Abstract

Endometrial receptivity is crucial for implantation and establishment of a normal pregnancy. The shift from proliferative to receptive endometrium is still far from being understood. In this paper, we comprehensively present the transcriptome of the human endometrium by comparing endometrial biopsies from proliferative phase with consecutive biopsies 7–9 days after ovulation. The results show a clear difference in expression between the two time points using both total and small RNA sequencing. A total of 3,297 messenger RNAs (mRNAs), 516 long noncoding RNAs (lncRNAs), and 102 small noncoding RNAs were identified as statistically differentially expressed between the two time points. We show a thorough description of the change in mRNA between the two time points and display lncRNAs, small nucleolar RNAs, and small nuclear RNAs not previously reported in the healthy human endometrium. In conclusion, this paper reports in detail the shift in RNA expression from the proliferative to receptive endometrium.

Introduction

Proper implantation of the human embryo in the maternal endometrium is crucial for the establishment of a normal pregnancy. This implantation requires finely tuned endometrial receptivity, which includes morphological and biochemical changes of the endometrium [1]. Window of implantation (WOI) occurs independent of pregnancy between 7 and 10 days after ovulation [1]. Although distorted receptivity is linked to poor pregnancy outcome [2], the process of receptivity is still far from being understood [3].

Earlier studies in humans have investigated the change in RNA expression from proliferative to receptive endometrium, almost all have used chip-based techniques [36] and only one study has used RNA sequencing [7]. No previous studies have used consecutive sampling in humans or studied the expression of both protein-coding messenger RNA (mRNA) and long noncoding RNA (lncRNA) and small noncoding RNA (sncRNA) in the endometrium using RNA sequencing.

A large number of mRNAs statistically significantly change their expression between early luteal phase and mid-luteal (receptive) phase [810]. Among these genes cell surface proteins, extracellular matrix components, and growth factors were the major groups [8]. However, a review by Horcajadas and coworkers found little consistency between microarray-based studies [11]. The different phases of the menstrual cycle can be identified with their genomic profile, which also have a strong correlation to the histopathological classification [12,13].

lncRNAs, defined as ncRNA > 200 nucleotides, have been shown to be functional as epigenetic regulators and regulate gene expression [14]. Recently, lncRNA expression in human endometrium was reported in endometrial carcinoma [15] and endometriosis [16], but expression patterns of lncRNA in healthy human endometrium have not been reported.

Small ncRNAs (<200 nucleotides) consist of several subgroups such as micro RNA (miRNA), small nuclear RNA (snRNA), and small nucleolar RNA (snoRNA). Of these, miRNA are the most extensively studied and have been shown to be involved in gene-regulating complexes, mainly in gene silencing through hybridization [17,18]. Altered miRNA in the human endometrium expression has been described in endometrial carcinoma, and associated with infertility and endometriosis [3,19,20]. Endometrial decidualization in vitro [3,21] and endometrial receptivity and implantation in vivo [4,22] are associated with changes in miRNA. Disturbed miRNA expression has been proposed as a causative factor in implantation failure [3]. Small nuclear RNAs are intrinsic to the spliceosome, have a function in Cajal bodies and telomerase activity [23], and have been linked to human diseases [24]. Small nucleolar RNA modulate ribosomal RNA (rRNA) and transfer RNA through methylations [25]. Small nuclear RNA and snoRNA have never been described in healthy endometrium, and only one recent study included snoRNA in their evaluation of small RNAs in endometrial carcinoma [26].

Using RNA sequencing, the primary aim of this study was to describe the difference in total RNA (mRNA, lncRNA, and sncRNA) expression pattern in human endometrium between the proliferative phase and 7–9 days after ovulation. We also compared the sncRNA expression in vivo with human endometrial stromal cell cultures, in vitro.

Materials and methods

Endometrial tissue

The endometrial tissue was collected after informed consent, in seven healthy donors after paracervical blockade anesthesia using a suction curette (Pipet Curet, CooperSurgical, USA). The first sample was taken 7–9 days after a positive urinary ovulation test, LH+ 7–9, (Clearblue, Proctor & Gamble, USA) from each donor, corresponding to WOI, the second biopsy from the same donor was taken 6–8 days after the start of the subsequent menstruation, corresponding to proliferative phase (cd 6–8). The donors were between 24 and 31 years, had regular menstrual cycles, had not used hormonal or other treatment for at least 3 months and did not have a history of chronic disease, Table 1. All women underwent normal general physical and gynecological examinations including transvaginal sonography. A gynecological pathologist confirmed the cycle days through histopathological examination. The endometrial tissue was immediately stored in RNA later solution (Ambion) and stored in –20°C until further analysis.

Table 1.

Characteristics of donors. All donors had regular menstrual cycles.

 Age BMI Pregnancies Childbirth 
Donor 1 28 33.2 
Donor 2 27 28.7 
Donor 3 29 22 
Donor 4 30 23.6 
Donor 5 29 19.8 
Donor 6 24 23.2 
Donor 7 24 26.6 
 Age BMI Pregnancies Childbirth 
Donor 1 28 33.2 
Donor 2 27 28.7 
Donor 3 29 22 
Donor 4 30 23.6 
Donor 5 29 19.8 
Donor 6 24 23.2 
Donor 7 24 26.6 

Cell cultures

After informed consent endometrial biopsies were taken from six healthy donors at cycle day 5–9. The donors had the same clinical features as described for the endometrial biopsies except age, which were between 19–39 years. Small noncoding RNA was analyzed in cell cultures before and after induced decidualization. For further information about the preparation of the stromal cells [27,28], see

. In vitro decidualization was performed as described previously, [29]. Briefly, cells were seeded in 6-well Costar plates in Dulbeccos modified eagle medium (DMEM)/F12 without phenol red, supplemented with 2% chalcoal stripped fetal bovine serum (Sigma-Aldrich, Sweden) and 0.2% penicillin-streptomicin. At 80–90% of confluence, cells were treated with 1 μM medroxyprogesterone-17-acetate (Sigma-Aldrich, Sweden) and 0.5 mM N6, 2`-O-dibutyryladenosine cyclic adenosine monophosphate (cAMP) (Sigma-Aldrich, Sweden) for 6 days to induce decidualization.

Sample preparation and sequencing

RNA extraction

Total RNA was extracted using miRNeasy Mini Kit (Qiagen) according to manufacturer's instruction. RNA quality was assessed using Agilent Bioanalyzer and RNA quantity was assessed using Qubit fluorometer. The amount of RNA, measured by Qubit, for the 14 samples ranged from 3 to 33 μg averaging on 19.1 μg. The RNA integrity number (RIN) for the samples ranged from 4.3 to 9.0, averaging on 6.4. See

for further details.

Sequencing libraries

Two types of sequencing libraries were made from each of the 14 RNA extractions; small RNA libraries and total RNA libraries to access the expression of both mRNA and ncRNA. For small RNA libraries, 1 μg of RNA from each of the 14 samples were subjected to library preparation using TruSeq Small RNA Sample Preparation Kit (Illumina) according to the manufacturer's protocol. This preparation enriches for shorter RNA transcripts; from 22 to 150 nucleotides including the adapters. The libraries were then sequenced on the Illumina HiSeq2500 generating 100-bp paired end reads. For the total RNA libraries, 1 μg of RNA from each of the 14 samples were first subjected to ribosomal removal using RiboZero kit before library preparation using the TruSeq Stranded Total RNA kit. The libraries were sequenced on the Illumina HiSeq2500 generating 125 bp paired end reads. All libraries were prepared and sequenced by the National Genomics Infrastructure, Science for Life Laboratory, Stockholm, Sweden. The sequencing of the small RNA libraries generated from 22.0 to 55.9 million reads per sample, averaging on 35.7 million reads. The sequencing of the total RNA libraries generated from 69.7 to 81.6 million reads per sample, averaging on 76.8 million reads. The RNA sequencing was validated with polymerase chain reaction (qPCR), for more information see

.

Data analysis

Preprocessing, mapping, and quality control

Before being mapped to the genome, using STAR [30], the raw sequencing reads were quality trimmed and potential adapter sequences were removed. For the quality trimming process, the tool TrimGalore [31] was used. Further information on mapping parameters and quality control for RNA data, preprocessing of sequencing reads, and ribosomal depletion can be found in the

; ; and . Every sample was prepared, sequenced, and analyzed individually, i.e., the samples were not pooled in any way.

Novel micro RNA expression in endometrium

miRDeep2 [32] was used to identify novel miRNAs in endometrium based on the small RNA from tissue. miRDeep2 was run individually for each of the 14 small RNA samples. All novel miRNA predicted by miRDeep2 that had very low or no expression in our aligned data as assessed through the Integrative Genomics Viewer (IGV) were discarded.

Expression quantification and differential expression

To quantify gene expression, the program htseq-count [33] was used to count how many reads match each gene in an annotation file in gene transfer format (gtf) (Ensemble version GRCh38.82). When quantifying gene expression for the small RNA libraries, the gtf file was altered so that it only included annotated miRNA, snRNA, snoRNA, and novel miRNA. Novel miRNA were identified with miRDeep2 [32] and their expression verified using IGV. Similarly, when quantifying gene expression of the total RNA libraries, all annotated miRNA, snRNA, and snoRNA were excluded from the gtf file. For further details on this editing of gtf files and the effects it has on quantification, see

.

The count tables generated by htseq-count were imported into R [34], and differential expression (DE) analysis between the two time points was performed using DESeq2 [35]. Prior to DE analysis, genes with low expression were discarded. Genes were defined as lowly expressed if it had zero or one read. The design model of the DE analysis took into account that the sample is paired which both increases statistical power and reduces confounding effects. P-values were corrected for multiple testing using the false discovery rate (FDR) method. The threshold of significance was set to FDR = 0.01. The DE analysis for the total RNA libraries was divided into two; one restricted to protein-coding mRNA, and one restricted to lncRNAs. Thus, in total, there were four DE analyses; (i) sncRNA from tissue, comparing proliferative phase and receptive endometrium (LH+ 7–9 days), (ii) lncRNA from tissue, comparing proliferative phase and receptive endometrium (LH +7–9 days), (iii) protein-coding mRNA from tissue, comparing proliferative phase and receptive endometrium (LH +7–9 days), and (iv) sncRNA from stromal cells comparing decidualized stromal cells with nondecidualized stromal cells. All DE results were cross validated using the leave-one-out cross validation (LOOCV) method. For the LOOCVs, the DE analyses were carried out additional times, each time leaving out one sample pair from the analysis. Each gene that did not meet the adjusted P-value threshold of 0.01 in these cross validation analyses was said to fail LOOCV and was not included in the final list of significantly differentially expressed (SDE) genes.

Micro RNA targeting in endometrium

Known miRNA were grouped into miRNA families according to TargetScan database [36], and a separate DE analysis was done on the small RNA dataset only using miRNA families, see

. Using SDE miRNA families, we evaluated three miRNA targeting methods: probability of conserved targeting (PCT) from TargetScan [37], cumulative weighted contexts ++ score (CWC) from TargetScan, and miRTarBase (MTB) [38]. For further description see .

Ethical approval

The study was approved by the board of ethics in biomedical research at the Karolinska Institute (dnr: 2008/865-32 and 2010/1094-31/2).

Results

Principal component analysis RNA expression in vivo

To analyze the variance in mRNA, lncRNA, and sncRNA expression between the proliferative phase and 7–9 days after ovulation in vivo, we performed principal component analyses of their expression values. Figure 1 shows the first two principal components of the log-transformed expression values for the four RNA analyses: mRNA, lncRNA, sncRNA in vivo, and sncRNA in vitro. Expression for all RNA types clustered according to the two time points of the menstrual cycle and not according to individuals (Figure 1A–C). Indicating a clear DE between the time points. For all three RNA types, one subject deviated from other samples during the proliferative phase (orange square, Figure 1 A–C). Small noncoding RNA in vitro expression differed before and after decidualization (Figure 1D), though the separation is not as clear as for the endometrial biopsies. When comparing the expression for sncRNA in vitro and in vivo, the samples cluster according to the type of source material, tissue or cells, and not according to the time points (Figure 2).

Figure 1.

First two principal components of expression values for mRNA in tissue (A), lncRNA in tissue (B), sncRNA in tissue (C), and sncRNA in cells (D). In the tissue (A–C), there is a clear separation of expression according to the two different time points of the menstrual cycle. Of note in the tissue is that one sample in the proliferative phase that deviates in expression from the other samples in the proliferative phase (orange square). For the cells (D), there is a separation of expression between before and after decidualization but the difference is not as clear as with the tissue. For the tissue, squares denote proliferative phase and triangles denote WOI. For the cells, squares denote before decidualization and triangles denote after decidualization.

Figure 1.

First two principal components of expression values for mRNA in tissue (A), lncRNA in tissue (B), sncRNA in tissue (C), and sncRNA in cells (D). In the tissue (A–C), there is a clear separation of expression according to the two different time points of the menstrual cycle. Of note in the tissue is that one sample in the proliferative phase that deviates in expression from the other samples in the proliferative phase (orange square). For the cells (D), there is a separation of expression between before and after decidualization but the difference is not as clear as with the tissue. For the tissue, squares denote proliferative phase and triangles denote WOI. For the cells, squares denote before decidualization and triangles denote after decidualization.

Figure 2.

First two principal components of expression values for sncRNA for tissue and cells. Here, the expression clusters according to source material. The separation of tissue into the two time points is still visible.

Figure 2.

First two principal components of expression values for sncRNA for tissue and cells. Here, the expression clusters according to source material. The separation of tissue into the two time points is still visible.

Messenger RNA expression patterns

We used RNA sequencing to display a detailed description of mRNA expression in healthy human endometrium and the difference in gene expression between the proliferative phase and 7–9 days after ovulation in consecutive samples from the same individuals. A total of 4,867 out of 17,446 mRNAs were SDE; 1,570 out of these 4,867 failed LOOCV (see Methods), which resulted in 3,297 SDE RNAs; of which 1,648 were higher expressed in the proliferative phase and 1,639 were higher expressed 7–9 days after ovulation. Overview of the results from the DE analysis is shown in

and in Figure 3. A complete list of SDE genes is supplied in .

Figure 3.

Volcano plots from DE analysis for mRNA in tissue (A), lncRNA in tissue (B), sncRNA in tissue (C), and sncRNA in cells (D).

Figure 3.

Volcano plots from DE analysis for mRNA in tissue (A), lncRNA in tissue (B), sncRNA in tissue (C), and sncRNA in cells (D).

A total of 745 mRNAs were differentially expressed with a log 2 fold change (log2FC) of two or more and 35 mRNA with a log2FC of five or more. The 10 SDE mRNAs with the highest fold change are presented in Table 2. Two genes encoding for solute carrier proteins, SLC30A2 (log2FC = 5.86, P = 8.55E–123) and SLC25A48 (log2FC = 7.64, P = 6.22E–43), have the lowest P-value and the highest fold change, respectively. In fact, 75 solute carrier genes are found to be SDE, 49 of which are higher expressed 7–9 days after ovulation. We noted that LEFTY1 (log2FC = 4.87, P = 2.63E–13) and IGFBP2 (log2FC = 1.66, P = 8.49E–08) were significantly higher expressed 7–9 days after ovulation, but the commonly used markers of decidualization, LEFTY2 and IGFBP1 as well as prolactin (PRL) remained unchanged in our material. In summary, these analyses provide a thorough description of the change in mRNA expression in human endometrium from proliferative phase to 7–9 days after ovulation.

Table 2.

SDE mRNAs in tissue with the highest fold change.

Gene log2FC P-value 
SLC25A48 7.64 6.22 E–43 
ATP12A 7.18 6.55 E–38 
SULT1E1 7.01 7.37 E–38 
CXCL13 6.94 2.33 E–56 
MT1H 6.84 3.77 E–44 
ZBTB16 6.25 1.51 E–90 
LHFPL3 6,18 3.92 E–35 
SLC5A1 6.10 2.86 E–49 
SCARA5 5.96 8.45 E–38 
MT1G 5.95 5.25 E–24 
Gene log2FC P-value 
SLC25A48 7.64 6.22 E–43 
ATP12A 7.18 6.55 E–38 
SULT1E1 7.01 7.37 E–38 
CXCL13 6.94 2.33 E–56 
MT1H 6.84 3.77 E–44 
ZBTB16 6.25 1.51 E–90 
LHFPL3 6,18 3.92 E–35 
SLC5A1 6.10 2.86 E–49 
SCARA5 5.96 8.45 E–38 
MT1G 5.95 5.25 E–24 

We then compared our mRNA data to the only previous RNA sequencing-based study [7]. In total, 1408 genes were identified as SDE in both studies, 1889 genes were uniquely identified in our study and 964 genes were uniquely identified in [7] (Figure 4A). Of the genes identified as SDE in both studies, there was a good agreement with regard to the direction of expression (Figure 4B). Of the 1408 genes found to be SDE in both studies, 1404 were expressed in the same direction.

Figure 4.

Comparison of SDE genes between RNA sequencing studies of the endometrium. (A) The overlap of SDE between the current study and Hu et al. [7]. In total, 1408 genes were identified as SDE in both studies, 1889 genes were uniquely identified in the current study, and 964 genes were uniquely identified in Hu et al. [7]. (B) Of the genes identifies as SDE in both studies, there is a good agreement with regards to the direction of expression. Of the 1408 genes found to be SDE in both studies, 1404 are expressed in the same direction. That is, 716 of these 1408 genes have a higher expression at WOI in both studies, and 688 genes have a lower expression at WOI in both studies.

Figure 4.

Comparison of SDE genes between RNA sequencing studies of the endometrium. (A) The overlap of SDE between the current study and Hu et al. [7]. In total, 1408 genes were identified as SDE in both studies, 1889 genes were uniquely identified in the current study, and 964 genes were uniquely identified in Hu et al. [7]. (B) Of the genes identifies as SDE in both studies, there is a good agreement with regards to the direction of expression. Of the 1408 genes found to be SDE in both studies, 1404 are expressed in the same direction. That is, 716 of these 1408 genes have a higher expression at WOI in both studies, and 688 genes have a lower expression at WOI in both studies.

Long noncoding RNA expression patterns

To analyze lncRNA, we restricted the total RNA sequences to lncRNA. Out of 17,832 lncRNAs, 875 were SDE; 359 out of these 875 failed LOOCV, which resulted in 516 SDE lncRNAs. Of these, 177 were higher expressed in the proliferative phase and 339 were higher expressed at 7–9 days after ovulation (

). Ten SDE lncRNAs with the highest fold change are presented in Table 3. Of the lncRNAs that were SDE between the time points, the Long intergenic noncoding RNA Nuclear enriched abundant transcript 1 (lincRNA NEAT1) (log2FC = 1.55, P = 8.99 E–12) was by far the most highly expressed (mean reads = 63,013 for both time points combined). The most SDE lncRNAs were RP11-627G23.1 (log2FC = 3.42, P = 7.24E–39) and PSORS1C3 (log2FC = 1.62, P = 1.62E–34). This analysis thus resulted in a comprehensive description of lncRNA expression in both the proliferative phase and 7–9 days after ovulation in human endometrium and describes lncRNA not previously reported in the endometrium.

Table 3.

SDE lncRNAs in tissue with the highest fold change.

Gene log2 FC P-value 
RP11-599B13.9 5.79 1.87 E–20 
LINC01411 −5.61 6.27 E–17 
RP11-203P23.2 5.49 2.98 E–14 
RP11-557H15.3 5.48 6.34 E–31 
LHFPL3-AS1 5.43 2.97 E–13 
LINC01016 −5,41 9,81 E–14 
RP11-411K7.1 5.31 4.93 E–22 
RP11-143E21.3 −5.27 1.12 E–14 
RP11-481J2.2 5.26 5.19 E–18 
RP11-874J12.3 5.22 9.21 E–19 
Gene log2 FC P-value 
RP11-599B13.9 5.79 1.87 E–20 
LINC01411 −5.61 6.27 E–17 
RP11-203P23.2 5.49 2.98 E–14 
RP11-557H15.3 5.48 6.34 E–31 
LHFPL3-AS1 5.43 2.97 E–13 
LINC01016 −5,41 9,81 E–14 
RP11-411K7.1 5.31 4.93 E–22 
RP11-143E21.3 −5.27 1.12 E–14 
RP11-481J2.2 5.26 5.19 E–18 
RP11-874J12.3 5.22 9.21 E–19 

Small noncoding RNA expression patterns

To describe the expression pattern of small RNA, we restricted the library to miRNA, snRNA, snoRNA, and novel miRNA (see Methods). A total of 156 out of 2,024 sncRNAs were found to be SDE and 54 out of these 156 failed LOOCV (see Methods). This resulted in 102 SDE sncRNAs. Of these, 52 were higher expressed in the proliferative phase and 50 were higher expressed 7–9 days after ovulation (see

). Ten SDE sncRNAs with the highest fold change are presented in Table 4. Ten sncRNAs were differentially expressed with a log2FC of two or more. Of the sncRNAs that were SDE between the time points, MIR449A, MIR449B, and MIR449C were the three most highly differently expressed sncRNAs. We noted that MIR30d (P = 9.69 E–16, log2FC = 1.8), MIR30b (P = 1.08 E–7, log2FC = 1.34), MIR10a (P = 9.37 E–7, log2FC = 0.78), MIR345 (P = 1.18 E–12, log2FC = 1.54), MIR210 (P = 6.28 E–8, log2FC = 1.39), MIR203 (P = 3.47 E–15, log2FC = 1.40), MIR105-1 (P = 1.01 E–8, log2FC = 1.59), and MIR200C (P = 1.72 E–8, log2FC = 1.11) were all significantly higher expressed 7–9 days after ovulation, and miRNA 135b (P-value 1,18 E–12, FC = –1.11) was higher expressed at the proliferative phase [36]. We also noted that MIR31 and MIR494 were not significantly different between the proliferative phase and 7–9 days after ovulation [4,6].

Table 4.

SDE sncRNAs in tissue with the highest fold change.

Gene log2FC P-value 
MIR449B 3.68 3.94 E–36 
MIR449C 3.64 3.01 E–27 
MIR449A 3.45 7.79 E–32 
MIR885 3.13 4.70 E–38 
MIR375 2.70 3.12 E–21 
MIR549A − 2.60 7.06 E–09 
AL139147.1 − 2.33 1.77 E–22 
MIR1298 − 2.32 8.34 E–11 
RNU4-4P 2.23 2.43 E–07 
MIR483 − 2.06 4.34 E–36 
Gene log2FC P-value 
MIR449B 3.68 3.94 E–36 
MIR449C 3.64 3.01 E–27 
MIR449A 3.45 7.79 E–32 
MIR885 3.13 4.70 E–38 
MIR375 2.70 3.12 E–21 
MIR549A − 2.60 7.06 E–09 
AL139147.1 − 2.33 1.77 E–22 
MIR1298 − 2.32 8.34 E–11 
RNU4-4P 2.23 2.43 E–07 
MIR483 − 2.06 4.34 E–36 

There were 31 SDE snoRNA, 12 of them belonged to the SNORD113 and SNORD114 families, all higher expressed in the proliferative phase. There were five snoRNAs higher expressed 7–9 days after ovulation and 26 higher expressed in the proliferative phase. SNORD 113-3 had the largest fold change (P = 1.14 E–11, log2FC = –1.05). Two snRNAs were statistically differentially expressed, RNU4-82P (log2FC = 1.83, P = 1.05 E–7) and RNU4-4P (log2FC = 2.23, P = 2.43 E–7). Both these snRNAs were higher expressed 7–9 days after ovulation, but their expression were low. One novel miRNA, novel_miRNA_12 (

), was statistically higher expressed 7–9 days after ovulation (log2FC = 1.13, P = 3.70 E–7).

To study the change in small RNA expression patterns after decidualization in a more controlled setting, we also performed RNA sequencing before and after treatment with cAMP and medroxyprogsterone in human endometrial stromal cells in vitro. A total of 235 out of 2,010 sncRNAs were found to be SDE, and 80 out of these 235 failed LOOCV. This resulted in 155 SDE sncRNAs (see

). The 10 most differentially expressed small RNAs with regard to fold change are presented in Table 5. Similar to sncRNA in the tissue, four SDE sncRNAs, MIR21, MIR143, MIR10B, and MIR30A, comprise over 80% of all the reads of SDE sncRNAs. These data provide a comprehensive description of the change in small RNA expression after decidualization in vitro and shows that small RNA expression in human endometrial stromal cells is influenced by progestin treatment. Of the 102 SDE sncRNAs in tissue and the 155 SDE sncRNA in stromal cells, only 33 are SDE in both. In addition, only 12 of these are SDE in the same direction, (see ). The five miRNAs, MIR30A, MIR30B, MIR30D, MIR148A, and AC114498.1, are significantly higher expressed during the WOI in tissue and after decidualization in cells. The seven miRNAs, MIR143, MIR145, MIR146B, MIR155, MIR424, MIR708, and MIR7974, are significantly higher expressed during the proliferative phase in tissue and before decidualization in cells.

Table 5.

SDE sncRNAs in stromal cells with the highest fold change.

Gene log2FC P-value 
MIR7974 − 2.61 7.76 E–18 
MIR483 2.59 8.16 E–20 
MIR2114 2.20 8.44 E–17 
SNORD114-24 − 1.93 3.45 E–47 
MIR218-2 − 1.90 7.59 E–17 
MIR549A 1.81 7.03 E–33 
SNORD88B − 1.58 4.79 E–17 
MIR146B − 1.53 3.23 E–54 
MIR490 − 1.42 3.27 E–47 
MIR3126 1.40 1.21 E–09 
Gene log2FC P-value 
MIR7974 − 2.61 7.76 E–18 
MIR483 2.59 8.16 E–20 
MIR2114 2.20 8.44 E–17 
SNORD114-24 − 1.93 3.45 E–47 
MIR218-2 − 1.90 7.59 E–17 
MIR549A 1.81 7.03 E–33 
SNORD88B − 1.58 4.79 E–17 
MIR146B − 1.53 3.23 E–54 
MIR490 − 1.42 3.27 E–47 
MIR3126 1.40 1.21 E–09 

Novel micro RNA expression in endometrium

miRDeep2 was applied to identify novel miRNAs in endometrium individually for each of the 14 samples. Some novel miRNAs turned out to have very low or even no expression in our data and were discarded as candidates for novel miRNAs. In total, there were 30 miRNAs that we identify as novel miRNAs and a complete list is shown in

. Of these 30 novel miRNAs, only novel_miRNA 12 on chromosome 7 was SDE in tissue between the proliferative phase and 7–9 days after ovulation ().

Predicting gene targets based on micro RNA expression

We applied three methods to predict the gene targets of the SDE miRNA families: PCT [37], CWC [36], and MTB [38] (see Methods).

shows the number of unique genes targeted by each method, how many of those are expressed in endometrium and how many are SDE in endometrium. shows the overlap of targeted genes between these three prediction methods. The three methods predict the same targets only to a limited extent and predicted gene targets have comparable amount of SDE as the general population of expressed genes in endometrium.

Discussion

The coordinate gene transcription in the receptive endometrium is probably of importance for successful implantation; yet, the transcriptomic shift from proliferative phase to receptive endometrium has not been described in detail. Our data show that not only mRNA but also lncRNA and sncRNA transcription differs significantly as the endometrium becomes receptive. We employed RNA sequencing and designed our study to allow analyses of mRNA, lncRNA, and sncRNA. We reported 31 snoRNA, 2 snRNA, and 1 novel miRNA to be SDE between the proliferative phase and 7–9 days after ovulation, a total of 34 molecules that have not been reported before to be involved in the physiological changes in the human endometrium. Additionally, we report complications and inconsistencies in using miRNA expression in gene target prediction. Taken together, our findings provide the most complete description of the coordination of the transcriptome 7–9 days after ovulation in human endometrium.

Most previous studies on gene expression in the endometrium have used microarray techniques [2,4,5,810,39] that ultimately limit the results [40]. There are several studies [810,39,41] performed at the WOI. The results from these studies have been compared [11,41], and stringly only one gene, osteopontin, was consistently upregulated in all five studies. These studies use different time points, study design, and methodology which can explain some of the differences. The only previous study using RNA sequencing [7] compared the transcriptome between three women 2 days after ovulation to three other women at 7 days after ovulation. In our study, SLC30A2, solute carrier family 30 member 2, was the most highly upregulated mRNA 7–9 days after ovulation, which is in agreement with the only previous study using RNA sequencing [7] but not with studies using microarray technics [8,12]. Of the 3,297 SDE genes in the current study, 1,408 found to be SDE by us were also found to be SDE in the previous RNA sequencing-based study [7], which considering the difference in design and different time points used is a surprisingly good agreement. Furthermore, 1,404 of these 1,408 genes agreed on the direction of expression (Figure 4B). The most likely reason for the divergence in the results is differences in technique and experimental design. For instance, the time points differ and some mRNA, such as many histone genes and others, lack polyA tails [42,43], meaning they would not be probed using polyA selection. In fact, some of the SDE genes reported here are histone genes (

). Compared to Ref. [7], our libraries are more refined in the sense that they are strand specific and enriched by ribosomal depletion instead of poly A selection. Finally, the reported inconsistency between the various microarray-based studies and the relative good correlation between the two RNA sequencing studies further highlights the advantage of sequencing over microarray-based approach in terms of both robustness and accuracy.

The exact role of lncRNA in the endometrium has not been elucidated. A total of 42 of these 516 lncRNAs found SDE in our material overlap with lncRNAs that have been implemented with endometrial cancer [15], among them urothelial cancer associated 1 (UCA1), an lncRNA, previously associated with cancer and embryonic growth [44].

Endometrial miRNA expression differs throughout the menstrual cycle and has been proposed to influence receptivity. MIR30D, MIR30B, MIR10A, MIR345, MIR210, MIR203, MIR105, MIR200C, and MIR193A were all highly differentially expressed between the time points that agree well with the results of earlier research [35]. We found MIR146B to be downregulated 7–9 days after ovulation, which has not been described in earlier studies. This miRNA regulates signal transduction of Transforming growth factor beta (TGF-β) by repressing SMAD4 [45]. TGF-β represses decidualization and prolactin expression, in endometrial stromal cells in vitro but SMAD4 inhibition restores this [45]. In addition, we found MIR449 to be expressed in the receptive endometrium. MIR449 is well described in other tissues, especially ciliated cells, bronchial mucociliary airway epithelia and testis [46]. It has also been described in bovine endometrium [47] and in endometrial cancer [48] but never in healthy human endometrium. The MIR449 cluster is located on chromosome 5, in a highly conserved region [46]. Three members of the MIR449 family, MIR449A, MIR449B, and MIR449C, were expressed in our study; all of them were among the most highly differentially expressed. The expression of the MIR449 family increased in all women 7–9 days after ovulation. The MIR449 family is important in the induction of cell cycle arrest and cell apoptosis [46], which is some of the hallmark in receptive endometrium [49].

Our findings have also uncovered the expression of snoRNA in the receptive endometrium. Small nucleolar RNAs are involved in post-transcriptional maturation of rRNA. Alteration in snoRNA expression can affect numerous vital cellular processes as, for example, the development of tumor cells or the response to viral diseases [25]. Future studies need to focus on the functional role of these RNA subtypes and other highly altered RNAs in endometrial receptivity.

The expression of small RNA in vitro and in vivo clustered according to type of tissue (cells or endometrial biopsies) and not according to the time point. This was particularly evident in the cell cultures, whereas the small RNA expression in the endometrial tissue samples differed distinctly between the time points (Figure 2). The endometrial biopsies contain several cell types, for example, epithelial cells, stromal cells, leucocytes, and fibroblasts [40] compared to the in vitro analyses that only contain stromal cells. The difference in sncRNA expression between the in vivo and in vitro analysis may also be explained by hormonal factors. The stromal cells were subjected to gestagen but not to other sex steroids or other hormones that may regulate gene transcription in vivo. It should also be noted that the major compartment of stromal cells are not decidualized 7–9 days post ovulation. Instead, only a proportion of the cells are decidualized or predecidualized [9], starting in patches surrounding the spiral arteries, and decidualization becomes apparent on day 9–10 after ovulation [50,51]

This study has some limitations. There is a small sample size, which is common in this kind of study with general invasive sample collection [4,5,810,12,41,52]. Biopsies from the two time points were taken from the same individuals, the first in the secretory phase and the second in the subsequent proliferative phase, which could potentially influence gene expression in the latter. However, the two biopsies were taken with one menstrual bleeding in between. The strength lies in that consecutive samples from the same individuals allowed paired statistical testing, increasing the power of the test, and reducing confounding factors.

Furthermore, LOOVC analyses (see methods) have been used to lower the effect from every single donor. Endometrium is heterogeneous and interindividual differences might have influenced the results, although the samples cluster according to time point instead of individuals. There are studies suggesting that interindividual differences might override differences in menstrual phases [53]. One subject deviated from other samples during the proliferative phase, orange square, we cannot find anything in the characteristics that differ from the other subjects.

In conclusion, this paper reports in detail the shift in RNA expression from the proliferative to receptive endometrium. Due to study design, we have been able to show the expression of mRNA, lncRNA, and sncRNA including snoRNA, snRNA, and miRNA. The data include descriptions of novel miRNAs never described before and shows that snoRNA is expressed in human healthy endometrium. Future studies should investigate the biological relevance of these new miRNAs and the role of snoRNA in endometrial receptivity.

Supplementary data

Supplementary data are available at BIOLRE online.

Supplemental Figure S1. Graphical representation of the preprocessing of endometrial RNA reads, both from total RNA and small RNA. The biggest decrease in total RNA reads is due to ribosomal contamination, but on average around 13.8% of the mapped reads were ribosomal sequences. The biggest decline in small RNA reads is due to mapping but on average around 74.5% of the reads mapped to the genome. See Supplemental Table S2 for further details.

Supplemental Figure S2. (A) All samples, despite varying RIN, show similar coverage over the gene body. The coverage is generally even albeit with a slight drop in the 5΄ end. (B) The RiboZero kit, used here for rimosomal removal, performs better on degraded samples than on intact samples.

Supplemental Figure S3. (A) Shown here is how the micro RNA MIR21 overlaps the protein-coding gene VMP1. If both genes are present in the annotation file, then all reads that fall within the overlapping region are labeled as ambiguous. This is essentially all the reads mapping to MIR21. Limiting the annotation file to only miRNA, snRNA, and snoRNA solves this problem. (B) Similar scenario as in A. Here, MIR33A overlaps the protein-coding gene SREPF2. Limiting the annotation file to only miRNA, snRNA, and snoRNA allows for a correct expression quantification for MIR33A. (C) Present as peaks in the top rows, two snoRNA families, SNORD113 and SNORD114, are visible. Nearly, all these snoRNAs overlap the two noncoding RNAs SNHG23 and SNHG24, shown at the bottom. Removing all annotations except for miRNA, snRNA, and snoRNA from the annotation file allows for expression quantification of these and other snoRNAs.

Supplemental Figure S4. Comparison of SDE genes from sncRNA in tissue and sncRNA in cells. (A) Of the 155 SDE sncRNA identified in cells and the 102 SDE sncRNA identified in tissue, only 33 are identified in both tissue and cells. (B) Of the 33 sncRNA identified as SDE in both tissue and cells, only 12 agree on the direction of expression (green dots). Five sncRNA are SDE at WOI in tissue and after decidualization in cells, and 7 sncRNA are SDE at proliferative phase in tissue and before decidualization in cells.

Supplemental Figure S5. Overlap of the different methods used to predict miRNA gene targets. (A) Overlap of all target genes expressed in endometrium. (B) Overlap of all SDE genes in endometrium. PCT: probability of conserved targeting, CWC: cumulative weighted contexts ++ score, MTB: miRTarBase.

Supplamental Table S1. Overview of results from RNA extraction of endometrial tissue.

Supplemental Table S2. Preprocessing and mapping of RNA. No. of reads (R1 + R2) in millions. Raw reads are the number of reads coming from the sequencing machine, quality reads is the number of reads after the trimming process, mapped reads is how many reads mapped to the genome, and useable reads is the number of reads remaining after ribosomal reads.

Supplemental Table S5a. Significantly differentially expressed mRNA in endometrial tissue between the proliferative phase and the window of implantation. Ordered according to log2 fold change, i.e., from most overexpressed in the receptive phase (LH+ 7–9) to the most overexpressed in the proliferative phase (cd 6–8).

Supplemental Table S5b. Significantly differentially expressed lncRNA in endometrial tissue between the proliferative phase and the window of implantation. Ordered according to log2 fold change, i.e., from most overexpressed in the receptive phase (LH+ 7–9) to the most overexpressed in the proliferative phase (cd 6–8).

Supplemental Table S5c. Significantly differentially expressed sncRNA in endometrial tissue between the proliferative phase and the window of implantation. Ordered according to log2 fold change, i.e., from most overexpressed in the receptive phase (LH+ 7–9) to the most overexpressed in the proliferative phase (cd 6–8).

Supplemental Table S5d. Significantly differentially expressed sncRNA in endometrial stromal cells between before and after decidualization. Ordered according to log2 fold change, i.e., from most overexpressed in decidualized cells to the most overexpressed in the naïve cells.

Supplemental Table S6. Novel miRNA expressed in endometrial tissue. Identified with miRDeep2 and cross referenced from alignment data. Of the novel miRNA only miRNA_novel12 was found to be SDE between the two time points.

Supplamental Table S7. Verification of RNAseq results using RT-qPCR.

Acknowledgments

We gratefully acknowledge Science for Life Laboratory (SciLifeLab, Stockholm), National Genomics Infrastructure (NGI—Sweden), and Uppmax for providing massive parallel sequencing and computational infrastructure.

References

1.
Gellersen
B
,
Brosens
IA
,
Brosens
JJ
.
Decidualization of the human endometrium: mechanisms, functions, and clinical perspectives
.
Semin Reprod Med
 
2007
;
25
:
445
453
.
2.
Li
R
,
Qiao
J
,
Wang
L
,
Li
L
,
Zhen
X
,
Liu
P
,
Zheng
X
.
MicroRNA array and microarray evaluation of endometrial receptivity in patients with high serum progesterone levels on the day of hCG administration
.
Reprod Biol Endocrinol
 
2011
;
9
:
1
9
.
3.
Hull
ML
,
Nisenblat
V
.
Tissue and circulating microRNA influence reproductive function in endometrial disease
.
Reprod Biomed Online
 
2013
;
27
:
515
529
.
4.
Altmae
S
,
Martinez-Conejero
JA
,
Esteban
FJ
,
Ruiz-Alonso
M
,
Stavreus-Evers
A
,
Horcajadas
JA
,
Salumets
A
.
MicroRNAs miR-30b, miR-30d, and miR-494 regulate human endometrial receptivity
.
Reprod Sci
 
2013
;
20
:
308
317
.
5.
Kuokkanen
S
,
Chen
B
,
Ojalvo
L
,
Benard
L
,
Santoro
N
,
Pollard
JW
.
Genomic profiling of microRNAs and messenger RNAs reveals hormonal regulation in microRNA expression in human endometrium
.
Biol Reprod
 
2010
;
82
:
791
801
.
6.
Kresowik
JD
,
Devor
EJ
,
Van Voorhis
BJ
,
Leslie
KK
.
MicroRNA-31 is significantly elevated in both human endometrium and serum during the window of implantation: a potential biomarker for optimum receptivity
.
Biol Reprod
 
2014
;
91
(
1
):17,
1
6
.
7.
Hu
S
,
Yao
G
,
Wang
Y
,
Xu
H
,
Ji
X
,
He
Y
,
Zhu
Q
,
Chen
Z
,
Sun
Y
.
Transcriptomic changes during the pre-receptive to receptive transition in human endometrium detected by RNA-Seq
.
J Clin Endocrinol Metab
 
2014
;
99
:
E2744
E2753
.
8.
Carson
DD
,
Lagow
E
,
Thathiah
A
,
Al-Shami
R
,
Farach-Carson
MC
,
Vernon
M
,
Yuan
L
,
Fritz
MA
,
Lessey
B
.
Changes in gene expression during the early to mid-luteal (receptive phase) transition in human endometrium detected by high-density microarray screening
.
Mol Hum Reprod
 
2002
;
8
:
871
879
.
9.
Kao
LC
,
Tulac
S
,
Lobo
S
,
Imani
B
,
Yang
JP
,
Germeyer
A
,
Osteen
K
,
Taylor
RN
,
Lessey
BA
,
Giudice
LC
.
Global gene profiling in human endometrium during the window of implantation
.
Endocrinology
 
2002
;
143
:
2119
2138
.
10.
Borthwick
JM
,
Charnock-Jones
DS
,
Tom
BD
,
Hull
ML
,
Teirney
R
,
Phillips
SC
,
Smith
SK
.
Determination of the transcript profile of human endometrium
.
Mol Hum Reprod
 
2003
;
9
:
19
33
.
11.
Horcajadas
JA
,
Pellicer
A
,
Simon
C
.
Wide genomic analysis of human endometrial receptivity: new times, new opportunities
.
Hum Reprod Update
 
2007
;
13
:
77
86
.
12.
Ponnampalam
AP
,
Weston
GC
,
Trajstman
AC
,
Susil
B
,
Rogers
PA
.
Molecular classification of human endometrial cycle stages by transcriptional profiling
.
Mol Hum Reprod
 
2004
;
10
:
879
893
.
13.
Talbi
S
,
Hamilton
AE
,
Vo
KC
,
Tulac
S
,
Overgaard
MT
,
Dosiou
C
,
Le Shay
N
,
Nezhat
CN
,
Kempson
R
,
Lessey
BA
,
Nayak
NR
,
Giudice
LC
.
Molecular phenotyping of human endometrium distinguishes menstrual cycle phases and underlying biological processes in normo-ovulatory women
.
Endocrinology
 
2006
;
147
:
1097
1121
.
14.
Kung
JT
,
Colognori
D
,
Lee
JT
.
Long noncoding RNAs: past, present, and future
.
Genetics
 
2013
;
193
:
651
669
.
15.
Yang
L
,
Zhang
J
,
Jiang
A
,
Liu
Q
,
Li
C
,
Yang
C
,
Xiu
J
.
Expression profile of long non-coding RNAs is altered in endometrial cancer
.
Int J Clin Exp Med
 
2015
;
8
:
5010
5021
.
16.
Wang
Y
,
Li
Y
,
Yang
Z
,
Liu
K
,
Wang
D
.
Genome-wide microarray analysis of long non-coding RNAs in eutopic secretory endometrium with endometriosis
.
Cell Physiol Biochem
 
2015
;
37
:
2231
2245
.
17.
Bartel
DP
.
MicroRNAs: target recognition and regulatory functions
.
Cell
 
2009
;
136
:
215
233
.
18.
Bartel
DP
.
MicroRNAs: genomics, biogenesis, mechanism, and function
.
Cell
 
2004
;
116
:
281
297
.
19.
Yanokura
M
,
Banno
K
,
Kobayashi
Y
,
Kisu
I
,
Ueki
A
,
Ono
A
,
Masuda
K
,
Nomura
H
,
Hirasawa
A
,
Susumu
N
,
Aoki
D
.
MicroRNA and endometrial cancer: roles of small RNAs in human tumors and clinical applications (Review)
.
Oncol Lett
 
2010
;
1
:
935
940
.
20.
Burney
RO
,
Hamilton
AE
,
Aghajanova
L
,
Vo
KC
,
Nezhat
CN
,
Lessey
BA
,
Giudice
LC
.
MicroRNA expression profiling of eutopic secretory endometrium in women with versus without endometriosis
.
Mol Hum Reprod
 
2009
;
15
:
625
631
.
21.
Qian
K
,
Hu
L
,
Chen
H
,
Li
H
,
Liu
N
,
Li
Y
,
Ai
J
,
Zhu
G
,
Tang
Z
,
Zhang
H
.
Hsa-miR-222 is involved in differentiation of endometrial stromal cells in vitro
.
Endocrinology
 
2009
;
150
:
4734
4743
.
22.
Hu
SJ
,
Ren
G
,
Liu
JL
,
Zhao
ZA
,
Yu
YS
,
Su
RW
,
Ma
XH
,
Ni
H
,
Lei
W
,
Yang
ZM
.
MicroRNA expression and regulation in mouse uterus during embryo implantation
.
J Biol Chem
 
2008
;
283
:
23473
23484
.
23.
Fu
D
,
Collins
K
.
Human telomerase and Cajal body ribonucleoproteins share a unique specificity of Sm protein association
.
Gene Dev
 
2006
;
20
:
531
536
.
24.
Hales
CM
,
Dammer
EB
,
Diner
I
,
Yi
H
,
Seyfried
NT
,
Gearing
M
,
Glass
JD
,
Montine
TJ
,
Levey
AI
,
Lah
JJ
.
Aggregates of small nuclear ribonucleic acids (snRNAs) in Alzheimer's disease
.
Brain Pathol
 
2014
;
24
:
344
351
.
25.
Stepanov
GA
,
Filippova
JA
,
Komissarov
AB
,
Kuligina
EV
,
Richter
VA
,
Semenov
DV
.
Regulatory role of small nucleolar RNAs in human diseases
.
Biomed Res Int
 
2015
;
2015
:
1
11
.
26.
Ravo
M
,
Cordella
A
,
Rinaldi
A
,
Bruno
G
,
Alexandrova
E
,
Saggese
P
,
Nassa
G
,
Giurato
G
,
Tarallo
R
,
Marchese
G
,
Rizzo
F
,
Stellato
C
et al
Small non-coding RNA deregulation in endometrial carcinogenesis
.
Oncotarget
 
2015
;
6
:
4677
4691
.
27.
Arnold
JT
,
Kaufman
DG
,
Seppala
M
,
Lessey
BA
.
Endometrial stromal cells regulate epithelial cell growth in vitro: a new co-culture model
.
Hum Reprod
 
2001
;
16
:
836
845
.
28.
Satyaswaroop
PG
,
Bressler
RS
,
de la Pena
MM
,
Gurpide
E
.
Isolation and culture of human endometrial glands
.
J Clin Endocr Metab
 
1979
;
48
:
639
641
.
29.
Frolova
AI
,
O’Neill
K
,
Moley
KH
.
Dehydroepiandrosterone inhibits glucose flux through the pentose phosphate pathway in human and mouse endometrial stromal cells, preventing decidualization and implantation
.
Mol Endocrinol
 
2011
;
25
:
1444
1455
.
30.
Dobin
A
,
Davis
CA
,
Schlesinger
F
,
Drenkow
J
,
Zaleski
C
,
Jha
S
,
Batut
P
,
Chaisson
M
,
Gingeras
TR
.
STAR: ultrafast universal RNA-seq aligner
.
Bioinformatics
 
2013
;
29
:
15
21
.
31.
Felix
K
.
Trim galore!
[Internet].
2013
. .
32.
Friedlander
MR
,
Mackowiak
SD
,
Li
N
,
Chen
W
,
Rajewsky
N
.
miRDeep2 accurately identifies known and hundreds of novel microRNA genes in seven animal clades
.
Nucleic Acids Res
 
2012
;
40
:
37
52
.
33.
Anders
S
,
Pyl
PT
,
Huber
W
.
HTSeq–a Python framework to work with high-throughput sequencing data
.
Bioinformatics
 
2014
;
31
:
166
169
.
34.
Team
RC
.
R: A Language and Environment for Statistical Computing
 .
Vienna, Austria
:
R Foundation for Statistical Computing
;
2014
.
35.
Love
MI
,
Huber
W
,
Anders
S
.
Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2
.
Genome Biol
 
2014
;
15
:
1
21
.
36.
Agarwal
V
,
Bell
GW
,
Nam
JW
,
Bartel
DP
.
Predicting effective microRNA target sites in mammalian mRNAs
.
Elife
 
2015
;
4
:
1
36
.
37.
Friedman
RC
,
Farh
KK
,
Burge
CB
,
Bartel
DP
.
Most mammalian mRNAs are conserved targets of microRNAs
.
Genome Res
 
2009
;
19
:
92
105
.
38.
Chou
CH
,
Chang
NW
,
Shrestha
S
,
Hsu
SD
,
Lin
YL
,
Lee
WH
,
Yang
CD
,
Hong
HC
,
Wei
TY
,
Tu
SJ
,
Tsai
TR
,
Ho
SY
et al
miRTarBase 2016: updates to the experimentally validated miRNA-target interactions database
.
Nucleic Acids Res
 
2016
;
44
:
D239
D247
.
39.
Riesewijk
A
,
Martin
J
,
van Os
R
,
Horcajadas
JA
,
Polman
J
,
Pellicer
A
,
Mosselman
S
,
Simon
C
.
Gene expression profiling of human endometrial receptivity on days LH+2 versus LH+7 by microarray technology
.
Mol Hum Reprod
 
2003
;
9
:
253
264
.
40.
Altmae
S
,
Esteban
FJ
,
Stavreus-Evers
A
,
Simon
C
,
Giudice
L
,
Lessey
BA
,
Horcajadas
JA
,
Macklon
NS
,
D’Hooghe
T
,
Campoy
C
,
Fauser
BC
,
Salamonsen
LA
et al
Guidelines for the design, analysis and interpretation of ‘omics’ data: focus on human endometrium
.
Hum Reprod Update
 
2014
;
20
(
1
):
12
28
.
41.
Mirkin
S
,
Arslan
M
,
Churikov
D
,
Corica
A
,
Diaz
JI
,
Williams
S
,
Bocca
S
,
Oehninger
S
.
In search of candidate genes critically expressed in the human endometrium during the window of implantation
.
Hum Reprod
 
2005
;
20
:
2104
2117
.
42.
Yang
L
,
Duff
MO
,
Graveley
BR
,
Carmichael
GG
,
Chen
LL
.
Genomewide characterization of non-polyadenylated RNAs
.
Genome Biol
 
2011
;
12
:
1
14
.
43.
Sigurgeirsson
B
,
Emanuelsson
O
,
Lundeberg
J
.
Sequencing degraded RNA addressed by 3΄ tag counting
.
PLoS One
 
2014
;
9
:
1
11
.
44.
Wang
F
,
Li
X
,
Xie
X
,
Zhao
L
,
Chen
W
.
UCA1, a non-protein-coding RNA up-regulated in bladder carcinoma and embryo, influencing cell growth and promoting invasion
.
FEBS Lett
 
2008
;
582
:
1919
1927
.
45.
Kane
NM
,
Jones
M
,
Brosens
JJ
,
Kelly
RW
,
Saunders
PT
,
Critchley
HO
.
TGFbeta1 attenuates expression of prolactin and IGFBP-1 in decidualized endometrial stromal cells by both SMAD-dependent and SMAD-independent pathways
.
PLoS One
 
2010
;
5
:
1
9
.
46.
Lize
M
,
Klimke
A
,
Dobbelstein
M
.
MicroRNA-449 in cell fate determination
.
Cell Cycle
 
2011
;
10
:
2874
2882
.
47.
Ponsuksili
S
,
Tesfaye
D
,
Schellander
K
,
Hoelker
M
,
Hadlich
F
,
Schwerin
M
,
Wimmers
K
.
Differential expression of miRNAs and their target mRNAs in endometria prior to maternal recognition of pregnancy associates with endometrial receptivity for in vivo- and in vitro-produced bovine embryos
.
Biol Reprod
 
2014
;
91
:
1
12
.
48.
Wu
W
,
Lin
Z
,
Zhuang
Z
,
Liang
X
.
Expression profile of mammalian microRNAs in endometrioid adenocarcinoma
.
Eur J Cancer Prev
 
2009
;
18
:
50
55
.
49.
Boeddeker
SJ
,
Hess
AP
.
The role of apoptosis in human embryo implantation
.
J Reprod Immunol
 
2015
;
108
:
114
122
.
50.
Teklenburg
G
,
Salker
M
,
Heijnen
C
,
Macklon
NS
,
Brosens
JJ
.
The molecular basis of recurrent pregnancy loss: impaired natural embryo selection
.
Mol Hum Reprod
 
2010
;
16
:
886
895
.
51.
Kajihara
T
,
Brosens
JJ
,
Ishihara
O
.
The role of FOXO1 in the decidual transformation of the endometrium and early pregnancy
.
Med Mol Morphol
 
2013
;
46
:
61
68
.
52.
Evans
GE
,
Martinez-Conejero
JA
,
Phillipson
GT
,
Simon
C
,
McNoe
LA
,
Sykes
PH
,
Horcajadas
JA
,
Lam
EY
,
Print
CG
,
Sin
IL
,
Evans
JJ
.
Gene and protein expression signature of endometrial glandular and stromal compartments during the window of implantation
.
Fertil Steril
 
2012
;
97
:
1365
1373
.
e1361–1362
.
53.
Sredni
ST
,
Gadd
S
,
Jafari
N
,
Huang
CC
.
A parallel study of mRNA and microRNA profiling of peripheral blood in young adult women
.
Front Genet
 
2011
;
2
:
1
6
.

Author notes

Accession Number: GEO (GSE86491)
These authors have contributed equally in the preparation of this paper.
Disclosure statement: The authors have nothing to disclose.