It has been well documented that aberrant expression of selected microRNAs (miRNAs) might contribute to the pathogenesis of disease. The aim of the present study is to compare miRNA expression by the most comprehensive locked-nucleic acid (LNA) miRNA microarray in eutopic endometrium of patients with endometriosis and control. In the study we recruited 21 patients with endometriosis and 25 were disease-free women. The miRNA expression profiles were determined using the LNA miRNA microarray and validated for selected molecules by real-time PCR. We identified 1198 human miRNAs significantly differentially altered in endometriosis versus control samples using false discovery rate of <5%. However only 136 miRNAs showed differential regulation by fold change of at least 1.3. By the use of selected statistical analysis we obtained 45 potential pathways that might play a role in the pathogenesis of endometriosis. We also found that natural killer cell mediated cytotoxicity pathway was found to be inhibited which is consistent with previous studies. There are several pathways that may be potentially dysregulated, due to abnormal miRNA expression, in eutopic endometrium of patients with endometriosis and in this way contribute to its pathogenesis.
Endometriosis is relatively common, benign gynecological disease in which endometrial tissue ectopically implants in a location outside the uterine cavity. The most common clinical symptoms include pelvic pain and infertility which can seriously influence the quality of life of affected patients [
Despite the current controversy regarding the pathophysiology of this disease, Sampson’s theory explains the existence of endometrial cells in the peritoneal cavity by retrograde menstruation. Several factors such as increased chemokines and matrix metalloproteinases in the peritoneal fluid (PF), angiogenesis, tumor suppressor, and oncogenes as well as upregulating of proinflammatory cytokines may facilitate the pathogenesis of endometriosis [
The molecular changes that could potentially lead to abnormal growth of endometrium outside uterus include microRNAs (miRNAs) expression fluctuations [
In our recent study we also showed that, out of 667 miRNAs, two molecules, namely, hsa-miR-483-5p and hsa-miR-
Patients (
Patients with endometriosis (Group I,
All women had regular menstrual cycles (28–30 days) and were not taking any medication for at least 3 months prior to operation. We excluded patients with autoimmune disease, pelvic inflammatory disease, adenomyosis, fibroids, and dysfunctional uterine bleeding. The study was approved by the Institutional Review Board of Medical University of Bialystok and informed consent was obtained from each patient.
The collected tissue was placed separately in buffered formalin for histopathological studies and in RNA later (Sigma-Aldrich, Poland) for molecular analysis. The latter was stored for 24 hours in +4°C and then tissues were transferred and stored in −80°C. Total RNA was extracted using the
For the purpose of miRNA microarray screening we chose 10 samples of patients with advanced endometriosis and 11 controls. Microarray analysis was conducted as single-channel Hy3 experiments on Exiqon’s miRCURY LNA (locked-nucleic acid) microRNA Array 7th generation—hsa, mmu and rno. Exiqon arrays contain 3100 capture probes, complementary to most human, mouse, rat, and their related viral sequences from the v.19.0 release of miRBase. The arrays also contain 25 proprietary human miRPlus sequences not yet in miRBase. 500 ng RNA sample was labelled with a Hy3 fluorophore (Exiqon, Denmark). Labelling reactions were performed using Exiqon’s miRCURY LNA microRNA Hi-Power Labeling Kit with the use of synthetic spike controls, Spike-in microRNA Kit v2 (Exiqon, Denmark), according to the manufacturer’s protocol. Hybridization of labeled RNA to the array was performed in SureHyb chambers (Agilent Technologies, USA) for 16 hours at 56°C. Slides were washed according to manufacturer’s instructions and scanned at 10
The expression levels of the selected microRNAs and the assay miRBase IDs (miRBase Accession Number) are hsa-miR-4714-5p (MIMAT0019822), hsa-miR-4284 (MIMAT0016915), hsa-miR-5193 (MIMAT0021124), hsa-miR-4454 (MIMAT0018976), hsa-miR-3680-5p (MIMAT0018106), hsa-miR-3667-5p (MIMAT0018089), hsa-miR-23a-3p (MIMAT0000078), hsa-miR-23b-3p (MIMAT0000418), hsa-miR-5187-3p (MIMAT0021118), hsa-miR-3152-5p (MIMAT0019207), and hsa-miR-30d-5p (MIMAT0000245) which were evaluated using the miRCURY LNA Universal RT microRNA PCR system (Exiqon, Denmark). In the first step, we have conducted the one first-strand cDNA synthesis reaction, which provided template for all microRNA real-time PCR assays. The cDNA for each RNA sample was obtained using the Universal cDNA Synthesis Kit II (Exiqon, Denmark) according to the manufacturer’s instructions. Levels of miRNAs were quantitated using individual miRCURY LNA Universal RT microRNA PCR Assays (Exiqon, Denmark).
The conditions for qPCR were as follows: 95°C for 10 min and 45 cycles of 95°C for 10 sec followed by 60°C for 1 min. Finally a melting curve analysis was performed with denaturation at 95°C for 15 s and 60°C for 15 s followed by a temperature gradient from 60 to 95°C for 20 min and a final denaturation at 95°C for 15 s. U6 snRNA was used as the endogenous control. LNA PCR amplification reactions for each miRNA molecule were repeated independently three times. Quantitative real-time PCR analysis was performed using the Applied Biosystems 7900HT System (Life Technologies, Foster City, CA).
Gene expression values were calculated based on the
All data analyses were performed in R statistical environment (
The raw microarray data were preprocessed with
In our study we used signaling pathway definitions from KEGG (
To assess perturbation of signaling pathways we applied SPIA (
While comparing two groups for quantitative data, Mann-Whitney-Wilcoxon test was used due to the nonnormal distribution of the tested variables. The significance level was equal to 0.05. The calculations have been carried out by means of Microsoft Excel spreadsheet and STATISTICA, StatSoft, Inc. Version 7.1. statistical package (data analysis software system).
Patients clinical characteristics are presented in Table
Clinical characteristics of patients.
Endometriosis ( |
Control ( | |
---|---|---|
Age (years) | 31.35 ± 0.90 | 30.73 ± 0.91 |
Infertility, |
9 (50) | 16 (64) |
Primary | 7 | 8 |
Secondary | 2 | 8 |
Duration of infertility (months) | 48.2 ± 19.4 (23–71) | 50.8 ± 15.8 (13–71) |
Ovarian cysts, |
||
Endometrial | 21 (100) | — |
Simple | — | 9 (36) |
Data are mean ± SEM. Ranges are provided for “duration of infertility.”
We identified 1198 human miRNAs significantly differentially altered in endometriosis versus control samples using false discovery rate of <5% (Table
Volcano plot (Figure
Volcano plot for all analyzed human miRNAs. The
Heatmap display of significantly expressed miRNAs. Heatmap representation of expression data for miRNAs significantly altered in endometriosis samples compared with control samples (adjusted
The associations between miRNA and mRNA are presented in the zipped supplement directory (called Table
We obtained 45 potential pathways from the KEGG (Kyoto Encyclopedia of Genes and Genomes) database and particularly mTOR and VEGF signaling pathway caught our attention due to its close potential relation to pathogenesis of endometriosis (Figures
The figure shows signaling pathways (3: mTOR and 4: VEGF) with marked hypothetical influence of miRNA. The colors present expression of miRNAs assigned to a particular gene. The blue color stands for higher expression in control samples, and the red color stands for higher expression endometriosis. The color scale represents the magnitude of changes. The green arrows represent connections start from transcription factor and which hypothetical expression changes are consistent with hypothetical expression changes of genes regulated by this transcription factor.
The figure shows signaling pathways (3: mTOR and 4: VEGF) with marked hypothetical influence of miRNA. The colors present expression of miRNAs assigned to a particular gene. The blue color stands for higher expression in control samples, and the red color stands for higher expression endometriosis. The color scale represents the magnitude of changes. The green arrows represent connections start from transcription factor and which hypothetical expression changes are consistent with hypothetical expression changes of genes regulated by this transcription factor.
The significantly overexpressed GO (gene ontology) terms are presented in Table
Gene ontology (GO) terms overrepresented in eutopic endometrium of patients with endometriosis and control.
GO term | Adjusted |
Adjusted |
---|---|---|
Translational elongation | 1 | 0.077 |
|
||
Translational termination | 1 | 0.077 |
|
||
Cotranslational protein targeting to membrane | 1 | 0.077 |
|
||
SRP- (signal-recognition particle-) dependent cotranslational protein targeting to membrane | 1 | 0.077 |
|
||
Protein targeting to endoplasmic reticulum | 1 | 0.077 |
|
||
Protein localization to endoplasmic reticulum | 1 | 0.077 |
|
||
Establishment of protein localization to endoplasmic reticulum | 1 | 0.077 |
The first column contains names of significantly overrepresented GO terms. Both the second and the third columns contain adjusted
The SPIA showed perturbation of potentially nine pathways in two analyzed groups. Most interestingly natural killer cell mediated cytotoxicity pathway was found to be inhibited in eutopic endometrium of patients with endometriosis as compared with control (Table
Signaling Pathway Impact Analysis (SPIA) showing perturbation of signaling pathways in eutopic endometrium of patients with endometriosis versus control.
The name of the pathway | Adjusted |
Status |
---|---|---|
Alcoholism | 0 | A |
Olfactory_transduction | 0 | A |
Viral_carcinogenesis | 0 | I |
Systemic_lupus_erythematosus | 0.01 | A |
Chronic_myeloid_leukemia | 0.01 | A |
Cytosolic_DNA-sensing_pathway | 0.01 | A |
Natural_killer_cell_mediated_cytotoxicity | 0.01 | I |
RNA_transport | 0.03 | I |
Neurotrophin_signaling_pathway | 0.04 | A |
The first column shows names of significantly impacted pathways, the second column contains adjusted
Following the selection of miRNAs by fold change filtering (fold change > 1.3), we found that there were 136 upregulated miRNAs and no downregulated miRNAs in the eutopic endometrium of patients with advanced ovarian endometriosis compared with the eutopic endometrium.
We then validated 11 selected miRNAs but we were not able to observe clear statistical differences as to the expression of any of the chosen molecule (Table
Relative quantity (RQ) values of selected miRNA as validated by real-time PCR in patients with endometriosis (
Number | miRNA | Endometriosis |
Control |
|
Fold change difference between endometriosis and control by miRNA microarrays |
---|---|---|---|---|---|
1 | miR-4454 | 0.34 (0.07–2.07) | 0.41 (0.14–3.5) | 0.34 | 1.51 |
2 | miR-4714-5p | 0.63 (0.05–9.9) | 0.9 (0.01–4) | 0.9 | 1.86 |
3 | miR-5193 | 0.29 (0.04–11.3) | 0.51 (0.04–22.9) | 0.13 | 1.93 |
4 | miR-4284 | 1.88 (0.1–10.3) | 1.6 (0.2–15.7) | 0.83 | 2.28 |
5 | miR-5187-3p | 0.28 (0.05–3.2) | 0.20 (0.07–0.69) | 0.051 | 1.41 |
6 | miR-3680-5p | 0.8 (0.07–31.6) | 1.2 (0.06–11.2) | 0.96 | 2.21 |
7 | miR-3667-5p | 0.68 (0.06–18.19) | 1.8 (0.07–8.5) | 0.63 | 2.3 |
8 | miR-3152-5p | 0.17 (0.03–1.4) | 0.27 (0.002–2.8) | 0.048 | 1.44 |
9 | miR-23a-3p | 1.22 (0.02–11.3) | 1.17 (0.03–12.3) | 0.48 | 1.55 |
10 | miR-23b-3p | 0.4 (0.06–1.7) | 0.5 (0.1–2.2) | 0.44 | 2.08 |
11 | miR-30d-5p | 0.22 (0.05–21) | 0.42 (0.1–40) | 0.05 | 1.37 |
Although in our study we found more than 1000 miRNAs that are significantly differentially regulated between eutopic endometrium of patients with and without endometriosis, the differences, either up- or downregulated, were not higher than 2.6-fold in any case. The problem with obtaining clearly significant qPCR results may point to reservations that we may have as to the expression array results which can indirectly imply that some of the results might be false positives and that there are a number of variables that must also be considered. However, we must emphasize that although qPCR is the method of choice of confirming gene expression changes, the current technology contains a vital limitation related to experiments of this nature [
Our study showed little concordance with the endometriosis-associated miRNAs identified by two previous published studies, which also had minimal convergence with one another.
In the first miRNA eutopic versus normal endometriosis study, Pan et al. profiled the expression of 287 miRNAs in paired early-mid secretory eutopic and ectopic endometrium and isolated endometrial cells from women with stage III endometriosis and compared to control by the use of mirVana miRNA Bioarray. They also aimed to compare the expression of selected miRNA between isolated endometrial stromal cell (ESC) and glandular epithelial cell (GEC).
It was found that 65 of these miRNAs were identified to be expressed above the threshold levels set during the analysis in the endometrium of women without endometriosis. By the use of ANOVA it was also identified that there are 48 miRNAs which are differentially expressed between different combinations of eutopic and ectopic endometria. Specifically it was observed that there exists a progressive decline in miRNAs numbers from endometrium of women without endometriosis to eutopic endometrium, ectopic endometrium, and ectopic endometrium without paired eutopic tissue (miR23a and miR23b). It is somewhat confusing since it shows that there might be potential differences between ectopic endometrium of patients in whom eutopic endometrium was collected as compared with patients where this procedure was not performed. It was also shown that 32 miRNAs are differentially expressed in ESC and GEC, a significantly lower numbers when compared with endometrium of women without endometriosis. The obvious limitation of the above study was very limited sample size, a total of 16 samples including those with and without endometriosis as well as those without paired eutopic tissue, and restriction to only second phase of the cycle [
On the other hand, it was also more recently shown in advanced (III-IV) disease that miR23a and miR23b were significantly decreased in proliferative ectopic and eutopic endometrium compared with normal endometrium [
It was also shown that in the early secretory endometrium (ESE), by the use of LNA microarray (which consisted of Tm-normalized capture probes for 1488 distinct miRNAs), there are only 6 miRNAs which were differentially expressed with a fold change of >1.5 in the ESE from women with versus without endometriosis [
The above study is strengthened by the stringency of the surgical confirmation of presence or absence of disease and by the inclusion of only biopsy confirmed, moderate-severe (rAFS III-IV) stage endometriosis among the affected cohort. On the other hand, our study is strengthened by the use of microarray technology employing locked-nucleic acid (LNA) probes. Mature miRNAs are approximately 22 nucleotides in length, and this short length presents problems of specificity in their detection and localization. Microarray analysis using LNA probes allows very sensitive detection of these short-coding sequences [
Only very recently there have been published two studies which yielded quite different results. Shi et al. studied endometriotic, eutopic, and normal endometrial tissues in the proliferative phase of the cycle [
Second recent study concentrated on miRNA expression profile in relation to selected angiogenesis factors [
In our own previous study we used different technique, based on TaqMan miRNA microfluidic cards and validated by TaqMan real-time PCR, and we found in similar samples that out of 667 miRNAs there are eventually two, that is, hsa-miR-483-5p and hsa-miR-62
Our
In one of our studies we found that out of 15 different mTOR related genes including
On the other hand, the most interesting finding of our study comes from Signaling Pathway Impact Analysis which showed that out of nine potentially perturbed pathways in two analyzed groups it is natural killer cell mediated cytotoxicity which is inhibited in eutopic endometrium of patients with endometriosis. The inhibition of natural killer cells activity has been widely accepted [
Briefly the problem may be the structure of control group, which is limited to infertile patients and simple ovarian cysts. It is therefore difficult to evaluate whether control group is a healthy one alone since infertility and cysts themselves might to some degree influence expression of genes in endometrium. We also think that the result could also represent the heterogenous nature of whole endometrium tissue which is composed of different cellular compartments which create the potential for a large amount of biological variation. On the other hand, we intentionally decided not to perform laser capture microdissection (LCM) since there are several reasons, described in the discussion section of the article by Borghese et al. [
As we already discussed it previously in our publications [
In summary, we found that there are at least 136 miRNAs with differential expression fold change of at least 1.3 that could be involved in the pathogenesis of endometriosis. It seems however that current quantitative PCR methods are not always sensitive enough to detect differences in the expression of miRNA of less than 2.5. It is hoped that with advent of digital PCR it will be possible to more effectively detect small miRNA expression differences which may have significant impact on the discovery of potential biomarkers. We also found that potential several pathways, including mTOR, VEGF, natural killer cell cytotoxicity, and at least 6 potential cellular processes, involved mainly in protein synthesis, may be regulated by abnormal miRNAs expression. The confirmation of their potential role requires further functional studies.
We identified several miRNAs and potentially new pathways that may be abnormally regulated in eutopic endometrium of patients with endometriosis which may contribute to the pathogenesis of this debilitating disease.
The authors declare that they have no competing interests.
P. Laudanski had an idea that designated the study, participated in the laboratory work, analysed the data, and wrote the paper. R. Charkiewicz performed molecular biology experiments, especially RNA and protein isolation as well as PCR array experiments. A. Tolwinska helped in the interpretation and analysis of data as well as performing PCR array experiments and writing the paper. J. Szamatowicz helped in the collection of samples and molecular biology interpretation and analysis of data. A. Charkiewicz performed molecular biology experiments, especially RNA and protein isolation as well as PCR array experiments. J. Niklinski coordinated molecular biology experiments and analysed the data.
This study was supported by National Science Centre within the funds of Grant no. N N407 571938. The authors are grateful to Dr. Lukasz Mieczkowski for his professional help in statistical analysis.