RESUMO
STUDY QUESTION: What is the transcriptome signature associated with poor performance of rescue IVM (rIVM) oocytes and how can we rejuvenate them? SUMMARY ANSWER: The GATA-1/CREB1/WNT signalling axis was repressed in rIVM oocytes, particularly those of poor quality; restoration of this axis may produce more usable rIVM oocytes. WHAT IS KNOWN ALREADY: rIVM aims to produce mature oocytes (MII) for IVF through IVM of immature oocytes collected from stimulated ovaries. It is not popular due to limited success rate in infertility treatment. Genetic aberrations, cellular stress and the absence of cumulus cell support in oocytes could account for the failure of rIVM. STUDY DESIGN, SIZE, DURATION: We applied single-cell RNA sequencing (scRNA-seq) to capture the transcriptomes of human in vivo oocytes (IVO) (n = 10) from 7 donors and rIVM oocytes (n = 10) from 10 donors. The effects of maternal age and ovarian responses on rIVM oocyte transcriptomes were also studied. In parallel, we studied the effect of gallic acid on the maturation rate of mouse oocytes cultured in IVM medium with (n = 84) and without (n = 85) gallic acid. PARTICIPANTS/MATERIALS, SETTING, METHODS: Human oocytes were collected from donors aged 28-41 years with a body mass index of <30. RNA extraction, cDNA generation, library construction and sequencing were performed in one preparation. scRNA-seq data were then processed and analysed. Selected genes in the rIVM versus IVO comparison were validated by quantitative real-time PCR. For the gallic acid study, we collected immature oocytes from 5-month-old mice and studied the effect of 10-µM gallic acid on their maturation rate. MAIN RESULTS AND THE ROLE OF CHANCE: The transcriptome profiles of rIVM/IVO oocytes showed distinctive differences. A total of 1559 differentially expressed genes (DEGs, genes with at least 2-fold change and adjusted P < 0.05) were found to be enriched in metabolic processes, biosynthesis and oxidative phosphorylation. Among these DEGs, we identified a repression of WNT/ß-catenin signalling in rIVM when compared with IVO oocytes. We found that oestradiol levels exhibited a significant age-independent correlation with the IVO mature oocyte ratio (MII ratio) for each donor. rIVM oocytes from women with a high MII ratio were found to have over-represented cellular processes such as anti-apoptosis. To further identify targets that contribute to the poor clinical outcomes of rIVM, we compared oocytes collected from young donors with a high MII ratio with oocytes from donors of advanced maternal age and lower MII ratio, and revealed that CREB1 is an important regulator. Thus, our study identified that GATA-1/CREB1/WNT signalling was repressed in both rIVM oocytes versus IVO oocytes and in rIVM oocytes of lower versus higher quality. Consequently we investigated gallic acid, as a potential antioxidant substrate in human rIVM medium, and found that it increased the mouse oocyte maturation rate by 31.1%. LARGE SCALE DATA: Raw data from this study can be accessed through GSE158539. LIMITATIONS, REASONS FOR CAUTION: In the rIVM oocytes of the high- and low-quality comparison, the number of samples was limited after data filtering with stringent selection criteria. For the oocyte stage identification, we were unable to predict the presence of oocyte spindle, so polar body extrusion was the only indicator. WIDER IMPLICATIONS OF THE FINDINGS: This study showed that GATA-1/CREB1/WNT signalling was repressed in rIVM oocytes compared with IVO oocytes and was further downregulated in low-quality rIVM oocytes, providing us the foundation of subsequent follow-up research on human oocytes and raising safety concerns about the clinical use of rescued oocytes. STUDY FUNDING/COMPETING INTEREST(S): This work was supported by the Collaborative Research Fund, Research Grants Council, C4054-16G, and Research Committee Funding (Research Sustainability of Major RGC Funding Schemes), The Chinese University of Hong Kong. The authors have no conflicts of interest to declare.