Introduction
Oocyte cryopreservation has become a clinically viable option for women who are facing age-related or iatrogenic decreases in both the quality and quantity of their oocytes because of declining oocyte quality, invasive cancer treatment, premature ovarian failure, and polycystic ovary syndrome [
1]. Oocyte cryopreservation is a breakthrough technique for women who elect to delay childbearing, require gonadotoxic therapy, or perform egg donation [
2–
4].
The first human oocyte reported to be cryopreserved was preserved via the slow-freezing technique, in which the rate of cooling can be programmed in low concentrations of cryoprotectant agents to minimize intracellular ice crystal formation. Over the last decade, vitrification has replaced slow-freezing as the technique of choice for cryopreserving oocytes. This method is performed with an ultrarapid cooling time and a very high cryoprotectant concentration to minimize ice crystal formation, thereby resulting in high rates of cell survival, fertilization, embryo development, and pregnancy [
5–
8]. At present, approximately 4000 live births have been obtained via oocyte vitrification since the first birth of a healthy baby following oocyte vitrification in 1999 [
9,
10].
As the largest human cells, oocytes have relatively lower surface area-to-volume ratios and are more likely to retain water during the freezing process than other cell types. Therefore, oocytes are more vulnerable to damage caused by intracellular ice crystal formation during cryopreservation procedures than other cells. Moreover, dramatic osmotic changes in cryoprotective agents may further seriously damage oocytes [
11]. Cryopreservation-induced oocyte damage includes zona pellucida thickening, premature cortical granule exocytosis, meiotic spindle disruption, and aneuploidy [
12–
16]. Additionally, studies on ovine [
17], bovine [
18], canine [
19], murine [
20], and human [
21,
22] oocyte cryopreservation using the vitrification method and human studies using slow freezing [
8,
22,
23] have revealed that cryopreservation deregulates a subset of genes that involved in oxidative stress, apoptosis, cell cycle, chromosomal organization, RNA splicing and processing, DNA repair, cellular response to DNA damage and to stress, calcium ion binding, malate dehydrogenase activity, and mitochondrial activity. Furthermore, Stigliani
et al. [
8] demonstrated that the length of storage time does not alter the gene expression profiles of human metaphase II (MII) oocytes following slow freezing. Hence, we can infer from available evidence that the potential damage resulting from oocyte cryopreservation might be attributed to the cryopreservation process per se rather than storage time.
Aside from inducing dysregulated gene expression, the slow-freezing cryopreservation technique reportedly affects cellular proteomics. Studies on mice have shown that exposure of mouse oocytes to 1,2-propanediol can alter oocyte proteins and thus affect fertilization and early embryo development [
24,
25]. Following the slow-freezing technique, free-radical production due to disruption of the oocyte plasma membrane may induce protein efflux, leading to downregulation of proteins. By contrast, the exposure of oocytes to cytotoxic cryoprotectant agents may induce upregulation of several stress proteins.
Among the indicators of the efficacy and safety of cryopreservation techniques, implantation ability and subsequent pregnancy rates are the most critical factors [
7,
26–
30]. However, at present, only limited information about the long-term effects of cryopreservation after thawing on human oocytes and embryos is available. Recently, Stigliani
et al. [
8] demonstrated that storage time does not alter the gene expression profiles of human MII oocytes following the slow-freezing method. Nevertheless, no study has been designed to evaluate whether storage time can affect the gene expression of vitrified-thawed human oocytes. Therefore, this study aimed to investigate the effects of different storage durations after open vitrification on the gene expression patterns in human MII oocytes at the single-cell level. Herein, we compared the transcriptomes of vitrified-thawed human MII oocytes after 1, 2, 3, and 12 months of storage in liquid nitrogen with those of fresh ones using single-cell RNA sequencing analysis.
Materials and methods
Ethical approval and informed consent
This study was reviewed and approved by the Reproductive Study Ethics Committee of Peking University Third Hospital (Research License 2013SZ017). All oocytes were obtained from patients undergoing in vitro fertilization (IVF) who decided to donate their surplus oocytes. Written informed consent was obtained from all participants.
Patient selection and oocyte collection
All MII oocytes used in the current study were donated by patients undergoing IVF at the Center for Reproductive Medicine of Peking University Third Hospital following standard clinical protocols. The data for a control group consisting of three fresh human MII oocytes were downloaded from our previously published research [
31]. Four patients aged 30–32 years who were diagnosed with tubal factor infertility without any ovarian pathology were enrolled in this prospective study. A total of 13 oocytes from these four patients were randomly divided into four groups according to storage time: 1-, 2-, 3-, and 12-month groups.
Oocyte vitrification and thawing
MII oocytes were vitrified and thawed as previously described [
10]. In brief, the oocytes were first equilibrated in 7.5% (v/v) ethylene glycol (EG, Sigma, cat. 293237, USA) + 7.5% (v/v) dimethyl sulfoxide (DMSO, Sigma, cat. D2650, USA) solution for 5 min at room temperature. These oocytes were then transferred into vitrification solutions composed of 15% (v/v) EG+ 15% (v/v) DMSO+ 0.5 mol/L sucrose (Sigma, cat. V900116, USA) for less than 1 min at room temperature. Finally, these oocytes were immediately loaded onto sterile iVitri straws (Reprobiotech Corp., cat. RBC-S-008, USA) and directly transferred into liquid nitrogen for storage lasting 1, 2, 3, or 12 months.
The frozen oocytes were thawed step by step by using different concentrations of sucrose solution. First, the oocytes were expelled from the straws into 1.0 mol/L sucrose and incubated for 3 min at 37 °C. These oocytes were then successively transferred into 0.5, 0.25, and 0 mol/L sucrose medium for 3 min at room temperature. Finally, the thawed oocytes were cultured in human tubal fluid (LifeGlobal, cat. LGGF-100, Belgium) medium at 37 °C under 5% CO2 in humidified air for 2 h.
Single-cell transcriptome library preparation
Individual oocytes were first isolated and added to lysate buffer by mouth pipetting. These single-cell samples were then directly amplified following the Smart-Seq2 method by Genome.cn. The DNA concentration of each library was measured using a Qubit 2.0 Fluorometer (Life Technologies, CA, USA). The fragment distribution of the amplified products was detected using an Agilent 2100 high-sensitivity DNA assay kit (Agilent Technologies, CA, USA).
RNA-Seq library preparation and sequencing
Amplified cDNA products were used as input for library construction of the single-cell transcriptomes. First, the sample cDNA was sheared into 300 bp fragments by using a Bioruptor® sonication system (Diagenode Inc.). Second, terminal repair, addition of A-tailing, and ligation of sequencing adapters to the fragmented cDNA were performed. The sequencing adapter products were amplified using PCR, and the samples were distinguished from each other by adding different index tags. Finally, electrophoresis was performed with the PCR amplification products in 2% agarose gels, the 350–450 bp DNA bands were extracted from the gels with a CWBIO gel extraction kit, and the DNA was dissolved to create the final libraries.
After library construction, the insertion size was assessed with an Agilent 2100 bioanalyzer (Agilent Technologies, CA, USA), and the accurate insertion size was quantified with a TaqMan fluorescent probe by using an AB StepOnePlus real-time PCR system (valid library concentration>10 nmol/L).
In addition, the index-coded samples were clustered using a HiSeq PE cluster kit v4-cBot-HS (Illumina) according to the manufacturer’s instructions. Finally, the sequences of the libraries were analyzed on an Illumina HiSeq platform with 150 bp paired-end sequencing.
Single-cell RNA-Seq data analysis
The single-cell RNA-Seq data was analyzed as previously described [
32,
33]. In brief, the raw sequencing data were processed to remove the library adapter sequences, amplification primers, and 3′ polyA sequences. Then, the clean data were aligned to the human genome 19 by using TopHat2. Afterward, Cufflinks was used to calculate the gene expression levels in the form of fragments per kilobase of transcript per million fragments mapped reads (FPKM) values. Only genes with FPKM values>1 in at least one sample were analyzed for their expression levels to ensure accuracy and remove the auxiliary data. The data were transformed by log
2(
x+1) for the subsequent analyses.
The expression patterns of different oocytes were then analyzed via principal component analysis (PCA). Transcript integrity number (TIN) was applied to measure RNA degradation in the oocytes. Differentially expressed genes for different groups were evaluated by multiple t-tests. False discovery rate (FDR)-corrected P values were calculated using the Benjamini–Hochberg method. The differentially expressed genes were defined on the basis of multiple t-tests with criteria of P<0.05, FDR<0.05, and a fold change (log2-converted FPKM)>2 or<0.5.
Functional annotation analyses of the differentially expressed genes were performed via Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. The dynamic expression patterns of known long noncoding RNAs (lncRNAs) were analyzed using the GENCODE lncRNA database.
Results
In the current study, we used 16 donated human MII oocytes to compare the transcriptomes of fresh and vitrified-thawed ones with different storage times at the single-cell level. Overall, 110 Gb of raw data were generated with an average of 28.1 million paired reads per oocyte, and the read length was 150 bp.
Transcriptional profiles of different MII oocytes
First, we evaluated the sequencing data quality. The average mapping rate was 91.33%±1.67%. On average, we detected the expression of 12 280 out of 26 364 RefSeq genes, indicating that almost 47% of known human genes were covered by our data.
Then, we analyzed all the data via PCA. Results showed that most of the cryopreserved oocytes clustered together, showing a pattern distinct from that of the fresh oocytes (Fig. 1A).
We applied TIN to measure RNA degradation during the process of oocyte cryopreservation. The TIN metric is a reliable measurement of RNA integrity at both the transcriptome and transcript levels. TIN scores can be calculated after the RNA-Seq data are produced. The relative median TIN of cryopreserved oocytes was approximately 40% that of fresh oocytes (Fig. 1B).
Storage time did not affect the gene expression profiles of mature human oocytes
Pairwise comparisons were performed to analyze the gene expression patterns among the 1-, 2-, 3-, and 12-month groups. However, no differentially expressed genes were found between any two cryopreserved groups (Fig. 1C).
Vitrification-thawing procedures altered the gene expression profiles of mature human oocytes
A total of 1987 differentially expressed genes were found in the 13 vitrified-thawed oocytes compared with the fresh MII oocytes. Among the differentially expressed genes, 1646 genes were downregulated, whereas 341 were upregulated (Fig. 2A).
To obtain insights into their potential functional roles, we further analyzed all 1987 differentially expressed genes via GO and KEGG analyses.
Among the downregulated genes, the main enriched GO terms included 65 different biological process terms. The upregulated genes were related to 29 unique biological processes (Table S1). The important biological processes of the downregulated and upregulated genes are presented in Fig. 2B and 2C, respectively. In particular, cell cycle and meiosis are the two important biological processes for the functions of oocytes. Notably, two meiosis-related genes, namely, TUBGCP5 and NCAPD2, were considerably downregulated (Supplementary Fig. S1A). In addition, cell cycle-related genes, including NCAPD2 and TUBB4B, were significantly downregulated in cryopreserved MII oocytes, whereas RANGAP1 and ANP32E were upregulated (Supplementary Fig. S1B). We also found that two genes related to cellular response to heat, namely, HSPA1A and HSPA1B, were significantly downregulated in cryopreserved MII oocytes (Supplementary Fig. S1C). Other important differentially expressed genes that were closely related to oocyte growth and development, such as microtubule-based process, methylation, ubiquinone biosynthetic process, the tricarboxylic acid cycle, and the MAPK cascade, were graphically displayed (Supplementary Fig. S1D–1H).
Moreover, we performed KEGG analysis to explore the related signaling pathways. We found that some important pathways, including carbon metabolism; metabolic pathways; gap junction; axon guidance; pathogenic Escherichia coli infection; the citrate cycle; pyrimidine metabolism; 2-oxocarboxylic acid metabolism; biosynthesis of antibiotics; alanine, aspartate, and glutamate metabolism; the GnRH signaling pathway; biosynthesis of amino acids; and RNA transport were enriched for the fresh oocytes (Fig. 2D, Supplementary Table S1). Interestingly, oxidative phosphorylation, lysosome pathways, Huntington’s disease pathways, regulation of lipolysis in adipocytes, the synaptic vesicle cycle, and the AMPK signaling pathway were upregulated in the cryopreserved oocytes (Fig. 2E, Table S1).
To further investigate the effects of storage time on gene expression in cryopreserved oocytes, we compared the gene expression profiles of oocytes in different storage time groups with those of fresh ones. Results showed 520 downregulated and 93 upregulated genes in the 1-month group, 499 downregulated and 95 upregulated genes in the 2-month group, 194 downregulated and 28 upregulated genes in the 3-month group, and 575 downregulated and 40 upregulated genes in the 12-month group (Fig. 3A, Tables S2–S5). Possible overlap among these aberrant genes was assessed using Venn diagrams. Results showed 52 overlapping downregulated genes. GO analysis revealed that these 52 genes were enriched for GO terms mainly related to cellular heat acclimation, regulation of protein ubiquitination, and ATP metabolic processes (Fig. 3B and 3D, Table S6). The aforementioned GO terms were mainly associated with HSPA1A and HSPA1B, which were described earlier. As shown in Fig. 3C, only four genes, namely, SCAND1, HERC2P7, PET100 and TRAPPC1, were located at the intersection of the upregulated gene sets. However, the upregulation of these genes might have been a stochastic event, as the genes were not enriched in any pathway.
Dynamic expression of lncRNAs
The expression patterns of lncRNAs were analogous to those of the aforementioned RefSeq genes, as unsupervised hierarchical clustering analysis indicated that the lncRNAs of most cryopreserved oocytes were clustered together, with two exceptions (one belonging to the 3-month group and one belonging to the 12-month group) that presented a distinct cluster pattern (Fig. 4A). A total of 440 lncRNAs were differentially expressed in the 13 vitrified-thawed oocytes compared with the fresh MII oocytes, of which 395 lncRNAs were downregulated and 45 were upregulated (Fig. 4B, Table S7).
A total of 62, 51, 75, and 75 differentially expressed lncRNAs were found in the 1- (58 downregulated and 4 upregulated), 2- (51 downregulated and 0 upregulated), 3- (67 downregulated and 8 upregulated), and 12-month groups (72 downregulated and 3 upregulated), respectively, compared with the control group (Fig. 4C, Tables S8–S11). Specifically, among the significantly downregulated lncRNAs in the vitrified-thawed oocytes, five lncRNAs, namely, CTB-180A7.6, AP000320.7, OOEP-AS1, RP11-59H7.3, and RP4-785G19.5, overlapped, whereas no upregulated lncRNAs were found to overlap among the four cryopreserved groups (Fig. 4D). No differentially expressed lncRNAs were found between any of the two cryopreserved groups, consistent with the results for the differentially expressed genes (Fig. 4E).
Discussion
As the largest cells in the human body, mature MII oocytes have unique biochemical and physical characteristics. Given that they contain a high proportion of water in the cytoplasm, MII oocytes are vulnerable to damage caused by intracellular ice crystal formation during cryopreservation. Currently, slow freezing and vitrification are the two most commonly used methods in freezing oocytes. Recently, with the introduction and advancement of vitrification techniques, notable improvements have been achieved in ensuring the efficacy and safety of oocyte preservation, as demonstrated by enhanced oocyte survival, fertilization, and pregnancy rates [
4,
6,
34–
36].
Prior animal studies have demonstrated that long-term storage of cryopreserved oocytes in liquid nitrogen does not have detrimental effects on the survival, fertilization, embryonic development potential, embryo quality, and gene expression profiles of thawed oocytes [
8,
37,
38]. In the present study, we found that different storage times of vitrified oocytes in liquid nitrogen did not alter the gene expression profiles of the subsequent surviving thawed oocytes. This result was consistent with that of a previous study that reported that the storage time of slow-frozen oocytes does not affect the gene expression profiles of thawed oocytes [
8]. These findings suggest the efficacy and safety of long-term storage of frozen oocytes. Collectively, our study and previous studies indicate that the potential damage resulting from oocyte cryopreservation might be due to the cryopreservation procedure per se rather than the storage time. In this regard, these studies have commercial and clinical applications for long-term oocyte banking that can be potentially used for gamete donation programs and for women suffering from fertility disorders.
In the current study, we found that vitrification-thawing procedures may alter the expression of several genes. Overall, fewer mRNAs were extracted from cryopreserved oocytes than from fresh oocytes, revealing that RNA degradation occurred during cryopreservation. The aberrant genes in vitrified oocytes were closely related to oogenesis and development processes, including meiosis, cell cycle, cellular response to heat, microtubule-based processes, methylation, ubiquinone biosynthetic processes, the tricarboxylic acid cycle, the MAPK cascade, sister chromatid biorientation, DNA repair, oxidative phosphorylation, and ATP metabolic processes. Given that high-quality oocytes rely on the precise expression and production of maternal RNAs and proteins, the negative effects of vitrification-thawing procedures on gene expression may affect oocyte quality, leading to poor embryonic development. These results are in agreement with those of previous reports, indicating that cryopreservation procedures may have detrimental effects on oocyte quality [
8,
17–
20,
22,
23].
HSPA1A and
HSPA1B were the two most representatively downregulated genes in vitrified-thawed oocytes compared with fresh oocytes.
HSPA1A and
HSPA1B are members of the heat shock protein family A (Hsp70). In combination with other heat shock proteins, HSPA1A and HSPA1B proteins are involved in diverse cellular functions, including the regulation of mitosis, the ubiquitin-proteasome pathway, RNA binding, mRNA stabilization, autoimmunity, and inflammatory diseases [
39–
44]. Intriguingly, the expression levels of
HSPA1A and
HSPA1B are correlated with the capacity of embryos to exhibit the heat-shock response, which plays an important role in regulating the successive steps of early embryonic development [
45,
46]. Another finding that merits our attention was the upregulation of
SCAND1 and
TRAPPC1 in all four groups of cryopreserved oocytes. Although such upregulation seemed to be a stochastic event as these genes were not enriched in any pathway,
SCAND1 and
TRAPPC1 have been shown to have regulatory roles in numerous biological processes. Notably,
SCAND1, a SCAN family protein that lacks a zinc finger region, is involved in the regulatory mechanism of transcription [
47–
49]. The gene product of
TRAPPC1, a component of the multisubunit transport protein particle complex, may play a role in regulating Golgi membrane trafficking [
50]. Therefore, alterations in these genes may affect the synthesis of several biological process-related proteins, leading to low oocyte developmental competence. In the present study, we obtained average values for cell-to-cell variability that pave the way for characterizing biological samples at the single-cell level.
In recent years, lncRNAs have attracted the attention of many researchers. In addition to demonstrating that gene expression profiles were affected by vitrification-thawing procedures, this study revealed that
CTB-180A7.6,
AP000320.7,
OOEP-AS1,
RP11-59H7.3, and
RP4-785G19.5 were downregulated in all four cryopreserved groups compared with the control group. Although the functions of lncRNAs are still uncategorized, studies on humans [
31,
51–
53], human-induced pluripotent stem cells [
54], pigs [
55], mouse embryonic stem (ES) cells [
56], and other subjects [
57] have revealed that certain lncRNAs may play important roles in various biological processes, such as chromatin-modifying complex functions, cellular phenotypes, transcriptional regulation, pre-implantation embryonic development, epigenetic regulation, and pluripotency maintenance. Thus, although functional studies are lacking, we speculate that the downregulated lncRNAs such as
CTB-180A7.6 and
RP4-785G19.5 might be molecular substrates that are involved in the harmful effects of vitrification procedures on cryopreserved oocytes.
However, we have to acknowledge several limitations of the present study. The number of analyzed oocytes was relatively low, and the longest cryostorage duration was only 12 months. Future studies with larger sample sizes and in which the oocytes are kept frozen for longer periods are warranted. Since Chen [
58] reported the first live birth of a child developed from cryopreserved oocytes more than 30 years ago, the clinical success rate of oocyte cryopreservation has lagged behind that of embryo cryopreservation [
16]. Optimization of oocyte cryopreservation is far from being finished.
In conclusion, our results indicated that the effects of vitrification on the transcriptomes of mature human oocytes are induced by the procedure itself rather than by the storage time. Furthermore, we found, for the first time, that several dysregulated genes and lncRNAs could be involved in the detrimental effects of the vitrification-thawing procedure on cryopreserved oocytes at the single-cell level. These findings provide insights into the molecular mechanisms by which freezing methods and the related storage times affect vitrified-thawed oocytes.