1 and PRM2). The histone-to-protamine transition is associated with hyperacetylation of histone H4 (Sonnack et al. 2002). Hyperacetylation promotes a looser nucleosomal structure, which allows for easy removal of the histones and incorporation of the TPs and subsequently the protamines (Sonnack et al. 2002; Fig. 3).
The replacement of histones for TPs then PRMs is essential as evidenced in several gene knockout models. Mice lacking either TP1 or TP2 have subtle abnormalities in chromatin condensation during spermiogenesis and were fertile with normal numbers of sperm (Yu et al. 2000, Zhao et al. 2001). However, mice lacking both TPs exhibited irregular chromatin condensation in all spermatids, many showed DNA breaks, PRM2 was not post-translationally processed and male mice were sterile (Zhao et al. 2004; Table 2).
Protamines are small basic proteins rich in arginine and cysteine and are found only in spermatids (reviewed in Wouters-Tyrou et al. 1998). Most mammals produce two forms of protamine, PRM1 and PRM2, which are responsible for the DNA being packaged into a very compact arrangement. Disruption of one copy of either gene, i.e. haploinsufficiency for either PRM1 or PRM2, disrupts nuclear formation, processing of PRM2 and sperm function (Cho et al. 2003; Table 2). Consequently, heterozygous males were sterile. The importance of protamines is also evident in humans where an absence of PRM2 or changes in the PRM1 to PRM2 ratio in the sperm have been associated with male infertility (Balhorn et al. 1988, de Yebra et al. 1993).
The importance of protamines and their regulation for fertility is further evidenced in the Jhdm2a knockout mouse line. Protamine expression is at least in part regulated by the H3K9me2/H3K9me1 demethylase JmjC-domain-containing histone demethylase 2A (JHDM2A). Expression of JHDM2A starts in the late pachytene stage and persists until the elongated spermatid period. Sperm do not contain JHDM2A (Okada et al. 2007). Disruption of Jhdm2a caused post-meiotic chromatin condensation defects including elongating spermatids with abnormal nuclear structure and elongation (Okada et al. 2007; Table 2). As a result, mutant males had reduced sperm counts and were infertile (Okada et al. 2007). JHDM2A functions to demethylate mono- and di-methylated H3K9 in the promoter regions of Tnp1 and Prm1 in round spermatids (Okada et al. 2007). An absence of JHDM2A results in promoter hyper-methylation and gene silencing and a lack of TP1 and PRM1, and infertility.
As indicated above, TP and PRM function, and as a consequence spermiogenesis, is regulated by epigenetic modifications. The loss of calmodulin-dependent protein kinase IV (CAMK4)-dependent serine/threonine phosphorylation of PRM2 results in sterility (Wu et al. 2000). Camk4 knockout spermatids showed a loss of PRM2 and prolonged retention of TP2 (Wu et al. 2000). As a consequence, Camk4 knockout males were infertile and displayed a reduced number of late elongating spermatids and no mature sperm in the epididymides (Wu et al. 2000; Table 2).
In addition to the histone to protamine transition, spermiogenesis is critically dependent on changes in H1 linker histones. H1T2 is a H1 variant specifically expressed in round through to elongated spermatids (Martianov et al. 2005, Tanaka & Baba 2005). Knockout males were sterile, or severely sub-fertile, and epididymal sperm showed abnormal morphology, cytoplasmic retention and impaired motility. Interestingly, sperm showed only trace amounts of PRM1 and PRM2 suggesting that H1T2 is also involved in the histone to protamine transition (Martianov et al. 2005, Tanaka & Baba 2005).
At least two other H1 variants are also involved in spermiogenesis. H1T, which is a testis-specific variant, is first detected in pachytene spermatocytes where it is integrated into the genome and replaces the majority of somatic H1 linker histones. H1T is not detected beyond the early elongating spermatid stage concordant with its replacement by TPs and PRMs (Meistrich et al. 1985, Yan et al. 2003). Despite such a compelling expression pattern, an absence of H1T in the H1t knockout line did not result in male infertility, suggesting that H1T does not significantly affect chromatin structure or that there is redundancy with other H1 testis variants (Lin et al. 2000).
HILS1 is another H1 linker histone variant that is detected within elongating spermatids. HILS1 protein localisation within elongating spermatids overlaps only minimally with H1T, but substantially with the TPs and PRMs (Yan et al. 2003, Iguchi et al. 2004). At least one study indicates that decreased HILS1 expression in sperm is associated with asthenozoospermia (Jedrzejczak et al. 2007).
| Transgenerational epigenetic inheritance |
|---|
The existence of transgenerational epigenetic inheritance has been most conclusively demonstrated using the agouti viable yellow (Avy) mouse model (Bultman et al. 1992, Morgan et al. 1999). The agouti gene is responsible for the production of yellow pigment in hair and is normally controlled by a promoter which is active at only one stage in the hair growth cycle resulting in a yellow band on black hair (Bultman et al. 1992). In the Avy line, a retrotransposon IAP has integrated upstream of the agouti promoter (Duhl et al. 1994; Fig. 4A). As a consequence, expression at the Avy locus is controlled by the LTR of the IAP retrotransposon. When the IAP retrotransposon is unmethylated, and therefore active, it overrides the endogenous promoter and results in constitutive agouti production which is visible as yellow fur. The IAP retrotransposon can be silenced (methylated), resulting in the production of agouti using the endogenous promoter and therefore brown fur. Genetically identical littermates carrying the Avy allele display variable expressivity of the agouti gene ranging from completely yellow to brown depending on the methylation state of the IAP in the promoter (Morgan et al. 1999; Fig. 4B). Indeed, the agouti gene within different cells can be differentially methylated to produce mottled mice. The coat colour phenotype of the mice can be classified by a trained observer as either yellow, mottled or agouti.
|
| Paternal effect genes in mice |
|---|
Specifically, when Smarca5 heterozygous males were mated to yellow Avy heterozygous females, heterozygous offspring were more likely to be brown than their wild-type littermates (Blewitt et al. 2005). This suggested an increased incidence of methylation at the IAP retrotransposon associated with the Avy allele in these offspring and is consistent with Smarca5 being an enhancer of variegation (Fig. 5). As an extension of this, Chong et al. (2007) observed that the range of coat colours of wild-type offspring produced from a Smarca5 heterozygous male was different to the range seen in wild-type offspring from a Smarca5 wild-type male (Table 2). The wild-type offspring produced from the heterozygous fathers were more likely to be yellow than the wild-type offspring produced from the wild-type fathers (Chong et al. 2007; Fig. 5). This is consistent with hypomethylation of the IAP retrotransposon in wild-type pups from heterozygous fathers and suggests that while the majority of paternally inherited epigenetic modifications are erased in the zygote, not all are.
|
Transgenerational paternal effects of another type have been observed in the Dnmt3l knockout mouse line. In the Dnmt3l knockout mouse line, when heterozygous males were mated to wild-type females, spontaneous X chromosome aneuploidy (XO) occurred with an increased frequency in offspring. Humans with an XO chromosome content have Turner's syndrome and are outwardly female, but have early degeneration of the ovaries and oestrogen insufficiency. They usually have short stature and may have a high arched palate, neck webbing, broad chest, cardiac and renal abnormalities, impaired cognitive skills and spatial intelligence (Skuse 2005).
When Dnmt3l heterozygous males were mated to wild-type females the frequency of XO offspring was 2–2.4%, depending on genetic background, compared with 0% from wild-type males (Chong et al. 2007). Out of the three Dnmt3l XO offspring studied, two had lost the paternal X or Y chromosome while one had lost the maternal X chromosome. The loss of the paternal X or Y chromosome suggested that the loss occurred either in the male germ line during spermatogenesis or post-fertilisation, while the loss of the maternal X chromosome suggested that the X chromosome is lost post-fertilisation (Chong et al. 2007). The occurrence of XO in the Dnmt3l mouse line suggests a previously unrecognised role for DNMT3L in regulating X chromosome alignment and segregation during meiosis and of the epigenetic state in general in the generation of aneuploid gametes. Aneuploidy in humans occurs in at least 5% of all live births and probably accounts for a large number of miscarriages (Lamb & Hassold 2004). These data suggest that hypomorphic epigenetic modifiers in a potential father may be a cause of sub-fertility. Clearly, this is an area of research that requires additional work.
| Evidence of transgenerational effects of environmental factors on the germ line in rodents |
|---|
While it is clear that exposure of the grandmother in these circumstances did have affects on the grandchild, these affects could all be directly due to the initial exposure; that is, the foetus and its primordial germ cells were all directly exposed via the mother's (or grandmother's) circulation. For a compound or incident to truly have a transgenerational affect it should affect at least the F3 generation. There is some evidence that such situations do occur. For example, exposure of pregnant mice to the carcinogen benzapyrene increased the risk of lung tumours in not only the F1 generation but also in the F2–F5 generations of mice (Turusov et al. 1990).
More recent studies have suggested transgenerational effects from the endocrine disruptor vinclozolin (Anway et al. 2005, Chang et al. 2006). Vinclozolin is a fungicide used on agricultural crops including in the wine industry and is an anti-androgenic (Kelce et al. 1994). Pregnant rats treated with vinclozolin at the time of gonadal sex determination (embryonic days 8–14) produced male offspring with reduced spermatogenic capacity which persisted in male offspring in the F2–F4 generations (Anway et al. 2005). When male and female progeny in the F1–F4 generations were maintained until adulthood they developed diseases or tissue abnormalities such as prostate disease, kidney disease, breast cancer, testis abnormalities and immune system abnormalities (Anway et al. 2006). These transgenerational disease states were only observed when colonies were maintained through the paternal line. When offspring were produced using an affected dam, subsequent generations of animals appeared normal.
The authors have proposed that these transgenerational affects were due to altered epigenetic state and involve aberrant regulation of the methyltransferases DNMT3A and DNMT3L (Anway et al. 2008). They propose that disease (and male sub-fertility) was the result of altered genome-wide methylation and as a consequence gene expression.
It is true to say that both the epigenetic mechanism and transgenerational consequences of vinclozolin are hotly debated. Nonetheless, these studies provide a fascinating area for future research with potentially widespread implications for public health.
| Evidence of transgenerational effects of environmental factors on the germ line in humans |
|---|
Assisted reproductive techniques (ART) have been associated with a statistically significant increased risk of AS and BWS in offspring (reviewed in Gosden et al. 2003). It is proposed that the frequency of imprinting disorders is increased as a consequence of in vitro culturing conditions, or chemical composition of the culture media, which can result in the loss of methylation on imprinted regions in either the oocyte or embryo as opposed to aberrant epigenetic state in male gametes (Khosla et al. 2001). However, as mentioned earlier, it has been recently shown that males with oligospermia have an increased frequency of defective methylation compared with normospermic males (Kobayashi et al. 2007). This raises the possibility that hypomethylated paternally imprinted genes may contribute to imprinting errors and disease in ART-conceived children.
Similar to the exposure of rodents to dioxins or caloric restriction, in the human populations there have been many documented cases of changes in sex ratios following chemical exposure or after a natural disaster which may be consistent with altered epigenetic state across a population. For example in 1976, the population of Seveso in Italy were exposed to the herbicide dioxin. Exposed males produced a higher percentage of female offspring (65%) compared with male offspring (35%; Mocarelli et al. 1996). Pesticide workers in several countries are still exposed to high levels of dioxin. An investigation conducted on Russian pesticide workers also found that male exposure to dioxin was associated with an increase in the number of female births (62%; Ryan et al. 2002). It was proposed that exposure to dioxin affected sperm production in the testis leading to changes in the sex ratio (Jongbloet et al. 2002).
Similarly, after the Kobe earthquake in Japan (Fakuda et al. 1998), the 10-day war in Slovenia (Zorn et al. 2002) and in California following the terrorist attacks of September 11 (Catalano et al. 2005) a decline in the sex ratio (male:female) at birth was observed. It has been proposed that stress related to a catastrophic event may have a negative effect on sperm production or function resulting in a decline in sex ratio (Fakuda et al. 1998, Zorn et al. 2002). Alternatively, it could be interpreted that these studies reveal paternal effects similar to those documented in rodents (Anway et al. 2005, Ikeda et al. 2005). It is of note that within these populations there is no documented evidence of elevated levels of sex reversal or infertility, suggesting that the transmission ratio distortion is a true change in X- versus Y-bearing sperm.
| Conclusion |
|---|
The studies discussed in this review suggest that environmental factors may influence the epigenetic state and that these epigenetic modifications may be inherited through the male germ line and passed onto more than one generation. The paternal effects described in this review have further highlighted the importance of research into epigenetic regulation and male fertility. The concept that untransmitted alleles passed through the male germ line can affect the phenotype of the next generation is a new and exciting area of research. The transgenerational and paternal effects displayed by these mice may be related to idiopathic cases of sporadic disease or infertility in humans.
| Declaration of interest |
|---|
| Funding |
|---|
Received 23 December 2007
Revision received 15 February 2008
1 April 2008.
Accepted 14 May 2008
| References |
|---|
Aapola U, Maenpaa K, Kaipia A & Peterson P 2004 Epigenetic modifications affect Dnmt3L expression. Biochemical Journal 380 705–713.[CrossRef][ISI][Medline]
Agger K, Christensen J, Cloos PA & Helin K 2008 The emerging functions of histone demethylases. Current Opinion in Genetics & Development 18 159–168.[ISI][Medline]
Alberts B, Bray D, Lewis J, Raff M, Roberts K & Watson JD In Molecular Biology of the Cell 1989New York:Garland Publishing, Inc:.
Alexeev A, Mazin A & Kowalczykowski SC 2003 Rad54 protein possesses chromatin-remodeling activity stimulated by the Rad51-ssDNA nucleoprotein filament. Nature Structural Biology 10 182–186.[CrossRef][ISI][Medline]
Allis CD, Jenuwein T & Reinberg D 2007 Overview and conceptsCD Allis, T Jenuwein & D Reinberg In Epigenetics 1New York:Cold Spring Harbour Laboratory Press:.
An W 2007 Histone acetylation and methylation: combinatorial players for transcriptional regulation. Sub-Cellular Biochemistry 41 351–369.[Medline]
Anway MD, Cupp AS, Uzumcu M & Skinner MK 2005 Epigenetic transgenerational actions of endocrine disruptors and male fertility. Science 308 1466–1469.
Anway MD, Leathers C & Skinner MK 2006 Endocrine disruptor vinclozolin induced epigenetic transgenerational adult-onset disease. Endocrinology 147 5515–5523.[CrossRef][ISI][Medline]
Anway MD, Rekow SS & Skinner MK 2008 Transgenerational epigenetic programming of the embryonic testis transcriptome. Genomics 91 30–40.[CrossRef][ISI][Medline]
Baarends WM & Grootegoed JA 2003 Chromatin dynamics in the male meiotic prophase. Cytogenetic and Genome Research 103 225–234.[CrossRef][ISI][Medline]
Baarends WM, Wassenaar E, Hoogerbrugge JW, van Cappellen G, Roest HP, Vreeburg J, Ooms M, Hoeijmakers JH & Grootegoed JA 2003 Loss of HR6B ubiquitin-conjugating activity results in damaged synaptonemal complex structure and increased crossing-over frequency during the male meiotic prophase. Molecular and Cellular Biology 23 1151–1162.
Baarends WM, Wassenaar E, van der Laan R, Hoogerbrugge J, Sleddens-Linkels E, Hoeijmakers JH, de Boer P & Grootegoed JA 2005 Silencing of unpaired chromatin and histone H2A ubiquitination in mammalian meiosis. Molecular and Cellular Biology 25 1041–1053.
Baarends WM, Wassenaar E, Hoogerbrugge JW, Schoenmakers S, Sun ZW & Grootegoed JA 2007 Increased phosphorylation and dimethylation of XY body histones in the Hr6b-knockout mouse is associated with derepression of the X chromosome. Journal of Cell Science 120 1841–1851.
Balhorn R, Reed S & Tanphaichitr N 1988 Aberrant protamine 1/protamine 2 ratios in sperm of infertile human males. Experientia 44 52–55.[CrossRef][ISI][Medline]
Berger SL 2002 Histone modifications in transcriptional regulation. Current Opinion in Genetics & Development 12 142–148.[CrossRef][ISI][Medline]
Blewitt ME, Vickaryous NK, Hemley SJ, Ashe A, Bruxner TJ, Preis JI, Arkell R & Whitelaw E 2005 An N-ethyl-N-nitrosourea screen for genes involved in variegation in the mouse. PNAS 102 7629–7634.
Bourc'his D & Bestor TH 2004 Meiotic catastrophe and retrotransposon reactivation in male germ cells lacking Dnmt3L. Nature 431 96–99.[CrossRef][ISI][Medline]
Bultman SJ, Michaud EJ & Woychik RP 1992 Molecular characterization of the mouse agouti locus. Cell 71 1195–1204.[CrossRef][ISI][Medline]