Introduction
Reproductive activity is theoretically possible from puberty to senescence; however, senescence declines this activity in all animals studied up to its complete interruption (Harman, Reference Harman2006). Cattle show a large variability in the age to conceive and being able to bring to term a pregnancy. Unfortunately, the low heritability (h2 = 0.09) for traditional fertility parameters makes the discrimination of more fertile genotypes difficult; consequently, selection for fertility is frustrated (Darwash et al., Reference Darwash, Lamming and Woolliams1997). Erickson et al. (Reference Erickson, Reynolds and Murphree1976) analysed reproductive performance of aged Hereford cows and established that 15 years old was the age at which over 50% of the herd became infertile. This threshold may decrease under stressful conditions, as in the case of high-producing dairy cows.
Cattle have been proposed as a suitable model for studying reproductive aging in humans as follicular wave emergence, follicle selection, ovulation, and age-associated endocrinal changes are similar in the two species (Baerwald et al., Reference Baerwald, Adams and Pierson2003; Malhi et al., Reference Malhi, Adams and Singh2005; Malhi et al., Reference Malhi, Adams, Pierson and Singh2006). Although age may affect many aspects of female reproduction, oocytes have been assumed as a major focus in reproductive aging studies. Malhi et al. (Reference Malhi, Adams, Mapletoft and Singh2007) analysed the superovulatory efficiency in old versus young cows and observed the comparative recovery of fewer embryos and a greater proportion of unfertilized oocytes in old cows. Therefore, they suggested that there was a reduced developmental competence of oocytes in old cows; although, there was no difference in embryo quality and in the proportion of calves born. Yamamoto et al. (Reference Yamamoto, Iwata, Goto, Shiratuki, Tanaka, Monji and Kuwayama2010) found that oocytes from old cows had a lower fertilizing ability and showed a faster meiosis progression and a rapid disappearance of gap junctions between oocytes and cumulus cells compared with those derived from young cows. This finding suggested that the premature meiosis progression in old cow oocytes was partly due to impaired cumulus–oocyte complex communication.
DNA damage has been demonstrated as one of the main consequences of aging (Bertram & Hass, Reference Bertram and Hass2008; Schumacher et al., Reference Schumacher, Garinis and Hoeijmakers2008; Maynard et al., Reference Maynard, Schurman, Harboe, de Souza-Pinto and Bohr2009). Genome integrity is an essential condition for assuring cell functions and division. In order to maintain this prerequisite, several DNA repair strategies have been established in cells (Bessho et al., Reference Bessho, Roy, Yamamoto, Kasai, Nishimura, Tano and Mitra1993; Wood et al., Reference Wood, Mitchell, Sgouros and Lindahl2001). Different types of damage, i.e. wrong nucleotide insertion during DNA replication, chemical modifications of nitrogenous bases, spontaneous hydrolysis of N-glycoside binding, pyrimidine alteration due to reactive oxygen species (ROS), ultraviolet (UV) and ionic radiations, chemotherapy drugs, genotoxic substances, oxidation or alkylation base, occur daily within the DNA of each individual (Wood et al., Reference Wood, Mitchell, Sgouros and Lindahl2001). These alterations are counteracted by DNA repair systems, such as nucleotide excision repair (NER), base excision repair (BER), mismatch repair (MMR) and double-stranded break (DSB) repair in DNA. Under genotoxic alert, these systems interfere with the cell cycle and, in some cases, activate specific checkpoints in order to repair the damage, or induce apoptosis when cytoprotection cannot be achieved (Jaroudi & SenGupta, Reference Jaroudi and SenGupta2007). Cell lifespan depends on the efficacy of DNA repair and aging has been proposed as a consequence of declining DNA repair activity (Best, Reference Best2009). In fact, the silencing of genes involved in DNA repair causes an accelerated aging and an early manifestation of aging-associated diseases in mouse (de Boer et al., Reference de Boer, Andressoo, de Wit, Huijmans, Beems, van Steeg, Weeda, van der Horst, van Leeuwen, Themmen, Meradji and Hoeijmakers2002; Espejel et al., Reference Espejel, Klatt, Ménissier-de Murcia, Martín-Caballero, Flores, Taccioli, de Murcia and Blasco2004). DNA repair is supposed to be one of the crucial activities established during gamete growth and maturation, fertilization and embryo development (El-Mouatassim et al., Reference El-Mouatassim, Bilotto, Russo, Tosti and Menezo2007; Ménézo et al., Reference Ménézo, Dale and Cohen2010). The role played by genotoxic stress has been deeply studied in somatic and differentiated cell types; however, little information is available on the activity of DNA repair genes during gametogenesis and embryo growth. In rhesus monkey, the expression of APEX-1, a component of BER genes, did not show significant variations in either immature (germinal vesicle stage, GV) or mature (metaphase II stage, MII) oocytes (Zheng et al., Reference Zheng, Schramm and Latham2005). MLH1, which is a component of the MMR gene family, is highly expressed in growing and maturing oocytes, but its expression decreases during early embryo development (Zheng et al., Reference Zheng, Schramm and Latham2005). Genes related to NER seem to be weakly expressed in oocytes (Zheng et al., Reference Zheng, Schramm and Latham2005; Jaroudi & SenGupta, Reference Jaroudi and SenGupta2007). In monkey, some DNA repair genes have been found to be up-regulated in oocytes in respect to early embryos (Zheng et al., Reference Zheng, Schramm and Latham2005). This observation may represent an adaptive strategy that allows oocytes to remain arrested for long time at first meiotic division to guarantee genome integrity (Jaroudi & SenGupta, Reference Jaroudi and SenGupta2007). RAD51, involved in the DSB repair mechanism (Hsieh, Reference Hsieh2001; Kelley & Parsons, Reference Kelley and Parsons2001), is an essential gene for embryo vitality and is up-regulated during embryo development (reviewed in Ozturk & Demir, Reference Ozturk and Demir2011). Several studies on mammalian germ cells have established a close relationship among DNA repair failure, embryo mortality, infertility and aging (Hart & Setlow, Reference Hart and Setlow1974; Grube & Bürkle, Reference Grube and Bürkle1992; Cortopassi & Wang, Reference Cortopassi and Wang1996; Hsieh, Reference Hsieh2001). However, to date, studies on the involvement of DNA repair systems in mammalian gametes are scarce and most of these are limited to monkeys and mice.
In this context, we hypothesized that defects in DNA repair gene expression could represent key factors affecting the donor age-related oocyte quality and fertility in cattle. Here, we present the transcriptional profile of three genes, APEX-1, RAD51, and MLH1, belonging to BER, DSB and MMR pathways, respectively. We compared their expression in bovine immature and in in vitro matured oocytes obtained from sexually mature heifers or old cows.
Materials and methods
If not otherwise stated, all chemicals were purchased from Sigma-Aldrich (Milano, Italy).
Oocyte source
Ovaries from slaughtered 17 ± 3-month-old Holstein Friesian heifers or from 94 ± 20-month-old Holstein Friesian cows were collected separately from the abattoir and transported in a thermal bag at 27–30°C to the laboratory within 3–4 h from collection. The laboratory temperature was 30°C. Immature oocytes were collected from 2–8 mm follicles by an 18-gauge needle under controlled pressure (50–70 mm Hg). Cumulus–oocyte complexes (COCs) with suitable morphological characteristics (Boni et al., Reference Boni, Cuomo and Tosti2002) were isolated from the follicular fluid and washed three times with TCM-199 supplemented with 0.05% (w/v) polyvinyl alcohol and 10 mM HEPES. The COCs were then divided randomly into homogeneous groups for in vitro maturation (IVM) or immediately prepared for treatment for RNA extraction in Trizol (Life Technologies, Monza, Italy). In the latter case, oocytes were freed from the cumulus cells by vortexing for 4 min. The resultant intact oocytes without cumulus traces were counted under a stereomicroscope and transferred into tubes with the lowest quantity of washing medium. Trizol was added to the tubes (1 μl Trizol/oocyte) and samples immediately stored at –80°C until analysis.
In vitro maturation
The COCs destined for IVM were transferred into 4-well maturing plates (Nuclon, Nunc, Roskilde, DK) containing a maturation medium (MM) that consisted of TCM-199 supplemented with 10% fetal bovine serum (FBS), 10 IU/ml luteinising hormone (LH), 0.1 IU/ml follicle-stimulating hormone (FSH), 1 μg/ml of 17β-estradiol (approximately 30 μl MM/COC). The plates were then transferred to an incubator at 39.0°C in 5% CO2 humidified air for 24 h. At the end of the incubation period, COCs were placed in 15-ml tubes that contained 1 ml of washing medium and vortexed, as described above. The IVM efficiency was assessed in retrieved oocytes by evaluating the presence of a polar body. The MII oocytes were counted and transferred in Trizol for RNA extraction.
RNA extraction, reverse transcription and real-time PCR quantification
Total RNA extraction from at least three independent pools (30–50) of oocytes was performed using the Trizol reagent. An amount of 1–4 μg glycogen was added as carrier to facilitate RNA precipitation. RNA was resuspended in an appropriate volume of RNase-free water. Each aliquot of extracted RNA was subjected to treatment with RNase-free DNase using 1 U/μl (BD-Clontech, Milan, Italy) in order to eliminate any possible contamination of residual genomic DNA. cDNA synthesis was performed by using the SuperScript® VILO™ cDNA Synthesis Kit (Life Technologies) with random hexamers, following the manufacturer's protocol in a final volume of 20 μl and carried out on five oocytes equivalent RNA. Real-time qPCR analysis was performed on randomized cDNA equivalent to 1.33 oocyte using the SYBR Green PCR Master Mix (Applied Biosystems, Monza, Italy) in a final volume of 25 μl according to the manufacturer's protocol in a 7000 Real-Time PCR System (Applied Biosystems). Quantification was in triplicate and samples for each gene were analysed on the same run in order to minimize inter-experiment variation. The expression levels of candidate genes were normalised using the internal standard glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as reported previously (Vallée et al., Reference Vallée, Robert, Méthot, Palin and Sirard2006). GAPDH was stable in both immature and in vitro maturated oocytes (data not shown; Somfai et al., Reference Somfai, Imai, Kaneda, Akagi, Watanabe, Haraguchi, Mizutani, Dang-Nguyen, Inaba, Geshi and Nagai2011). The gene sequences used in the experiments have been retrieved from GenBank. Primers were designed using the Plexor™ Primer Design system (http://www.promega.com) and sequences reported in Table 1. The amplification protocol included: denaturation (95°C for 10 min), 40 amplification cycles (15 s at 95°C, 1 min at 60°C). The specificity of each amplification reaction was verified by melting curve analysis. The ΔΔCt method was used to calculate the expression level of each gene (Pfaffl, Reference Pfaffl2001) and PCR efficiency was 83–95% (calculated from the slope of a standard curve based on the formula: efficiency = –1 + 10(–1/slope)).
qPCR primer sequences for bovine APEX-1, MLH1 and RAD51 genes
*Vallèe et al., Reference Vallée, Robert, Méthot, Palin and Sirard2006.
Statistical analysis
Values are presented as means ± standard deviation (SD). Differences in gene expression between the age and meiotic stage groups were analysed by analysis of variance (ANOVA) (Systat 11.0 release). Pairwise comparison of the means was made with Fisher's least significant difference (LSD) test. COVARIANCE procedure was applied in order to analyse the age and the meiotic stage effects within each gene transcript singly, as well as to compare the relative transcript abundance between the analysed genes independently of age and meiotic stage effects. A P-value < 0.05 indicated significance.
Results
This study analysed data of combined batches, including three replicates of 30–50 bovine oocytes for a cumulative number of 230 at the immature (92 and 138 from young and old individuals, respectively) and 217 (87 and 130 from young and old individuals, respectively) at the in vitro matured stage. We did not find any significant difference in maturation rate between young and old oocytes on the basis of polar body evaluation (80 ± 7% versus 76 ± 6% in young and old oocytes, respectively). The three transcripts analysed were differently expressed in bovine oocytes, independently from meiotic stage and donor age. The relative amount measured indicated the following order of abundance: RAD51 > APEX-1 > MLH1 (P < 0.01). However, the amount of APEX-1 and MLH1 transcripts did not significantly differ between them. At the end of IVM, all transcripts were up-regulated in comparison with the immature stage (0.117 ± 0.05 vs 0.505 ± 0.05; P < 0.01, data not shown). The relative abundance of each single transcript increased after IVM from two-fold (MLH1), up to four-fold (APEX-1) compared with the immature stage. In contrast in the meiotic stage, the effect of donor age did not produce homogeneous changes of gene transcript patterns, hence when the analysis was carried out on overall data, no significant differences were found between the two age classes (0.267 ± 0.047 versus 0.355 ± 0.058; P > 0.05, data not shown).
APEX-1 transcript patterns are shown in Fig. 1 in relation to the age and meiotic stage. An up-regulation of gene transcripts following in vitro maturation was clearly found in oocytes from young animals (P < 0.01), but it was not statistically significant (P = 0.066) in oocytes from old animals. Similarly, MLH1 transcripts showed a significant increase (P < 0.02) after in vitro maturation in oocytes from young donors, but not in old animals (Fig. 2). In the case of RAD51, oocytes from both young and old donors showed a significant increase of transcript abundance (Fig. 3). In addition, in immature oocytes a significant (P < 0.01) difference between the two age classes was found with an up-regulation of this gene in the old donor cows.
Relative expression profiles of APEX-1 gene in immature (GV) and in vitro matured (MII) bovine oocytes in relation to the donor age. The mRNA abundance of the candidate gene was normalised against the internal standard (GAPDH). Relative abundance values are expressed as 2^–ΔCt and show the mean value ± standard deviation (SD). A,B indicates P < 0.01.
Relative expression profiles of MLH1 gene in immature (GV) and in vitro matured (MII) bovine oocytes in relation to the donor age. The mRNA abundance of the candidate gene was normalised against the internal standard (GAPDH). Relative abundance values are expressed as 2^–ΔCt and show the mean value ± standard deviation (SD). a,b indicates P < 0.02.
Relative expression profiles of RAD51 gene in immature (GV) and in vitro matured (MII) bovine oocytes in relation to the donor age. The mRNA abundance of the candidate gene was normalised against the internal standard (GAPDH). Relative abundance values are expressed as 2^-ΔCt and show the mean value ± standard deviation (SD). A,B indicates P < 0.01 between GV from the young and old groups; a,b indicates P < 0.05 between GV and MII from the old group. C,D indicates P < 0.01 between GV and MII from the young group.
Discussion
Data obtained in the present study showed a clear up-regulation of all three DNA repair genes evaluated, i.e. RAD51, APEX-1 and MLH1, in bovine oocytes after IVM. Among these genes, only the expression of RAD51 significantly increased in immature oocytes collected from old individuals compared with young ones. This finding suggests that RAD51 may represent a helpful candidate in discriminating bovine immature oocytes in relation to the donor age.
During oocyte maturation, differences in transcriptional activity result in either over-expression, underexpression or stable expression of genes that concur to establish the maternal cytoplasm inherence (Memili et al., Reference Memili, Dominko and First1998; Tomek et al., Reference Tomek, Torner and Kanitz2002). The complex genetic remodelling that occurs during oocyte maturation, in which homologous chromosome pairing, synapse and recombination take place, requires a perfectly active and functional DNA repair machinery. This factor may explain why all three transcripts (and possibly others) quantified in the present study increased their expression significantly from GV to MII. Conversely, following a dated, but still current, theory the adaptive advantage of meiosis consists in facilitating recombinational repair of DNA damage that is otherwise difficult to repair, and occurs especially as the result of oxidative stress (Bernstein et al., Reference Bernstein, Byerly, Hopf and Michod1985). In mouse, a massive destruction of transcripts has been described during oocyte maturation. Of the estimated 85 pg of mRNA present at the GV stage, polyadenylated mRNA declines during oocyte maturation to 50 pg; about half undergoes deadenylation, while the remaining part is subject to degradation (Paynton et al., Reference Paynton, Rempel and Bachvarova1988). Su et al. (Reference Su, Sugiura, Woo, Wigglesworth, Kamdar, Affourtit and Eppig2007) using the Affymetrix Mouse Genome 430 v2.0 GeneChip observed that this degradation is a selective process that involves transcripts associated with meiotic arrest at the GV stage and during oocyte maturation. An up-regulation of some genes that are likely to be involved in DNA replication, amino acid metabolism and signaling molecules has been described during oocyte maturation, with 1682 genes under-expressed and 1936 over-expressed out of 12,164 analysed in in vivo matured oocytes (Cui et al., Reference Cui, Li, Yin, Kong, Kang and Kim2007).
In bovine, Mamo et al. (Reference Mamo, Carter, Lonergan, Leal, Naib, McGettigan, Mehta, Evans and Fair2011) using the Affymetrix GeneChip Bovine Genome Array, analysed global mRNA expression in immature and in vitro matured oocytes. They found that approximately 25% of the 8489 analysed transcripts significantly changed their abundance from immature to in vitro matured stage. In particular, 72% of them were under-expressed and 28% were over-expressed throughout IVM. Microarray analysis was further supported by qPCR, which showed in all but one of the 25 analysed genes a similar expression pattern to the microarray data. Furthermore, treatment with α-amanitin significantly decreased transcript abundance for all the genes analysed (Leal et al., Reference Leal, Mamo, Fair and Lonergan2012). The same authors compared the transcriptional activity of 35 selected genes related to cell cycle regulation and oocyte competence in in vitro matured bovine oocytes after temporary inhibition of meiosis using the cyclin-dependent kinase inhibitor butyrolactone I. Also in this study, after IVM, most of the examined transcripts (54%) were down-regulated, whereas 20% of them were up-regulated in control and inhibited oocytes (Leal et al., Reference Leal, Mamo, Fair and Lonergan2012). More exhaustive information can be obtained when the transcriptional analysis is completed by protein expression. Recently, we analysed NFKBIA (NFκB inhibitor α) gene and its product, IκBα, during the oocyte-to-embryo transition period in bovine. Interestingly, during meiosis progression, we found a decrease in maternal NFKBIA mRNA and a parallel increase of IκBα, indicating a different regulation of the two processes (Paciolla et al., Reference Paciolla, Boni, Fusco, Pescatore, Poeta, Ursini, Lioi and Miano2011).
DNA repair genes should play a crucial role in the fertilized egg due to the extensive chromatin remodelling that occurs at the time of genome decondensation. This event may result in the modified expression of several members of the DNA repair machinery. Rad51 protein is involved in the homologous repair of DNA DSBs, playing a relevant role in the cell fate. Perez et al. (Reference Perez, Acton, Jurisicova, Perkins, White, Brown, Trbovich, Kim, Fissore, Xu, Ahmady, D'Estaing, Li, Kagawa, Kurumizaka, Yokoyama, Okada, Mak, Ellisman, Casper and Tilly2007) found that apoptosis in MII oocytes from inbred mouse strains is affected by genetic background, as demonstrated by numerous DNA DSBs in freshly isolated oocytes. Microinjection of Rad51 protein reduced DNA damage, suppressed apoptosis and improved embryonic development (Kujjo et al., Reference Kujjo, Laine, Pereira, Kagawa, Kurumizaka, Yokoyama and Perez2010). RAD51 was also related to the halting cell death programme induced by ionizing radiation in bovine oocytes (Kujjo et al., Reference Kujjo, Ronningen, Ross, Pereira, Rodriguez, Beyhan, Goissis, Baumann, Kagawa, Camsari, Smith, Kurumizaka, Yokoyama, Cibelli and Perez2012). In particular, these authors found that either inhibition of Rad51 or increase in caspase-3 activity, before irradiation, caused an increased cytoplasmic fragmentation and DNA damage in mouse and bovine oocytes. Conversely, both types of cell damage markedly decreased when recombinant human Rad51 was microinjected in oocytes. Bovine oocytes were more resistant than mouse oocytes to these environmental stimuli and showed a faster protection rate following Rad51 treatment. In the present study, RAD51 expression increased significantly both in in vitro matured oocytes and in aged donor oocytes. At the present time, it is not easy to formulate hypotheses on why only RAD51 expression increased in aged donors. The larger abundance of the RAD51 transcript may reflect a biological strategy towards conditions of high sensitivity to DNA damage. Future studies will be devoted to clarify and extend the primary observation reported in the present work. Our initial attempts to investigate this issue have been focused on the possibility that an increased accumulation of ROS in aged oocytes could represent the molecular requirement for the increased synthesis of the RAD51 transcript. However, preliminary observations obtained in our laboratory on oocytes treated with different concentrations of hydrogen peroxide seem to exclude this possibility (data not shown). Alternatively, a recent work suggests a novel role for human RAD51 in mitochondria. The authors provided the first demonstration that Rad51 exists in human mitochondria and interacts physically with mitochondrial DNA (mtDNA), which suffers significant damage resulting from the production of ROS via the oxidative phosphorylation pathway (Sage et al., Reference Sage, Gildemeister and Knight2010). Considering the importance of mitochondrial function in mammalian oocytes and embryos and their role in developmental competence (reviewed in Dumollard et al., Reference Dumollard, Carroll, Duchen, Campbell and Swann2009; Van Blerkom, Reference Van Blerkom2011), the hypothesis that high levels of RAD51 transcript and protein are required to ensure the maintenance of the mitochondrial genome during aging cannot be excluded, a concept that deserves further study in the near future.
Apex/Ref-1 (apurinic/apyrimidic endonuclease) is an enzyme involved in the DNA BER pathway to remove abasic sites, the most common type of DNA decay. These gene transcripts increase greatly following IVM, but did not differ between oocytes of young and old individuals. In a previous comparative study, we analysed the expression patterns of APEX-1 in human gametes and embryos and in sea squirt Ciona intestinalis (ascidian), an organism studied intensively in developmental biology and that has been proposed as a model to study meiotic regulation (Russo et al., Reference Russo, Wilding, Marino and Dale1998, Reference Russo, Bilotto, Silvestre, Tosti and Boni2011). qPCR analysis clearly indicated that APEX transcripts were clearly detectable in oocytes and embryos until the larva stage, but not in spermatozoa, a finding that suggested the appearance of paternal contribution to DNA repair during development (El-Mouatassim et al., Reference El-Mouatassim, Bilotto, Russo, Tosti and Menezo2007).
MLH1 belongs to the DNA MMR family that is involved in the process of meiotic recombination in germ cells. In particular, MMR gene mutations in mice result in meiotic disruption during prophase I (Edelmann et al., Reference Edelmann, Cohen, Kane, Lau, Morrow, Bennett, Umar, Kunkel, Cattoretti, Chaganti, Pollard, Kolodner and Kucherlapati1996). MMR function is required for the formation and stabilization of crossovers in mammalian oocytes (Hunter & Borts, Reference Hunter and Borts1997) and, in the absence of a functional MMR system, the failure to maintain chiasmata results in a reduced ability to proceed normally through the first and second meiotic divisions, despite near-normal levels of meiotic resumption after dictate arrest (Kan et al., Reference Kan, Sun, Kolas, Avdievich, Kneitz, Edelmann and Cohen2008). The larger transcript abundance of this gene in MII oocytes, found in the current study, supports its role during meiosis progression in bovine oocytes. This finding, however, conflicts with data from Mamo et al. (Reference Mamo, Carter, Lonergan, Leal, Naib, McGettigan, Mehta, Evans and Fair2011) who described an opposite pattern of MLH1 expression in bovine oocytes following IVM. There is no clear explanation for this discrepancy, which may be due to epigenetic effects associated with different IVM protocols used in the two studies.
Conclusions
The analysis of RAD51, APEX-1 and MLH1 gene expression reveals a clear up-regulation of these genes in bovine oocytes from the immature to the in vitro matured stage. Among these patterns, only RAD51 expression differs significantly in relation to donor age with an up-regulation in immature oocytes of aged donor.
On the basis of the role played by these genes, it is likely that several additional sources of variability affecting oocyte quality (size and grade of atresia of the follicle, environment, etc.) may interfere with their expression in order to protect the oocytes’ fate and allow fertility preservation.
Acknowledgements
We thank Dr Elisabetta Tosti for her helpful comments. Stefania Bilotto was supported by a grant from University of Basilicata (D.R. n. 534, 17/11/2009).
