Dear Editor,
Previously, the Mendelian phenotypes in human oocyte maturation arrest, fertilization failure and early embryonic arrest, are largely underestimated. In recent years, “missing” Mendelian phenotypes and genes in these processes are beginning to be uncovered by us and others (
Huang et al., 2014;
Alazami et al., 2015;
Feng et al., 2016;
Xu et al., 2016;
Chen et al., 2017;
Sang et al., 2019). However, the genetic basis for majority of patients resulting from abnormalities in these phenotypes remains to be elucidated.
The cell division cycle 20 (CDC20, HGNC: 1723) is the coactivator of anaphase-promoting complex/cyclosome (APC/C) during mitosis, and plays a role in maintaining the genome by regulating spindle assembly checkpoints (
Musacchio and Hardwick, 2002). In oocytes, the activation of APC/C by CDC20 is a key step in homologue disjunction and in transition from meiosis I to meiosis II (
Jones, 2011). CDC20 is therefore an essential component of the mammalian cell cycle mechanism regulating both miotic and meiotic exit. Although
CDC20 is an extensively studied gene, until now, no solid evidence has been provided to establish the causal relationship between
CDC20 mutations and human diseases. The only report was an association study relating
CDC20 mutations with idiopathic azoospermia (
Li et al., 2017).
Here, we identified biallelic CDC20 mutations in five infertile individuals with oocyte maturation arrest, fertilization failure, and early embryonic arrest. We investigated the effects of the corresponding mutations in cell lines and mouse oocytes and explored a potential therapeutic treatment by direct CDC20 cRNA injection.
The probands in families 1 and 2 had normal menstrual cycles and had been diagnosed as primary infertility with unknown reasons for several years (Table S1). Both of the probands came from consanguineous families (Fig. 1A). The probands in family 1 had undergone two failed in vitro fertilization (IVF) attempts and one failed intracytoplasmic sperm injection (ICSI). In her ICSI cycle, eleven oocytes were obtained, and ten oocytes were arrested at the germinal vesicle (GV) or metaphase I (MI) stage (Fig. 1B), only one was matured and fertilized, but was arrested at three-cell stage (Table S1). The proband in family 2 had undergone two failed ICSI cycles. All of the retrieved oocytes were arrested at the MI stage (Table S1). In summary, both two probands had phenotypes of oocyte maturation arrest in their ICSI attempts. Whole-exome sequencing and homozygosity mapping was performed (Fig. S1). By a recessive inheritance model, we identified two homozygous missense mutations (c.683A>G, p.Tyr228Cys and c.1316T>G, p.Leu439Arg) in CDC20, respectively (Fig. 1A). Because the parents of both probands were unavailable, in order to rule out the possibility that the homozygous status of CDC20 resulted from the deletion of one allele, we performed a copy number variation (CNV) analysis. There was no CNV for CDC20 in the probands of both families, which confirmed that the mutations in the two individuals were indeed homozygous (Fig. S2A and S2B).
Mutational screening of CDC20 were pursed in a large cohort of 1,250 infertile individuals with abnormalities in oocyte maturation, fertilization and early embryonic development by whole exome sequencing. Additional biallelic mutations in CDC20 were detected in another three individuals from families 3–5 (Fig. 1A). The compound heterozygous mutations in CDC20 in the affected individual in family 3 consisted of a missense mutation c.965G>A (p. Arg322Gln) and a nonsense mutation c.544C>T (p. Arg182*). The proband in family 4 carried a compound heterozygous mutation consisting of a missense mutation c.965G>A (p.Arg322Gln) and a 4 bp insertion c.813_814ins AGTG (p.Gly272Serfs*24). The proband in family 5 carried a compound heterozygous mutation consisting of a missense mutation c.965G>A (p.Arg322Gln) and a 4 bp deletion c.1176_1179del TCTG (p.Cys392*). Sanger sequencing were performed to confirm the mutations in these three families (Fig. 1A). Information about the mutations is shown in Fig. 1A and Table S2, and the positions of the mutations and their conservation in different species are shown in Fig. 1C. In family 3, oocytes were immature when retrieved. Most of immature oocytes could develop into the first polar body (PB1) oocytes after 2 hours' in vitro culture, but showed fertilization failure or early embryonic arrest. The proband in family 4 and family 5 had phenotypes of fertilization failure or early embryonic arrest. The specific clinical information is indicated in Fig. 1B and Table S1.
In transfected Chinese hamster ovary (CHO) cells, the missense mutations p.Tyr228Cys and p.Leu439Arg resulted in a reduction in CDC20 protein level, while mutations p. Arg182*, p.Gly272Serfs*24, and p.Cys392* resulted in truncated proteins (Fig. 1D and 1E). For mutation p. Arg322Gln, though there was no obvious effect on CDC20 protein level in CHO cells, the Western blot analysis of lymphoblastoid cell line (LCL) of the affected individual II in family 3 with a missense mutation c.965G>A (p. Arg322Gln) and a nonsense mutation c.544C>T (p. Arg182*) showed significantly reduced protein level (Figs. 1F and S3A). In addition, the mRNA expression of
CDC20 in the LCLs of individual II-1 in family 3 and in the granulosa cells (GCs) of individual II-1 in family 4 were also reduced significantly (Fig. S3B and S3C). We also explored the effects of the mutations (c.683A>G, c. 813_814ins AGTG, c.1176_1179del TCTG, and c.1316T>G) on mRNA expression in transfected CHO cells, and the results showed that all these four mutations caused significantly reduced mRNA expression (Fig. S4). All these results indicate that mutations in
CDC20 lead to unstable protein and degraded RNA
per se. During the metaphase to anaphase transition, APC/C is activated through the release from CDC20 inhibition, and this leads to the breakdown of cyclin B1 (HGNC: 1579) (
Nasmyth and Haering, 2005). We therefore determined the effects of
CDC20 mutations on cyclin B1 degradation. As shown in Fig. 1D and 1E, overexpression of wild-type CDC20 significantly decreased the endogenous protein level of cyclin B1, while most of mutations affected the degradation of cyclin B1.
We next investigated the effects of the mutations on human CDC20 localization in mouse oocytes. For the three missense mutations (p.Tyr228Cys, p.Arg322Gln, p. Leu439Arg), CDC20 showed normal kinetochore localization as the wild-type. In contrast, CDC20 failed to localize to the kinetochore for the other three nonsense or frameshift mutations (p.Arg182*, p.Gly272Serfs*24, p.Cys392*) (Fig. 2A), indicating these are loss-of-function mutations.
Knockdown of the
Cdc20 gene exclusively blocks PB1 extrusion in mouse oocytes and the PB1 extrusion can be rescued by injecting
Cdc20 cRNA into the oocytes (
Reis et al., 2007). To further explore the effect of mutations on CDC20 function in oocytes, we first knocked down the endogenous
Cdc20 in mouse oocytes by using
Cdc20 siRNA, and we observed an MI arrest phenotype that could be rescued by supplementation with human wild-type
CDC20 cRNA (Fig. 2B). We then performed a rescue experiment using patient-derived mutant cRNAs in the same way. Compared with wild-type, all mutations significantly reduced the ability of CDC20 to rescue PB1 extrusion (Fig. 2B). These results indicate the overall impaired effect on CDC20 function of homozygous and compound heterozygous mutations.
We then explored a potential therapeutic treatment for two patients by CDC20 cRNA injection. For the proband in family 3, compared with control, all four oocytes injected with CDC20 cRNA were successfully fertilized as indicated by the formation of two pronuclei on day 1, and two of the oocytes developed into blastocysts on day 6 (Fig. 2C and 2D). For the proband in family 4, all three oocytes injected with CDC20 cRNA were successfully fertilized on day 1, and one developed into an eight-cell stage embryo (Fig. 2E). Preimplantation genetic screening showed that one of blastocysts in family 3 had normal numbers of chromosomes and no obvious large repetition/deletion fragments (Fig. 2F). These results provide a potential treatment for these patients in the future.
We found phenotypic variability among the affected individuals with
CDC20 mutations. In brief, the proband in family 1 and 2 had the phenotype of oocyte maturation arrest. Although the proband in family 3 had a slight delay in oocyte maturation, the ultimate phenotype in family 3 and 4 was characterized by fertilization failure. The proband in family 5 showed early embryonic arrest. It has been reported in mice that
Cdc20 has the highest expression in metaphase II (MII) oocytes compared to GV and MI oocytes as well as early embryos (
Amanai et al., 2006), and this expression pattern was also observed in our qRT-PCR results in human oocytes and early embryos (Fig. S5). It is therefore likely that different amounts of CDC20 are needed at different stages of oocyte maturation, fertilization, and early embryo development. As for the mutations we identified, both family 1 and 2 had homozygous missense mutations, while families 3–5 harbored heterozygous compound mutations including one missense mutation and either a nonsense mutation or a frameshift mutation. The combinations of various types of mutations may result in different degrees of impairment of CDC20 and thus could lead to phenotypic variability.
Although CDC20 plays an important role in both mitosis and meiosis (
Musacchio and Hardwick, 2002;
Jones, 2011), all affected individuals in the study only exhibited the phenotype of female infertility without any abnormalities in somatic tissues or organs. This might due to the different thresholds of CDC20 amount required between mitosis and meiosis (Fig. S5). In addition, previous studies showed that
Cdc20 knockout mice were embryonic lethal, while
Cdc20 hypomorphic mice with graded reduction of CDC20 protein level from 60%–27% were healthy and had a normal lifespan compared to wild type mice, but only female hypomorphic mice with 27% CDC20 protein level were infertile or subfertile (
Jin et al., 2010;
Malureanu et al., 2010). In contrast, in mitosis even profound reduction of CDC20 levels to 10% of normal still supports the onset of anaphase and the completion of mitosis (
Wolthuis et al., 2008). All of these results indicate that female infertility resulting from CDC20 reduction is dosage dependent and that mitosis is more tolerant than meiosis to CDC20 reduction.
In summary, we identified biallelic mutations in CDC20 responsible for variable phenotypes of female infertility characterized by abnormalities in oocyte maturation, fertilization and early embryonic development and implicated cRNA injection strategy for a potential therapeutic treatment for these patients.
PDF
(4526KB)
