Embryo implantation 7

12
The genomics of the human endometrium , ☆☆ Maria Ruiz-Alonso a , David Blesa b , Carlos Simón c, a IVIOMICS, Parc Cientic Universitat de Valencia, Calle Catedrático Agustín Escardino, 9, 46980 Paterna, Valencia, Spain b IVI Foundation, Parc Cientic Universitat de Valencia, Calle Catedrático Agustín Escardino, 9, 46980 Paterna, Valencia Spain c University of Valencia, Department of Paediatrics, Obstetrics and Gynaecology, Faculty of Medicine, Avenida Blasco Ibañez, 15, 46010, Valencia, Spain abstract article info Article history: Received 14 December 2011 Received in revised form 4 April 2012 Accepted 6 May 2012 Available online 24 May 2012 Keywords: Endometrium Transcriptomic Menstrual cycle Receptivity Infertility The endometrium is a complex tissue that lines the inside of the endometrial cavity. The gene expression of the different endometrial cell types is regulated by ovarian steroids and paracrine-secreted molecules from neighbouring cells. Due to this regulation, the endometrium goes through cyclic modications which can be divided simply into the proliferative phase, the secretory phase and the menstrual phase. Successful em- bryo implantation depends on three factors: embryo quality, the endometrium's state of receptivity, and a synchronised dialogue between the maternal tissue and the blastocyst. There is a need to characterise the endometrium's state of receptivity in order to prevent reproductive failure. No single molecular or histolog- ical marker for this status has yet been found. Here, we review the global transcriptomic analyses performed in the last decade on a normal human endometrium. These studies provide us with a clue about what global gene expression can be expected for a non-pathological endometrium. These studies have shown endometri- al phase-specic transcriptomic proles and common temporal gene expression patterns. We summarise the biological processes and genes regulated in the different phases of natural cycles and present other works on different conditions as well as a receptivity diagnostic tool based on a specic gene set prole. This article is part of a Special Issue entitled: Molecular Genetics of Human Reproductive Failure. © 2012 Elsevier B.V. All rights reserved. 1. Introduction The endometrium is a complex tissue that lines the inside of the en- dometrial cavity. It is morphologically divided into functional and basal layers. The functional layer occupies two thirds of the endometrial thickness and it presents different cellular compartments: the luminal epithelium, the glandular epithelium, stroma and the vascular compart- ments. The epithelium is composed of epithelial cells that may be on the surface or coat epithelial glands, and the stroma consists of an extracel- lular matrix (ECM) and broblasts that differentiate during the decidualization process. In the stroma, blood vessels (spiral arteries) are also present, along with immune resident cells such as macrophages and NK cells [13]. Regarding their function, the functional layer is re- sponsible for proliferation, secretion and tissue degeneration, while the regenerative capacity of this organ lies in the basal layer [24]. The gene expression of the different endometrial cell types is reg- ulated by ovarian steroids and paracrine-secreted molecules from neighbouring cells. Due to this endocrine/paracrine regulation, the endometrium goes through cyclic modications which compose the endometrial cycle. An endometrial cycle can be divided simply into the proliferative phase, corresponding to the follicular phase in the ovary, the secretory phase, corresponding to the luteal phase in the ovary, and the menstrual phase. The rst day of the endometrial cycle corresponds to the beginning of the menstrual phase (the rst day of menstrual bleeding) and desquamation of the functional layer takes place during this phase. From the menstrual phase to ovu- lation, the proliferative phase takes place and, at this time, oestrogen levels begin to rise, leading to not only a proliferation of stromal cells and glands, but to the elongation of spiral arteries. Finally, the secre- tory phase is dened as the time between ovulation and menstrua- tion. The beginning of this phase is characterised by a rise in progesterone levels [4]. The secretory phase has been the most stud- ied because, in this phase, the endometrium acquires a receptive phe- notype which allows the implantation of the blastocyst. This period of receptivity is known as the Window of implantation(WOI). The WOI opens on day 19 or 20 of the cycle, and lasts 45 days [5]. If im- plantation does not occur, the progesterone and oestrogen levels de- crease, accompanying a vasoconstriction of the spiral arteries and leading to involution of the endometrium; then the cycle starts again [4]. It is generally accepted that successful implantation depends on three factors: embryo quality, the endometrium's state of receptivity, and a synchronised dialogue between the maternal tissue and the Biochimica et Biophysica Acta 1822 (2012) 19311942 This article is part of a Special Issue entitled: Molecular Genetics of Human Reproductive Failure. ☆☆ Disclosures: The authors declare that they have competing interests. CS is the inventor of the patent application, AX090139WO, covering the Endometrial Recep- tivity Array (ERA). All the authors work full- or part-time for IVIOMICS, which com- mercialises the ERA. Corresponding author. Tel.: + 34 963903305; fax: + 34 963902522. E-mail addresses: [email protected] (M. Ruiz-Alonso), [email protected] (D. Blesa), [email protected] (C. Simón). 0925-4439/$ see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.bbadis.2012.05.004 Contents lists available at SciVerse ScienceDirect Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbadis

Transcript of Embryo implantation 7

Page 1: Embryo    implantation 7

Biochimica et Biophysica Acta 1822 (2012) 1931–1942

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bbad is

The genomics of the human endometrium☆,☆☆

Maria Ruiz-Alonso a, David Blesa b, Carlos Simón c,⁎a IVIOMICS, Parc Cientific Universitat de Valencia, Calle Catedrático Agustín Escardino, 9, 46980 Paterna, Valencia, Spainb IVI Foundation, Parc Cientific Universitat de Valencia, Calle Catedrático Agustín Escardino, 9, 46980 Paterna, Valencia Spainc University of Valencia, Department of Paediatrics, Obstetrics and Gynaecology, Faculty of Medicine, Avenida Blasco Ibañez, 15, 46010, Valencia, Spain

☆ This article is part of a Special Issue entitled: MReproductive Failure.☆☆Disclosures: The authors declare that they have cinventor of the patent application, AX090139WO, covtivity Array (ERA). All the authors work full- or part-timercialises the ERA.⁎ Corresponding author. Tel.: +34 963903305; fax: +

E-mail addresses: [email protected] (M. Ruiz(D. Blesa), [email protected] (C. Simón).

0925-4439/$ – see front matter © 2012 Elsevier B.V. Alldoi:10.1016/j.bbadis.2012.05.004

a b s t r a c t

a r t i c l e i n f o

Article history:Received 14 December 2011Received in revised form 4 April 2012Accepted 6 May 2012Available online 24 May 2012

Keywords:EndometriumTranscriptomicMenstrual cycleReceptivityInfertility

The endometrium is a complex tissue that lines the inside of the endometrial cavity. The gene expression ofthe different endometrial cell types is regulated by ovarian steroids and paracrine-secreted molecules fromneighbouring cells. Due to this regulation, the endometrium goes through cyclic modifications which canbe divided simply into the proliferative phase, the secretory phase and the menstrual phase. Successful em-bryo implantation depends on three factors: embryo quality, the endometrium's state of receptivity, and asynchronised dialogue between the maternal tissue and the blastocyst. There is a need to characterise theendometrium's state of receptivity in order to prevent reproductive failure. No single molecular or histolog-ical marker for this status has yet been found. Here, we review the global transcriptomic analyses performedin the last decade on a normal human endometrium. These studies provide us with a clue about what globalgene expression can be expected for a non-pathological endometrium. These studies have shown endometri-al phase-specific transcriptomic profiles and common temporal gene expression patterns. We summarise thebiological processes and genes regulated in the different phases of natural cycles and present other works ondifferent conditions as well as a receptivity diagnostic tool based on a specific gene set profile. This article ispart of a Special Issue entitled: Molecular Genetics of Human Reproductive Failure.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

The endometrium is a complex tissue that lines the inside of the en-dometrial cavity. It is morphologically divided into functional and basallayers. The functional layer occupies two thirds of the endometrialthickness and it presents different cellular compartments: the luminalepithelium, the glandular epithelium, stroma and the vascular compart-ments. The epithelium is composed of epithelial cells thatmay be on thesurface or coat epithelial glands, and the stroma consists of an extracel-lular matrix (ECM) and fibroblasts that differentiate during thedecidualization process. In the stroma, blood vessels (spiral arteries)are also present, alongwith immune resident cells such asmacrophagesand NK cells [1–3]. Regarding their function, the functional layer is re-sponsible for proliferation, secretion and tissue degeneration, whilethe regenerative capacity of this organ lies in the basal layer [2–4].

The gene expression of the different endometrial cell types is reg-ulated by ovarian steroids and paracrine-secreted molecules from

olecular Genetics of Human

ompeting interests. CS is theering the Endometrial Recep-me for IVIOMICS, which com-

34 963902522.-Alonso), [email protected]

rights reserved.

neighbouring cells. Due to this endocrine/paracrine regulation, theendometrium goes through cyclic modifications which compose theendometrial cycle. An endometrial cycle can be divided simply intothe proliferative phase, corresponding to the follicular phase in theovary, the secretory phase, corresponding to the luteal phase in theovary, and the menstrual phase. The first day of the endometrialcycle corresponds to the beginning of the menstrual phase (the firstday of menstrual bleeding) and desquamation of the functionallayer takes place during this phase. From the menstrual phase to ovu-lation, the proliferative phase takes place and, at this time, oestrogenlevels begin to rise, leading to not only a proliferation of stromal cellsand glands, but to the elongation of spiral arteries. Finally, the secre-tory phase is defined as the time between ovulation and menstrua-tion. The beginning of this phase is characterised by a rise inprogesterone levels [4]. The secretory phase has been the most stud-ied because, in this phase, the endometrium acquires a receptive phe-notype which allows the implantation of the blastocyst. This period ofreceptivity is known as the “Window of implantation” (WOI). TheWOI opens on day 19 or 20 of the cycle, and lasts 4–5 days [5]. If im-plantation does not occur, the progesterone and oestrogen levels de-crease, accompanying a vasoconstriction of the spiral arteries andleading to involution of the endometrium; then the cycle startsagain [4].

It is generally accepted that successful implantation depends onthree factors: embryo quality, the endometrium's state of receptivity,and a synchronised dialogue between the maternal tissue and the

Page 2: Embryo    implantation 7

Table 1Reviews on endometrial transcriptomics.

Authors Date Reference

Giudice, L.C. 2003 [19]Horcajadas et al. 2004 [13]White and Salamonsen 2005 [14]Sherwin et al. 2006 [12]Giudice, L.C. 2006 [16]Simmen and Simmen 2006 [15]Horcajadas et al. 2007 [20]Aghajanova et al. 2008 [17]Aghajanova et al. 2008 [21]Haouzi et al. 2011 [18]

1932 M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

blastocyst (as cited in [6]). This dialogue is the result of a series ofcomplex autocrine, paracrine and/or juxtacrine interactions of differ-ent molecules and their receptors, which are expressed in a coordi-nated manner in the endometrium and/or the embryo [7,8]. Theembryo implantation process can be summarised as follows: the em-bryo, in the morula stage, enters the uterine cavity where it continuesto divide to form the blastocyst. This happens in a defined time periodof between 72 and 96 h after fertilisation. Once inside the uterine cav-ity, a dialogue is established between the blastocyst and the receptiveendometrium, mediated by hormones and growth factors (reviewedin [9]), and the blastocyst is oriented correctly (apposition). The blas-tocyst then comes into contact with the endometrial epithelium andadheres to its surface (adhesion). Following this, penetration of theendometrium by the blastocyst occurs (invasion) (reviewed in[2,10]). This process is mediated by immune cells, cytokines, growthfactors, chemokines and adhesion molecules (reviewed in [9]). Toreach this point, the endometrium needs to display a proper, coordi-nated gene expression regulation during the whole endometrialcycle.

1.1. Transcriptomics of the human endometrium

During the endometrial cycle, there are morphological and func-tional changes that are specific to each menstrual cycle phase and

Table 2Original papers on endometrial transcriptomics.

Authors Date Time of biopsya Comparativ

Kao et al. 2002 CD 8–10 vs. LH+(8–10) LP vs. MSBorthwick et al. 2003 CD 9–11 vs. LH+(6–8) LP vs. MSKuokkanen et al. 2010 CD 11–13 vs. CD 19-23 LP vs. MS w

expressionCarson et al. 2002 LH+(2–4) vs. LH+(7–9) ES vs. MSRiesewijck et al. 2003 LH+2 vs. LH+7 ES vs. MSMirkin et al. 2005 LH+3 vs. LH+8 ES vs. MSHaouzi et al. 2009 LH+2 vs. LH+7 ES vs. MSCritchley et al. 2006 Dating by Noyes MS vs. LSTseng et al. 2010 Dating by Noyes ES vs. MS vPunyadeera et al. 2005 CD 2–5 vs. CD 11–14 M vs. LPPonnampalam et al. 2004 Complete cycle, dating by Noyes EP vs. MP v

MS vs. LS vTalbi et al. 2006 Complete cycle, dating by Noyes EP vs. MP v

MS vs. LSDiaz-Gimeno et al. 2011 (LH+1 vs. +3 vs. +5 vs. +7)

(LH+(1–5) vs. LH+7 vs. CD 8–12)LP vs. ES v

Koler et al. 2009 CD 21 Fertility vs

Altmae et al. 2010 LH+7 Fertility vs

Revel et al. 2011 CD 20–24 Fertility vs

Yanahaira et al. 2005 CD 9–11 Epithelial vproliferativ

Van Vaerenbergh et al. 2010 LH+(5–7) MS vs. pre

a CD: Cycle day; LH+: LH surge+days.b EP: Early-proliferative; MP: mid-proliferative; LP: late-proliferative; ES: early-secretory

which are caused by the response of the different endometrial celltypes to steroid hormones and paracrine molecules. Since 1950, thehistological point of view has been used for endometrial dating [11].Afterwards, the need to understand the genetic mechanism underly-ing the histological changes emerged. Before the genomic era, re-searchers were limited to studying “gene by gene” to determine themolecular changes responsible for the alterations observed. Howeverin the “genomic” era, the general trend is a global screening of all thegenes transcribed and their interactions [12]. So in the last decade,the transcriptional mechanisms underlying endometrial biologyhave been broadly investigated.

For global transcriptomic analyses, the preferred technique hasbeen microarray-based gene expression (for a review see: [12–21]).Table 1 summarises some of the most representative reviews on en-dometrial transcriptomics in natural cycles analysed by microarraytechnology. Nevertheless despite its obvious advantages, this tech-nology is not without its limitations [12,14]. Ponnampalam et al.[22] were the first authors to propose the transcriptomic characteri-sation of the human endometrium throughout the menstrual cycleby using this technology as an alternative to histological dating. How-ever, other groups had previously used this tool to search for a molec-ular signature of receptivity [23–26].

Endometrial transcriptomics studies could be classified accordingto their main objective: firstly, identification of the transcriptomicphysiological profile in each endometrial cycle phase (with emphasisplaced on endometrial receptivity) [22–38]; secondly, a comparisonbetween the endometrial profiles of fertile patients and those withrecurrent implantation failure [39–41]; thirdly, comparisons of pro-files between healthy women and women with endometrial pathol-ogies [42,43]; and finally, studies on the altered transcriptomicprofiles following different ovarian stimulation protocols [44,45].Table 2 lists some of the most representative original papers on en-dometrial transcriptomics in natural cycles analysed by microarraytechnology.

There is some disagreement among transcriptomic studies, whichmay be attributed to differences in experimental designs, in the typeof array used (distinct probes in different arrays), sampling conditions,sample selection criteria, sample size, day of the cycle when the sample

eb Array ID Ref

HG U95A (Affymetrix) [23]HG U95A–E (Affymetrix) [25]

ith RNA and miRNA HG U133 Plus 2.0 (Affymetrix) andmiRCHIP V1 Array

[36]

HG U95A (Affymetrix) [24]HG U95A (Affymetrix) [26]HG U95Av2 (Affymetrix) [33]HG U133 Plus 2.0 (Affymetrix) [27]HG U133A (Affymetrix) [35]

s. LS HG U133 Plus 2.0 (Affymetrix) [37]HG U133A (Affymetrix) [28]

s. LP vs. ES vs.s. M

Home made (Peter MacCallum Cancer Institute) [22]

s. LP vs. ES vs. HG U133 Plus 2.0 (Affymetrix) [34]

s. MS vs. LS HG U133A (Affymetrix) andHomemade “ERA”

[38]

. infertility Array-Ready Oligo Set for the HumanGenome Version 3.0 (Operon)

[40]

. infertility Whole Human Genome Oligo Microarray(Agilent Technologies)

[39]

. infertility with miRNA Taqman Human MiRNA Array Card A(Applied Biosystems)

[29]

s. stromal cells ine phase

BD Atlas Nylon cDNA Expression Array;BD Biosciences (Clontech)

[31]

gnant HG U133 Plus 2.0 (Affymetrix) [30]

; MS: mid-secretory; LS: late-secretory; M: menstrual.

Page 3: Embryo    implantation 7

1933M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

is collected, the statistical analysis applied to the results, separation ornot of tissue compartments, among others [13,15,27,34].

To our knowledge, there are very few human studies that aim todefine the transcriptomic profile throughout all the menstrual cyclephases [22,34]. Most research has focused on finding a receptivetranscriptomic signature and on comparing the expression profileduring the WOI to other phases [23–26,32,33,35–38,46] for the pur-pose of developing a tool that dates endometrial receptivity [38], orto identify abnormalities in the endometrium [35,37,38].

1.2. Discovery of specific endometrial transcriptomic profiles

The most widely used analysis of gene expression data is hierar-chical clustering and principal component analyses (PCA) with visualheatmap representations. With these analyses, the grouping of thevarious types of samples based on their transcriptomic profiles canbe detected. In all the studies conducted [22,26,34,35,37,38], samplesare clearly grouped according to the stage to which they belong. Gen-erally in these studies, phase assignment is completed based on pre-vious histological dating, and is usually performed by at least twoindependent pathologists.

Grouping samples in accordance with the endometrial cycle isclearly observed with the hierarchical clustering method, especiallyin the case of those studies that analyse the entire menstrual cycle,which is the case of Ponnampalam et al. [22] and of Talbi et al. [34].In these two studies, samples are grouped into two main brancheswhich are divided into other sub-branches. In both studies, a majorbranch contains the proliferative and early-secretory phases (andthe menstrual phase in the case of Ponnampalam et al. [22]), whilethe other main branch groups include the samples in the mid-secretory and the late-secretory phases. The PCA analysis reportedby Talbi et al. [34] detects four clusters of samples corresponding tothe predominant proliferative, early-secretory, mid-secretory orlate-secretory phases. Despite the sets of genes they used in thisstudy for PCA and hierarchical clustering being different, the sameclusters were found by both methods. In addition, although six sam-ples were dated histologically as “ambiguous”, both PCA and hierar-chical clustering categorise them in the same phase. These factsfavour the existence of well-defined transcriptomic profiles for eachmenstrual phase. In addition, Ponnampalam et al. [22] found differ-ences between the clustering based on the expression profiles ofsome samples and the group they were classified as by the histologi-cal criteria. The authors suggested for the first time that trans-criptome profiling can detect those abnormalities, which are notidentified by histology and, even more importantly, the expressionof the specific gene clusters were found to be delayed by 2 dayswith specific induction ovulation treatments [44].

The cited studies conclude that it is possible to accurately cata-logue endometria at different stages based on their transcriptomicprofiles, thus facilitating the transition from the anatomical to themolecular medicine of the human endometrium [16].

1.3. Identification of common temporal gene expression patterns

A classic gene expression data analysis is the detection of clusters ofgenes with similar expression patterns throughout the endometrialcycle. Based on the data from samples taken at different cycle phases(early-proliferative, mid-proliferative, late-proliferative, early-secretory,mid-secretory, late-secretory and menstrual), Ponnampalam et al. [22]found for those genes with a fold change greater than or equal to 2,seven main groups of genes with the same expression pattern through-out the menstrual cycle, and with a peak of expression in all seven de-fined cycle phases (menstrual, early-proliferative, mid-proliferative,late-proliferative, early-secretory, mid-secretory and late-secretory).This finding agrees with those obtained by Talbi et al. [34], who foundgroups of genes with a peak of expression in all their studied phases

(proliferative, early-secretory, mid-secretory, late-secretory). Diaz-Gimeno et al. [38] also detected clusters of different genes betweenprereceptive and receptive samples.

Based on these results, the existence of groups of genes that be-have similarly in terms of the expression level in all the cycle phasescan be established. This should have implications for the physiologicalprocesses that characterise each phase. However, it is difficult to finda functional association with the phase in which they are over- orunderexpressed.

1.4. Specific gene expression profiles in endometrial tissue compartments

Theendometrium is a complex tissuewithdifferent cell compartmentsand, in general, the endometrial samples analysed for transcriptomic pro-filing consist in a combination of the different compartments or tissues.Therefore, the reported expressionmeasured for each gene is a weightingof the expression of the given gene in each cell compartment analysed.This weighting task canmask large differences in expression between dif-ferent compartments, and it has been proposed to be a main cause of theresulting differences reported in apparently similar studies [24]. To solvethis issue, laser capture microdissection to separate different tissue com-partments has been tested in mice [47] and the human endometrium[31,46,48,49].

Yanaihara et al. [31] conducted a study in the proliferative cyclephase (days 9–11) by separating stroma and the epithelium. By com-paring the various compartments, they found differentially expressedup-regulated genes (including enzymes, growth-related proteins, andintra/extracellular matrix proteins), most of which were present inthe stroma with functions relating to cell growth and remodelling.Another study, conducted in the menstrual phase, found different,specific profiles for stromal cells or glands, and for different depthsof the endometrium [49]. Nevertheless, this technique is not withoutits disadvantages as it requires prior manipulation using frozen sec-tions that extend the processing time and, in some cases, may intro-duce errors given the need for an RNA amplification of the epithelialcompartment [20].

2. Transcriptomics of the different phases of menstrual cycle

2.1. Menstruation

Onset of menstruation corresponds to day 1 of the cycle, with anapproximate duration of 3–5 days, and includes loss of the functionallayer of the endometrium due to a drop in oestrogen levels [4].

This is one of the stages where major changes occur at the trans-criptomic level. The main processes involved are apoptosis and tissuebreakdown. In different studies, the analysis of temporal gene expres-sion patterns has revealed a peak of expression of those genes relatedto apoptosis, inflammation, signal transduction, transcription andDNA repair [22,28,35]. These include natural cytotoxicity-triggeringreceptor 3 (NCR3) [22,28], Wnt family members (Wnt5a andWnt7a) [28] and the family of matrix metalloproteinases (MMPs)[28,35,49].

NCR3 was reported for the first time by Ponnampalam et al. [22],which increased towards the late-secretory phase and was sustainedin the menstrual phase, then its levels lowered in the proliferativephase [22,28]. This gene is involved in the inflammatory response inrecognition of no-HLA ligands by natural killer cells [50]. NCR3 is ac-tivated by the presence of NK cells and mimics the expression profileof NK cells in the endometrium [22]. Wnt5a and Wnt7a are other im-portant genes involved in this stage which belong to the Wnt path;their high expression level during the menstrual phase lowers inthe proliferative phase, probably due to an increase in the expressionof the specific inhibitors of Wnt action. In mice, they have provenmost important in uterine glandular development [51]; thus, perhaps,they may be implicated in endometrial gland development [28].

Page 4: Embryo    implantation 7

1934 M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

In general, the expression of MMPs begins to increase towards thelate-secretory phase [35] before dropping back to the late-proliferative phase [28]. These metalloproteinases are involved in tis-sue desquamation [52]. This pattern has been observed for MMP1, -3and ‐10, up-regulated in the menstrual phase vs. the proliferativephase. In contrast, MMP26 (endometase) shows increased expressionin the proliferative phase; this is possibly because it is involved in tis-sue remodelling, which occurs during this phase [28].

Other genes detected in the premenstrual phase with maximum ex-pression levels are protease-activated receptor type 1 (F2R (PAR-1)) andlysyl oxidase (LOX) [35]. The thrombin receptor (PAR-1) is responsiblefor mediating many thrombin functions in the cardiovascular system,and is located on the surface of endothelial cells and platelets. Activationof this molecule may lead to platelet aggregation, vascular smooth mus-cle mitogenesis and vasoconstriction. It is likely that PAR-1 plays an im-portant role in the events preceding and is, at least partly, responsible formenstrual bleeding [35]. On the other hand, the LOX gene encodes anamino oxidase, which is crucial in the collagen deposition process andvarious related wound-healing functions [53]; therefore, it seems likelythat it plays an important role in endometrial repair [35].

Despite all these studies, the menstrual phase is the least studiedfrom the global gene expression perspective.

2.2. Proliferative phase

The proliferative phase lasts from the end of the menstrual phase(days 3–5) until ovulation (days 12–14). During this phase, oestrogenlevels increase, leading to stromal and glandular proliferation, and tothe elongation of spiral arteries to create a new functional layer [4].

The proliferative phase involves several processes, including cellu-lar proliferation and differentiation, extracellular matrix remodelling,angiogenesis and vasculogenesis [15,34,36,54]. Moreover, there isalso a high expression of the genes related to adhesion and ion chan-nels [34]. The cellular proliferation in this stage is evidenced by mitot-ic activity, DNA synthesis and increased tissue weight (as cited in[34]).

It is noteworthy that there is a global decrease in gene expressionin the transition from the menstrual to the proliferative phase be-cause up to 64% of the genes that are differentially expressed betweenthese two phases are down-regulated in the proliferative phase [28].The processes that increase in this transition (with implicate up-regulated genes) are tissue remodelling (MMP26, TFF3) [25,28], celldifferentiation (Hoxa10, Hoxa11) and vasculogenesis (Connexin-37,Hoxb7, sFRP) [28].

MMPs are repressed in this stage, except for MMP26, whose ex-pression is activated and maintained until the WOI. This expressionpattern suggests that MMP26 could be implicated in both tissueremodelling and blastocyst invasion [28]. In addition, TFF3 is anothercandidate gene that falls in the endometrial remodelling factor cate-gory [25]. The trefoil family of peptides (TFF) are mucin-associatedpeptides and their high expression in the proliferative vs. secretoryphases [23,25] makes them candidates as participants in endometri-um regeneration after the menstrual phase [25].

Hox family genes are up-regulated in this stage if compared to themenstrual phase. This overexpression appears to be caused by in-creased oestrogen levels that inhibit the Wnt5a and Wnt7a genes toallow the expression of Hoxa10 and Hoxa11. The Hox gene expres-sion completes glandular development and differentiation, whichwere initiated by Wnt5a and Wnt7a [28]. This context explains theup-regulation of numerous Wnt inhibitors (like sFRP andWIF). More-over one of these inhibitors, sFRP, has been associated with vascularremodelling and maturation, while Hoxb7 has been related with vas-cular development induction (cited in [28]).

Towards the end of the proliferative phase and during the transi-tion to the secretory phase, there is a change in the predominant pro-cesses within the endometrium in relation to the early-secretory

phase, which include: cell motility, communication and cell adhesion,Wnt receptor signalling, the I-κB kinase/nuclear factor (NF)-κB cas-cade, cellular signalling, cellular cycle regulation, cellular division,cellular proliferation, collagen metabolism, extracellular matrix regu-lation, signal transduction and regulation of enzymes and ion chan-nels. Some highly represented molecular functions are those relatedto cell cycle regulation, DNA synthesis and steroid binding, and themost prominent cellular components involved include chromosomes,replication fork, DNA polymerase, microtubule skeleton, extracellularmatrix, organelles not bound to membranes and collagen [34].

The genes described as being overexpressed are consistent withmitotic activity and tissue remodelling, which occur in this phase[34]: Olfactomedin 1 [25,34], SOX4, MMP11, tPA (PLAT), ADAM12,retinoic acid, members of the IGF and TGF family, FGF1, HGF, FGFR 3and 1.2 and specific activator protein-kinase activities [34].

Olfactomedin1 (OLFM1) is suppressed when comparing the mid-secretory phase with the early-secretory one [38], suggesting thatthis gene is regulated by progesterone, which would inhibit its ex-pression [34]. In contrast, SOX4 shows not only a down-regulationin the proliferative vs. the menstrual phase [28], but a down-regulation in the early secretory vs. the proliferative phase, implyingthat it is inhibited by oestrogen [34]. The aforementioned cytokinesare the factors involved in immune reactions, but are also involvedin cell communication processes. Therefore during the proliferativephase, these molecules act as paracrine, or even autocrine, moleculesby participating in oestrogen modulation and progesterone control ofendometrial cell differentiation. IL1, TNF alpha, TGF alpha, EGF or Csf1(M-CSF) are involved in establishing a specific regional distribution ofcell populations and their differential proliferative activity [1].

On the other hand, high levels of DNA replication licensing mem-bers (such as MCM2, MCM3, MCM4, MCM5, MCM6, MCM7, and originrecognition complex proteins) have been found. These data suggestthat, in addition to cell cycle regulation, DNA replication licensing islikely to be another key regulatory point in the hormonal regulationof epithelial cell proliferation in the human endometrium [36].

2.3. Secretory phase

The secretory phase takes place between ovulation (days 12–14)and onset of menstruation in a new cycle (days 28–30). It is sub-divided into early-, mid- and late-secretory phases, and correspondsto the luteal phase of the ovary [4]. Traditionally, this phase has beenthemost studied because the endometrium becomes receptive duringthe mid-secretory subphase, allowing for embryo implantation.

During the secretory phase, the endometrium is subjected to in-creasing levels of oestrogen and progesterone. A rise in progesteroneinvolves glycogen and mucus secretion. In the late-secretory phase,oestrogen and progesterone levels drop in the absence of implanta-tion, implying that the menstrual phase undergoes preparation [4].

The endometrium's state of receptivity includes a limited time pe-riod in which the endometrial epithelium acquires a transient statethat is dependent on steroid hormones during blastocyst adhesion,and penetration occurs [9,55]. It is the result of the synchronised, in-tegrated interaction of ovarian hormones, endometrial factors andsignals from the embryo [9]. Moreover to achieve this state of recep-tivity, the endometrium responds to a process mediated by growthfactors, lipid mediators, transcription factors and cytokines with para-crine action (reviewed in [8]). This process is tightly controlled interms of space and time. This period of receptivity, or the WOI, wassuggested in the 1970's [10]. The WOI opens on either day 19 or 20of the menstrual cycle [5] or on days 4 or 5 after progesterone pro-duction or administration [56]. This window expands for 4–5 days,corresponding to days 19–24 of the menstrual cycle [5,56]. This oc-curs for most women with regular cycles, with implantation on days7–10 after the LH surge. However in a small percentage of women,implantation can occur from days 6 to 12 after the LH surge [57]. In

Page 5: Embryo    implantation 7

1935M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

assisted reproductive treatments, it is possible to induce a clinicalWOI by supplying progesterone and oestradiol to synchronise withthe embryo transfer [58]. WOI duration has been proven by severalstudies which followed the embryo implantation process (cited in[23]), onset of structural changes and the emergence of pinopodes[55], or were based on implantation success rates on different daysduring or after the WOI [57].

Given the interest shown in this phase, several studies have inves-tigated its transcriptomic profile. The main objective has been to finda molecular signature that is characteristic of a receptive endometri-um to gain insight into the implantation phenomenon Some ofthese studies were designed by comparing samples taken in the re-ceptive phase vs. samples taken in the prereceptive (early-secretory)phase [24,26,33,37,38,46,59]. Other studies were designed to useproliferative phase [23,25,36] or late-secretory phase samples[35,37] (reviewed in [20]). Different studies also vary in terms ofthe many other parameters used, such as number of patients, takingsamples from the same patient or not, using pools of samples, daysof the cycle when samples are taken, or types of data analysis[13,15,16,60]. Despite these differences, they all agree on the exis-tence of a transcriptomic profile that is typical of the WOI. However,and perhaps due to differences in design, no consensus has beenreached on the genes making up this transcriptomic signature.According to the review by Giudice et al. [16] if compared withsome studies like that of Talbi et al. [34], we see that of the 75 up-regulated genes and the 56 down-regulated ones in Riesewijck et al.[26], only 41 and 11 respectively match. Of the 74 up-regulated andthe 76 down-regulated genes in the study of Carson et al. [24], only11 and 1 respectively agree with the study of Talbi et al. [34]. Of the49 up-regulated and the 58 down-regulated genes in the study ofMirkin et al. [33], only 14 and 3 respectively agree. However, thesestudies are complementary to each other and shed light on the com-plexity of endometrial receptivity and the molecules (known andnew) involved in the implantation process [13].

Despite the variability in transcriptomic studies, potentially im-portant factors in endometrial receptivity have been found: growthfactors, cytokines, lipoproteins, transcription factors and other mole-cules regulated by steroid hormones [9,19,24]. Furthermore, the over-all set of results suggests that transcriptional activation has to occurto achieve a state of receptivity. This is because most genes are up-regulated if compared to their expression in the prereceptive phase[25–27,38,44,59]. For this reason, the term “the transcriptional awak-ening process” was coined [44]. However, this activation process isdetected when samples were taken on LH +7 because, in those stud-ies in which samples were taken on LH+8–10 [23,24,33], more geneswere found to be down-regulated than up-regulated.

2.3.1. Early-secretoryThe early-secretory phase is characterised by the predominance of

the processes related to cell metabolism (fatty acids, lipids, eicosa-noids, and amino alcohols), transport (with a large representationof transporters for the biological molecules involved in these meta-bolic processes), germ cell migration (which could facilitate thetransport of sperm and ensure an aseptic environment) and negativecell proliferation regulation. An increase in metabolism is consistentwith the fact that this phase is biosynthetically highly active, probablyin preparation for embryo implantation, and mitosis inhibition issupported by the down-regulation of numerous growth factors [34].

Some of the prominent genes up-regulated in the early secretoryphase are MSX1 [26,33], PIP5K1B [33], Keratin 8, 17βHSD, Osteopontin(SSP1), secretoglobins, metallothionein, arginase 2, phospholipase C,DKK1, Aquaporin 3 and MUC1 [34]. Of the genes listed, the followingare highlighted; MSX1, which belongs to the family of HOX genes thatcorrelate temporally and spatially with critical morphogenesis events[33]; 17βHSD, which regulates oestradiol availability in the endometri-um [34]; PIP5K1B, which regulates a wide range of cellular processes,

including proliferation, survival and cytoskeleton organisation [33];and MUC1, which maintains cell surface hydration by lubricating andprotecting cells against microorganisms and enzymes [34,61].

Wnt pathway regulation is striking. Some ligands, but not others,are down-regulated. Furthermore some Wnt inhibitors, such assFRP1, are repressed, but DKK1 (another inhibitor of Wnt action) ishighly up-regulated in this phase vs. the proliferative phase [34] tofurther increase in the mid-secretory phase [24].

2.3.2. Mid-secretoryThis phase is characterised by being highly active metabolically,

secretory, non-proliferative, with great importance being placed onthe innate immune response, stress response and response towounding [15,16,34].

The secretory phase is when more gene expression changes aredetected due to the presence of the progesterone peak. These changescoincide with the time that the endometrium becomes receptive. Thegenes whose expression changes during the transition between theearly- and the mid-secretory phases, and between the mid- and thelate-secretory phases, are potential candidate genes to be regulatedby progesterone [23,25,34]. In fact, Ponnampalam et al. [22] detecteda cluster of genes that follows a temporal regulation pattern duringthe endometrial cycle which mimics the increase in circulating pro-gesterone. These genes have been identified in some of the major bio-logical processes taking place during implantation, such as signalling,growth, differentiation and cell adhesion. However, there are no sig-nificant changes associated with the peak of oestrogen.

The genes whose expression alters during the transition from theearly- to the mid-secretory phases are mainly involved in cell cycleregulation, ion binding, transport of signalling proteins, and membersof the family of immunomodulators [33]. The genes that are repressedin this subphase transition are mainly sFRP, PR, PR membrane compo-nent 1, ER-α, MUC1, 17βHSD-2, MMP11 [34], Endothelin 3 (EDN3)[23,24,26] and, especially, Olfactomedin-1 [23,24,26,34], which ispresent in most studies.

On the other hand, there are those genes that are overexpressed inthe mid-secretory vs. the early-secretory. Here there are some geneswhose known function seems to be involved in implantation prepara-tion, and they are involved in processes related to cell adhesion, me-tabolism, response to external stimuli, signalling, immune response,cell communication and negative regulation of proliferation and de-velopment [34,38]. The processes related to the immune system in-clude stress response, defence response, humoral immune response,innate immune response and response to wounding [38]. The trans-criptomic detection of activation of the immune response in mid-stage secretory is consistent with previous studies which have al-ready linked it to implantation [62,63]. The overexpressed molecularfunctions in this stage include three terms: oxidoreductase activity,carbohydrate binding and receptor binding, and cellular componentsand spindle microtubules [38]. It is worth noting some of the genesinvolved in these processes: cysteine-rich secretory stress protein 3(CRISP3) [34,38], glutathione peroxidase 3 (GPX3) [22,26,34,38],metallothionein 1G, 1H, 1E, 1F, 1L, 1X and 2A [23,25,33,34,38], leukae-mia inhibitory factor (LIF) (in contrast to other studies: [24,26,33], butsupported by others: [34,38,64]), granulysin, Glycodelin (PAEP, PP-14)[23,25,26,34,38], IGFBP-3 [26,34], Apolipoprotein D [23,24,26] andDickkopf 1 (DKK1) [23,24,26,36].

A highly represented process during the mid-secretory phase iscell adhesion. Entrance to the blastocyst in the receptive endometriumtriggers the production of cytokines by endometrial epithelial cells.These cytokines modulate receptivity by regulating the expression ofthe adhesion molecules in mammals [65]. Deregulation of cytokinesand their signals leads to implantation failure or abnormalities in pla-centa formation [66]. Although activation of LIF in this stage is disput-ed in some studies (possibly due to the use of different microarrayversions, according to [21]), it is a cytokine that induces proliferation

Page 6: Embryo    implantation 7

1936 M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

and differentiation of many tissue types. Some women with infertilityor recurrent abortions present low levels of LIF or mutations in thecoding region of this gene [67–69]. Recently, LIF has been shown toplay an important role in the adhesion and invasion stages during im-plantation due to an anchor effect in the trophoblast [70]. These find-ings suggest that LIF plays an important role in implantation andmanystudies have focused on this (reviewed in [21]). According to the re-view by Rashid et al. [9], optimal levels of LIF are necessary for im-plantation since the neutralisation of LIF in vitro models inhibitsimplantation of the blastocyst. Homogeneous LIF production in preg-nant women seems to be consistent with its importance in the re-productive process [71]. Another important molecule involved inadhesion is Osteopontin (SPP1), a structural protein of the extracellu-lar matrix [33]. SPP1 has been found to be up-regulated in most of thestudies presented. This glycoprotein is a ligand for αvβ3 integrin, itmediates cellular adhesion and migration during implantation, andis regulated by progesterone [72]. In addition, SPP1 shows its peak ofexpression in endometrial epithelial cells and has also proven its pres-ence on the surface of pinopodes [73].

The cellular metabolism is a well represented process in the secre-tory phase. Apolipoprotein D is a multifunctional transporter involvedin both lipid metabolism and the transport of cholesterols. The activa-tion of its expression detected in most studies is consistent with theactivation of lipid metabolism in the mid secretory phase [17].

The immune response also plays an important role throughout thesecretory phase. In the mid-secretory phase, an up-regulation of thegenes involved in the activation of the innate immune responsetakes place (complement, antimicrobial peptides, Toll-like receptorexpression), as well as increased chemotaxis of monocytes, T cellsand NK cells (CXCL14, granulysin, IL15, carbohydrate sulfotransferase2, and suppression of NK and T-cell activation) [34]. CXCL14 is a che-mokine that is thought to be the major recruiter of monocytes duringthe WOI due to its strong overexpression in this stage [34]. In addi-tion, a role for this chemokine has been recently proposed in that itstimulates the chemotaxis of natural killer cells for clustering aroundepithelial glands [74]. IL15 is thought to play important roles in uNKcell proliferation and differentiation [75], and has been suggested tobe involved in the recruitment of peripheral blood cells CD16-NK[76]. Glycodelin has been broadly studied in relation to implantationand it also participates in immune response regulation [23] by an im-munosuppressive effect promoting the opening of the WOI bylowering maternal immunological responses to the implanting em-bryo (as cited in [21]). The complement gene family is also regulatedin this stage, with some genes being overexpressed and othersunderexpressed in relation to the early-secretory phase [26,33,34].

Some overexpressed genes are involved in protecting theendometrium and/or the embryo in this phase. Metallothioneinsand GPXs (antioxidants) protect the cell from free radicals andheavy metals; this function can prove important in protecting theembryo against these molecules [34]. GPX3 is a selenium-dependentprotein and, interestingly, it has been associated with infertility inselenium-deficient women [77]. It protects the cell from oxidativedamage by catalysing the reduction of hydrogen peroxide, lipid per-oxides and organic hydroxyperoxide by glutathione [26]. On theother hand, the decay accelerating factor (DAF) is a complement reg-ulatory protein for which two possible functions are postulated: pro-tection of the embryo from the maternal complement-mediatedattack and prevention of epithelium destruction by an increased ex-pression of the complement at the time of implantation [46]. In thestudy into expression by Franchi et al. [46], which separately analyseddifferent tissues, it is noted that although DAF is expressed in bothcompartments, the expression is greater in the epithelium. In addi-tion, this protein was found to decrease in the endometrium of pa-tients with recurrent abortions associated with the antiphospholipidsyndrome [78]. Another protective mechanism activated in thisstage is response to stress, which is also the case with GADD45 (a

member of a group of stress-inducible genes). Its up-regulationshows an increase in the protective mechanisms to anticipate implan-tation and blastocyst invasion [17].

Embryo implantation in the endometrium and its apposition havebeen compared to what happens between leukocytes and the endo-thelium [79], where L-selectin is important in this process. The ex-pression of this molecule occurs in the blastocyst [80] while that ofits ligand occurs in endometrial epithelial cells [81]. Although no dif-ferences were found between the early‐ and the mid-secretory sub-phases, L-selectin ligand levels have been seen to vary betweengroups of women with and without successful implantation (withhigher levels in the first group) [81]. IVF patients with progesteroneand oestradiol supplementation present an increased expression ofthe L-selectin ligands in the endothelium of the endometrium [82].

In the transition between the mid-secretory and the late-secretoryphases, the genes that mostly change their expression are related notonly to the immune response of the innate immune system, but alsoto the humoral and cellular immune response [34]. The study byTseng et al. [37] identified 126 up-regulated genes in the mid-secretory and the late-secretory phases. Overexpressed processes in-clude coagulation cascades and complex metabolism, including car-bohydrates, glucose, lipids, cofactors, vitamins, xenobiotics andamino acids, all of which suggest extracellular remodelling in themid-secretory phase [37].

Of the overexpressed genes in this subphase, some peak in themid-secretory expression to then gradually decrease in the late-secretory.This is the case of S100 calcium-binding protein P (S100P), aquaporin3 (AQP3), chemokine (CXC motif) ligand 14 (CXCL14) [22,34,38],Claudin-4 [24,26], Annexin 4 (ANXA4) [22,23,26,33,34,37], Osteopontin(SPP1) [23–26,33,34,46], FoxO1 [33,36], the DAF for complement CD55,monoamine oxidase A (MAOA), GADD45 [23,25,26,33,34], MAPKKK5[23,25,33,34], IL15 [23,24,26,33,46], ID4 [23,26,33], SERPING 1 and CIR[33]. Other genes maintain a high expression in the late-secretoryphase, such as Apo E and Stanniocalcin 1 [34].

2.3.3. Late-secretoryDuring the late secretory phase, oestrogen and progesterone

levels decrease and the main processes that take place include extra-cellular matrix degradation, inflammatory response and apoptosis[15,16].

In the transition from the mid-secretory to the late-secretoryphases, the changes taking place in the extracellular matrix and cyto-skeleton undergo preparation with favoured processes such as vaso-constriction, smooth muscle contraction, haemostasis, and thetransition of the immune response to an inflammatory response[37,83]. The genes that are regulated in this transition mostly relateto the immune response of the innate immune system, the humoraland cellular immune response [34], haemostasis, blood coagulation,steroid biosynthesis and the metabolism of prostaglandins [35],which is consistent with previous contributions. The processes repre-sented in this late-secretory stage, such as matrix degradation, in-flammatory response and cell apoptosis, do not favour implantation.Thus, the transition from the mid- to the late-secretory phase definethe WOI closure and a return to the non-receptive endometrial phe-notype [34].

The prominent genes repressed in the transition to the late-secretory phase are GABARAPL1, GABARAPL3, DKK1 [37]. GABARAPL1and GABARAPL3 are involved in autophagy, thus suggesting thatautophagic degradation may be involved in the mechanism responsi-ble for the formation of new blood vessels [37].

In contrast, the genes exhibiting a peak expression at the end ofthe secretory phase include SOX4, ADAMTS5, GNG4, Integrin α2, pro-lactin receptor, MMPs, ADAMS, PLAU, PLAT, EBAF [34], endothelin re-ceptor B (EDNRB) and MMP10 [35]. It is known that progesteroneinhibits stromal cell-derived proteases (uPA, MMP3) (as cited in[34]). So a drop in progesterone levels is important for extracellular

Page 7: Embryo    implantation 7

1937M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

matrix degradation. Accordingly, the up-regulation in late-secretorymetalloproteinases (MMPs and Adams), serine proteases (PLAU andPLAT) and associated inhibitors is consistent with the progressivedrop in progesterone levels. The overexpression of PLAU allows theactivation of TGF [84] and converts plasminogen into plasmin, a po-tential activator of MMPs [49]. Hence, PLAU plays a key role in prepar-ing tissue for desquamation [34]. TGF family members are alsoinvolved in tissue breakdown, and EBAF stimulates pro-MMP3 andMMP7 production in the absence of progesterone (as this inhibitsthe expression of EBAF and MMPs) (as cited in [34]). Furthermore,the drop in progesterone allows the overexpression of inflammatorymediators [83], as well as local mediators of function endometrial re-ceptor endothelial B (EDNRB) and MMP10 [35].

This stage witnesses intense immune system activity [34], whichis consistent with the histological observations of leukocyte extrava-sation [85]. Regarding the immune activation, to highlight the over-expression of Fc receptors, MHC molecules, NK molecules and T-cell-secreted molecules, it corresponds to the preparation of an innateand adaptive immunity response. Through the up-regulation of the Fcreceptors, monocytes and granulocytes are prepared to respond toantibodies, and by expressing MHC-II molecules, antigen presentingcells are more effective [34]. TNF alpha and IL beta are secreted bythe white blood cells in the stromal compartment at the end of thecycle by stimulating the release of matrix-degrading enzymes, thuscontributing to the scaling of the vascular basal membrane and con-nective tissue [86]. All that is described above reveals that the pre-dominant activities in the late-secretory phase correspond to anendometrium preparing for scaling in the next menstrual phase,which is when the process starts again.

The most significant biological processes regulated, and the geneshighlighted in this manuscript which change among the different men-strual cycle phases, are summarised in Fig. 1 and Table S1, respectively.

2.4. Transcriptomics in pregnancy

There is a unique in vivo study of the comparative transcriptomicprofile during pregnancy in humans [30]. Verenbergh et al. [30] stud-ied a case in which a biopsy was taken in the mid-secretory phasefrom a patient whowas subsequently found to be pregnant. This valu-able sample was compared with non-pregnant patient samples fromthe same phase and a total of 394 differentially expressed geneswere found. The major networks represented by these genes includethose of post-translational modification, cell signalling, cell move-ment, cell development and haematologic system functioning. Thesignificant canonical routes involved in the endometrium of the preg-nant patient were oxidative phosphorylation, ephrin receptor signal-ling, neurotrophin/tyrosine kinase receptor (TRK) signalling, thepurine metabolism and peroxisome proliferator-activated receptor(PPAR) signalling. These networks and canonical pathways formpart of the molecular mechanisms known to be involved in both theembryo–endometrial dialogue and implantation in mice and humans.However, this study does not indicate the distance between samplingand the implantation site location, so the question remains as towhether the observed changes are due to paracrine and/or to endo-crine effects.

The aforementioned study is the only in vivo study conducted inhumans to date. Nevertheless, there are other in vitro studies withcells in co-culture with blastocysts that help gain a better under-standing of the embryo–endometrium relationship [21].

2.5. miRNA expression in the endometrium

MicroRNAs (miRNAs) are RNA molecules of 21–23 nucleotidesthat regulate the expression of other genes through the complemen-tary binding sites in their mRNA targets. MiRNAs can lead to the

degradation or repression of their targets [87]. Recently, the numberof publications whose aim is to study the expression of miRNAs inthe endometrium has increased, particularly during the WOI [36],but also by comparing the natural cycle vs. the ovarian stimulationcycles [88]. Furthermore, studies have also been done to profile themiRNA expression in women with implantation failure vs. womenwith proven fertility, and different profiles were found in both groups[29]. As result of all these studies, the existence of an miRNA expres-sion profile of each cycle phase has been proposed [36,88,89].

Kuokkannen et al. [36] compared the miRNA expression profilebetween the late-proliferative and the mid-secretory phases to findthat the miRNAs up-regulated in the mid-secretory phase weremiRNAs whose target genes are involved in DNA replication licensingand cell cycle regulators. These results are consistent with the sup-pression of cell proliferation during the secretory phase. Amongthese up-regulated miRNA, we find MIR31, MIR29B, MIR29C,MIR30B, MIR30D and MIR210, whose targets are key regulators ofthe cell cycle, cyclins and their cyclin-dependent kinase (CDK) part-ners; for example, replication-licensing factors MCM2 and MCM3,and E2F transcript factor 3 (a transcription factor that induces the ex-pression of those genes regulating the cell cycle and promoting cellcycle progression). The overexpression of these miRNAs in the mid-secretory phase is consistent with the repression detected in someof its target genes, but not in all of them. Kuokkannnen et al. [36]suggested that these targets could be regulated at the translationlevel without, therefore, any effect on the gene expression.

Shah et al. [88] also found a different profile between the differentsub-phases, in this case between the early-secretory and the mid-secretory phases (8 up-regulated miRNAs and 12 down-regulatedmiRNAs on LH +7 vs. LH +2). The bioinformatics analysis revealedthat these miRNAs regulate a large number of genes, some of whichare known to be differentially expressed during the WOI. For exam-ple, SPP1 is a target of MIR424 and, during the receptive phase, theexpression of SPP1 increases [23–26,33,34,46], while that of MIR424decreases [88]. Therefore, as expected, the differential expression ofmiRNAs may play an important role in establishing and maintainingendometrial receptivity [88].

3. Gene expression in ovarian stimulation regimes

Ovarian stimulation is a usual process in assisted reproductivetreatments which is becoming increasingly common in societies ofdeveloped countries. However, some researchers argue that ovarianstimulation may affect endometrial receptivity if compared to a natu-ral cycle [44,45,59,64,90–93], suggesting that it may be due to highconcentrations of steroids [92]. In fact, clinical data have associatedstimulated cycles with a lower implantation rate than in naturalcycles [91], and different studies have addressed the fact thatsupraphysiological concentrations of oestradiol on the day of humanchorionic gonadotrophin (HCG) administration are deleterious to em-bryonic implantation [94–97].

Several studies have investigated the effect of stimulated cycles onendometrial receptivity transcriptomics [98]. Some studies havecompared stimulated (agonist or antagonist) vs. natural cycles[44,59,64,92,93], and others have compared the effect between ago-nists and antagonists [45,99,100]. In those studies that focus on thesecretory phase, it is generally observed that while the expressionpatterns in the prereceptive phase are similar between natural andstimulated cycles, notable differences are found when compared inthe WOI [44,45,59,64,92,93]. The principal differentially expressedgenes are involved in processes of angiogenesis, negative regulationof proliferation and up-regulation of apoptosis, among others [44].Horcajadas' group [44] emphasised the fact that the transition be-tween the prereceptive and the receptive phase is not as obvious instimulated cycles (hCG +5 vs. hCG +7) as in natural cycles (LH +5versus LH +7). This group crossed the secretory phase by sampling

Page 8: Embryo    implantation 7

Fig. 1. Biological processes along the endometrial cycle. Cartoon depicting the endometrium tissue evolution along the menstrual cycle. Underneath are represented the relativelevels of the main regulatory hormones affecting the endometrium. Principal biological processes are indicated at the cycle phase when they are more significantly regulated.

1938 M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

on days LH+1, +3, +5, +7, +9 and hCG+1, +3, +5, +7, +9, andnoted that the hCG +9 profile is more similar to the LH +7 than thehCG +7 profile, thus concluding that the endometrium in stimulatedcycles does not reach the state of receptivity in the same way as innatural cycles [44]. Moreover, Haouzi et al. [59] analysed samples inboth the prereceptive and receptive phases in stimulated (hCG +2and hCG +5) and natural cycles (LH +2 and LH +7) with thesame patient. They observed moderate to severe expression changesat the time of endometrial receptivity in stimulated vs. natural cycles.Nevertheless, Mirkin et al. [99] found no significant differences in thetranscriptome of endometrial receptivity in stimulated cycles whencomparing LH +8 and hCG +9. The fact that Mirkin et al. [99]found no significant changes when comparing LH +8 and hCG +9if compared with other studies that compare LH +7 and hCG +7[44,45,59,64,92,93] can be explained by the delay described inHorcajadas' study [44] by which the hCG +9 profile presents a great-er similarity to that of receptivity in natural cycles.

On the other hand, there have been comparisons made betweentreatment with antagonists and treatment with agonists. Simon etal. [45] compared samples in LH +2 and LH +7 (natural cycle) vs.stimulated cycle samples with antagonists and agonists. On day +2,the results were similar in all three groups, but on day +7, thetranscriptomic profiles of the groups in the stimulated cycles differedfrom the natural cycle. Endometrial dating and the pinopode expres-sion in the agonist-stimulated cycle suggested arrested endometrial

development if compared with the other regimes. These results areconsistent with those recently obtained by another group, Haouzi etal. [100], which found more genes in common in the stimulatedcycle with antagonist vs. a natural cycle than in a stimulated cyclewith agonists vs. a natural cycle in the transition between theprereceptive and the receptive phases (43% and 15%, respectively).

Regarding miRNA, Sha et al. [88] found that 22 miRNAs were sig-nificantly deregulated between the natural and the stimulated cycle.Notably, half of them had at least one putative ERE or PRE in the pro-moter region, suggesting that the altered miRNA expression wasprobably caused by suboptimal progesterone and oestrogen levelsduring stimulation. In addition, the miRNA expression was more sim-ilar between LH +7 and hCG +4 than with hCG +7, and they pro-posed a change in endometrial maturation due to the stimulationtherapy [88]. Table 3 lists some of the most representative originalpapers on endometrial transcriptomics in controlled ovarian stimula-tion cycles analysed by microarray technology.

The presence of oestradiol and progesterone responsive elements(ERE and PRE) was noteworthy in the promoter regions of differen-tially expressed genes, suggesting that oestradiol and progesteronedirectly modulate their expression and may affect the receptivitystate [92]. The alteration in the gene expression at the time of recep-tivity with a stimulated cycle, which attempts to mimic the naturalcycle, suggests a shift in time in differentiation towards a receptiveendometrium caused by these treatments. This could have negative

Page 9: Embryo    implantation 7

Table 3Original papers on endometrial transcriptomics in controlled ovarian stimulation regimes.

Authors Date Time of biopsya Comparativeb Study ID Ref

Horcajadas et al. 2008 LH+(1–9) vs. hCG+(1–9) NC vs. COS Gene expression [44]Haouzi et al. 2009 LH (+2; +7) vs. hCG+(+2; +5) NC vs. COS Gene expression [59]Liu et al. 2008 LH+7 vs. hCG+5 NC vs. COS Gene expression [92]Macklon et al. 2008 LH+5 vs. hCG+2 NC vs. COS Gene expression [93]Sha et al. 2011 LH (+2; +7) vs. hCG (+4; +7) NC vs. COS miRNA expression [88]Horcajadas et al. 2005 LH (+2; +7) vs. hCG+7 NC vs. COH Gene expression [64]Mirkin et al. 2004 LH+8 vs. hCG+9 Ag vs. Atg vs. NC Gene expression [99]Simon et al. 2005 LH (+2; +7) vs. hCG (+2; +7) Ag vs. Atg vs. NC Gene expression [45]Haouzi et al. 2010 LH (+2; +7) vs. hCG (+2; +5) Ag vs. Atg vs. NC Gene expression [100]

a hCG+: hCG administration+days; LH+: LH surge+days.b NC: Natural cycle; COS: controlled ovarian stimulation; COH: controlled ovarian hyperstimulation; Ag: agonist; Atg: antagonist.

1939M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

effects on the implantation process and could, therefore, be one of themain causes with less success rates in ovarian stimulation using thisprotocol [91].

4. Making the most of microarrays [115]

As we have noticed throughout the manuscript, the main objec-tives of microarray studies are to: firstly, determine discrete homoge-neous subsets of an entity based on gene expression profiles (classdiscovery); secondly, identify the genes differentially expressedamong predefined classes of samples with different characteristics(class comparison); thirdly, develop a mathematical function basedon the expression profiles that can accurately predict the biologicgroup, diagnostic category or prognostic stage (class prediction)[113,114].

The MicroArray Quality Control (MAQC) consortium is acommunity-wide effort that sought to experimentally address thekey issues surrounding the reliability of DNA microarray data.MAQC brought together more than a hundred researchers at 51 aca-demic, government and commercial institutions to assess the perfor-mance of seven microarray platforms in profiling the expression oftwo commercially available RNA sample types. The results were com-pared not only at different locations and between different micro-arrays formats, but also in relation to three more traditionalquantitative gene expression assays. MAQC's main conclusions con-firmed that, with careful experimental design and appropriate datatransformation and analysis, microarray data can indeed be reproduc-ible and comparable among different formats and laboratories. Thedata also demonstrated that the fold change results of microarray ex-periments correlated closely with the results of assays like quantita-tive reverse transcription PCR [115,116]. The MAQC effort endedlong-lasting controversy about the reliability of microarrays, buttheir conclusions rely on “careful experimental design and appropri-ate data transformation and analysis”.

Most of the transcriptomic studies reviewed in this manuscriptshow, through unsupervised analyses, that discrete subsets of endo-metrial phase-specific profiles can be identified (class discovery).General agreement has been reached on this, and it shoulders thestandpoint that the endometrial cycle can be subclassified from a mo-lecular point of view. However, class comparisons identify differen-tially expressed genes in different studies, gene lists that do notcompletely agree. As stated above, careful experimental design andproper analysis are needed to compare transcriptome data. Themain reasons for these discrepancies are related to the experimentaldesign. Microarray platforms were a problem in the early stages ofthis technology, but nowadays, commercial arrays have overcomethis, although the MAQC project detected that some laboratories pro-duced lower quality data. Surely, the main factors for discrepanciesare those associated with samples; sampling conditions, selectioncriteria, different criteria on the day of collection of samples, etc. Ulti-mately, these criteria depend on the study objectives and how these

objectives were achieved. This introduces subtle differences, evenfor apparently identical objectives that render different gene lists.Sample sizes should vary depending on the number of truly differen-tially expressed genes or when fold changes are smaller, whichshould also be addressed in the experimental design. Given that thou-sands of genes are tested in a single experiment, the proportion offalse positives among the genes declared significant can be veryhigh and the false discovery rate (FDR) must be controlled duringthe data analysis. Likewise when identifying differentially expressedgene lists, should there be many genes with more or less the samecorrelation with classes, the reported restricted molecular signaturecan strongly depend on the samples selection [114]. All these aspectshave contributed to the different gene lists shown in the papersreviewed in this manuscript. Another fundamental aspect to maintainthe confidence in reported data and in class predictors is validation. Abasic way of performing validation is to divide the initial group of pa-tients into two identical samples: one for the class comparison andone to confirm the results using the same platform and predictionrule if any has been developed to predict a clinical outcome. Cross-validation methods, like leave-one-out, are an alternative to the splitsample methods [113]. Out of the 18 original papers listed inTable 2, Riesewijk et al. [26], Punyadeera et al. [28], Revel et al. [29],Yanaihara et al. [31], have validated the expression level of some oftheir differentially expressed genes, with a different technique, in asecond set of samples from the same population. Haouzi et al. [27]and Diaz-Gimeno et al. [38] validated their class predictors throughthe leave-one-out method. These last two and the remaining authorsverified the level of expression of some reported genes, with a differ-ent technique, in the same set of samples.

Finally, it is obvious that receptivity it is not only regulated bygene expression for there are multiple downstream steps from geneto function. Therefore, it can be assumed that even in carefullydesigned, analysed and validated microarray experiments, part ofthe receptivity characteristics will not be represented in a trans-criptomic profile.

5. Transcriptomics as a diagnostic tool

Infertility is a major problem for many couples undergoing fertilitytreatments to solve it. Despite the success of these treatments, em-bryo implantation failures still occur. The main causes of implantationfailure include low endometrial receptivity, embryonic defects andother multifactorial effects (diseases or disorders of the endometri-um) [101]. It is currently assumed that two thirds of these implanta-tion failures are caused by poor endometrial receptivity or defects inthe embryo–endometrial dialogue [59]. For many years, the histolog-ical criteria established by Noyes [11] were considered the gold stan-dard for endometrial dating to determine the receptive state [2,6,34].Nevertheless, this method has been criticised in recent years becauseit is unable to discriminate between fertility and infertility because,

Page 10: Embryo    implantation 7

1940 M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

although histologically endometrial morphology may submit a mid-secretory phase, this does not imply that it is receptive [22,102,103].

For a long time now, endometrial receptivity foundations havebeen studied from the morphological, biochemical, molecular andcellular points of view to propose possible markers of receptivity.However, none of the proposed markers have become a diagnostictool because they offer a low predictive value [5,17,18,56]. Some au-thors have suggested that “bad receptivity” could be caused by an al-tered gene expression [101,102,104], and that the search for markersof receptivity should focus on analysing and understanding thesegenes as a possible underlying cause of infertility [105]. With this ob-jective in mind, several studies linking gene expression with implan-tation failure have been conducted. This comparison is performed byanalysing endometrial samples in the prereceptive versus the recep-tive phases in healthy women compared to womenwith implantationfailure. It was concluded that at least part of this implantation failureis due to the endometrium being unable to enter the receptive phase[102]. In this way, many studies that search for endometrial markershave focused on a single molecule or on a family of specific molecules,such as integrins [106,107], mucins [108], cytokines [109], mono-amine oxidase A [110], C4BPA, CRABP2 and OLFM1 [111], amongothers. Although the results of different research works relate a defi-cient expression of a gene or of a reduced group of genes with im-plantation failure, the molecules studied have not succeeded asmarkers in the clinical practice, as demonstrated by LIF [112]. Theconclusion reached was that a single molecule does not suffice to de-scribe such a complex phenomenon like receptivity. By acknowledg-ing this fact, transcriptomic profiles can prove to be complexenough to classify the different endometrial cycle phases, includingreceptivity.

Ponannampalam's group [22] used the “GeneRaVe” algorithm toidentify a group of genes (Pyrophosphate (inorganic), outer densefibre of sperm tails 2, Ectonucleoside triphosphate diphosphohydrolase3, Membrane cofactor protein and Zinc finger protein 549), which arecapable of dating the correct cycle phase for 36 of the 37 samples theyanalysed. In their search for marker genes involved in implantation,Allegra et al. [71] conducted a study with samples from women whohad become pregnant after one IVF/ICSI cycle. Heavily regulated genes(FC>23 or b−11) selected from the studies of Carson et al. [24], Kaoet al. [23], Borthwick et al. [25], Riesewijck et al. [26] and Mirkin et al.[33], plus 13 genes involved in the apoptosis pathway, were analysed.Of the total of 43 genes, only six showed a homogeneous expressionamong all the samples (VEGF, LIF, PLA2G2A, ALPL, NNMT and STC1);of these, only the first two showed any involvement with the implanta-tion process. In order to establish a list of possibly relevant genes for en-dometrial receptivity, Horcajadas et al. [20] compared the resultsobtained from three different situations: a natural cycle in fertilewomen, a stimulated cycle, and refractory conditions (IUD). The resultof this comparison was that the three works only shared 25 WOIgenes. The 25 genes were regulated in one sense in the natural cycle,and conversely in COS and IUD (dysregulated). Among them, somegenes are classically involved in endometrial receptivity, such asglycodelin, LIF or α-catenin.

Given the need for reliable, objective molecular endometrium dat-ing, our institution developed a specific tool based on the trans-criptomic signature of different menstrual cycle phases. We namedthis tool Endometrial Receptivity Array (ERA), which has been de-scribed in the work of Diaz-Gimeno et al. [38]. The ERA tool consistsin a custom array containing 238 genes that are differentiallyexpressed in different endometrial cycle phases, as well as a compu-tational predictor capable of cataloguing samples depending onwhether the expression profile of these 238 genes is similar, or not,to the profiles of a set of training samples obtained from normal en-dometria. These 238 genes were selected by means of an exhaustivestudy, which compared the expressions of the samples obtained inthe prereceptive and the receptive phases of women with proven

fertility, and genes were selected with a fold change greater than 3.Of the 238 genes, 134 were a specific transcriptomic signature ofthe receptivity stage. This signature was determined by selectingthe common genes between comparisons made in the early vs. mid-secretory and mid-secretory vs. proliferative phases (74 genes wereup-regulated and 60 were down-regulated). The group of selectedgenes included those genes independently proposed in other publica-tions (18 of the 25 proposed by Horcajadas et al. [20], 61 of the 126proposed by Tseng et al. [37], and laminin beta 3 and microfibril-associated protein 5 reported as being potential markers of receptiv-ity [27]).

The endometrial dating results using the molecular tool proposedby Diaz-Gimeno et al. [38] indicated an ACC of 95% and an AUC of0.94, with a specificity and sensitivity of 0.8857 and 0.99758, respec-tively, which shows the analytic power of this molecular tool for en-dometrial dating. This is the first time that a molecular tool basedon microarray technology can be used clinically in reproductive med-icine to assess the endometrium. Specifically, it has been proven todiscriminate among patients with an implantation failure of endome-trial origin those that are non-receptive, which accounts for approxi-mately 23% of cases. We view the possibility that this test will beexpanded to be used routinely as an endometrium diagnostic tool inthe basic infertility work-up. Moreover, ERA is also a new molecularresearch tool for endometrial research as it contains a finite numberof genes involved in endometrial receptivity, thus avoiding the useof whole genomemicroarrays to cut costs and to simplify data mining[38].

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbadis.2012.05.004.

Acknowledgements

Thanks to Francisco Domínguez and to José A. Martínez-Conejerofor their useful comments, and to Tamara Garrido for the Fig. 1cartoon.

References

[1] H.M. Beier, K. Beier-Hellwig, Molecular and cellular aspects of endometrial re-ceptivity, Hum. Reprod. Update 4 (1998) 448–458.

[2] K. Diedrich, B.C. Fauser, P. Devroey, G. Griesinger, Evian Annual Reproduction(EVAR) Workshop Group, The role of the endometrium and embryo in humanimplantation, Hum. Reprod. Update 13 (2007) 365–377.

[3] E.Y.A.S.E. Bulun, The physiology and pathology of the female reproductive axis,in: H.M. Kronenberg, S. Melmed, K.S. Polonsky, P.R. Larsen (Eds.), Williams Text-book of Endocrinology, 11th ed., Elsevier, 2009, pp. 549–622.

[4] S.M. Hawkins, M.M. Matzuk, The menstrual cycle: basic biology, Ann. N. Y. Acad.Sci. 1135 (2008) 10–18.

[5] B.A. Lessey, Assessment of endometrial receptivity, Fertil. Steril. 96 (2011)522–529.

[6] H. Cakmak, H.S. Taylor, Implantation failure: molecular mechanisms and clinicaltreatment, Hum. Reprod. Update 17 (2011) 242–253.

[7] C. Simon, J.C. Martin, A. Pellicer, Paracrine regulators of implantation, BaillieresBest Pract. Res. Clin. Obstet. Gynaecol. 14 (2000) 815–826.

[8] M. Singh, P. Chaudhry, E. Asselin, Bridging endometrial receptivity and implan-tation: network of hormones, cytokines, and growth factors, J. Endocrinol. 210(2011) 5–14.

[9] N.A. Rashid, S. Lalitkumar, P.G. Lalitkumar, K. Gemzell-Danielsson, Endometrialreceptivity and human embryo implantation, Am. J. Reprod. Immunol. 66(Suppl. 1) (2011) 23–30.

[10] B.A. Lessey, Endometrial receptivity and the window of implantation, BaillieresBest Pract. Res. Clin. Obstet. Gynaecol. 14 (2000) 775–788.

[11] R.W. Noyes, A.T. Hertig, J. Rock, Dating the endometrial biopsy, Fertil. Steril. 1(1950) 3–17.

[12] R. Sherwin, R. Catalano, A. Sharkey, Large-scale gene expression studies of theendometrium: what have we learnt? Reproduction 132 (2006) 1–10.

[13] J.A. Horcajadas, A. Riesewijk, F. Dominguez, A. Cervero, A. Pellicer, C. Simon, De-terminants of endometrial receptivity, Ann. N. Y. Acad. Sci. 1034 (2004)166–175.

[14] C.A. White, L.A. Salamonsen, A guide to issues in microarray analysis: applica-tion to endometrial biology, Reproduction 130 (2005) 1–13.

[15] F.A. Simmen, R.C. Simmen, Orchestrating the menstrual cycle: discerning themusic from the noise, Endocrinology 147 (2006) 1094–1096.

[16] L.C. Giudice, Application of functional genomics to primate endometrium: in-sights into biological processes, Reprod. Biol. Endocrinol. 4 (Suppl. 1) (2006) S4.

Page 11: Embryo    implantation 7

1941M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

[17] L. Aghajanova, A.E. Hamilton, L.C. Giudice, Uterine receptivity to human embry-onic implantation: histology, biomarkers, and transcriptomics, Semin. Cell Dev.Biol. 19 (2008) 204–211.

[18] D. Haouzi, H. Dechaud, S. Assou, J. De Vos, S. Hamamah, Insights into human en-dometrial receptivity from transcriptomic and proteomic data, Reprod. Biomed.Online (2011), http://dx.doi.org/10.1016/j.rbmo.2011.09.009.

[19] L.C. Giudice, Elucidating endometrial function in the post-genomic era, Hum.Reprod. Update 9 (2003) 223–235.

[20] J.A. Horcajadas, A. Pellicer, C. Simon, Wide genomic analysis of human endome-trial receptivity: new times, new opportunities, Hum. Reprod. Update 13 (2007)77–86.

[21] L. Aghajanova, C. Simon, J.A. Horcajadas, Are favourite molecules of endometrialreceptivity still in favour? Expert Rev. Obstet. Gynecol. 3 (2008) 487–501.

[22] A.P. Ponnampalam, G.C. Weston, A.C. Trajstman, B. Susil, P.A. Rogers, Molecularclassification of human endometrial cycle stages by transcriptional profiling,Mol. Hum. Reprod. 10 (2004) 879–893.

[23] L.C. Kao, S. Tulac, S. Lobo, B. Imani, J.P. Yang, A. Germeyer, K. Osteen, R.N. Taylor,B.A. Lessey, L.C. Giudice, Global gene profiling in human endometrium duringthe window of implantation, Endocrinology 143 (2002) 2119–2138.

[24] D.D. Carson, E. Lagow, A. Thathiah, R. Al-Shami, M.C. Farach-Carson, M. Vernon,L. Yuan, M.A. Fritz, B. Lessey, Changes in gene expression during the early tomid-luteal (receptive phase) transition in human endometrium detected byhigh-density microarray screening, Mol. Hum. Reprod. 8 (2002) 871–879.

[25] J.M. Borthwick, D.S. Charnock-Jones, B.D. Tom,M.L. Hull, R. Teirney, S.C. Phillips, S.K.Smith, Determination of the transcript profile of human endometrium, Mol. Hum.Reprod. 9 (2003) 19–33.

[26] A. Riesewijk, J. Martin, R. van Os, J.A. Horcajadas, J. Polman, A. Pellicer, S.Mosselman, C. Simon, Gene expression profiling of human endometrial recep-tivity on days LH+2 versus LH+7 by microarray technology, Mol. Hum. Reprod.9 (2003) 253–264.

[27] D. Haouzi, K. Mahmoud, M. Fourar, K. Bendhaou, H. Dechaud, J. De Vos, T. Reme,D. Dewailly, S. Hamamah, Identification of new biomarkers of human endome-trial receptivity in the natural cycle, Hum. Reprod. 24 (2009) 198–205.

[28] C. Punyadeera, H. Dassen, J. Klomp, G. Dunselman, R. Kamps, F. Dijcks, A.Ederveen, A. de Goeij, P. Groothuis, Oestrogen-modulated gene expression inthe human endometrium, Cell. Mol. Life Sci. 62 (2005) 239–250.

[29] A. Revel, H. Achache, J. Stevens, Y. Smith, R. Reich, MicroRNAs are associatedwith human embryo implantation defects, Hum. Reprod. 26 (2011) 2830–2840.

[30] I. Van Vaerenbergh, R. McIntire, L. Van Lommel, P. Devroey, L. Giudice, C.Bourgain, Gene expression during successful implantation in a natural cycle,Fertil. Steril. 93 (2010) 268.e15–268.e18.

[31] A. Yanaihara, Y. Otsuka, S. Iwasaki, T. Aida, T. Tachikawa, T. Irie, T. Okai, Differ-ences in gene expression in the proliferative human endometrium, Fertil. Steril.83 (Suppl. 1) (2005) 1206–1215.

[32] J.A. Horcajadas, A. Riesewijk, J. Martin, A. Cervero, S. Mosselman, A. Pellicer,C. Simon, Global gene expression profiling of human endometrial receptivi-ty, J. Reprod. Immunol. 63 (2004) 41–49.

[33] S. Mirkin, M. Arslan, D. Churikov, A. Corica, J.I. Diaz, S. Williams, S. Bocca, S.Oehninger, In search of candidate genes critically expressed in the human endome-trium during the window of implantation, Hum. Reprod. 20 (2005) 2104–2117.

[34] S. Talbi, A.E. Hamilton, K.C. Vo, S. Tulac, M.T. Overgaard, C. Dosiou, N. Le Shay,C.N. Nezhat, R. Kempson, B.A. Lessey, N.R. Nayak, L.C. Giudice, Molecularphenotyping of human endometrium distinguishes menstrual cycle phasesand underlying biological processes in normo-ovulatory women, Endocrinology147 (2006) 1097–1121.

[35] H.O. Critchley, K.A. Robertson, T. Forster, T.A. Henderson, A.R. Williams, P. Ghazal,Gene expression profiling of mid to late secretory phase endometrial biopsiesfrom women with menstrual complaint, Am. J. Obstet. Gynecol. 195 (2006)406.e1–406.16.

[36] S. Kuokkanen, B. Chen, L. Ojalvo, L. Benard, N. Santoro, J.W. Pollard, Genomicprofiling of microRNAs and messenger RNAs reveals hormonal regulation inmicroRNA expression in human endometrium, Biol. Reprod. 82 (2010) 791–801.

[37] L.H. Tseng, I. Chen, M.Y. Chen, H. Yan, C.N. Wang, C.L. Lee, Genome-based ex-pression profiling as a single standardized microarray platform for the diagnosisof endometrial disorder: an array of 126-gene model, Fertil. Steril. 94 (2010)114–119.

[38] P. Diaz-Gimeno, J.A. Horcajadas, J.A. Martinez-Conejero, F.J. Esteban, P. Alama,A. Pellicer, C. Simon, A genomic diagnostic tool for human endometrial recep-tivity based on the transcriptomic signature, Fertil. Steril. 95 (2011) 50–60(60.e1–15).

[39] S. Altmae, J.A. Martinez-Conejero, A. Salumets, C. Simon, J.A. Horcajadas, A.Stavreus-Evers, Endometrial gene expression analysis at the time of embryo implan-tation in womenwith unexplained infertility, Mol. Hum. Reprod. 16 (2010) 178–187.

[40] M. Koler, H. Achache, A. Tsafrir, Y. Smith, A. Revel, R. Reich, Disrupted gene pat-tern in patients with repeated in vitro fertilization (IVF) failure, Hum. Reprod. 24(2009) 2541–2548.

[41] A. Tapia, L.M. Gangi, F. Zegers-Hochschild, J. Balmaceda, R. Pommer, L. Trejo, I.M.Pacheco, A.M. Salvatierra, S. Henriquez, M. Quezada, M. Vargas, M. Rios, D.J.Munroe, H.B. Croxatto, L. Velasquez, Differences in the endometrial transcriptprofile during the receptive period between women who were refractory to im-plantation and those who achieved pregnancy, Hum. Reprod. 23 (2008)340–351.

[42] S. Matsuzaki, DNA microarray analysis in endometriosis for development ofmore effective targeted therapies, Front. Biosci. (Elite Ed.) 3 (2011) 1139–1153.

[43] J.K. Habermann, N.K. Bundgen, T. Gemoll, S. Hautaniemi, C. Lundgren, D.Wangsa, J. Doering, H.P. Bruch, B. Nordstroem, U.J. Roblick, H. Jornvall,

G. Auer, T. Ried, Genomic instability influences the transcriptome andproteome in endometrial cancer subtypes, Mol. Cancer 10 (2011) 132.

[44] J.A. Horcajadas, P. Minguez, J. Dopazo, F.J. Esteban, F. Dominguez, L.C. Giudice, A.Pellicer, C. Simon, Controlled ovarian stimulation induces a functional genomicdelay of the endometrium with potential clinical implications, J. Clin.Endocrinol. Metab. 93 (2008) 4500–4510.

[45] C. Simon, J. Oberye, J. Bellver, C. Vidal, E. Bosch, J.A. Horcajadas, C. Murphy, S.Adams, A. Riesewijk, B. Mannaerts, A. Pellicer, Similar endometrial developmentin oocyte donors treated with either high- or standard-dose GnRH antagonistcompared to treatment with a GnRH agonist or in natural cycles, Hum. Reprod.20 (2005) 3318–3327.

[46] A. Franchi, J. Zaret, X. Zhang, S. Bocca, S. Oehninger, Expression of immunomod-ulatory genes, their protein products and specific ligands/receptors during thewindow of implantation in the human endometrium, Mol. Hum. Reprod. 14(2008) 413–421.

[47] A.L. Niklaus, J.W. Pollard, Mining the mouse transcriptome of receptive endome-trium reveals distinct molecular signatures for the luminal and glandular epi-thelium, Endocrinology 147 (2006) 3375–3390.

[48] H.P. Gaide Chevronnay, P. Lemoine, P.J. Courtoy, E. Marbaix, P. Henriet, Ovariansteroids, mitogen-activated protein kinases, and/or aspartic proteinases cooper-ate to control endometrial remodeling by regulating gene expression in the stro-ma and glands, Endocrinology 151 (2010) 4515–4526.

[49] H.P. Gaide Chevronnay, C. Galant, P. Lemoine, P.J. Courtoy, E. Marbaix, P. Henriet,Spatiotemporal coupling of focal extracellular matrix degradation and recon-struction in the menstrual human endometrium, Endocrinology 150 (2009)5094–5105.

[50] D. Pende, S. Parolini, A. Pessino, S. Sivori, R. Augugliaro, L. Morelli, E. Marcenaro,L. Accame, A. Malaspina, R. Biassoni, C. Bottino, L. Moretta, A. Moretta, Identifica-tion and molecular characterization of NKp30, a novel triggering receptor in-volved in natural cytotoxicity mediated by human natural killer cells, J. Exp.Med. 190 (1999) 1505–1516.

[51] C. Miller, D.A. Sassoon, Wnt-7a maintains appropriate uterine patterning duringthe development of the mouse female reproductive tract, Development 125(1998) 3201–3211.

[52] L.A. Salamonsen, D.E. Woolley, Matrix metalloproteinases in normal menstrua-tion, Hum. Reprod. 11 (Suppl. 2) (1996) 124–133.

[53] H.M. Kagan, W. Li, Lysyl oxidase: properties, specificity, and biological roles in-side and outside of the cell, J. Cell. Biochem. 88 (2003) 660–672.

[54] T. Maruyama, Y. Yoshimura, Molecular and cellular mechanisms for differentiationand regeneration of the uterine endometrium, Endocr. J. 55 (2008) 795–810.

[55] A. Psychoyos, Hormonal control of uterine receptivity for nidation, J. Reprod.Fertil. Suppl. 25 (1976) 17–28.

[56] J. Martin, F. Dominguez, S. Avila, J.L. Castrillo, J. Remohi, A. Pellicer, C. Simon,Human endometrial receptivity: gene regulation, J. Reprod. Immunol. 55(2002) 131–139.

[57] A.J. Wilcox, D.D. Baird, C.R. Weinberg, Time of implantation of the conceptus andloss of pregnancy, N. Engl. J. Med. 340 (1999) 1796–1799.

[58] J. Remohi, B. Gartner, E. Gallardo, S. Yalil, C. Simon, A. Pellicer, Pregnancy andbirth rates after oocyte donation, Fertil. Steril. 67 (1997) 717–723.

[59] D. Haouzi, S. Assou, K. Mahmoud, S. Tondeur, T. Reme, B. Hedon, J. De Vos, S.Hamamah, Gene expression profile of human endometrial receptivity: compar-ison between natural and stimulated cycles for the same patients, Hum. Reprod.24 (2009) 1436–1445.

[60] D. Haouzi, H. Dechaud, S. Assou, C. Monzo, J. de Vos, S. Hamamah, Transcriptomeanalysis reveals dialogues between human trophectoderm and endometrialcells during the implantation period, Hum. Reprod. 26 (2011) 1440–1449.

[61] M. Brayman, A. Thathiah, D.D. Carson, MUC1: a multifunctional cell surface com-ponent of reproductive tissue epithelia, Reprod. Biol. Endocrinol. 2 (2004) 4.

[62] L.A. Salamonsen, N.J. Hannan, E. Dimitriadis, Cytokines and chemokines duringhuman embryo implantation: roles in implantation and early placentation,Semin. Reprod. Med. 25 (2007) 437–444.

[63] N.J. Hannan, L.A. Salamonsen, Role of chemokines in the endometrium and inembryo implantation, Curr. Opin. Obstet. Gynecol. 19 (2007) 266–272.

[64] J.A. Horcajadas, A. Riesewijk, J. Polman, R. van Os, A. Pellicer, S. Mosselman, C.Simon, Effect of controlled ovarian hyperstimulation in IVF on endometrialgene expression profiles, Mol. Hum. Reprod. 11 (2005) 195–205.

[65] C. Simon, M.J. Gimeno, A. Mercader, J.E. O'Connor, J. Remohi, M.L. Polan, A. Pellicer,Embryonic regulation of integrins beta 3, alpha 4, and alpha 1 in human endome-trial epithelial cells in vitro, J. Clin. Endocrinol. Metab. 82 (1997) 2607–2616.

[66] O. Guzeloglu-Kayisli, U.A. Kayisli, H.S. Taylor, The role of growth factors and cy-tokines during implantation: endocrine and paracrine interactions, Semin.Reprod. Med. 27 (2009) 62–79.

[67] E. Hambartsoumian, Endometrial leukemia inhibitory factor (LIF) as a possiblecause of unexplained infertility and multiple failures of implantation, Am. J.Reprod. Immunol. 39 (1998) 137–143.

[68] H.D. Tsai, C.C. Chang, Y.Y. Hsieh, H.Y. Lo, Leukemia inhibitory factor expression indifferent endometrial locations between fertile and infertile women throughoutdifferent menstrual phases, J. Assist. Reprod. Genet. 17 (2000) 415–418.

[69] Z. Kondera-Anasz, J. Sikora, A. Mielczarek-Palacz, Leukemia inhibitory factor: animportant regulator of endometrial function, Am. J. Reprod. Immunol. 52 (2004)97–105.

[70] E. Dimitriadis, G. Nie, N.J. Hannan, P. Paiva, L.A. Salamonsen, Local regulation ofimplantation at the human fetal–maternal interface, Int. J. Dev. Biol. 54 (2010)313–322.

[71] A. Allegra, A. Marino, F. Coffaro, A. Lama, G. Rizza, P. Scaglione, F. Sammartano, A.Santoro, A. Volpes, Is there a uniform basal endometrial gene expression profile

Page 12: Embryo    implantation 7

1942 M. Ruiz-Alonso et al. / Biochimica et Biophysica Acta 1822 (2012) 1931–1942

during the implantation window in women who became pregnant in a subse-quent ICSI cycle? Hum. Reprod. 24 (2009) 2549–2557.

[72] K.B. Apparao, M.J. Murray, M.A. Fritz, W.R. Meyer, A.F. Chambers, P.R. Truong,B.A. Lessey, Osteopontin and its receptor alphavbeta(3) integrin are coexpressedin the human endometrium during the menstrual cycle but regulated differen-tially, J. Clin. Endocrinol. Metab. 86 (2001) 4991–5000.

[73] B.A. Lessey, Two pathways of progesterone action in the human endometrium:implications for implantation and contraception, Steroids 68 (2003) 809–815.

[74] N.M. Mokhtar, C.W. Cheng, E. Cook, H. Bielby, S.K. Smith, D.S. Charnock-Jones,Progestin regulates chemokine (C–X–C motif) ligand 14 transcript level inhuman endometrium, Mol. Hum. Reprod. 16 (2010) 170–177.

[75] H. Okada, T. Nakajima, M. Sanezumi, A. Ikuta, K. Yasuda, H. Kanzaki, Progester-one enhances interleukin-15 production in human endometrial stromal cellsin vitro, J. Clin. Endocrinol. Metab. 85 (2000) 4765–4770.

[76] K. Kitaya, T. Yamaguchi, H. Honjo, Central role of interleukin-15 in postovulatoryrecruitment of peripheral blood CD16(−) natural killer cells into human endo-metrium, J. Clin. Endocrinol. Metab. 90 (2005) 2932–2940.

[77] P.D. Kingsley, J.C. Whitin, H.J. Cohen, J. Palis, Developmental expression of extra-cellular glutathione peroxidase suggests antioxidant roles in deciduum, visceralyolk sac, and skin, Mol. Reprod. Dev. 49 (1998) 343–355.

[78] J. Francis, R. Rai, N.J. Sebire, S. El-Gaddal, M.S. Fernandes, P. Jindal, A.Lokugamage, L. Regan, J.J. Brosens, Impaired expression of endometrial differen-tiation markers and complement regulatory proteins in patients with recurrentpregnancy loss associated with antiphospholipid syndrome, Mol. Hum. Reprod.12 (2006) 435–442.

[79] F. Dominguez, M. Yanez-Mo, F. Sanchez-Madrid, C. Simon, Embryonic implanta-tion and leukocyte transendothelial migration: different processes with similarplayers? FASEB J. 19 (2005) 1056–1060.

[80] O.D. Genbacev, A. Prakobphol, R.A. Foulk, A.R. Krtolica, D. Ilic, M.S. Singer, Z.Q.Yang, L.L. Kiessling, S.D. Rosen, S.J. Fisher, Trophoblast L-selectin-mediated ad-hesion at the maternal–fetal interface, Science 299 (2003) 405–408.

[81] B. Wang, J.Z. Sheng, R.H. He, Y.L. Qian, F. Jin, H.F. Huang, High expression ofL-selectin ligand in secretory endometrium is associated with better endometri-al receptivity and facilitates embryo implantation in human being, Am. J.Reprod. Immunol. 60 (2008) 127–134.

[82] N.F. Vlahos, C.W. Lipari, B. Bankowski, T.H. Lai, J.A. King, I. Shih, K. Fragakis, Y.Zhao, Effect of luteal-phase support on endometrial L-selectin ligand expressionafter recombinant follicle-stimulating hormone and ganirelix acetate for in vitrofertilization, J. Clin. Endocrinol. Metab. 91 (2006) 4043–4049.

[83] H.O. Critchley, R.W. Kelly, R.M. Brenner, D.T. Baird, The endocrinology of menstru-ation—a role for the immune system, Clin. Endocrinol. (Oxf) 55 (2001) 701–710.

[84] B. Casslen, T. Sandberg, B. Gustavsson, R. Willen, M. Nilbert, Transforminggrowth factor beta1 in the human endometrium. Cyclic variation, increased ex-pression by estradiol and progesterone, and regulation of plasminogen activa-tors and plasminogen activator inhibitor-1, Biol. Reprod. 58 (1998) 1343–1350.

[85] D.C. Daly, N. Tohan, T.J. Doney, I.A. Maslar, D.H. Riddick, The significance oflymphocytic–leukocytic infiltrates in interpreting late luteal phase endometrialbiopsies, Fertil. Steril. 37 (1982) 786–791.

[86] L.A. Salamonsen, D.E. Woolley, Menstruation: induction by matrixmetalloproteinases and inflammatory cells, J. Reprod. Immunol. 44 (1999) 1–27.

[87] D.P. Bartel, MicroRNAs: genomics, biogenesis, mechanism, and function, Cell116 (2004) 281–297.

[88] A.G. Sha, J.L. Liu, X.M. Jiang, J.Z. Ren, C.H. Ma, W. Lei, R.W. Su, Z.M. Yang, Geno-me-wide identification ofmicro-ribonucleic acids associatedwith human endome-trial receptivity in natural and stimulated cycles by deep sequencing, Fertil. Steril.96 (2011) 150–155.e5.

[89] B.A. Lessey, Fine tuning of endometrial function by estrogen and progesteronethrough microRNAs, Biol. Reprod. 82 (2010) 653–655.

[90] R.J. Paulson, M.V. Sauer, R.A. Lobo, Embryo implantation after human in vitro fertil-ization: importance of endometrial receptivity, Fertil. Steril. 53 (1990) 870–874.

[91] J.A. Martinez-Conejero, C. Simon, A. Pellicer, J.A. Horcajadas, Is ovarian stimula-tion detrimental to the endometrium? Reprod. Biomed. Online 15 (2007) 45–50.

[92] Y. Liu, K.F. Lee, E.H. Ng, W.S. Yeung, P.C. Ho, Gene expression profiling of humanperi-implantation endometria between natural and stimulated cycles, Fertil.Steril. 90 (2008) 2152–2164.

[93] N.S. Macklon, M.H. van der Gaast, A. Hamilton, B.C. Fauser, L.C. Giudice, The im-pact of ovarian stimulation with recombinant FSH in combination with GnRHantagonist on the endometrial transcriptome in the window of implantation,Reprod. Sci. 15 (2008) 357–365.

[94] C. Simon, F. Dominguez, D. Valbuena, A. Pellicer, The role of estrogen in uterine recep-tivity and blastocyst implantation, Trends Endocrinol. Metab. 14 (2003) 197–199.

[95] C. Simon, J.J. Garcia Velasco, D. Valbuena, J.A. Peinado, C. Moreno, J. Remohi, A.Pellicer, Increasing uterine receptivity by decreasing estradiol levels during thepreimplantation period in high responders with the use of a follicle-stimulatinghormone step-down regimen, Fertil. Steril. 70 (1998) 234–239.

[96] D. Valbuena, J. Martin, J.L. de Pablo, J. Remohi, A. Pellicer, C. Simon, Increasinglevels of estradiol are deleterious to embryonic implantation because they di-rectly affect the embryo, Fertil. Steril. 76 (2001) 962–968.

[97] C. Simon, F. Cano, D. Valbuena, J. Remohi, A. Pellicer, Clinical evidence for adetrimental effect on uterine receptivity of high serum oestradiol concentra-tions in high and normal responder patients, Hum. Reprod. 10 (1995)2432–2437.

[98] J.A. Horcajadas, J.A. Martinez-Conejero, C. Simon, Endometrial receptivity in nat-ural and controlled ovarian-stimulated cycles, Biennial Rev. Infertility (2011)43–55, http://dx.doi.org/10.1007/978-1-4419-8456-2_5 (Part 1).

[99] S. Mirkin, G. Nikas, J.G. Hsiu, J. Diaz, S. Oehninger, Gene expression profiles andstructural/functional features of the peri-implantation endometrium in naturaland gonadotropin-stimulated cycles, J. Clin. Endocrinol. Metab. 89 (2004)5742–5752.

[100] D. Haouzi, S. Assou, C. Dechanet, T. Anahory, H. Dechaud, J. De Vos, S.Hamamah, Controlled ovarian hyperstimulation for in vitro fertilization al-ters endometrial receptivity in humans: protocol effects, Biol. Reprod. 82(2010) 679–686.

[101] E.J. Margalioth, A. Ben-Chetrit, M. Gal, T. Eldar-Geva, Investigation and treat-ment of repeated implantation failure following IVF-ET, Hum. Reprod. 21(2006) 3036–3043.

[102] A.M. Sharkey, S.K. Smith, The endometrium as a cause of implantation failure,Best Pract. Res. Clin. Obstet. Gynaecol. 17 (2003) 289–307.

[103] C. Coutifaris, E.R. Myers, D.S. Guzick, M.P. Diamond, S.A. Carson, R.S. Legro, P.G.McGovern, W.D. Schlaff, B.R. Carr, M.P. Steinkampf, S. Silva, D.L. Vogel, P.C.Leppert, NICHD National Cooperative Reproductive Medicine Network, Histo-logical dating of timed endometrial biopsy tissue is not related to fertility status,Fertil. Steril. 82 (2004) 1264–1272.

[104] N.A. Bersinger, D.M. Wunder, M.H. Birkhauser, M.D. Mueller, Gene expression incultured endometrium fromwomen with different outcomes following IVF, Mol.Hum. Reprod. 14 (2008) 475–484.

[105] L.C. Kao, A. Germeyer, S. Tulac, S. Lobo, J.P. Yang, R.N. Taylor, K. Osteen, B.A.Lessey, L.C. Giudice, Expression profiling of endometrium from women with en-dometriosis reveals candidate genes for disease-based implantation failure andinfertility, Endocrinology 144 (2003) 2870–2881.

[106] R.O. Hynes, Integrins: versatility, modulation, and signaling in cell adhesion, Cell69 (1992) 11–25.

[107] B.A. Lessey, L. Damjanovich, C. Coutifaris, A. Castelbaum, S.M. Albelda, C.A.Buck, Integrin adhesion molecules in the human endometrium. Correlationwith the normal and abnormal menstrual cycle, J. Clin. Invest. 90 (1992)188–195.

[108] M. Meseguer, J.D. Aplin, P. Caballero-Campo, J.E. O'Connor, J.C. Martin, J. Remohi,A. Pellicer, C. Simon, Human endometrial mucin MUC1 is up-regulated by pro-gesterone and down-regulated in vitro by the human blastocyst, Biol. Reprod.64 (2001) 590–601.

[109] S. Tabibzadeh, X.Z. Sun, Cytokine expression in human endometrium through-out the menstrual cycle, Hum. Reprod. 7 (1992) 1214–1221.

[110] S. Henriquez, A. Tapia, M. Quezada, M. Vargas, H. Cardenas, M. Rios, A.M.Salvatierra, H. Croxatto, P. Orihuela, F. Zegers-Hochschild, D.J. Munroe, L.Velasquez, Deficient expression of monoamine oxidase A in the endometriumis associated with implantation failure in women participating as recipients inoocyte donation, Mol. Hum. Reprod. 12 (2006) 749–754.

[111] J. Lee, J. Oh, E. Choi, I. Park, C. Han, H. Kim do, B.C. Choi, J.W. Kim, C. Cho, Differ-entially expressed genes implicated in unexplained recurrent spontaneousabortion, Int. J. Biochem. Cell Biol. 39 (2007) 2265–2277.

[112] P.R. Brinsden, V. Alam, B. de Moustier, P. Engrand, Recombinant human leuke-mia inhibitory factor does not improve implantation and pregnancy outcomesafter assisted reproductive techniques in women with recurrent unexplainedimplantation failure, Fertil. Steril. 91 (2009) 1445–1447.

[113] R. Simon, Diagnostic and prognostic prediction using gene expression profiles inhigh-dimensional microarray data, Br. J. Cancer 89 (2003) 1599–1604.

[114] S. Michiels, S. Koscielny, C. Hill, Interpretation of microarray data in cancer, Br. J.Cancer 96 (2007) 1155–1158.

[115] Editorial, Making the most of microarrays, Nat. Biotechnol. 24 (2006) 1039.[116] M.A.Q.C. Consortium, The MicroArray Quality Control (MAQC) project shows the

inter- and intraplatform reproducibility of gene expression measurements, Nat.Biotechnol. 24 (2006) 1151–1161.