The Role of the Cytoskeleton in Oocyte Maturation
The process of oocyte maturation is complex and still not well understood in humans. New interest in human oocyte maturation has been stimulated, as infertility and subfertility is increasing worldwide and new technological approaches are in demand to assist couples who are seeking assisted reproductive technology (ART) in IVF clinics.
Oocyte maturation refers to the time spanning germinal vesicle breakdown (GVBD) to oocyte arrest in metaphase of second meiosis (MII) (reviewed in Voronina and Wessel 2003; Fan et al. 2003, 2009; Brunet and Maro 2005; Liang et al. 2007; Jones, 2008; Swain and Pool 2008; Ai et al. 2008a,b, 2009; Gosden and Lee, 2010; Schatten and Sun, 2011b). In humans, as in most mammals, oocytes are arrested at birth in the ovary at the diakinesis stage of prophase I, the germinal vesicle (GV) stage. At puberty, follicle-stimulating hormone (FSH) induces development of small antral follicles into the preovulatory stage, followed by one oocyte and its follicle to become ovulatory. Stimulation by luteinizing hormone (LH) surge results in germinal vesicle breakdown (GVBD) and the beginning of meiotic resumption.
Accurate oocyte maturation is critically important for the oocyte to achieve fertilization competence and developmental potential (Eppig 1996; Kim et al., 1998; Combelles et al., 2002; Sirard et al., 2006; Liu et al., 2010a; Nogueira et al., 2012; Mao et al., 2014). The process includes nuclear and cytoplasmic maturation, referring to chromatin organization from GVBD to meiosis II (nuclear maturation; reviewed in detail in Schatten and Sun 2011a,b; 2015a,b) and to numerous processes in the ooplasm to prepare the oocyte for activation and preimplantation development (cytoplasmic maturation; Liu et al., 2010a; reviewed in detail in Sirard et al., 2006; Nogueira et al., 2012; Mao et al., 2014; Combelles et al., 2016) which includes redistribution of organelles such as cortical granules and mitochondria, storage of mRNAs, proteins, lipids, and transcription factors needed for subsequent development. Although these events are frequently addressed separately nuclear maturation and cytoplasmic maturation are interdependent and well-coordinated to achieve successful oocyte maturation.
Protein phosphorylation and dephosphorylation are required for several aspects during oocyte maturation, including the formation of the meiotic spindle during meiosis I (MI) and II (MII) (reviewed in Swain and Pool, 2008; Ai et al., 2009; Fan et al., 2009; Gosden and Lee, 2010). Meiotic spindle formation and maintenance of spindle integrity is achieved by a complex set of regulatory kinases including PKA, AKT, MAPK, polo and Aurora A kinases (Barr and Gergely 2007), and cdc2/cyclin B kinase (Jackman et al., 2003), CaMKII, the phosphatases CDC5, CDK14s, and others that participate in the meiotic process. Failure or inaccuracies in nuclear and cytoplasmic maturation becomes manifested in absence or failure of MI and MII spindle formation with consequences for chromosomal segregation errors (reviewed in Miao et al., 2009a; Wang et al., 2011; Qiao et al., 2014).
Two major regulatory components are required for MI and MII spindle formation and dynamics, which includes the maturation promoting factor (MPF; a CDK1/cyclin B1 complex) and the mitogen-activated protein kinase (MAPK). Decreased activity of MPF is associated with spindle instability and premature chromosomal centromere separation (Jeffreys et al., 2003; Pellestor et al., 2003) as well as centrosome instability (Miao et al., 2009a,b; Wang et al., 2011; Qiao et al., 2014), resulting in an inability of centrosomes to properly interact with microtubules, thereby increasing the probability of aneuploidy in zygotes. We will address spindle stability/instability in Section 10.4 on oocyte aging.
In humans, it is estimated that about 15-20% of oocytes undergo chromosomal segregation errors (Pellestor et al., 2005; reviewed in Miao et al., 2009a; Wang et al., 2011; Qiao et al., 2014) and that 5% of all pregnancies are aneuploid (Hassold and Hunt, 2001), which in many cases is related to molecular defects during oocyte maturation. The Society for Assisted Reproductive Technologies (SART) reported in 2012 that the implantation rate even in women under 35 years of age is lower than 40% and it is well known that the rate decreases with increasing age of the woman. While meiotic chromosome segregation has been studied well on cell and molecular levels (reviewed by Jones, 2008; Holt and Jones, 2009; Wang et al., 2011; Qiao et al., 2014) and is not included in the present review, our knowledge of cell and molecular mechanisms underlying cytoskeletal dynamics during meiosis is still limited. Animal models have allowed some insights into microtubule and microfilament dynamics but research on human meiosis has been slow, in part because of the limited research material and also in part because many laboratories had focused on the mouse model but it has become clear that the mouse is an inadequate model for understanding human oocyte meiosis (and fertilization, as detailed in Section 10.5), as the mouse uses entirely different mechanisms compared to humans and other non-rodent mammalian species (Schatten et al., 1985, 1986; Schatten and Sun 2011a,b; 2015a,b). For these reasons and others, it is important to distinguish between rodent and non-rodent mammals rather than referring to “mammalian oocytes" or “mammalian fertilization". However, because of the large research community that is studying reproduction in the mouse, research on other species is still not pursued as much as needed. More recently, several groups have now focused on the porcine and bovine systems and important new insights have been gained from research on these non-rodent mammalian species allowing close analogies with the human (Prather 1993, 2007; Schatten and Sun 2009a,2009b; Schatten and Sun 2011a,b; 2015a,b). Several excellent papers have been published on human oocyte maturation (Kim et al., 1998; Combelles et al., 2002; Liu et al., 2010b; Nogueira et al., 2012; Mao et al., 2014; Combelles 2016), which has increased our knowledge resulting in new possibilities for improved human oocyte maturation culture. These studies were aimed at more accurately mimicking natural in vivo oocyte maturation. A number of previous studies had employed experimental approaches such as including the microtubule drug taxol. However, taxol interferes with the dynamic instability process of microtubule biochemistry by promoting microtubule polymerization while preventing microtubule depolymerization, thereby artificially creating microtubules that are not present under normal non-experimental conditions, perhaps influencing accurate assessment and interpretation of microtubule dynamics. For example, cytoplasmic asters are induced by taxol in non-rodent mammalian oocytes that are not present in non-treated non-rodent mammalian oocytes (Schatten and Sun 2011a,b; 2015a,b). They are normally present in mouse oocytes that do use a different mechanism for microtubule organization and functions. They are also present in non-rodent aging oocytes, as will be discussed in Section 10.4.
Newer studies have focused on live cell imaging using different imaging modalities which has provided a better understanding of human oocyte maturation. These studies have revealed some new insights into cytoskeletal dynamics in the process of oocyte maturation, meiosis, cell division, and early preimplantation human development (Wang et al., 2001a,b,c; Cohen et al., 2004; Mandelbaum 2000; Wong et al., 2010; Miao et al 2012; Rienzi et al., 2012; Li and Albertini 2013; Holubcova et al., 2015). These studies have been complemented by fixed cell analysis of control and experimentally treated oocytes (Kim et al., 1998; Combelles et al., 2003; Goud et al., 2004; Chatzimeletiou et al., 2005, 2008; Hussein et al., 2006; Christopikou et al., 2010; Xu et al., 2011; Alvarez-Sedo et al., 2011; Huang et al., 2012; Nogueira et al., 2012; Coticchio et al., 2014, 2015). A number of these studies have recognized the importance of centrosomes for oocyte maturation and MI and MII spindle formation.
There are still different opinions about the terminology used for meiotic centrosomes in oocytes which has been discussed in detail in previous reviews (Schatten and Sun 2011a,b; 2015a,b). We use the term acentriolar centrosomes, as it best reflects the functions that are similar to centrosomes in mitotic cells except that centrioles are absent in meiosis (Calarco-Gillam et al., 1983; Calarco 2000; Lee et al., 2000; reviewed in Schatten 2008, Schatten and Sun 2009a,b, 2010; Schatten et al., 2011a,b). As is the case for microtubule dynamics, the mechanisms underlying centrosome dynamics in humans are similar to those in non-rodent systems such as the pig, cow, and horse (reviewed in Schatten and Sun 2015a), which can serve as suitable animal models for humans while non-rodent mammals such as the mouse use different centrosome dynamics and mechanisms for centrosome organization as well as different compositions of centrosomal proteins to comprise the centrosome in different cell cycle stages. For example, Lee et al., (2000) showed different NuMA distribution patterns in rodent (mouse) and non-rodent (pig) species and other differences have also been noted (reviewed in Schatten and Sun 2011a, 2015a). Therefore, similar to microtubule dynamics, in regard to centrosome dynamics the mouse is not an adequate animal model for comparison studies with humans.
Cytoskeletal formation in non-rodent mammalian oocytes after GVBD begins when microtubules form and start to get organized into a centrally located meiotic spindle that then migrates to the oocyte cortex and becomes anchored to the cortex by an actin filament cap (Kim et al., 1998; reviewed in Ai et al., 2009; Sun and Kim 2013). The formation of the spindle at the oocyte center has been described for several animal species including the mouse (Can et al., 2003; Brunet and Maro, 2005), the pig (reviewed in Ai et al., 2008a,b; Fan et al., 2003), and several others (reviewed in Ai et al., 2009). In all species it has been shown that spindle migration and anchoring to the oocyte cortex requires microfilament and microtubule functions. As first shown by Kim et al. (1998), the formation of a meiotic spindle in human oocytes starts when a small microtubule aster is formed near the condensed chromatin after GVBD. The aster enlarges around the condensed chromatin before forming the complete meiotic spindle. After translocation to the oocyte cortex microtubules are organized into the MI and subsequent MII spindle. The paper by Kim et al. (1998) describes the 18-21 ^m-sized MII spindle in human oocytes as a symmetric, barrel-shaped structure containing broad anastral poles. The MII human oocyte spindle is located peripherally and displays a perpendicular orientation (Figure 10.1), which can easily be visualized by non-invasive Polscope microscopy (Wang et al., 2001a,b,c; Cohen et al., 2004). Microfilaments are detected around the cell cortex as a thick uniform area and they are further found near the GV, surrounding the female chromatin after GVBD. The paper by Kim et al., (1998) also discussed the role of centrosomes in oocyte maturation that had been described in a previous paper by Battaglia et al., (1996). However, in the paper by Battaglia et al., (1996) taxol was used that most likely influenced the organization and dynamics of microtubules and centrosomes, as addressed earlier.
Figure 10.1 (a-d) From left to right: (a) MII meiotic spindle in unfertilized human oocyte is oriented perpendicularly to the oocyte surface. Insets: enlarged meiotic spindle and enlarged oocyte centrosome without centrioles (acentriolar centrosomes) of one meiotic pole. (b) Sperm incorporation. Inset shows sperm centrioles before fertilization. After fertilization the sperm aster forms from the proximal centriole of the centriole pair while the distal centriole disintegrates.
(c) Zygote aster formation and two apposed pronuclei. (d) Mitotic apparatus of first embryonic cell division. Inset: enlarged mitotic apparatus and enlarged centrosome of one mitotic pole containing centrioles. (See color plate section for the color representation of this figure.)
Figure 10.1 (Continued)
While we do not yet have a full understanding of centrosome dynamics in human oocytes after GVBD, some data have been generated in the mouse, porcine, and bovine models as well as in human oocytes. Our understanding on centrosome formation in human oocytes is especially poor and we do not yet have sufficient information on the formation and specific role of centrosomes and specific centrosome proteins in maturing human oocytes. Much of our basic knowledge of centrosome biology is still mainly derived from somatic cells (reviewed in Schatten 2008). The centrosomal proteins у-tubulin, pericentrin and the centrosome-associated Nuclear Mitotic Apparatus (NuMA) protein have been studied to some extent and their essential roles in MI and MII spindle formation has been reported (George et al., 1996; Alvarez-Sedo et al., 2011). These proteins require nuclear and cytoplasmic regulation. As outlined in the introduction and reviewed previously (Schatten 2008; Schatten and Sun 2011a,b; 2015a,b), у-tubulin is a component of the centrosome core structure and has a major role in microtubule nucleation (reviewed in Moritz et al., 2004; Stearns, 2004). In addition to its association with the centrosomal core structure, у-tubulin is also distributed throughout the ooplasm. Pericentrin, along with у-tubulin, organizes microtubules into the spindles (Doxsey et al., 1994; Dictenberg et al., 1998; Young et al., 2000). It depends on the microtubule motor protein dynein for assembly onto centro- somes (Young et al., 2000). NuMA is both a nuclear protein and a centrosome-associated protein in meiotic and mitotic spindles (reviewed in Sun and Schatten, 2006). In human oocytes, we have shown that NuMA is dispersed across the nucleoplasm in GV-stage oocytes and disperses into the ooplasm after GVBD to relocate to the meiotic spindle poles in MI and MII oocytes (Alvarez-Sedo et al., 2011). We further found that aberrant NuMA localization patterns were associated with aberrant in vitro maturation. After IVF, NuMA was localized to the pronuclei. We further showed that NuMA translocation from the GV to the meiotic spindles is essential for the maturation process and developmental potential of the embryo (Alvarez-Sedo et al., 2011). These experiments and others indicate that human oocyte maturation does depend on accurate composition of centrosomes in which NuMA plays a critical role. Other studies have shown that accurate association of у-tubulin with the meiotic spindle poles is essential for accurate spindle formation in human oocytes (George et al., 1996) and that dispersion of у-tubulin from the spindle poles is characteristic for aging oocytes or for oocytes that are not competent to be fertilized. We have obtained similar results for pig oocytes (Miao et al., 2009a,b). The mechanisms underlying oocyte aging are discussed in more detail in Section 10.4. Recently, a paper has been published in which у-tubulin could not be detected at the MI and MII spindle poles in human oocytes (Holubcova et al., 2015), which is different from other studies on human oocytes in which centrosomes were clearly detected as indicated by у-tubulin staining (George et al., 1996). The differences in results may be related to different culture conditions, different preparation methods, or other unknown factors. It is also worth mentioning that absence of antigen immunostaining does not necessarily translate to absence of antigens. The study by Holubcova et al. (2015) did not include antigen recovery experiments to determine possible masking of antigens as explanation for the absence of у-tubulin detection in human oocyte meiosis. This study suggested that spindle assembly in human oocytes is mediated by chromosomes and dependent on Ran-GTP. This finding falls in line with previous studies in different cell systems suggesting that chromatin-bound Ran-GEF, RCC1, is able to catalyze the Ran-GDP/Ran GTP transition and to generate high local concentration of Ran-GTP around chromosomes that could play a role in microtubule growth (reviewed in Schatten and Sun, 2011a). Interestingly, there may be differences for MI and MII spindles, as Dumont et al. (2007) reported a Ran-GTP-independent spindle assembly pathway for MI Xenopus and live mouse oocytes but MII spindle assembly required Ran-GTP. The mechanisms for microtubule nucleation and organization can be diverse and it is likely that centrosomes and Ran-GTP-mediated microtubule formation takes place and is well coordinated in human oocytes. In Drosophila (Moritz et al., 1995) and Xenopus egg extracts (Zheng et al., 1995), у-tubulin served as microtubule-nucleating component aided by Ran-GTP in the vicinity of chromosomes during the early stages of meiotic spindle formation. The mechanisms for spindle assembly in different species are reviewed in Schatten and Sun (2011a).
New research has focused on determining more optimal culture conditions for human oocyte maturation to increase the success rates in IVF clinics (Christopikou et al., 2010; Combelles 2016). We do not yet have detailed comparison studies of in vivo compared to in vitro human oocyte maturation; studies in other species have shown that significant differences exist regarding cytoskeletal dynamics and organization for in vivo and in vitro obtained oocytes (Combelles et al., 2003).