15 years of transcriptomic analysis on endometrial receptivity: what have we learnt?

In 15 years of transcriptomic analysis on endometrial receptivity, we have learnt that there is a specific transcriptomic profile during the WOI which is affected or delayed by COS in fertile and infertile patients. COS can slightly or strongly affect ER compared with natural cycles. However, ER under the GnRH antagonist protocol corresponds more closely with natural cycle receptivity than under the GnRH agonist protocol. The endometrial gene expression pattern at the time of embryo implantation in infertile women differs from that of fertile women; indeed the proportion of WOI displacement is higher in infertile compared with fertile patients. These data highlight the necessity to detect the WOI of IVF patients and eventually adjust the embryo transfer day according to the shift in the WOI, particularly in patients with multiple implantation failures. In this case, a clinician could chose to adapt the IVF stimulation protocol or counsel patients to either a fresh or frozen embryo transfer, depending on the impact level of the COS protocols on the endometrium. The large range of results of genes expression profiles associated with the WOI lead us to conclude that we do not yet have a complete understanding of all the interactions between the embryo and the endometrium at the time of implantation . Regardless of all these transcriptomic studies, only two signatures became clinically validated tests. The ERA Test [ 20 ] and the Win-Test [ 21 ], were designed to evaluate ER and to personalize the FET, however, these two genomic diagnostic tools are not comparable from an analytical point of view. ERA is a customized array based on the expression of 238 genes coupled to a computational predictor capable of diagnosing a functionally receptive endometrium, whereas the Win-Test is based on the quantitative expression of 11 predictive genes of ER coupled with an algorithm allowing for the identification of the receptive state. Available data on fresh (stimulated cycles) and frozen cycles (natural/hormonal replacement cycles) showing no differences in pregnacy or implantation rates.A recent studies comparing the fertility success rate in fresh versus frozen embryo transfer, reported that no significant difference in clinical pregnancy rates, implantation rates, or ongoing pregnancy rate between the fresh and the frozen embryo transfer groups [ 51 – 53 ]. These findings suggest that new biomarkers for exploration of endometrial receptiveness are highly conserved that’s why they are proposed in natural and stimulated cycles. Expression of this genes is not affect by COS is just delayed hence the delay of WOI. In their review on ER in 2012, Haouzi et al. emphasized the existence of a common transcriptomic profile during the endometrial shifts from the pre-receptive to the receptive stage between natural and stimulated cycles. Highly conserved endometrial receptivity biomarkers have been identified from this common transcriptomic signature between natural and stimulated cycles. Although, these two studies have observed modifications in gene expression profile associated to the transition of the human endometrium from a pre-receptive to a receptive in fertile women without considering the fact that these women might have a variation in the WOI. Moreover, a recent study demonstrated that pregnancy rate in patients who underwent FET is the same following an adjustment in timing of FET according to the ERA test [ 54 ]. Other studies questioned the consistency of the diagnosis of the WOI using the ERA test after variability observed from month to month following various biopsies in one patient [ 55 , 56 ]. Successful implantation is a complex process requiring a receptive endometrium, a viable embryo and synchronized dialogue between maternal and embryonic tissues. Despite 15 years of transcriptomic analysis on endometrial receptivity more studies are needed to optimize the selection of biomarkers of ER and maternal-fetal dialogue that influence pregnancy outcome. Without forgetting that there are many other factors that can influence the pregnancy rate such as the physician or the embryologist performing the transfer, difficulties in inserting the transfer catheter, endometrial thickness and pattern, and subendometrial contractions. Those factors were not controlled in those studies [ 57 ].

Over the course of the last four decades, IVF has allowed an increasing number of infertile couples the chance to conceive. Considering the extensive research and advances in ART, too many IVF attempts still do not result in a successful pregnancy [ 1 , 2 ]. Embryo implantation is a crucial event in the establishment of a pregnancy. It is now clear that embryo implantation relies upon cross-talk and synchronicity between the implanting embryo and a receptive endometrium [ 3 ]. This embryo-maternal cross-talk involves an elaborate and coordinated network of communication via timely released embryonic, maternal-derived signals, and well-targeted actions [ 4 ]. If the clinical and culture conditions to obtain a “good quality” embryo are well advanced today, ER remains the last barrier in ART. When a high-quality embryo is transferred, impaired uterine receptivity is believed to be one of the major reasons behind failure of the establishment of pregnancy [ 5 , 6 ]. It has been suggested in a few studies that up to two-thirds of implantation failures are due to defects in ER whereas the quality of the embryo itself is responsible for only one-third of failures [ 7 , 8 ]. An endometrium is receptive to an embryo in a spatially and temporally restricted period called the window of implantation (WOI). In natural cycles, this period, occurring during the mid-secretory phase, is limited to approximately 48 h, starting around the seventh day after the LH (luteinizing hormone) surge [ 9 , 10 ]. The dynamic transition from a non-receptive to a receptive endometrium is still poorly understood. Several reports have shown that ER is defined by specific changes in factors involved in adhesion, invasion, survival, growth, differentiation, decidualization and immuno-modulation. The correct spatio-temporal synthesis and balance of these various factors are thought to play an important role in human uterine preparation for implantation [ 4 , 11 ]. Extensive efforts have been made to understand and characterize a receptive endometrium, from the first histological dating methods to the ‘omics’ technologies [ 3 , 7 , 12 ]. Several endometrial dating criteria have been commonly used in clinical practice. One or more panels of biomarkers, predictive of optimal ER, have been analyzed in blood and uterine fluid. Leukaemia inhibitor factor (LIF) is an example of a potential biomarker of the WOI [ 13 ]. However, this approach has been judged unsatisfactory by several studies because LIF measurements in serum do not reflect fertility status and similar conclusions were reported for other biomarkers [ 3 , 14 ]. Cervical mucus has also been used to date ER by analyzing cytokines and growth factors produced by a receptive endometrium and their transport to the cervical mucus [ 15 ]. However, other studies have been unable to detect these growth factors in cervical secretions throughout the menstrual cycle and in these studies no correlation was observed between cytokine levels in cervico-vaginal secretions and serum and between the cytokines gene expression level in the secretory endometrium and the concentrations in serum [ 3 , 16 ]. Another group evaluated ER during IVF cycles using three-dimensional power Doppler ultrasound [ 17 ]. However, Sterzik et al. concluded that ultrasonography is an inadequate method to predict ER in IVF cycles since neither the endometrial thickness nor the echogenic pattern correlate with histological findings [ 18 ]. As demonstrated, endometrial dating criteria have been questioned in various randomized studies. This has encouraged further investigation and application of new technologies to try to objectively diagnose ER. Omics technologies, such as transcriptomics, have been used to identify biomarkers of human endometrium [ 7 , 19 ]. Based on the transcriptomic signature identified in these studies, only two ER diagnostic tools have been commercialized in order to personalize the frozen embryo transfer (FET): the ERA test (Endometrial Receptivity Array) [ 20 ] and the Win-Test (Window Implantation Test) [ 21 ]. In this review, we screened publications of the transcriptomic profiles of fertile and infertile women during the secretory phase of natural and stimulated cycles in order to understand lessons learned from endometrial gene profiling.

In the last few years some groups have been interested in learning whether women with unexplained infertility and/or repeated IVF failure have a different genetic profile during the WOI in natural cycles or stimulated cycles using microarray technologies [ 25 , 31 , 34 , 36 , 42 , 49 , 50 ]. Few studies have compared the gene-expression profile of human ER at the time of embryo implantation of infertile versus fertile patients in natural cycle. All these studies demonstrated a different endometrial gene expression pattern between the two groups. Moreover, a clear distinction is present between the two groups in the up-regulated genes, as well as with those that were down-regulated [ 31 , 34 , 42 ]. As shown in Table 2, three specific transcriptomic signatures that can predict recurrent implantation failure (RIF) were identified. Altmäe et al. reported a specific transcriptomic signature of 260 genes including 145 up- and 115 down-regulated genes in the endometria of infertile versus fertile women during the WOI under natural cycles [ 31 ]. In contrast, Koler’sgroup determined that more than 300 genes exhibited a modified expression profile in RIF, 288 of these genes were down-regulated and 25 genes were up-regulated [ 42 ]. This is consistent with a recent study of RIF that also reported a molecular signature containing 303 genes [ 34 ]. All these studies detected a substantial number of dysregulated genes and pathways in the endometrium of infertile women [ 31 , 34 , 42 ]. Indeed, Altmäe et al. detected a substantial number of dysregulated genes in the endometrium of infertile women, the latter being involved primarily in cellular localization, transport, transporter activity. Also, they showed dysregulation of pathways involved in leukocyte extravasation signalling, lipid metabolism and detoxification in the endometrium of infertile women [ 31 ]. Furthermore, the classification of the down-regulated genes to biological pathways by Koot et al. revealed dysregulation of specific interest pathways for implantation, the cell cycle, the cellular adhesion and the Wnt pathways [ 42 ]. These findings are also in accordance with Koot et al. who concluded that RIF is primarily associated with reduced cellular proliferation [ 34 ]. However until now, none of these specific transcriptomic signatures predicting recurrent implantation have been commercialized. There are few recent studies comparing the gene expression of the endometrium in stimulated cycles of infertile versus fertile women at the time of embryo implantation [ 25 , 36 , 50 ]. As demonstrated in fertile women with stimulated cycles, the majority of these studies showed that COS modulates the gene expression profiles of human endometrium, affect ER and causes a displacement of the WOI of infertile women. Liu et al. [ 25 ] compared the effect of COS on the gene expression patterns of endometrium in infertile patients with normal menstrual cycles in stimulated cycles (hCG + 7) versus natural cycles (LH + 7) [ 25 ]. They demonstrated that COS caused aberrant expression of a number of genes in human endometrium affecting the regulation of several biological pathways. Some of these pathways have been shown to play important roles in the implantation process. They observed changes in the expression of the genes involved in adherent junctions, complement and coagulation cascades, cytokine-cytokine interaction, eicosanoid synthesis, Wnt signaling, MAPK signaling pathway, and nicotinamide metabolism. This is consistent with a recent study of Altmäe et al. [ 50 ] which analyzed gene expression profiles in infertile women undergoing two different endometrial preparation protocols; stimulated and natural cycle with FET versus fertile women in natural cycle. They demonstrated that natural cycles are associated with a better ER than stimulated cycles. Stimulated cycles seemed to have a stronger negative effect on gene expression implicated in crucial pathways of ER. Moreover, using the ERA prediction tool, Ruiz-Alonzo et al. confirmed that the displacement of the WOI is more frequent in RIF patients both in COS or natural cycles than in fertile women [ 36 ]. These findings are also in accordance with Haouzi’s study [ 33 ], using the Win-test, they found that endometrium was non-receptive (29 and 43% in oocyte donation (OD) recipient and RIF OD recipient patients, respectively) or partially receptive (71 and 43% in OD recipient and RIF OD recipient patients, respectively), at Pg + 5/ + 6, in the majority of hormone replacement therapy treated patients. Overall, in natural or COS cycles, transcriptomic modification of the endometrium occurs in RIF patients during the WOI which could be due to a displacement of the WOI and/or pathologic alteration. Furthermore, there is a clearly increased proportion of WOI displacement in RIF patients compared with fertile patients.