Blastocyst activation engenders transcriptome reprogram affecting X-chromosome reactivation and inflammatory trigger of implantation

Significance

Our data systematically and comprehensively reveal the genetic programming of blastocyst for implantation in a delayed implantation mouse model. By analyzing the transcriptional profile, we demonstrate dynamic changes in the biosynthesis, metabolism, and cell proliferation during blastocyst reactivation from diapause. We have analyzed the global remodeling of ICM and TE, respectively, and disclosed the critical cellular and molecular events during implantation. Furthermore, our results suggest that an activating blastocyst functions as a proinflammatory body to trigger an inflammatory-like attachment response. Thus, we demonstrate a previously undefined role of blastocyst during implantation. We believe that the findings in this study substantially advance scientific understanding of the inflammatory-attachment response during implantation.

Abstract

Implantation of the blastocyst into the uterus is the gateway for further embryonic development in mammals. Programming of blastocyst to an implantation-competent state known as blastocyst activation is the determining factor for implantation into the receptive uterus. However, it remains largely unclear how the blastocyst is globally programmed for implantation. Employing a delayed implantation mouse model, we show here that the blastocyst undergoes extensive programming essential for implantation. By analyzing the transcriptional profile of blastocysts with different implantation competency, we reveal the dynamic change in the biosynthesis, metabolism, and proliferation during blastocyst reactivation from diapause. We also demonstrate that reactivation of the X chromosome, one of the most important events during periimplantation of female embryonic development, is not completed even in blastocysts under conditions of dormancy, despite long term suspension in the uterus. Moreover, the mural trophectoderm (TE), but not the polar TE, differentiates to be more invasive through the weakened cell-cell tight junctions and extracellular matrices (ECMs). By analyzing the differentially expressed profile of secretory proteins, we further demonstrate that the blastocyst functions as a proinflammatory body to secrete proinflammatory signals, such as TNFα and S100A9, thereby triggering embryo-uterine attachment reaction during implantation. Collectively, our data systematically and comprehensively disclose the programming of blastocyst reactivation from diapause for implantation and uncover previously undefined roles of blastocyst during implantation.

The implantation of the blastocyst into the uterus is a developmental milestone in mammalian embryonic development that determines the outcomes of pregnancy (1). In humans, natural conception per menstrual cycle is limited to approximately 30%, and most failed pregnancies are due to implantation defects (2). Among them, one-third of implantation failures are attributed to a dysfunctional blastocyst (3). In mice, previous studies have shown that the state of activity of the blastocyst is a determining factor in initiating implantation in the receptive uterus (4). In fact, the blastocyst functions as an active unit with distinct molecular programs of cell growth and differentiation during implantation. Under the guidance of ovarian estrogen (E2) and progesterone (P4), the blastocyst is programmed to an implantation-competent state during implantation, a process known as blastocyst activation (4). However, how this global programming of the blastocyst is achieved remains largely unclear, since this process is a progressive event dependent on changing levels of ovarian hormones. Thus, the state of competence relative to implantation is difficult to judge.

The naturally existing phenomenon of embryonic diapause makes it possible to address this issue, as the dormant blastocyst shows almost all characteristics of the normal blastocyst except for implantation competency and metabolic activity (4–6). For example, the components of the dormant blastocyst are survival competent and pluripotent, evidenced by the fact that the first embryonic stem cell line was derived from a dormant blastocyst (7). Blastocyst dormancy or embryonic diapause, a condition also referred to as delayed implantation, can be experimentally induced in mice. This is achieved by ovariectomy before preimplantation estrogen secretion on the morning of D4 (D1 = vaginal plug) thereby initiating a state of blastocyst dormancy (implantation incompetency), which can be maintained by daily injection of P4 from D5. Under the delayed condition, implantation can be induced by a single injection of 17β-estradiol, thereby bringing about the transition of the blastocyst to implantation-competent status (4). Using the delayed model, studies have endeavored to understand how the blastocyst is programmed for implantation, including a number of genes and signal pathways by which this is achieved. By comparing the microarray profile and proteomic profile between the dormant blastocyst and activated blastocyst, various cellular and molecular events, including cell cycle, metabolism, biosynthesis, translation, cytoskeleton organization, among others, were shown to be differentially regulated. Due to technical constraints, only 229 differentially expressed genes (DEGs) and 599 differentially expressed proteins were identified in these 2 studies, respectively (5, 6). Thus, the information regarding the genetic programming of blastocyst for implantation remains limited. In the present study, next-generation RNA sequencing was performed in about 10,000 dormant (implantation incompetent), reactivating, and reactivated (implantation competent) blastocysts to explore the differentially expressed genes while the blastocyst acquires implantation competency.

Results

Transcriptomic Profiles of Blastocysts with Different Implantation Competency.

To reveal the dynamic changes of blastocyst activation before implantation, we employed a delayed implantation mouse model. Dormant blastocysts were recovered from P4-primed mice, and reactivating as well as reactivated blastocysts were recovered 6 h and 12 h, respectively, after the E2 injection in P4-primed mice. RNA sequencing was then performed using more than 3,000 dormant (3,034), reactivating (3,266), and reactivated (3,611) blastocysts, respectively (Fig.1A). Spearman correlation coefficient analysis of the global gene expression indicates that the blastocyst is reprogrammed more drastically at the late 6-h stage during blastocyst activation (Fig. 1B), which is further confirmed by the results of gene differential expression analysis (Fig. 1 C and D). Gene ontology (GO) enrichment analysis of the DEGs reveals not only the known biological pathways involved in blastocyst activation, such as cell cycle, metabolism, transcription, and Wnt signaling (5, 6, 8), but also the novel pathways, including regulation of inflammation and cell migration (Fig. 1E). Furthermore, the specific RNA-sequencing data are confirmed by quantitative real-time PCR analysis (SI Appendix, Fig. S1 A and B).

Fig. 1.

Transcriptomic signatures of the dormant, reactivating, and reactivated blastocyst. (A) Experimental design to collect dormant, reactivating, and reactivated blastocysts. The numbers within parentheses indicate the number of blastocysts collected for RNA sequencing. (B) Spearman correlation heatmap of transcriptome of blastocysts with varying competence to implant. The color key from white to black indicates the correlation coefficient from low to high, respectively. Numerals indicate the correlation coefficient. (C) Volcano plots showing the DEGs through pairwise comparison. Numerals indicate the number of genes up-regulated (red) or down-regulated (green). (D) Heatmap showing the expression level of all differentially expressed genes. The value was normalized by z-score. (E) GO analysis of the DEGs during blastocyst activation.

Reduced Metabolism and Proliferation Arrest Were Broken during Blastocyst Activation.

During delayed implantation, embryos develop into blastocysts and undergo zona hatching, albeit at a slower pace, with reduced cell growth and biosynthesis. The reduced metabolism and proliferation arrest are the most distinctive feature in the dormant blastocyst (9). Consistently, GO enrichment analysis of the DEGs reveals that the most significantly enriched signal pathways are related to metabolic processes and cell proliferation during blastocyst activation (Fig. 1E). Furthermore, gene set enrichment analysis (GSEA) showed that when the blastocyst is reactivated, the genes involved in the biosynthetic processes, including the lipid synthesis, protein synthesis, and nucleic acid synthesis, are consistently up-regulated, whereas the catabolic processes are consistently down-regulated (Fig. 2A and SI Appendix, Fig. S1 A–C and Table S1). In addition, the genes, which are related to the glycolysis, tricarboxylic acid cycle, respiratory electron transport, and ATP synthesis, are up-regulated significantly in the reactivated blastocyst (SI Appendix, Fig. S2D and Table S1), which is consistent with a previous study (5).

Fig. 2.

Reduced metabolism and proliferation arrest are reversed during blastocyst activation. (A and B) GSEA of genes involved in the biosynthetic process (A) and cell proliferation (B). The gene sets are presented in SI Appendix, Tables S1A and S2B, respectively. (C) Confocal imaging of Ki67 (red) staining showing increased cell proliferation in the reactivated blastocyst. The nuclei are stained blue by 4′,6-diamidino-2-phenylindole (DAPI). (Scale bar, 50 µm.) (D) Bubble plots showing expression of genes involved in mTOR, Myc, and polyamine synthesis. The size of each dot represents RPKM, while difference is shown by color: blue, down-regulated significantly (P < 0.05); black, no significance (ns); red, up-regulated significantly (P < 0.05). (E) Western blot analysis showing the mTOR signaling pathway is significantly up-regulated during blastocyst activation.

Along with the reactivation of biosynthesis, cell proliferation tends to be resumed, which is evidenced by the enhanced expression of positive regulators for cell proliferation and the decreased expression of negative regulators (Fig. 2B). Specifically, genes, involved in cell cycle regulation, cell cycle transition, DNA replication, chromosome segregation, and cytokinesis, are significantly up-regulated, such as the cyclin E family members that regulate the G1/S phase transition, while the negative regulators for cell proliferation, such as Cdkn1a (p21) and Cdkn1b (p27), are conversely down-regulated (SI Appendix, Table S2). Immunofluorescence staining of Ki67 further confirms the increased cell proliferation, and shows that the proliferative cells are mainly the inner cell mass (ICM) and polar trophectoderm (TE), but not mural TE (Fig. 2C).

Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis indicates that the signal pathways that regulate biosynthesis and proliferation, such as mechanistic target of rapamycin (mTOR) pathway, Myc pathway, and polyamine synthesis pathway, are differentially regulated during blastocyst activation (Fig. 2D and SI Appendix, Fig. S3A). To be specific, the negative regulators for these signal pathways, such as Akt1s1/Tsc1/Fnip1 for mTOR signaling, Mnt/Mxd4 for Myc signaling, and Oaz1 for polyamine synthesis, are significantly down-regulated during blastocyst activation, suggesting that these signal pathways are activated during blastocyst activation. Taking the mTOR signal pathway as an example, Western blot analysis shows that the reactivated blastocyst has enhanced levels of mTORC1 and mTORC2 activity (Fig. 2E). In addition to genes coding for biosynthesis and proliferation, the genes involved in autophagy, which is negatively regulated by the mTOR signaling pathway, are significantly down-regulated during blastocyst activation (SI Appendix, Fig. S3B). As the previous studies have reported that inhibition of the mTOR signaling, Myc signaling, or polyamine biosynthesis may induce a blastocyst to enter into a dormant state with reduced metabolism and proliferation arrest (10–12), our findings that the negative regulators for these signal pathways that are differentially expressed during blastocyst activation may provide insight into the physiological mechanism of the blastocyst diapause.

Programming of the ICM to Acquire Implantation Competency.

The blastocyst is composed of ICM and TE. To explore how the ICM versus TE would be globally programmed during the blastocyst activation, we first defined the genes highly expressed in ICM or TE via the published data from single cell RNA sequencing on ICM and TE cells (13), respectively. More than 1,000 and 1,400 genes are highly expressed in ICM and TE cells, respectively (SI Appendix, Fig. S4A). By comparing ICM or TE highly expressed genes with the DEGs of dormant versus reactivated blastocyst, 239 ICM highly expressed DEGs and 306 TE highly expressed DEGs can be identified during blastocyst activation (SI Appendix, Fig. S4 B and C).

To further define the ICM-expressed DEGs for implantation, the published RNA sequencing data in D4 and D4.5 ICM cells were analyzed (14), and more than 4,000 DEGs are identified during ICM development (SI Appendix, Fig. S5A). By comparing the DEGs during ICM development with the ICM highly expressed DEGs during blastocyst activation, 79 DEGs are shared (SI Appendix, Fig. S5B). Most of these shared DEGs (75.95%) show consensus expression alteration, implying that the critical events related to these genes in ICM development are conserved in both on-time and delayed implantation. GO enrichment analysis further revealed that these genes are involved in regulation of metabolism and embryo development (Fig. 3A). These results suggest that the further embryonic development program is initiated when the blastocyst progressively achieves implantation competency.

Fig. 3.

Programming of the ICM for acquiring implantation competency. (A) Venn diagrams showing the number of the up-regulated and down-regulated ICM highly expressed DEGs during ICM development and blastocyst activation. Intersections represent the number of genes with consensus expression pattern in ICM development and blastocyst activation. The bar chart shows GO analysis of the consensus-expressed genes. (B) The average expression level of the X- or Y-chromosome genes during blastocyst activation. The logarithmic geometric mean expression levels were normalized to the Gapdh control. Kruskal-Wallis test, ***P < 0.001; ns, no significance. (C) Bubble plot showing expression of the genes that regulate X-chromosome reactivation. The size of each dot represents RPKM, while difference is shown by color: blue, down-regulated significantly (P < 0.05); black, ns. (D and E) Confocal imaging showing the expression of repressive histone modification H3K27me3 (D) and the active histone modification H3K4me3 (E). The red signal is for indicated histone modification, and DAPI (blue) for nuclei. (Scale bars, 50 µm.) (F) Heatmap showing expression of the X-linked genes during blastocyst activation. Red arrows represent the up-regulated genes at P < 0.05, while blue arrows represent the down-regulated genes at P < 0.05. Dor, dormant blastocyst; Rea, reactivated blastocyst.

A prominent change during periimplantation is X-chromosome reactivation in the ICM of female blastocyst (15). Although the presence of male embryos might decrease the difference, the average expression of X-chromosome genes, but not the Y-chromosome genes, was significantly increased when the blastocyst is reactivated (Fig. 3B). Furthermore, we analyzed the known regulators of X-chromosome reactivation. Consistent with the increased expression of X-chromosome genes, we find that most negative regulators, especially Xist, are significantly down-regulated, while the positive regulators were not significantly changed, despite the RPKM (reads per kilobase per million mapped reads) of some genes being slightly increased (Fig. 3C). Immunofluorescence staining consistently showed that the repressive histone modification H3K27me3 was significantly decreased in the ICM of reactivated blastocyst, but not in the TE (Fig. 3D), whereas the activating histone modification H3K4me3 maintains a high level in both states (Fig. 3E). For the periimplantation female blastocysts, the designation of early-, late-, and very late-reactivated X-linked genes refers to the genes that are reactivated in the D4, D4.5, and D5 ICM cells (14), respectively. As mentioned above, the blastocyst dormancy is initiated at D4 when the X-chromosome early-reactivation occurs. However, the gene expression profile of the X-linked genes shows that the consequent late reactivation fails to be completed even though the dormant blastocyst has survived for several days. Furthermore, the RNA-seq results show that most late-reactivated X-linked genes are significantly up-regulated when the blastocyst is reactivated after E2 treatment (Fig. 3F). All these findings suggest that the acquisition of implantation competency is indispensable for the reactivation of the X chromosome at late-blastocyst stage, and the erasure of the negative regulators may be the mechanism of X-chromosome reactivation during blastocyst activation.

Programming of the TE to Acquire Implantation Competency.

We further analyzed the DEG programming of TE cells, which directly contact the uterus during implantation. By analyzing the highly expressed DEGs in the TE during blastocyst activation (SI Appendix, Fig. S4C), we observed that the genes involved in the maintenance of trophoblast stem cell status, such as Cdx2 and Erf, are significantly down-regulated, and the genes involved in the trophoblast differentiation, such as Plac1 and Krt19, are significantly up-regulated (Fig. 4A). Immunofluorescence staining further showed that the decreased Cdx2 mainly occurred in mural TE cells, and the increased Plac1 was correspondingly up-regulated in these cells (Fig. 4B). This specific expression pattern suggests that the mural TE, but not the polar TE, was differentiated during blastocyst activation, which is consistent with the low-proliferative activity in the mural TE (Fig. 2C). Considering that Plac1 is a positive regulator for the migration and invasion of TE cells (16), we speculate that the mural TE cells differentiate to a more invasive state while the blastocyst acquires implantation competency. Consistently, analysis of the regulators for trophoblast migration also suggests that migratory activity is increased (Fig. 4C). All these results suggest that the TE cells, especially mural TE cells, not unexpectedly, differentiate to a more invasive state. Moreover, we find that the expression of the tight-junction genes is down-regulated during blastocyst activation (Fig. 4D). For example, the tight-junction protein Cldn7 is drastically decreased in mural TE cells of the reactivated blastocyst (Fig. 4E). Accordingly, a disturbed localization of the cell-cell tight-junction protein ZO-1 in the mural TE can be observed during blastocyst activation (SI Appendix, Fig. S6A). Ultrastructure of polar and mural TE cells in dormant and reactivated blastocysts further verify the observations that tight junctions have been weakened in the mural TE, but not polar TE (Fig. 4F and SI Appendix, Fig. S6B). Furthermore, GSEA profiles show a reduced expression of the extracellular matrix (ECM) component and increased expression of ECM-degrading enzymes (Fig. 4G and SI Appendix, Table S4), pointing toward a dynamic ECM remodeling during blastocyst activation. For example, Laminin, one of the important ECM components, is decreased accordingly in reactivated blastocysts (Fig. 4H). As apical tight junction and basal ECM are important regulators of epithelial cell migration, the weakened tight junction and ECM in the mural TE indicates an enhanced migration capacity for invading the uterine bed during implantation. However, this differentiation process may not be the same with the epithelial-mesenchymal transition since there are no obvious changes of these regulators at least at the mRNA level (SI Appendix, Table S5).

Fig. 4.

Programming of the TE during acquisition of implantation competency. (A) Bubble plot showing the expression of the TE highly expressed DEGs related to placenta development from GO analysis. In this and subsequent panels, the size of each dot represents RPKM, while difference is shown by color: blue, down-regulated significantly (P < 0.05); black, ns (no significance); red, up-regulated significantly (P < 0.05). (B) Confocal imaging showing increased expression of Cdx2 (Upper) and Plac1 (Bottom) in TE during blastocyst activation. The red signal indicates the specific protein in each analysis, and DAPI (blue) for nuclei (Scale bar, 50 µm.) (C) Bubble plots showing the expression of genes that regulate the trophoblast migration. (D) Bubble plot showing the expression of tight-junction related genes. (E) Confocal imaging showing decreased protein expression level of Claudin7 (red) in mural TE during blastocyst activation, DAPI (blue). (Scale bar, 50 µm.) (F) Ultrastructure of mural TE in dormant and reactivated blastocysts showing a remarkable decrease of tight junction (red arrows) in the mural TE of reactivated blastocysts. (Scale bars, 1 µm). (G) GSEA analysis of ECM component (Upper) and the enzymes degrading ECM component (Bottom) by comparing reactivated with dormant blastocysts. The gene sets are from the GO database with indicative GO identification. (H) Confocal imaging showing decreased Laminin (red) expression during blastocyst reactivation and DAPI (blue) for nuclei. (Scale bar, 50 µm.)

Furthermore, trophoblast cell differentiation is also observed during normal implantation. To characterize trophoblast cell differentiation during the physiological process, day 4 early (1000 h) and late (2200 h) blastocysts were analyzed. The mRNA was extracted from the day 4 early and later blastocysts for real-time quantitative PCR analysis. The well-characterized embryonic-derived signaling molecule Hb-egf is significantly up-regulated in day 4 later blastocysts, which confirms the natural differentiation of the trophoblast cell (SI Appendix, Fig. S7A). This result is also corroborated by the down-regulated stemness marker Cdx2 and Gata3 in the day 4 later embryo (SI Appendix, Fig. S7A). The immunofluorescence results further indicate that the differentiation mainly occurs in the mural TE, which mediated the attachment with uterus in the mouse (SI Appendix, Fig. S7B). The tight-junction related gene Tjp2 is also reduced in the day 4 later blastocyst.

Blastocyst Functions as a Proinflammatory Body to Trigger Inflammatory-Like Response in the Endometrium.

Implantation involves an intricate discourse between the embryos and uterus (17). As it is conceivable that the secretory proteins orchestrate communication between the blastocyst and endometrium, we further analyze the secretory proteins via the secretome database (18). Interestingly, we find that many up-regulated secretory molecules during blastocyst activation are proinflammatory regulators (SI Appendix, Tables S3 and S4). The GSEA profile consistently shows that most of the genes which positively regulate inflammation are up-regulated (Fig. 5A). For example, S100A9, a secretory factor that plays a prominent role in the regulation of inflammatory processes and immune response (19), is expressed in both ICM and TE and increased significantly in the reactivated blastocyst (Fig. 5B). Since implantation is an inflammatory-like response (17), we speculate that the proinflammatory proteins secreted by the blastocyst may act as paracrine factors to trigger the inflammatory-like attachment response. To test this hypothesis, Affi-Gel beads soaked with recombinant S100A9 were transferred into the uterine lumen on D4 of pseudopregnancy, and the blue band reaction is checked 24 h after transfer. We find that S100A9, but not BSA, can induce the implantation-like response efficiently, which is judged by the distinct blue band ascribed to locally increased vessel permeability (Fig. 5 C and D). Furthermore, S100A9 can successfully induce the expression of COX2 in stromal cells surrounding the beads (Fig. 5E), an inflammatory factor essential for implantation (20). And this implantation-like blue band response can also induce uterine stroma decidualization when examined 48 h after the transfer, as the decidualization marker Dtptp is expressed in the uterine stromal cell around the beads and Cox2 is still maintained in the stromal cells underneath the epithelium (SI Appendix, Fig. S8).

Fig. 5.

The blastocyst acts as a proinflammatory body to trigger implantation. (A) GSEA analysis of genes which positively regulate the inflammatory response by comparing reactivated to dormant blastocysts. The gene set is from the GO database with indicative GO identification. (B) Confocal imaging showing increased protein expression of S100A9 (red) in the reactivated blastocyst. DAPI (blue) for nuclei. (Scale bar, 50 µm.) (C) Representative uterus sections showing response to beads loaded with S100A9 compared with beads loaded with BSA. Blue bands indicate the implantation-like sites induced by the beads. (D) Quantitative results of bead transfer. The numbers on Top of the bars indicate the number of mice responding versus the total number of mice used, whereas the numbers within parentheses indicate the number of blue bands over the total number of beads transferred. Student’s t test, ***P < 0.001. (E) COX2 staining (red) in the implantation-like site; arrows indicate the Affi-Gel beads. (Scale bars, 100 µm.)

Since S100A9 is one of the up-regulated inflammatory molecules during the blastocyst activation, we wonder whether this molecule is indispensable for attachment in the implantation. Day 4 embryos exposed to S100A9 antibody were transferred into the pseudopregnant mice. The results show that the antibody cannot block the embryo attachment, while the S100A9 antibody could block attachment-like reaction induced by the S100A9 soaked beads (SI Appendix, Fig. S9), suggesting the potential redundancy of other inflammatory molecules during this process. This evidence suggests that a competent blastocyst might function as a proinflammatory body to release proinflammatory signals for implantation.

Lif, as a member of the interleukin 6 (IL6) family, is critical for normal implantation, as the Lif knockout female mice are infertile due to implantation failure (21). A previous study reported that LIF can be used to replace an estrogen injection to induce implantation in the delayed model (22). We notice that the embryo they used in transfer experiments is day 4 normal blastocysts, but not dormant blastocysts. Therefore, we also conducted a similar experiment to explore whether LIF could induce the blastocyst implantation in the delayed model. The bioactivity of LIF recombinant protein was first confirmed in the following experiment: Day 3 mice were ovariectomized and injected with LIF on the morning of day 4. In this model, LIF protein can induce downstream STAT3 phosphorylation in the absence of estrogen-induced endogenous Lif expression (23) (SI Appendix, Fig. S10A). Then, we analyzed the potential efficiency of LIF in reinitiation of implantation in the delayed model. We found that LIF only failed to reactivate blastocyst implantation in the delayed model, while it can facilitate the estrogen to induce this process when combined with a suboptional dose of estrogen injection as reported in our previous study (24) (SI Appendix, Fig. S10B). Since LIF expression in the uterus is tightly regulated by preimplantation estrogen levels (25, 26), this finding indicates that LIF as an estrogen downstream signaling molecule in the uterus can replace estrogen to prepare the uteri for implantation, but may not participate in the reactivation of blastocysts for implantation in the delayed condition.

Regarding the status of embryonic-derived LIF signaling, our transcriptomic data reveal that Lif expression is down-regulated in the activated blastocysts (SI Appendix, Table S4), which is confirmed by real-time PCR analysis (SI Appendix, Fig. S11A). Lif expression is also down-regulated in blastocyst from day 4 early to late stage during normal implantation (SI Appendix, Fig. S11B). The fact that Lif is expressed in a high level in the dormant blastocyst suggested that embryonic Lif signaling may be involved in dormant blastocyst viability. Indeed, there is evidence that the Lif coreceptor gp130 knockout embryos cannot survive in the delayed model (27).

TNFα Expressed in the Reactivating Blastocyst Up-Regulates the Expression of S100A9.

Although a relatively modest remodeling can be observed in the first 6 h during blastocyst activation (Fig. 1 B–D), differential expression analysis unravels more than 300 genes that are expressed specifically in the reactivating blastocyst (Fig. 6A). Among those specifically expressed genes, the transcript for Tnfa, a well-known proinflammatory gene, was transiently induced in the reactivating blastocyst (Fig. 6 B and D). However, the TNFα protein was constantly present in the reactivated blastocyst (Fig. 6E). Consistently, NFκB, a down-stream signal protein of TNFα, was translocated into the nuclei in the reactivated blastocysts (Fig. 6E). It has been shown that TNFα can induce the expression of S100 family proteins through the NFκB signaling pathway (28). Our RNA sequencing data also revealed that a rapid induction of S100A9 was followed by the temporal expression of TNFα (Fig. 6 B and C). This sequential expression pattern provoked us to surmise that the temporally expressed TNFα may boost the proinflammatory signals of blastocyst at the late 6 h after the activation signal in an autocrine manner. To test this hypothesis, a dormant blastocyst in vitro culture experiment was performed in the presence or absence of recombinant TNFα. In line with the in vivo observations, TNFα induced nuclear localization of NFκB and the expression of S100A9 in dormant blastocysts in culture (Fig. 6F). To further support the hypothesis of inflammation in blastocyst activation in normal condition, Tnfα and S100A9 were also detected in day 4 early and later blastocysts. It was found that Tnfα mRNA was highly expressed in day 4 early embryos, accompanied by the high protein expression in late stage (SI Appendix, Fig. S12 A and B), and both mRNA and protein levels of S100A9 were high in day 4 later embryos (SI Appendix, Fig. S12 A and C). This temporal expression pattern of Tnfα and S100A9 from day 4 early to late stage is similar to the observation in blastocyst from dormancy to activation.

Fig. 6.

Temporally expressed TNFα in the reactivating blastocyst up-regulates the expression of S100A9. (A) Venn diagram showing the number of genes specifically expressed in dormant, reactivating, and reactivated blastocyst. (B and C) The RPKM of Tnfa and S100a9 during blastocyst activation. (D) Quantitative real-time PCR analysis of embryonic Tnfa expression during blastocyst activation. The values are normalized to the Gapdh and indicated as the mean ± SEM, n = 3. Student’s t test, ***P < 0.001. (E) Confocal imaging showing expression of TNFα and NFκB in the dormant and reactivated blastocysts. The image is displayed in a composed 3D view, red signal for indicated protein and DAPI (blue) for nuclei. (Scale bar, 50 µm.) (F) Confocal imaging showing NFκB (Upper) and S100A9 (Bottom) in the dormant blastocyst cultured with or without recombinant TNFα. The image is displayed in a composed 3D view, red signal for indicated protein and DAPI (blue) for nuclei. (Scale bar, 50 µm.)

Taken together, these results suggest that temporally expressed TNFα is responsible for the rapidly induction of proinflammatory signals in blastocysts, which contributes to the inflammatory-like implantation response.

Discussion

Embryonic diapause and reactivation, an evolutionary phenomenon ensuring successful pregnancy, involve reversible reduced metabolism and proliferation arrest (9). The reduced metabolism and proliferation arrest are the 2 major characteristics of dormant blastocyst. Previous studies have revealed that genes related to these biological processes are differentially regulated during blastocyst activation (5, 6). However, only a limited number of genes have been identified in these studies. Here, we describe the transcriptional profiles of these processes, clearly demonstrating that the genes involved in anabolism and energy synthesis, including lipid biosynthesis, protein biosynthesis, nucleic acid biosynthesis, and aerobic glycolysis for energy synthesis, are up-regulated upon blastocyst reactivation. In contrast, and not unexpectedly, the genes involved in catabolism, including lipid catabolism, protein catabolism, and nucleic acid catabolism, were down-regulated. These results provide insight into metabolic activation during blastocyst activation, which should be more accurately defined as biosynthetic activation. In addition, the cell cycle positive related genes, such as cyclins, cyclin-dependent kinases, checkpoint genes, DNA replication genes, chromosome segregation genes, and so on, are consistently up-regulated. But the negative regulators, such as Cdkn1a (p21) and Cdkn1b (p27), are down-regulated. Correspondently, the staining of Ki67 shows that the ICM cells and polar TE cells begin to proliferate in the reactivated blastocyst. All these results more comprehensively delineated the events during which the reduced metabolism and growth arrest of the dormant blastocyst are disrupted during blastocyst activation. Recent studies have reported that the inhibition of mTOR signaling and Myc signaling could induce a blastocyst entry into a state of reduced metabolism and proliferation arrest (10, 11). However, these studies have not shown how these signal pathways are physiologically inhibited in the dormant blastocyst. Our results show that the negative regulators for these signal pathways are highly expressed in dormant blastocysts, which may be the physiological mechanism for sustaining the dormant state. Interestingly, another study shows that inhibition of polyamine synthesis in uterus also causes entry of the blastocyst into dormancy (12). Our data also show that the negative regulator for polyamine synthesis pathway is highly expressed in dormant blastocysts. Conversely, the negative regulators of mTOR signaling, Myc signaling, and the polyamine biosynthesis pathway are consistently down-regulated during blastocyst activation. The autophagy, which is inhibited by the mTOR signal pathway, is reported to regulate the embryonic survival of dormant blastocysts (29). Consistent with the activation of the mTOR signal pathway, the autophagy-related genes are simultaneously down-regulated during blastocyst activation. All these results suggest that the derepression of the mTOR signaling, Myc signaling, and polyamine biosynthesis combines to permit the dormant blastocyst breakout of reduced metabolism and proliferation arrest.

Blastocyst implantation is the rate-limiting step in embryonic development, as unimplanted embryos cannot continue development in vivo. Previous studies usually emphasized the programming of trophoblast during implantation (30, 31), but ignored the relationship of ICM development and the acquisition of implantation competency. Here, by comparing the DEGs in ICM development with the ICM-expressed DEGs during blastocyst activation, we find that the ICM is programmed for further development when the blastocyst becomes implantation competent. In mice, a prominent event in the ICM of female blastocyst during implantation is the process of X-chromosome reactivation (15). X-chromosome reactivation is a successive process, which is observed at D4 and completed at D5, during which blastocyst implantation occurs. However, much less is known about the relationship between X-chromosome reactivation and acquirement of implantation competency. Here, we find that the late and very late reactivated X-linked genes are up-regulated during blastocyst activation, which suggests that X-chromosome reactivation cannot be completed until the blastocyst is implantation competent. Previous studies have demonstrated that X-chromosome reactivation is associated with loss of Xist, a noncoding RNA which has a pivotal role in silencing the X chromosome (14, 32). And studies on the reprogramming of somatic cells into induced pluripotency further showed that the pluripotency factors, such as Nanog, Pou5f1, Sox2, and others, can partially initiate the process of X-chromosome reactivation by down-regulating the expression of Xist (33). Concurrent with down-regulated Xist, loss of the repressive epigenetic marks, such as H3K27me3, is another important mechanism for X-chromosome reactivation (34). A recent study further identified positive and negative regulators for X-chromosome reactivation (35). Consistent with these results, our data show that negative regulators, but not positive regulators, are significantly down-regulated during blastocyst activation, suggesting that the inactivation state of the X chromosome is most likely sustained by negative regulators for X-chromosome reactivation in the dormant blastocyst.

During the process of implantation, the trophoblast cells are closely apposed and attached to luminal epithelium of uterus, and then invade into the uterus (1). Despite multiple efforts in unearthing the molecular mechanism of the TE-uterus dialogue (30, 31), there are still many unknown aspects of trophoblast programming for implantation. Here, we show that the trophoblast nonhomogeneously differentiates in implantation-competent blastocyst, marked by the differentially expressed Ki67 and Cdx2 in polar and mural TE. Specifically, the expression of Cdx2, an important molecular marker of trophoblast stem cells, is significantly decreased in the mural TE cells. And Ki67 staining shows that the proliferative capacity of these cells is obviously compromised. The similar differentiation of mural TE during the attachment was also observed during normal implantation, suggesting that a certain molecular program for implantation competency acquiring is similar for both the delayed implantation and normal implantation models. However, in the cell cycle regulator analysis, we find that the expression of negative regulator expression was comparable between the day 4 early and later embryos (SI Appendix, Fig. S7A), in contrast to the down-regulation of these genes in dormant blastocysts. This discrepancy is mainly due to the lack of quiescence in normal developed day 4 early and late embryos; but in delayed implantation, the embryos must recover from the quiescent cell cycle to exit from dormant status. In brief, most of the observations from delayed implantation such as TE differentiation and up-regulated inflammatory molecule production are conserved in normal implantation.

Concurrent with the differentiation of mural TE, the cell-cell tight junctions are disrupted and the ECM is degraded in the reactivated blastocyst. As a result, the motility and invasiveness of the mural TE is significantly increased in the implantation-competent blastocyst, which is essential for the invasion of the blastocyst into the uterus. Interestingly, the genes involved in disassembling and degrading the ECM are up-regulated in the implantation-competent blastocyst. With the exception of the degradation of the ECM in the blastocyst, these secretory enzymes may also degrade the ECM in the uterus, which could facilitate the TE to invade into the uterus.

It is worth mentioning that activated mTOR signaling and Myc signaling may play distinct roles in each of the ICM, polar TE, and mural TE during blastocyst activation. Evidence from the blastocyst culture has shown that inhibition of mTOR signaling could block effects of exogenous amino acids in regulating the trophoblast cell differentiation (36, 37), suggesting mTOR signaling is responsible for the trophoblast differentiation. Consistent with mTOR signaling, accumulation of active β-catenin could induce the expression of c-Myc in mural TE cells in activated blastocysts (8), essential for the differentiation of mural TE. Studies on mouse embryonic stem cells (ESCs) derived from the ICM of blastocysts have shown that inhibition of mTOR or Myc signaling leads to a reversible proliferation arrest but without affecting the pluripotency. As a result, the ESCs enter into a state of dormancy similar to embryonic diapause (10, 11). All these results suggest that mTOR signaling and Myc signaling play distinct roles in the ICM and trophoblast. Our results concur with these studies, showing that the ICM cells and polar TE cells begin to proliferate, while the mural TE cells begin to differentiate. These changes are accompanied by the activation of mTOR and Myc signaling during blastocyst activation. Consequently, we suggest mTOR and Myc signaling may direct the proliferation of the ICM and polar TE cells, but regulate the differentiation of mural TE cells, which is essential for the acquisition of implantation competence.

Implantation is considered an inflammatory-like attachment response, which is conserved among multiple species (38), and inflammation is believed to be essential for implantation. For example, women who take antiinflammatory drugs in early pregnancy have a higher risk of miscarriage due to defective implantation (39). In addition, local injury to the endometrium, such as endometrial biopsy, promotes the implantation rate in women who have repeated implantation failure (RIF), which is attributed to the increased inflammation level in the injured site (40). The mechanisms by which the inflammatory response is initiated during implantation still remains controversial (41). There is increasing evidence that the blastocyst plays critical roles in the initiation of inflammation of the uterus during implantation. For example, RNA sequencing of the pregnant and pseudopregnant uterus shows that the inflammation level is obviously higher in the pregnant uterus, which contains floating embryos in the uterus (41). Here, we find that the implantation-competent blastocyst secretes a large number of proinflammatory proteins, and the expression patterns may well be coordinated to initiate the inflammatory response during implantation. Thus, we suggest that the activating blastocyst functions as a proinflammatory body to trigger embryo-uterine attachment reaction at implantation. Interestingly, a recent study disclosed that the chromatin accessibility of genes related to the inflammatory response tends to open for a short time in the mouse and human blastocyst at a later embryonic stage when the attachment response occurs (42), which provides further evidence to support our findings.

In summary, we applied next-generation high-throughput RNA sequencing at a high level of resolution to explore the genetic landscape of blastocysts for implantation in a delayed implantation model. Our study makes a great number of observations. First, we revealed the dynamic change of the biosynthesis, metabolism, and proliferation during blastocyst activation. And the differentially expressed inhibitory regulators of mTOR signaling, Myc signaling, and polyamine biogenesis may be the physiological mechanisms for reduced metabolism and cell cycle arrest. Second, the reactivation of the X chromosome in the ICM is not completed until the blastocyst is competent for implantation. Third, the TE is nonhomogeneously differentiated during implantation. Specifically, the mural TE, but not the polar TE, was differentiated into a more invasive state in the implantation-competent blastocyst, accompanied by the disturbed cell-cell tight junctions and degraded ECM. Fourth, the implantation-competent blastocyst functions as a source of proinflammatory signals to initiate the inflammatory-like response in the receptive endometrium.

Methods

Animal Use and Embryo Collection.

Mice were housed in the animal care facility of Xiamen University according to the guidelines for the care and use of laboratory animals. The animal use protocol was approved by the Animal Care Committee of Xiamen University. Experimental procedures to treat mice and analyze implantation sites are provided in SI Appendix, SI Materials and Methods.

RNA Extraction, Library Preparation, Sequencing, and Data Analysis.

Total RNA was isolated using the RNeasy mini kit (Qiagen). Sequencing libraries were generated following the manufacturer’s recommendations, and data analysis detailed information is provided in SI Appendix, Materials and Methods.

Data Availability.

The Gene Expression Omnibus (GEO) accession no. for the RNA-Seq data produced in this study is GSE112906.