Human oocyte maturation in vitro is improved by co-culture with cumulus cells from mature oocytes

Accepted Manuscript Title: Human oocyte maturation in vitro is improved by co-culture with cumulus cells from mature oocytes Author: Irma Virant-Klun, Chris Bauer, Anders Ståhlberg, Mikael Kubista, Thomas Skutella PII: DOI: Reference:

S1472-6483(18)30046-4 https://doi.org/10.1016/j.rbmo.2018.01.011 RBMO 1891

To appear in:

Reproductive BioMedicine Online

Received date: Revised date: Accepted date:

26-6-2017 14-1-2018 16-1-2018

Please cite this article as: Irma Virant-Klun, Chris Bauer, Anders Ståhlberg, Mikael Kubista, Thomas Skutella, Human oocyte maturation in vitro is improved by co-culture with cumulus cells from mature oocytes, Reproductive BioMedicine Online (2018), https://doi.org/10.1016/j.rbmo.2018.01.011. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Short title: Human oocyte transcriptomics and maturation

Human oocyte maturation in vitro is improved by co-culture with cumulus cells from mature oocytes Irma Virant-Klun,a* Chris Bauer,b Anders Ståhlberg,c Mikael Kubista,d Thomas Skutellae a

Department of Obstetrics and Gynaecology, University Medical Centre Ljubljana, 1000 Ljubljana, Slovenia b MicroDiscovery, 10405 Berlin, Germany c Sahlgrenska Cancer Center, Department of Pathology and Genetics, Institute of Biomedicine, Sahlgrenska Academy at University of Gothenburg, 41390 Gothenberg, Sweden d TATAA Biocenter AB, 40530 Gothenburg, Sweden e Institute for Anatomy and Cell Biology, Medical Faculty, University of Heidelberg, 69120 Heidelberg, Germany Corresponding author. E-mail address: [email protected]

*

Key message Co-culture of immature (germinal vesicle) oocytes with cumulus cells from mature oocytes of the same patient improves oocyte maturation in vitro in terms of maturation rate and expression of genes, which is more comparable to oocytes matured in vivo than in oocytes matured in vitro without co-culture.

Page 1 of 55

2

Author biography Dr Irma Virant-Klun is Senior Clinical Embryologist and Director of the IVF Lab, Department of Obstetrics and Gynaecology, University Medical Centre Ljubljana. She is also Associate Professor at the School of Medicine, University of Ljubljana. Her major interests are oocyte biology and ovarian stem cells, and she also has considerable experience in different techniques of assisted conception.

Abstract The conventional method of human oocyte maturation in vitro in the presence of gonadotrophins continues to be a relatively low-success procedure in the assisted conception programme owing to suboptimal maturation conditions in the absence of an ovarian ‘niche’ and poor understanding of this procedure at the molecular level in

Page 2 of 55

3 oocytes. In this study, the gene expression profiles of human oocytes were analysed according to their manner of maturation: in vivo (in the ovaries) or in vitro (matured either by the conventional method or by a new approach – co-cultured with cumulus cells of mature oocytes from the same patient). Our results show that the in-vitro maturation procedure strongly affects the gene expression profile of human oocytes, including several genes involved in transcriptional regulation, embryogenesis, epigenetics, development, and the cell cycle. The in-vitro maturation of oocytes cocultured with cumulus cells from mature oocytes provides an ovarian ‘niche’ to some degree, which improves oocyte maturation rates and their gene expression profile to the extent that they are more comparable to oocytes that naturally mature in the ovarian follicle.

KEYWORDS: human, oocyte, in vitro maturation, co-culture, cumulus, gene expression

Introduction The conventional human oocyte in-vitro maturation (IVM) procedure is a relatively low-success procedure owing to suboptimal maturation conditions in the absence of an ovarian ‘niche’ and the poor understanding of this process at the molecular level in oocytes. Some investigators have reported on structural and morphological differences in human oocytes after IVM compared with standard IVF (Coticchio et al., 2016; Walls et al., 2016). For in-vitro matured oocytes, worse IVF clinical results, i.e., embryonic development, pregnancy and live birth, are achieved compared with oocytes that naturally mature in the ovarian niche and that are supported by

Page 3 of 55

4 surrounding cumulus cells, especially granulosa cells, during their growth and maturation in the follicle (Roesner et al., 2012; Walls et al., 2015). There is no generally accepted procedure for oocyte maturation in vitro, and the techniques used for IVM differ substantially across infertility clinics, resulting in extremely variable clinical outcomes (Dahan et al., 2016). Therefore, the procedure for oocyte maturation in vitro continues to be an experimental procedure (ASRM, 2013). This procedure, however, represents an interesting model for studying the molecular mechanisms involved in the oocyte maturation process, and the IVM procedure should be optimized for the future. It is of great interest because it could be beneficial for some infertile women who have polycystic ovaries (Siristatidis et al., 2015), for women who have poor ovarian response to hormonal stimulation (Lee et al., 2016) and for oncological patients who cryopreserve their oocytes before oncotherapy to preserve their fertility (Revelli et al., 2012).

The ovarian follicle is the major functional unit of the ovary and female reproductive system. Bidirectional crosstalk between the oocyte and surrounding cumulus cells is essential for oocyte growth, for enabling nutrients and other small molecules to transfer between them, together ensuring that the oocyte properly acquires the molecular machinery required for subsequent early embryo development (Zuccotti et al., 1998; Albertini et al., 2001; Gilchrist et al., 2008). A highly coordinated interplay between the oocyte and surrounding cumulus cells (cranulosa cells) in the follicle depends on functional gap junctions (Granot and Dekel, 2002), which are mainly composed of connexin 37 (Cx37) (Furger et al., 1996; Tsai et al., 2003; Gershon et al., 2008) and directly mediate the cell–cell communication by allowing the passage of small molecules such as ions, metabolites, nutrients, and small signalling molecules between the neighbouring cells (Caspar et al., 1977; Makowski et al.,

Page 4 of 55

5 1977). Studies in knockout mice clearly demonstrated that the ablation of Cx37 leads to an interruption of intercellular coupling between oocytes and cumulus cells, disruption of follicle development, incompetent oocytes and ovulatory dysfunction (Simon et al., 1997).

Experimental data on cumulus cell transcriptome revealed that bidirectional communication via gap junctions between the human oocyte and cumulus cells is essential for the maturation and production of a competent oocyte (Huang and Wells, 2010; Ouandaogo et al., 2011; 2012). Among important substances exchanged between the oocyte and cumulus cells (granulosa cells) are bone morphogenetic protein 15 (BMP15) and growth differentiation factor 9 (GDF9), which play key roles in promoting follicle growth beyond the primary stage (Chang et al., 2002; Knight and Glister, 2006); the absence of these two substances causes infertility, as shown in the animal models (Dong et al., 1996; Galloway et al., 2000). Cumulus cell–oocyte gap junctions have also been reported to regulate the oocyte reduced glutathione synthesis and accumulation, which plays an important role as a reducing agent and important antioxidant in oocyte maturation, fertilization, and embryonic development (de Matos et al., 1997; 2002; Sutovsky and Schatten, 1997; Nagai, 2001; Curnow et al., 2008; Takeo et al., 2015); oocytes are quite susceptible to oxidative stress during the maturation process, and their most important antioxidant glutathione is mostly supplied by surrounding cumulus cells via gap junctions.

The human oocytes are usually matured in vitro without cumulus cells and ovarian ‘niche’. The human oocyte maturation process involves several known components, but there are still many unknown components that may be lacking in the present maturation media owing to the absence of cumulus cells, which may have negative

Page 5 of 55

6 consequences at the genetic level. Several differences in the global mRNA transcript profile have been reported between mature (metaphase II) and immature (GV) oocytes in both mice and humans (Wang et al., 2004; Assou et al., 2006; Yoon et al., 2006; Cui et al., 2007; Gasca et al., 2007; Zhang et al., 2007). Published data show that immature human oocytes express a higher number of genes than mature oocytes (Assou et al., 2006; Wells and Patrizio, 2008). A set of genes whose expression gradually increased during human oocyte maturation was identified, including phosphatase CDC25A, PCNA and SOCS7 (Assou et al., 2006). Fewer data are available on the molecular status of human oocytes during the IVM process. Existing data show that more than 2000 genes were differentially expressed in human oocytes matured in vitro compared with those matured in vivo (Jones et al., 2008); this group included several genes involved in transcriptional regulation. Reduced HDAC1 expression and insufficient histone deacetylation are associated with metaphase defects in human oocytes matured in vitro (Huang et al., 2012). Moreover, it has been found that inefficient SIRT3 expression induced decreased mitochondrial DNA copy number and biogenesis, and therefore impaired the developmental competence of human oocytes matured in vitro (Zhao et al., 2016).

The aims of this study were to elucidate how the IVM procedure affects the molecular status of oocytes compared with oocytes that matured in vivo and to elucidate the effects of oocyte co-culture with cumulus cells, which provides some degree of an ovarian niche during the maturation procedure. The gene expression profile of human oocytes was analysed according to whether they matured in vivo or in vitro (either by the conventional or in co-culture with cumulus cell). We evaluated the effects of the

Page 6 of 55

7 two IVM procedures on the gene expression profiles of human oocytes compared with oocytes that matured in vivo.

Materials and methods This study was approved by the Slovenian National Medical Ethical Committee (Ministry of Health RS; KME 23k/10/12) on 30 October 2012. The oocytes included in this study would have otherwise been discarded during daily medical practice because they were immature and, consequently, had no natural ability to fertilize, or they were mature but did not fertilize during the IVF procedure. The oocytes were included in the study only after informed consent forms were signed by women who gave donations, indicating their voluntary participation.

Oocyte in-vitro maturation In a prospective study, the germinal vesicle oocytes retrieved from the IVF programme in the Department of Obstetrics and Gynaecology, University Medical Centre Ljubljana, were matured in vitro in two different ways: using the conventional method (CONV-IVM) in maturation medium (MediCult IVM®System, Origio, Denmark) containing the reproductive hormones FSH (Puregon, Merck Sharp & Dohme Gmbh, Germany, 75 mIU/ml) and HCG (Pregnyl, Organon, Netherland, 100 mIU/ml) or 2) using a new approach (CC-IVM) in the same maturation medium with added FSH and HCG but co-cultured with cumulus cells retrieved by denudation of mature oocytes in the same patients. The MediCult IVM®System consisted of two different media: conditioning LAG medium (containing human albumin solution and recombinant human insulin) and in-vitro maturation (IVM) medium. In both cases,

Page 7 of 55

8 the immature oocytes were first exposed to the LAG medium of the MediCult IVM®System for 2 h and then cultured in the IVM medium with added gonadotrophins in an EGPS-010 Embryo GPS dish (SunIVF, USA) covered by paraffin oil; in CC-IVM oocytes a small piece of cumulus mass from mature oocytes in the same patient was added to the IVM medium just at the beginning of oocyte maturation. The oocytes were matured in vitro for 40–42 h to have a higher chance to achieve both the nuclear and cytoplasmic maturation; those oocytes that did not mature in 48 h lacked the ability to mature in vitro.

Oocyte maturation rate A group of 174 germinal vesicle oocytes were cultured in vitro (79 oocytes by CCIVM and 95 oocytes by CONV-IVM) to evaluate the proportion of oocytes that matured in vitro (maturation rate). Immature oocytes were germinal vesicle oocytes that had clearly visible nuclei and with no polar body. After the IVM procedure, they were considered mature, if they extruded a primary polar body that is characteristic of the metaphase II) stage.

Collection of mature oocytes for microarray analysis Four groups of oocytes were collected after the complete removal of cumulus cells by enzyme hyaluronidase and three-time rinsing in the sterile phosphate-buffered saline (PBS): (1) in vivo-matured oocytes that were retrieved as mature (metaphase II) oocytes by ultrasound-guided aspiration from ovarian follicles in women with no indications of infertility and partners of infertile men with impaired semen quality and

Page 8 of 55

9 with previous unsuccessful attempt of fertilization in 24 h after the intracytoplasmic sperm injection (ICSI) procedure; (2) oocytes that were matured in vitro by CONVIVM (in the presence of FSH and HCG); (3) oocytes that were matured in vitro by CC-IVM, in the presence of FSH and HCG and co-cultured with cumulus cells from mature oocytes of the same patient; and (4) oocytes that did not mature regardless of the IVM procedure that was used. Only oocytes with completely removed cumulus cells were included. Each group of oocytes consisted of three oocyte samples (biological replicates), and each of these samples consisted of 10 oocytes, which means that there were 12 oocytes samples and 120 oocytes altogether. All oocytes were collected step-by-step and stored at –90°C before analysis.

Microarray analysis of oocyte samples The Miltenyi Biotec Microarray service in Germany carried out the analysis of all four groups (12 samples) of oocytes. SuperAmp RNA amplification was conducted according to Miltenyi Biotec’s undisclosed procedure. Briefly, the amplification was based on a global polymerase chain reaction protocol using mRNA-derived cDNA. mRNA was isolated via magnetic bead technology, and amplified cDNA samples were quantified using an ND-1000 spectrophotometer (NanoDrop Technologies). Fifty nanograms of each cDNA was used as a template for Cy3 labelling, which was carried out according to Miltenyi Biotec’s undisclosed protocol. The Cy3-labelled cDNAs were hybridized to an Agilent Whole Human Genome Oligo Microarray 8x60K v2 platform overnight (17 h, 65°C) and analysed using microarrays. Resolver

Page 9 of 55

10 software (Rosetta Biosoftware, USA) made it possible to perform principal component analysis (PCA) of the oocyte groups.

Normalization of microarray data The (log2) raw microarray signal distributions of all 12 samples is shown in Supplementary Figure 1A. Pairwise correlation and scatterplots (scatterplots are shown as two-dimensional densities) of the (log2) raw signal are shown in Supplementary Figure 1B. The Pearson correlation coefficients were within a range of 0.70–0.89. The group of in vivo-matured oocytes especially showed the lowest correlation to other samples. Clear outliers were not present; therefore, all samples were used in the bioinformatics analyses. Normalization was performed with the multi lowess algorithm because it effectively removes systematic errors, keeping the maximum original signal. Boxplots of the normalized signal distributions of the 12 samples are shown in Supplementary Figure 2A. After normalization, the minimal correlation between samples was increased from 0.7 to 0.75 (Supplementary Figure 2B). Statistical analysis of microarray data To search for differentially expressed genes between the different groups of oocytes, two strategies were followed: the identification of genes differentially expressed in at least one of the groups using F-test statistics; and conducting T-tests of one group of oocytes versus the remaining groups. With the F-test, we identified genes that were different in at least one of the four groups, but this test did not report which of the groups was different. The null hypothesis was that μ1 = μ2 = μ3 = μ4. Therefore, if at

Page 10 of 55

11 least one of the groups was different, the null hypothesis was rejected. P-values were corrected for multiple testing with Benjamini Hochberg procedure. As the number of oocyte samples was relatively small (three replicates per group), the P-value may not be the best ranking criterion (since it greatly depends on standard deviation). For the T-tests, we hypothesized that the fold change could rank the differentially expressed genes better. In the T-test, the equivalent of the fold for the F-test was the mean squared error, calculated as the variance between group means. Statistical significance was set at P < 0.05.

Clustering and heatmaps of microarray data were created with complete-linkage clustering using Euclidean distance as a measure of distance.

Principal component analysis The normalized dataset, also used in other analyses, was used for PCA. The background genes that had expression lower than 2.2 were removed, and only the genes that had expression above the background (≥2.2) were included. A scatter plot of the first two principal components from PCA was created.

Gene ontology analysis comparing in-vivo and in-vitro matured oocytes The gene ontology analysis compared the in-vivo matured oocytes and the in-vitro matured oocytes of both groups (CONV-IVM and CC-IVM), and CC-IVM and CONV-IVM oocytes. Genes upregulated in the in vivo matured oocytes To this end, the genes that were upregulated in the group of in-vivo matured oocytes

Page 11 of 55

12 (three samples) compared with the two groups of in-vitro matured oocytes (six samples) were selected. For each gene, the log2 fold change and the T-test P-value were calculated. A group of genes had higher fold changes: 73 genes were expressed more than 16-fold higher, 384 genes more than eight-fold higher and 1552 genes more than four-fold higher in the in-vivo matured oocytes compared with in-vitro matured oocytes. We selected the set of genes that had greater than eight-fold upregulation and performed an enrichment analysis with Fisher statistics. The gene ontology analysis was performed with the topGO R-package (Alexa and Rahnenfuhrer, 2010). We used the ‘elim’ algorithm to decorrelate the GO graph structure (Alexa et al., 2006). The analysis was carried out for all three GO trees (biological process-BP, molecular function-MF, and cellular component-CC) separately. For each of the three GO trees, we showed the top 10 GO term results. The terms were annotated with the number of genes associated with the GO term (#Associated Genes), the number of genes with a fold change three or over (#Significant Genes), the number of genes expected to have a fold change three or over by chance (#Expected Genes), and the Fisher test P-value (P-value). Genes upregulated in the in-vitro matured oocytes: the genes that were upregulated in the two groups of in-vitro matured oocytes (six samples) compared with the group of in-vivo matured oocytes (three samples) were selected. There was a group of genes with higher fold changes: 125 genes were expressed more than 32-fold higher, 450 genes more than 16-fold higher, 1285 genes more than eight-fold higher, and 3144 genes more than four-fold higher in the in-vitro matured oocytes compared with invivo matured oocytes. We selected the set of genes that had greater than eight-fold upregulation and carried out an enrichment analysis with Fisher statistics. The gene ontology analysis was performed as described above.

Page 12 of 55

13 Genes upregulated in the CC-IVM oocytes The genes that were upregulated in the CC-IVM oocytes (three samples) compared with the group of CONV-IVM oocytes (three samples) were selected. There was a group of genes with higher fold changes: 49 genes more than 16-fold higher, 200 genes more than eight-fold higher, and 883 genes more than four-fold higher in the CC-IVM oocytes compared with CONV-IVM oocytes. We selected the set of genes that had greater than eight-fold upregulation and performed an enrichment analysis with Fisher statistics. The gene ontology analysis was performed as described above. Functional annotation To identify the associated pathways, we carried out an enrichment analysis. As a starting point, we selected the top 300 genes that had the highest mean squared error from the F-test. We used KEGG (www.genome.jp/kegg) and WikiPathways (www.wikipathways.org) as pathway repositories.

Results In this study, immature (germinal vesicle) oocytes from an IVF programme were matured in vitro in two diverse ways. The first way used the conventional approach (CONV-IVM), which involved adding FSH and HCG to the maturation medium. The second method (CC-IVM) used a new co-culture approach that involved using the same maturation medium, but Cumulus cells (Figure 1) from mature oocytes of the same patient were also added to the medium to provide a follicular/ovarian niche. We attempted to elucidate the differences in the gene expression profiles of oocytes matured in vitro by the CONV-IVM or by the CC-IVM compared with oocytes that matured in vivo (retrieved as mature oocytes by ultrasound-guided aspiration of

Page 13 of 55

14 ovarian follicles that were not fertilized during IVF). Oocytes that failed to mature in vitro regardless of the IVM procedure that was used (conventional or new) were also compared. Using microarrays, four groups of oocytes were analysed: oocytes that matured in vivo; oocytes matured in vitro by the CONV-IVM; oocytes matured in vitro by the CC-IVM; and oocytes that did not mature. The microarray data are available at ArrayExpress under the code E-MTAB-5400.

Oocyte maturation rates The maturation of immature germinal vesicle oocytes with the CC-IVM resulted in a significantly higher proportion of mature oocytes than the CONV-IVM (77.2% (61/79) versus 62.1% (59/95), respectively; P < 0.05, as revealed by the chi-squared test).

Gene expression profiles of the analyzed groups of oocytes as revealed by microarrays To identify whether a gene or transcript was expressed in oocytes, we needed to identify a background value. To this end, Supplementary Figure 3 depicts a histogram that shows all the normalized expression values that were less than 5 (all transcripts from all chips). We considered that transcripts with expression less than 2.2 were non-expressed, while transcripts 2.2 or over were expressed. The numbers of expressed and non-expressed genes and transcripts are shown in Table 1. The oocytes that matured in vivo showed a higher number of expressed genes (Table 1).

Gene expression profile of in vivo matured oocytes compared with IVM oocytes using Gene Ontology analysis

Page 14 of 55

15 To better elucidate the effect of IVM procedure on molecular status of oocytes, we compared the gene expression profile of in vivo-matured oocytes with that of IVM oocytes (CC-IVM and CONV-IVM). We selected a set of genes that were upregulated more than eight-fold in the in-vivo matured oocytes compared with IVM oocytes, and genes that were upregulated more than eight-fold in the IVM oocytes compared with in-vivo matured oocytes. The Gene Ontology (GO) analysis of these genes was carried out to determine their associated biological processes, molecular functions, and cellular components separately. The differentially expressed genes were related to several biological processes and molecular functions. Most of genes which were upregulated in oocytes matured in vivo were related to the biological processes of positive regulation of transcription, placenta development, fatty acid beta-oxidation, histone H3 acetylation, and cortical actin cytoskeleton organization (Figure 2A) with predominating molecular functions of nucleic acid binding transcription factor activity and microtubule motor activity, and cellular component of intracellular organelle (Supplementary Table 1). On the other hand, the genes upregulated in the IVM oocytes were related to other biological processes such as mRNA metabolic process, translation, mitochondrion organization, and respiratory electron transport chain (Figure 2) with the main molecular functions of RNA binding and structural constituent of ribosome, and intracellular membrane-bounded organelle as a cellular component (Supplementary Table 2). In the IVM oocytes, two genes were significantly upregulated compared with in-vivo matured oocytes: Prefoldin Subunit 2 (PFDN2), one of subunits of prefoldin, a molecular chaperone complex that binds and stabilizes newly synthesized proteins/polypeptides, thereby allowing them to fold correctly, and Heat Shock 10kD Protein 1 (Chaperonin 10; HSPE1), which acts as a co-chaperonin implicated in

Page 15 of 55

16 mitochondrial protein import and their correct folding according to GeneCards (http://www.genecards.org/); the multiple testing corrected P-values (correction with Benjamini Hochberg procedure) were 0.0324 for PFDN2 and less than 0.0001 for HSPE1. The gene expression profiles of oocytes according to the in-vitro maturation procedure used The data showed that oocytes matured in vitro using CONV-IVM had a notably higher number of upregulated genes and a higher number of downregulated genes than oocytes matured in vitro by the CC-IVM compared with oocytes matured in vivo (Table 1). Non-mature oocytes expressed a lower number of upregulated and downregulated genes than oocytes matured in vitro when compared with in vivomatured oocytes. Principal component analysis (PCA) showed that the gene expression profile for oocytes matured in vitro by the CC-IVM was more similar to the oocytes matured in vivo (in the ovary) compared with oocytes that were matured in vitro by the CONV-IVM conventional approach (Figure 3A). Two samples of oocytes that were matured in vitro by the CC-IVM were more similar to the in vivomatured oocytes than other groups and clustered with them, whereas the other groups of oocytes were different. Interestingly, oocytes that were matured in vivo and in vitro by the CC-IVM had gene expression profiles that were more heterogeneous than those of oocytes that were matured in vitro by the CONV-IVM and non-mature oocytes (Figure 3A). The heatmap clusters for genes with the highest differences in all samples (highest SD) (Figure 3B) and genes with the lowest expression in the in vivo-matured oocytes (Figure 3C) showed that the oocytes matured in vitro by the CC-IVM had more genes in common with the in vivo-matured oocytes than oocytes matured in vitro by

Page 16 of 55

17 the CONV-IVM; the genes involved with oocytes that did not mature in vitro were completely different (Figure 3B, C). On the other hand, the gene expression profile of oocytes matured in vitro by the CONV-IVM was quite comparable to that of oocytes that did not mature in vitro (non-mature). For both groups of genes, three samples of the in vivo-matured oocytes were clustered with two samples of oocytes matured in vitro by the new approach (CC-IVM), whereas all other samples differed from the aforementioned samples and were clustered together.

The Gene Ontology analysis of selected genes, which were at least eight-fold upregulated in the CC-IVM oocytes compared with CONV-IVM oocytes, showed that these genes were mostly associated with biological processes such as positive regulation of sequence-specific DNA binding transcription factor activity, protein homotetramerization and histone A4-K16 acetylation (Figure 4 and Supplementary Table 3). The molecular function of these genes were mostly related to the cytoskeletal protein binding, growth factor activity, and nucleotidyltransferase activity with microtubule and early endosome as the main cell components (Supplementary Table 3). Genes related to transcriptional regulation, embryogenesis, epigenetics, development, and the cell cycle in the analysed groups of oocytes We were interested in comparing the expression of genes related to transcriptional regulation, embryogenesis, epigenetics, development, and the cell cycle among the four analysed groups of oocytes. We identified genes that were differentially expressed in the in vivo-matured oocytes compared with the other groups of oocytes. As the genes related to transcriptional regulation, embryogenesis, epigenetics, development and the cell cycle are important for human oocytes, we used functional

Page 17 of 55

18 annotation provided by the Gene Ontology database to identify relevant genes among the analysed oocyte samples. To this end, we searched all GO terms containing: ‘transcription factor’, ‘embryo’, ‘epigenetic’, ‘development’, and ‘cell cycle’. Then, we identified all genes that were annotated with one of these terms in the analysed groups of oocytes. The obtained volcano plots (Figure 5) identified several genes, including those related to transcriptional regulation: e.g., FIGLA, ROQCD1, GTF2H5, DNMT3B, NPM2, POU5F1, and H1FOO; embryogenesis, e.g., SOX7, HOXA1, and NOTCH1, epigenetics, e.g., DNMT3B and POU5F1, development, e.g., SAT1 and SOD2 and the cell cycle, e.g., SKA3, BUB1, and MSH5, that were upregulated or downregulated in oocytes matured in vivo compared with groups of oocytes matured in vitro. In oocytes matured in vivo, there were fewer differentially expressed upregulated or downregulated genes compared with oocytes that were matured in vitro by the new CC-IVM than oocytes matured by the CONV-IVM (Figure 5A and Figure 5B). In oocytes matured in vitro by the CONV-IVM, several genes that were upregulated or downregulated compared with oocytes matured in vitro by the CC-IVM (Figure 5C). A mass of genes were upregulated or downregulated in oocytes matured in vivo (Figure 5D) and in non-mature oocytes (Figure 5E) compared with the rest of the oocytes. We found differentially expressed genes related to transcriptional regulation, embryogenesis, epigenetics, development, and the cell cycle according to the manner of oocyte maturation (Table 2). In oocytes matured in vivo, there were fewer differentially expressed upregulated or downregulated genes compared with oocytes matured in vitro by the CC-IVM than oocytes matured by the CONV-IVM approach.

Page 18 of 55

19 The oocytes matured in vitro by the CONV-IVM compared with the in vivo-matured oocytes were characterized by a high number of differentially expressed genes versus CC-IVM oocytes compared with the in vivo-matured oocytes. This situation was especially evident for genes related to transcriptional regulation and development.

Heatmap clusters of the most differentially expressed genes in the analysed groups of oocytes Differences in expression of genes related to transcription factor activity, embryogenesis, epigenetics, development and the cell cycle between the different groups of oocytes are clearly shown in heatmaps (Figure 6). We can see that oocytes that were matured in vitro by the CC-IVM resembled the oocytes matured in vivo more than oocytes matured in vitro by the CONV-IVM and non-mature oocytes, especially for genes related to transcription factors (Figure 6A), embryogenesis (Figure 6B), development (Figure 6D), and the cell cycle (Figure 6E), but less for epigenetics-related genes (Figure 6C). This finding was especially demonstrative for genes related to the cell cycle (Figure 6E); for these genes, the oocytes matured in vitro by the CC-IVM clustered together with the in vivo-matured oocytes, as revealed by dendrogram clustering (Figure 6F), whereas the oocytes matured in vitro by the CONV-IVM and non-mature oocytes represented separated groups of oocytes; two samples of oocytes matured in vitro by the new co-culture approach (CC-IVM) were clustered together with in vivo-matured oocytes, as seen in hierarchical clustering.

Top differentially expressed genes in oocytes matured in vivo compared with in vitro-matured and non-mature oocytes

Page 19 of 55

20 The top 25 differentially expressed genes comparing in vivo-matured oocytes versus all other oocyte groups are presented in Figure 7: two groups of in-vitro matured and non-mature oocytes were ordered according to their absolute fold changes. All these genes were upregulated in oocytes matured in vitro and in non-mature oocytes compared with in vivo-matured oocytes. The most differentially expressed genes were mitochondrial ribosomal protein L20 (MRPL20) and mitochondrial ribosomal protein S21 (MRPS21), which are nuclear genes that encode proteins involved in protein synthesis (translation) that occurs in the mitochondria, according to GeneCards (human gene database). Transcription factors, such as GTF2H5 and RQCD1, were among these genes. According to KEGG and WikiPathway, these genes are associated with different significant pathways, such as oxidative phosphorylation (energy metabolism, electron transport chain, TCA cycle, one carbon donor, phase I non-P450

biotransformations),

proteasome

(degradation,

Parkin-ubiquitin

proteasomal system, protein export), ribosome (structure), steroid biosynthesis, mismatch repair, TOR signaling, alpha linolenic acid and glycine metabolism (Table 3). Supplementary Tables 4–6 show that the most differentially expressed genes identified by the comparisons of in vivo-matured oocytes to the other groups of oocytes belong to the same pathways, which indicates that these pathways are deeply involved in the maturation of human oocytes.

Discussion Our results show that the IVM procedure strongly affects the gene expression profile of human oocytes, including affecting genes involved in transcriptional regulation,

Page 20 of 55

21 embryogenesis, epigenetics, development, and the cell cycle. Moreover, the IVM of oocytes co-cultured with cumulus cells from mature oocytes provides an ovarian ‘niche’, to some degree, which improves oocyte maturation rates and their gene expression profile to the extent that they are more comparable to oocytes that naturally mature in the ovary than oocytes matured in vitro in a conventional way.

Only scarce data have been retrieved from human oocytes until now showing the altered gene expression profile of human oocytes matured in vitro (Jones et al., 2008); however, more knowledge on animal oocytes, especially bovine oocytes has been obtained. Similar to the present study, studies of bovine oocytes have shown that genes differentially expressed in oocytes that were matured in vitro were related to the cell cycle, development, embryogenesis, energy metabolism, epigenetics, DNA replication and repair and regulation activities, including transcription and translation initiation (Dalbiès-Tranand and Mermillod, 2003; Fair et al., 2007; Katz-Jaffe et al., 2009; Mamo et al., 2011).

As with the present study, the global gene expression analysis during the bovine oocyte IVM showed that most of these genes were associated with regulation activities, especially regulation of transcription and translation (Fair et al., 2007). Our results showed that in vivo-matured oocytes were predominantly characterized by upregulation of genes involved in the positive regulation of transcription along with other biological processes such as placenta development, fatty acid beta-oxidation, histone H3 acetylation, and cortical actin cytoskeleton organization, which are deeply involved in the oocyte maturation process and may be negatively affected by the IVM procedure (Somfai et al., 2011; Franciosi et al., 2012; Huang et al., 2012; Valsangkar

Page 21 of 55

22 and Downs, 2013; Coticchio et al., 2014; Dunning et al., 2014; Sanchez-Lazo et al., 2016). On the other hand, the genes that were upregulated in the IVM oocytes were mostly related to translation, mRNA metabolic process, mitochondrion organization, and respiratory electron transport chain, thus indicating that these oocytes were more metabolically active and intensely producing the proteins and chemical energy, the efficiency ‘tools’ for the maturation process, than in vivo-matured oocytes; all these processes, especially mitochondria and their activity, may be affected by the IVM procedure and result in a less competent oocyte, as experienced by other research groups (Ge et al., 2012; Dalton et al., 2014; Jeseta et al., 2014; Milakovic et al., 2015; Coticchio et al., 2016; Zhao et al., 2016; Sowinska et al., 2017).

In the present study, the IVM of immature oocytes in a co-culture with cumulus cells of mature oocytes had beneficial effect on oocytes in making their maturation rate and gene expression profile more comparable to in-vivo matured oocytes by providing some degree of a follicular niche. This finding is reasonable because, under natural conditions, the antral follicle provides a complex microenvironment for oocyte differentiation in the ovary that includes three distinct populations of somatic cells (theca, granulosa and cumulus) that support the oocyte during its growth and maturation (Hennet and Combelles, 2012). The dialogue between oocytes and the surrounding cumulus cells has a major contribution to oocyte meiotic and developmental competence in vivo (Feuerstein et al., 2016) and in vitro, including nuclear and cytoplasmic maturation (Goud et al., 1998; Hassan et al., 2001). When cat oocytes were matured in vitro in a co-culture with cumulus cells, the cleavage rate of resulting embryos was increased compared with the oocytes that were matured in vitro as denuded (Sowinska et al., 2017). In addition, it was found that there is a

Page 22 of 55

23 correlation between cumulus mucification patterns (Thanaboonyawat et al., 2016) and the number of cumulus cell layers (Sato et al. 2007) during human oocyte maturation in vitro. Therefore, the conventional oocyte IVM procedure performed only in the presence of reproductive hormones (Wang et al., 2015) is suboptimal because of the lack of a follicular niche. To avoid this, we induced oocyte IVM in a co-culture with cumulus cells. We matured the oocytes with a new approach that involved coculturing with cumulus cells from mature oocytes from the same woman because the previous attempts to mature the oocytes in a co-culture with cumulus cells of the same immature oocytes or partially embedded cumulus-oocyte complexes were not successful (Johnson et al., 2008; Lin et al., 2009; Zhang et al., 2012). The main reason may be that the cumulus cells of immature oocytes possess abnormal receptors for gonadotrophins (FSH and LH), which prevents their maturation, as experienced in mouse oocytes (Vigone et al., 2015).

Our data show that the gene expression profile of CC-IVM oocytes was more similar to in-vivo matured oocytes than in the CONV-IVM oocytes, including genes associated with transcriptional regulation, embryogenesis, epigenetics, development, and the cell cycle. The genes, which were upregulated in the CC-IVM oocytes compared with CONV-IVM oocytes, were related to transcription regulation, cytoskeleton (spindle) structure, and, especially, histone deactylation, similar to invivo matured oocytes. It is not excluded that these biological processes make the CCIVM oocytes more comparable to in-vivo matured oocytes than CONV-IVM oocytes.

The results of other research groups have shown that epigenetics-related histone deacetylation in murine (Huang et al., 2017) and equine (Franciosi et al., 2012)

Page 23 of 55

24 oocytes was inadequate during the IVM procedure. Post-translational modifications of histones by acetylation are involved in the epigenetics-related regulating chromatin remodeling and gene expression. The insufficient histone deacetylation during the IVM procedure is associated with aberrant meiosis, and lowered genetic stability and developmental competence of human (Huang et al., 2012) and animal (Wang et al., 2010; Franciosi et al., 2012; 2015) oocytes, which may be transited to the embryo (Wang et al., 2010). Moreover, the results from the bovine model suggest that cumulus cells and granulosa cells, added to the culture medium, play a role in the development of oocyte–granulosa cell complex and the growth of oocyte, and also in the histone acetylation and chromatin remodelling in oocytes developed and matured in vitro (Sugiyama et al., 2016). Finally, the cumulus cell contact during the IVM of mice oocytes regulated the spatial organization and function of the meiotic spindle by actin-dependent mechanisms that enhance their development (Barrett and Albertini, 2010). Perhaps, this can partially explain the beneficial role of cumulus cell co-culture in our research, taking into account the genes that were upregulated in the in-vivo matured oocytes and CC-IVM oocytes compared with CONV-IVM oocytes.

The cumulus cells from mature oocytes that were added to the maturation medium to aid maturation of immature (granulosa cells) oocytes were efficient and beneficial despite possibly being developmentally out of step with the immature oocyte stage. The oocyte is not passive and has been shown to be a central regulator of follicular cell function in mammals by secreting soluble growth factors that act on surrounding follicular cells to regulate a broad range of granulosa cell and cumulus cell functions, including differentiation, proliferation, apoptosis and luteinization (Gilchrist et al., 2008; Huang et al., 2010; Wells, 2010). We assume that ‘cumulus cells’ adjusted to

Page 24 of 55

25 the oocyte.

The co-culture of in-vitro maturing oocytes with cumulus cells from mature oocytes is possible when mature oocytes are available in a patient. In most of the patients included in the IVF programme, including cancer patients who cryostore their oocytes or embryos before oncotherapy, at least one mature oocyte is available after ovarian stimulation. The CC-IVM is not possible in rare situations when no mature oocytes are available in a patient after hormonal ovarian stimulation and when immature oocytes are retrieved from non-stimulated ovaries of patients with polycystic ovary syndrome to prevent ovarian hyperstimulation or directly from the surgically retrieved ovarian tissue before oncotherapy to preserve the fertility; in such cases the oocytes could be matured in vitro in the presence of donated cumulus cells from mature oocytes in other patients included in the IVF programme (after written informed consent).

In conclusion, in-vitro human oocyte maturation (maturation rate and gene expression profile) may be significantly improved by co-culturing with cumulus cells from mature oocytes.

Acknowledgements The authors would like to thank the embryologists and gynaecologists of the Reproductive Unit in the Department of Obstetrics and Gynaecology, University Medical Centre Ljubljana, for collaboration with the clinical programme involving IVF. The authors would also like to thank all the patients who kindly donated their

Page 25 of 55

26 oocytes for this research. This work was financially supported by the German Federal Ministry of Education and Research (BMBF; pilot study German–Slovenian in cooperation with Professor Skutella) and the University Medical Centre Ljubljana (tertiary research grant No. 20150005 to Professor Virant-Klun). References

Albertini, D.F., Combelles, C.M., Benecchi, E., Carabatsos, M.J., 2001. Cellular basis for paracrine regulation of ovarian follicle development. Reproduction 121, 647 –653.

Alexa, A., Rahnenfuhrer, J., 2010. topGO: Enrichment analysis for Gene Ontology. R package version 2.22.0. Alexa, A., Rahnenfuhrer, J., Lengauer, T., 2006. Improved scoring of functional groups from gene expression data by decorrelating GO graph structure. Bioinformatics 22, 1600–1607. American Society for Reproductive Medicine (ASRM), 2013. In vitro maturation: a committee opinion. The Practice Committees of the American Society for Reproductive Medicine and the Society for Assisted Reproductive Technology. Fertil. Steril. 99, 663–666. Assou, S., Anahory, T., Pantesco, V., Le Carrour T, Pellestor F, Klein B, Reyftmann L, Dechaud H, De Vos J., Hamamah, S., 2006. The human cumulus-oocyte complex gene-expression profile. Hum. Reprod. 21, 1705–1719.

Caspar, D.L., Goodenough, D.A., Makowski, L., Phillips, W.C., 1977. Gap junction structures. I. Correlated electron microscopy and x-ray diffraction. J. Cell. Biol. 74, 605–628.

Page 26 of 55

27 Chang, H., Brown, C.W., Matzuk, M.M., 2002. Genetic analyses of the mammalian transforming growth factor beta superfamily. Endocr. Rev. 23, 787-823. Coticchio, G., Guglielmo, M.C., Albertini, D.F., Dal Canto, M., Mignini Renzini M., De Ponti, E., Fadini, R., 2014. Contributions of the actin cytoskeleton to the emergence of polarity during maturation in human oocytes. Mol. Hum. Reprod. 20, 200-207. Coticchio, G., Dal Canto, M., Fadini, R., Mignini Renzini, M., Guglielmo, M.C., Miglietta, S., Palmerini, M.G., Macchiarelli, G., Nottola, S.A., 2016. Ultrastructure of human oocytes after in vitro maturation. Mol. Hum. Reprod. 22, 110-118. Cui, X.S., Li, X.Y., Yin, X.J., Kong, I.K., Kang, J.J., Kim, N.H., 2007. Maternal gene transcription in mouse oocytes: genes implicated in oocyte maturation and fertilization. J. Reprod. Dev. 53, 405–418. Curnow, E.C., Ryan, J., Saunders, D., Hayes, E.S., 2008. Bovine in vitro oocyte maturation as a model for manipulation of the gamma-glutamyl cycle and intraoocyte glutathione. Reprod. Fertil. Dev. 20, 579 –588. Dahan, M.H., Tan, S.L., Chung, J., Son, W.Y., 2016. Clinical definition paper on in vitro maturation of human oocytes. Hum. Reprod., 31, 1383-1386. Dalbiès-Tran, R., Mermillod, P., 2003. Use of heterologous complementary DNA array screening to analyze bovine oocyte transcriptome and its evolution during in vitro maturation. Biol. Reprod. 68, 252-261. Dalton, C.M., Szabadkai, G., Carroll, J., 2014. Measurement of ATP in single

Page 27 of 55

28 oocytes: impact of maturation and cumulus cells on levels and consumption. J. Cell. Physiol. 229, 353-361.

de Matos, D.G., Furnus, C.C., Moses, D.F., 1997. Glutathione synthesis during in vitro maturation of bovine oocytes: role of cumulus cells. Biol. Reprod. 57, 1420 –1425. de Matos, D.G., Gasparrini, B., Pasqualini, S.R., Thompson, J.G., 2002. Effect of glutathione synthesis stimulation during in vitro maturation of ovine oocytes on embryo development and intracellular peroxide content. Theriogenology 57, 1443 –1451.

Dong, J., Albertini, D.F., Nishimori, K., Kumar, T.R., Lu, N., Matzuk, M.M., 1996. Growth Differentiation factor 9 is required during early ovarian folliculogenesis. Nature 383, 531-535.

Dunning, K.R., Russell, D.L., Robker, R.L., 2014. Lipids and oocyte developmental competence: the role of fatty acids and β -oxidation. Reproduction 148, R15-27. Fair, T., Carter, F., Park, S., Evans, A.C., Lonergan, P., 2007. Global gene expression analysis during bovine oocyte in vitro maturation. Theriogenology 68 (Suppl 1), S91-97. Feuerstein, P., Cadoret, V., Dalbies-Tran, R., Guérif, F., Royère, D., 2006. Oocyte cumulus dialog. Gynecol. Obstet. Fertil. 34, 793-800. Franciosi, F., Lodde, V., Goudet, G., Duchamp, G., Deleuze, S., Douet, C.,

Page 28 of 55

29 Tessaro, I., Luciano, A.M., 2012. Changes in histone H4 acetylation during in vivo versus in vitro maturation of equine oocytes. Mol. Hum. Reprod. 18, 243252.

Furger, C., Cronier, L., Poirot, C., Pouchelet, M., 1996. Human granulosa cells in culture exhibit functional cyclic AMP-regulated gap junctions. Mol. Hum. Reprod. 2, 541–548.

Galloway S.M., McNatty, K.P., Cambridge, L.M., Laitinen, M.P., Juengel, J.L., Jokiranta, T.S., McLaren, R.J., Luiro, K., Dodds, K.G., Montgomery, GW., Beattie, A.E., Davis, G.H., Ritvos, O., 2000. Mutations in an oocytederived growth factor gene (BMP15) caused increased ovulation rate and infertility in a dosage sensitive manner. Nat. Genet. 25, 279 -283.

Gasca, S., Pellestor, F., Assou, S., Loup, V., Anahory, T., Dechaud, H., De Vos, J., Hamamah, S., 2007. Identifying new human oocyte marker genes: a microarray approach. Reprod. BioMed. Online 14, 175–183. Ge, H., Tollner, T.L., Hu, Z., Da, M., Li, X., Guan, H., Shan, D., Lu, J., Huang, C., Dong, Q., 2012. Impaired mitochondrial function in murine oocytes is associated with controlled ovarian hyperstimulation and in vitro maturation. Reprod. Fertil. Dev. 24, 945-952.

Gershon, E., Plaks, V., Dekel, N., 2008. Gap junctions in the ovary: expression, localization and function. Mol. Cell. Endocrinol. 282, 18–25.

Gilchrist, R.B., Lane, M., Thompson, J.G., 2008. Oocyte-secreted factors: regulators of cumulus cell function and oocyte quality. Hum. Reprod.

Page 29 of 55

30 Update 14, 159-177.

Gilchrist, R.B., 2011. Recent insights into oocyte-follicle cell interactions provide opportunities for the development of new approaches to in vitro maturation. Reprod. Fertil. Dev. 23, 23 –31.

Goud, P.T., Goud, A.P., Qian, C., Laverge, H., Van der Elst, J., De Sutter, P., Dhont, M., 1998. In-vitro maturation of human germinal vesicle stage oocytes: role of cumulus cells and epidermal growth factor in the culture medium. Hum. Reprod. 13, 1638-1644.

Granot, I., Dekel, N., 2002. The ovarian gap junction protein connexin43: regulation by gonadotropins. Trends Endocrinol. Metab. 13, 310–313.

Grossman, D., Kalo, D., Gendelman, M., Roth, Z., 2012. Effect of di-(2 ethylhexyl) phthalate and mono-(2-ethylhexyl) phthalate on in vitro developmental competence of bovine oocytes. Cell. Biol. Toxicol. 28, 383-396. Hassan, H.A., 2001. Cumulus cell contribution to cytoplasmic maturation and oocyte developmental competence in vitro. J. Assist. Reprod. Genet. 18, 539-543. Hennet, M.L., Combelles, C.M., 2012. The antral follicle: a microenvironment for oocyte differentiation. Int. J. Dev. Biol. 56, 819-831.

Huang, Z., Wells, D., 2010. The human oocyte and cumulus cells relationship: new insights from the cumulus cell transcriptome. Mol. Hum. Reprod. 16, 715–725.

Huang, J., Li, T., Ding, C.H., Brosens, J., Zhou, C.Q., Wang, H.H., Xu, Y.W.,

Page 30 of 55

31 2012. Insufficient histone-3 lysine-9 deacetylation in human oocytes matured in vitro is associated with aberrant meiosis. Fertil. Steril. 97, 178-184. Jeseta, M., Ctvrtlikova Knitlova, D., Hanzalova, K., Hulinska, P., Hanulakova, S., Milakovic, I., Nemcova, L., Kanka, J., Machatkova, M., 2014. Mitochondrial patterns in bovine oocytes with different meiotic competence related to their in vitro maturation. Reprod. Domest. Anim. 49, 469-475. Jones, G.M., Cram, D.S., Song, B., Magli, M.C., Gianaroli, L., Lacham-Kaplan, O., Findlay, J.K., Jenkin, G., Trounson, A.O., 2008. Gene expression profiling of human oocytes following in vivo or in vitro maturation. Hum. Reprod. 23, 1138-1144. Johnson, J.E., Higdon, H.L. 3rd, and Boone, W.R., 2008. Effect of human granulosa cell co-culture using standard culture media on the maturation and fertilization potential of immature human oocytes. Fertil. Steril. 90, 1674-1679. Katz-Jaffe, M.G., McCallie, B.R., Preis, K.A., Filipovits, J., Gardner, D.K., 2009. Transcriptome analysis of in vivo and in vitro matured bovine MII oocytes. Theriogenology 71, 939-946. Knight, P.G., Glister, C., 2006. Focus on TGF-β Signalling: TGF-β superfamily members and ovarian follicle development. Reproduction 132, 191 -206.

Lee, H.J., Barad, D.H., Kushnir, V.A., Shohat-Tal, A., Lazzaroni-Tealdi, E., Wu, Y.G. and Gleicher, N., 2016. Rescue in vitro maturation (IVM) of immature oocytes in stimulated cycles in women with low functional ovarian reserve (LFOR). Endocrine 52, 165-171.

Li, H.J., Sutton-McDowall, M.L., Wang, X., Sugimura, S., Thompson, J.G.,

Page 31 of 55

32 Gilchrist, R.B., 2016. Extending prematuration with cAMP modulators enhances the cumulus contribution to oocyte antioxidant defence and oocyte quality via gap junctions. Hum. Reprod. 31, 810-821.

Lin, Y.H., Hwang, J.L., Seow, K.M., Huang, L.W., Chen, H.J., Tzeng, C.R., 2009. Effects of growth factors and granulosa cell co-culture on in-vitro maturation of oocytes. Reprod. Biomed. Online 19, 165-170.

Makowski, L., Caspar, D.L., Phillips, W.C., Goodenough, D.A., 1977. Gap junction structures. II. Analysis of the x-ray diffraction data. J. Cell Biol. 74, 629–645.

Mamo, S., Carter, F., Lonergan, P., Leal, C.L., Al Naib, A., McGettigan, P., Mehta, J.P., Evans, A.C., Fair, T., 2011. Sequential analysis of global gene expression profiles in immature and in vitro matured bovine oocytes: potential molecular markers of oocyte maturation. BMC Genomics 12, 151. Milakovic, I., Jeseta, M., Hanulakova, S., Knitlova, D., Hanzalova, K., Hulinska, P., Machal, L., Kempisty, B., Antosik, P., Machatkova, M., 2015. Energy Status Characteristics of Porcine Oocytes During In vitro Maturation is Influenced by Their Meiotic Competence. Reprod Domest Anim. 50, 812-819.

Nagai, T., 2001. The improvement of in vitro maturation systems for bovine and porcine oocytes. Theriogenology 55, 1291–1301.

Ouandaogo, Z.G., Haouzi, D., Assou, S., Dechaud, H., Kadoch, I.J., De Vos, J., Hamamah, S., 2011. Human cumulus cells molecular signature in relation to oocyte nuclear maturity stage. PLoS One 6, e27179.

Page 32 of 55

33 Ouandaogo, Z.G., Frydman, N., Hesters, L., Assou, S., Haouzi, D., Dechaud, H., Frydman, R., Hamamah, S., 2012. Differences in transcriptomic profiles of human cumulus cells isolated from oocytes at GV, MI and MII stages after in vivo and in vitro oocyte maturation. Hum. Reprod. 27, 2438-2447.

Revelli, A, Molinari, E., Salvagno, F., Delle Piane, L., Dolfin, E., Ochetti, S., 2012. Oocyte cryostorage to preserve fertility in oncological patients. Obstet. Gynecol. Int. 1912, 525896. Roesner, S., Von Wolff, M., Eberhardt, I., Beuter-Winkler, P., Toth, B., Strowitzki, T., 2012. In vitro maturation: a five-year experience. Acta Obstetricia et Gynecologica Scandinavica 91, 22–27. Sanchez-Lazo, L., Brisard, D., Elis, S., Maillard, V., Uzbekov, R., Labas, V., Desmarchais, A., Papillier, P., Monget, P., Uzbekova, S., 2016. Fatty acid synthesis and oxidation in cumulus cells support oocyte maturation in bovine. Mol. Endocrinol. 28, 1502-1521. Sato, C., Shimada, M., Mori, T., Kumasako, Y., Otsu, E., Watanabe, H., Utsunomiya, T., 2007. Assessment of human oocyte developmental competence by cumulus cell morphology and circulating hormone profile. Reprod. Biomed. Online 14, 49-56.

Simon, A.M., Goodenough, D.A., Li, E., Paul, D.L., 1997. Female infertility in mice lacking connexin 37. Nature 385, 525–529.

Siristatidis, C., Sergentanis, T.N., Vogiatzi, P., Kanavidis, P., Chrelias, C., Papantoniou, N., Psaltopoulou, T., 2015. In vitro Maturation in Women with vs. without Polycystic Ovarian Syndrome: A Systematic Review and

Page 33 of 55

34 Meta-Analysis. PLoS One 10, e0134696. Somfai, T., Kikuchi, K., Kaneda, M., Akagi, S., Watanabe, S., Mizutani, E., Haraguchi, S., Dang-Nguyen, T.Q., Inaba, Y., Geshi, M., Nagai, T., 2011. Cytoskeletal abnormalities in relation with meiotic competence and ageing in porcine and bovine oocytes during in vitro maturation. Anat. Histol. Embryol. 40, 335-344.

Sowinska, N., Müller, K., Niżański ,W., Jewgenow, K., 2017. Mitochondrial characteristics in oocytes of the domestic cat (Felis catus) after in vitro maturation and vitrification. Reprod. Domest. Anim. 52, 806-813.

Sutovsky, P., Schatten, G., 1997. Depletion of glutathione during bovine oocyte maturation reversibly blocks the decondensation of the male pronucleus and pronuclear apposition during fertilization. Biol. Reprod. 56, 1503 – 1512. Takeo, T., Horikoshi, Y., Nakao, S., Sakoh, K., Ishizuka, Y., Tsutsumi, A., Fukumoto, K., Kondo, T., Haruguchi, Y., Takeshita, Y., Nakamuta, Y., Tsuchiyama, S., Nakagata, N., 2015. Cysteine analogs with a free thiol group promote fertilization by reducing disulfide bonds in the zona pellucida of mice. Biol. Reprod. 92, 90. Thanaboonyawat, I., Makemaharn, O., Petyim, S., Laokirkkiat, P., Choavaratana, R., 2016. The correlation of cumulus mucification patterns with oocyte maturation rate in vitro in FSH + LH-primed IVM cycles: a prospective study. Arch. Gynecol. Obstet. 293, 681-686.

Tsai, M.Y., Lan, K.C., Huang, K.E., Huang, F.J., Kung, F.T., Chang, S.Y., 2003.

Page 34 of 55

35 Significance of mRNA levels of connexin37, connexin43, and connexin45 in luteinized granulosa cells of controlled hyperstimulated follicles. Fertil. Steril. 80, 1437–1443.

Valsangkar, D.S., Downs, S.M., 2013. A requirement for fatty acid oxidation in the hormone-induced meiotic maturation of mouse oocytes. Biol. Reprod. 89, 43. Vigone, G., Merico, V., Redi, C.A., Mazzini, G., Garagna, S., Zuccotti, M., 2015. FSH and LH receptors are differentially expressed in cumulus cells surrounding developmentally competent and incompetent mouse fully grown antral oocytes. Reprod. Fertil. Dev. 27, 497-503. Walls, M.L., Hart, R., Keelan, J.A., Ryan, J.P., 2016. Structural and morphologic differences in human oocytes after in vitro maturation compared with standard in vitro fertilization. Fertil. Steril. 106, 1392-1398. Walls, M.L, Hunter, T., Ryan, J.P., Keelan, J.A., Nathan, E., Hart, R.J., 2015. In vitro maturation as an alternative to standard in vitro fertilization for patients diagnosed with polycystic ovaries: a comparative analysis of fresh, frozen and cumulative cycle outcomes. Hum. Reprod. 30, 88-96.

Wang, Q.T., Piotrowska, K., Ciemerych, M.A., Milenkovic, L., Scott, M.P., Davis, R.W., Zernicka-Goetz, M., 2004. A genome-wide study of gene activity reveals developmental signaling pathways in the preimplantation mouse embryo. Dev. Cell. 6, 133–144.

Wang, X., Tsai, T., Qiao, J., Zhang, Z., Feng, H.L., 2014. Impact of

Page 35 of 55

36 gonadotropins on oocyte maturation, fertilisation and developmental competence in vitro. Reprod. Fertil. Dev. 26, 752-757. Wells, D., Patrizio, P., 2008. Gene expression profiling of human oocytes at different maturational stages and after in vitro maturation. Am. J. Obstet. Gynecol. 198, 455.e1–455.e9. Yoon, S.J., Kim, K.H., Chung, H.M., Choi, D.H., Lee, W.S., Cha, K.Y., Lee, K.A., 2006. Gene expression profiling of early follicular development in primordial, primary, and secondary follicles. Fertil. Steril. 85, 193–203.

Zhang, P., Kerkelä, E., Skottman, H., Levkov, L., Kivinen, K., Lahesmaa, R., Hovatta, O., Kere, J. 2007. Distinct sets of developmentally regulated genes that are expressed by human oocytes and human embryonic stem cells. Fertil. Steril. 87, 677-690. Zhang, A., Xu, B., Sun, Y., Lu, X., Niu, Z., Chen, Q., Feng, Y., Xu, C. J., 2012. The effect of human cumulus cells on the maturation and developmental potential of immature oocytes in ICSI cycles. J. Assist. Reprod. Genet. 29, 313-319. Zhao HC, Ding T, Ren Y, Li TJ, Li R, Fan Y, Yan J, Zhao Y, Li M, Yu Y, Qiao J., 2016. Role of Sirt3 in mitochondrial biogenesis and developmental competence of human in vitro matured oocytes. Hum. Reprod. 31, 607622. Zuccotti, M., Giorgi Rossi, P., Martinez, A., Garagna, S., Forabosco, A., Redi, C.A., 1998. Meiotic and developmental competence of mouse antral oocytes. Biol. Reprod. 58, 700-704.

Page 36 of 55

37 Declaration The authors report no financial or commercial conflicts of interest.

Figure 1 – Oocytes matured in vitro by a new approach that involves co-culturing with cumulus cells (arrows) retrieved by denudation of mature oocytes in the same patient. (A–C) In-vitro maturation of oocytes in droplets of culture medium under paraffin oil; (D–F) cumulus cells; (G–O) maturing oocytes in a co-culture with cumulus cells under an inverted microscope. Scale bars: (D, G–L) 100 m, (E, M–O) 50 m, (F) 10 m.

Figure 2 – Gene ontology analysis of differentially expressed genes based on their biological process: (A) for genes upregulated in the in-vivo matured oocytes; and (B) genes upregulated in the in-vitro matured oocytes (CC-IVM and CONV-IVM combined). CC, cumulus cell; conv, conventional; IVM, in-vitro maturation.

Figure 3 – Gene expression profiles of the analysed groups of oocytes. (A) Principal component analysis: scatterplot of the first two principal components. All genes above the background were included. The different groups of oocytes are shown in different colours: in-vivo matured oocytes (three samples S01-S03) are black, oocytes matured in vitro in a conventional way (CONV-IVM; three samples S04-S06) are red, oocytes matured in vitro by the new co-culture approach (CC-IVM; three samples S07-S09) are green, and non-matured oocytes (three samples S10-S12) are blue; (B) a heatmap of the 250 genes with the highest differences in all samples (highest SD); (C)

Page 37 of 55

38 a heatmap of the 250 genes with the lowest expression in the in vivo-matured oocytes. Genes and samples are clustered using hierarchical clustering with complete linkage. CC, cumulus cell; conv, conventional; IVM, in-vitro maturation.

Figure 4 – Gene expression profile of oocytes matured in vitro with cumulus cell coculture (CC-IVM) compared with oocytes matured in vitro in the conventional way (CONV-IVM). Gene ontology analysis of genes that were upregulated in the CC-IVM oocytes on (A) biological process and (B) cellular component. CC, cumulus cell; conv, conventional; IVM, in-vitro maturation.

Figure 5 – Volcano plots of T-tests comparing the different groups of oocytes: in vivo-matured (IN VIVO) oocytes with (A) oocytes matured in vitro by the new coculture approach (CC-IVM); (B) oocytes matured in vitro in the conventional way (CONV-IVM), (C) oocytes matured in vitro in the conventional way (CONV-IVM) and the new co-culture approach (CC-IVM); (D) in vivo-matured oocytes with the other three groups combined; and (E) non-mature oocytes with the three matured oocyte groups. Genes assigned to desired functions are in all plots coloured: greenembryogenesis, blue-transcription factors, red-epigenetic, and grey-rest. The horizontal line (many similar P-values) results from a multiple testing correction with the Benjamini–Hochberg procedure. CC, cumulus cell; conv, conventional; IVM, invitro maturation.

Page 38 of 55

39 Figure 6 – Heatmap clustering for genes related to (A) transcription factor activity; (B) embryogenesis; (C) epigenetics; (D) development; and (E) the cell cycle. (F) Hierarchical clustering for cell cycle-related genes. CC, cumulus cell; conv, conventional; IVM, in-vitro maturation.

Figure 7 – The top 25 differentially expressed genes comparing in vivo-matured oocytes (in-vivo) versus all the other oocyte groups: CONV-IVM (matured in vitro in a conventional way), CC-IVM (new co-culture maturation method) and not matured ordered according to their absolute fold change. CC, cumulus cell; conv, conventional; IVM, in-vitro maturation.

Page 39 of 55

Table 1 – Number of transcripts and genes, which were expressed and not expressed in analysed samples of oocytes (to be expressed as a gene/transcript had a normalized expression value of 2.2 or more for all three biological replicates) and upregulated and downregulated genes compared with in-vivo matured oocytes. Groups of oocytes In-vivo matured CONV-IVM CC-IVM Not matured Transcripts and genes Transcripts

28117

29610

29717

28882

Not expressed genes

25391

26649

26626

25979

Not expressed transcripts

22566

21073

20966

21801

17895

18204

18734

942

663

104

3248

862

92

Expressed genes

Upregulated genes compared with in-vivo matured oocytes

19585

Downregulated genes compared with in-vivo matured oocytes

Page 40 of 55

41 Table 2 – Genes related to epigenetics, transcription factor activity, embryogenesis, cell cycle and development according to the manner of oocyte maturation and in not matured oocytes. Functional groups of genes and their up- or down-regulation Epigenetics (184 transcripts)

Maturation of oocytes

UP

Transcription factors

Embryogenesis

Cell cycle

Development

(1025 transcripts)

(1683 transcripts)

(6019 transcripts)

(2127 transcripts)

DOWN

UP

DOWN

UP

DOWN

UP

DOWN

UP

DOWN

In vivo versus CCIVM

1

0

42

16

20

10

32

29

109

80

In vivo versus CONVIVM

3

4

112

56

60

28

76

75

326

178

1

33

28

14

23

33

30

66

117

5

59

32

44

19

55

48

223

120

CONVIVM versus CC-IVM

In vivo versus not matured

1

6

CC, cumulus cell; conv, conventional; IVM, in-vitro maturation.

Page 41 of 55

42 Table 3 – Most significant pathways from KEGG and WikiPathway using the differently expressed genes from T-test comparing in-vivo matured oocytes with in-vitro matured (CONV-IVM and CC-IVM) and not matured oocytes. For the enrichment analysis we selected the top 300 genes with the highest fold change. Source

Pathway

P-value

Adjusted P-value

Genes

KEGG

Proteasome

0.0003

0.0295

PSMA6, PSMB1, PSMB7, PSMC1, PSMD1

KEGG

Oxidative phosphorylation

0.0011

0.0414

ATP5G1, NDUFA4, ATP5H, ATP5G2, ATP5E, ATP6V0E1, UQCRQ

KEGG

Ribosome

0.0093

NS

MRPL20, MRP521, MRPL24, MRP56, RP58, RPL21

KEGG

Steroid biosythesis

0.0309

NS

LIPA, HSD17B7

KEGG

Mismatch repair

0.0396

NS

M5H6, SSBP1

KEGG

Alpha-Linolenic acid metabolism

0.0426

NS

PLA2G12A, PLA2G4C

KEGG

Protein export

0.0426

NS

SEC61G, SPCS1

WikiPathway

Oxidative phosphorylation

<0.0001

0.0046

ATP5G1, NDUFA4, ATP5H, ATP5G2, ATP5E

WikiPathway

Proteasome degradation

0.0012

0.0608

PSMA6, PSMB1, PSMB7, PSMC1, PSMD1

WikiPathway

Electron chain

0.0044

NS

ATP5G1, NDUFA4, ATP5H, ATP5G2, ATP5E

WikiPathway

One carbon donor

0.0132

NS

MTHFR

WikiPathway

TCA cycle

0.0231

NS

SUCLG1, IDH3A

WikiPathway

Parkin-Ubiquitin proteosomal system pathway

0.0239

NS

TUBA1B, PSMC1, PSMD1

WikiPathway

Glycine metabolism

0.0392

NS

MTHFR

WikiPathway

Phase I biotransformation, non P450

0.0519

NS

LIPA

WikiPathway

TOR signaling

0.0628

NS

AKT151, PRKAA1

WikiPathway

Energy metabolism

0.0856

NS

TFAM, PRKAA1

transport

NS-non-significant.

Page 42 of 55

43

Page 43 of 55

44

RBMO Figure 1_Virant-Klun.jpg

Page 44 of 55

45

RBMO Figure 2_Virant-Klun.jpg

Page 45 of 55

46

RBMO Figure 3_Virant-Klun.jpg

Page 46 of 55

47

RBMO Figure 4_Virant-Klun.jpg

Page 47 of 55

48

RBMO Figure 5_Virant-Klun.jpg

Page 48 of 55

49

Page 49 of 55

50 RBMO Figure 6_Virant-Klun.jpg

Page 50 of 55

51

RBMO Figure 7_Virant-Klun.png

Page 51 of 55

52

RBMO Suppl. Figure S3_Virant-Klun.png

Page 52 of 55

53

RBMO Supplementary Figure S1_Virant-Klun.jpg

Page 53 of 55

54

RBMO Supplementary Figure S2_Virant-Klun.jpg

Page 54 of 55

55

RBMO-photo-Irma Virant-Klun.JPG

Page 55 of 55