Analysis of DNA methyltransferases in human oocytes and pre-implantation embryos

Introduction:

We investigated the spatiotemporal expression pattern of different DNA methyltransferases (DNMTs): DNMT3L, the de novo DNA methyltransferase regulator, and DNMT1, an enzyme responsible for maintenance of the methylation patterns during DNA replication. There are two forms of DNMT1: DNMT1o, found only in oocytes and cleavage stage preimplantation embryos and the somatic form DNMT1s. DNMT1o lacks the N-terminus of DNMT1s. These proteins are crucial for the establishment and maintenance of epigenetic marks during gametogenesis and embryonic stages. DNMT deficiencies will lead to embryonic developmental defects, cancer and other diseases. Not much is known about the expression and regulation of these proteins in human oocytes and preimplantation embryos and we will compare our results to data obtained in mouse.

Material and Methods:

Spare human oocytes and preimplantation embryos from infertility treatments at our centre were donated after informed consent of the patients. Oocytes and embryos were stained by immunocytochemistry with antibodies against DNMT3L and two different epitopes of DNMT1 to distinguish both forms. One DNMT1 antibody is specific for the N-terminus, and thus binds only to DNMT1s, the other binds to the C-terminus of both proteins. Immunofluorescence was visualised with a confocal laser scanner for 3D images.

Results:

DNMT3L was absent in oocytes (n = 13), zygotes (n = 2) and 4-cell stage embryos (n = 9). It was observed in the cytoplasm of blastomeres from the 8-cell stage onwards. In expanding and expanded blastocysts (10/18) DNMT3L was found in both the cytoplasm and nucleus of cells from trophectoderm and inner cell mass.

DNMT1o and DNMT1s could not be detected in the majority of human oocytes (n = 37) while a weak DMNT1o expression could be detected in most zygotes (n = 7). DNMT1s was mainly found in the cytoplasm from the 2-cell stage on whereas DNMT1o was constantly present in the nucleus from the 8-cell stage onwards. This pattern seems opposite to mice, where Dnmt1s shows a nuclear localization during all preimplantation embryonic stages and Dnmt1o remains cytoplasmic with only a transient nuclear switch at the 8-cell stage.

Conclusion:

In human preimplantation embryos there is a gradual increase of DNMT3L in the cytoplasm from the 8-cell stage onwards, which may correspond with the activation of the embryonic genome. After blastocyst expansion DNMT3L appears in the nucleus of the cells where it may assist DNMT3a and DNMT3b in de novo methylation processes. These results are in line with mice where a sharp rise in Dnmt3L was found before the time of implantation, which is the time window for genome-wide de novo methylation processes. In mice, Dnmt3L has been found at early stages of oogenesis. We have not been able to examine this in humans because we do not have access to growing human oocytes yet.

The absence of DNMT1o in oocytes of different maturity stages together with its weak expression in zygotes could reflect the superior quality of oocytes suitable for fertilization and bring forward the hypothesis that DNMT1 expression is only associated with maturation competency. Alternatively, maternally stored DNMT1 mRNA transcripts are translated at the zygote stage. The absence of nuclear DNMT1 expression during the first two cleavage divisions may help the passive genome-wide demethylation process that takes place in the early embryo. After the 4-cell stage, the protein was also detected in the nucleus.

The differences in spatiotemporal expression patterns between human and mouse for DNMT1 may be due to differences in the demethylation reprogramming process while the similarities found for DNMT3L point to a better homology for the de novo methylation process.