This article is mentioned in:

Introduction

Due to the vast functional multiplicity, microRNAs (miRNA, miRs) shaped up as the most important class of small non-coding RNAs (ncRNAs) (1) playing a crucial role in post-transcriptional regulation of gene expression and signaling pathways. Mature miRNA is incorporated into the RNA-induced silencing complex (RISC) which usually leads to mRNA degradation and translational repression (2,3). Using this mechanism, miRNAs influence cellular processes of differentiation, proliferation and apoptosis. Due to their ability to participate in modulation of oncogenic pathways miRNAs are important key regulators of tumor suppressor genes and oncogenes (4–6). A widespread dysregulation of miRNAs is observed in a variety of human cancers (7–9).

miRNA molecules can be localized intracellularly or, after secretion, circulate in the extracellular compartment (10). Release of miRNA in the extracellular compartment can take place as passive leakage from cells or can include active export by apoptotic bodies, high-density lipoproteins, microvesicles, exosomes and protein complexes (11–13). Argonaute complexes, which are regular complexes of the RNA silencing machinery, and Nucleosphosmin-1 play a predominant role in protein-complex dependent secretion of miRNAs (14). In this context, miRNA packaging and extracellular export is mediated by Nucleophosmin-1, which also acts as a protector from RNA degradation (11,15). For exosomal miRNA secretion, Zhang et al reported trafficking and sorting mechanisms which play an important role in cell-cell communication processes (16).

Endometrial cancer (EC) is the most common gynecologic malignancy, accounting for 7% of all cancers in women in the United States (17). Incidence as well as death rate of EC has increased during recent years (18). EC can be subclassified into two categories: type I cancers are usually well to moderately differentiated, estrogen-related endometrial carcinomas (EEC) with good prognosis, which are associated with endometrial hyperplasia, whereas type II cancers are non-estrogenic high grade serous or clear cell carcinomas with poor prognosis, which are associated with atrophic endometrium. Molecular characteristics of EC cells do not always correlate with the traditional classification systems which have been based on clinical features or on histopathological findings (19–21).

MiRNAs have been shown to participate in regulation of endometrial growth and differentiation (22–26) and carcinogenesis of the endometrium has been associated with alterations in miRNA expression pattern. Furthermore, distinct miRNA signatures characterizing EC subtypes have been described (24,27–30). Since miRNAs have been shown to be stable molecules which are well preserved in serum and other body fluids (21,31,32) their use as promising non-invasive biomarkers in EC is under investigation (21). Jia et al reported in 2013 the identification of four serum-based miRNA types as potential non-invasive biomarkers for endometrioid endometrial cancer (32).

Alterations in miRNA expression patterns were also found in tumor cells exposed to stressors such as extracellular hypoxia and acidosis (33,34). However, there are no reports on miRNA expression under hypoxic or acidic conditions in EC cells. To our knowledge, to date there are only few studies investigating miRNA expression aberrations under the influence of hypoxia in endometriosis (35–37).

Hypoxia is an important regulatory factor in tumor growth and induces transcriptional cascades that promote a more aggressive tumor phenotype. In order to survive and proliferate in a hypoxic microenvironment tumor cells undergo genetic and adaptive changes and release substances that affect the microenvironment to promote tumor angiogenesis (38,39). These processes contribute to a malignant phenotype and aggressive tumor behavior as well as metastasis (38).

Kulshreshtha et al suggested that hypoxia may represent a key contributing stress factor for microRNA alterations in cancer. In colon and breast cancer cell lines, expression of specific miRNAs was altered under hypoxia (40). Hua et al described a miRNA directed regulation of VEGF and other angiogenetic factors under hypoxia in cells from a human nasopharyngeal carcinoma cell line (41).

Acidosis of the extracellular microenvironment is a consequence of increased glycolysis due to the adaption of tumor cells to prolonged periods of hypoxia. Cancer cells suffering oxygen deprivation use the glycolytic pathways to maintain their ATP level for survival and proliferation. Acidification of the peritumoral microenvironment induces necrosis and apoptosis of normal cells, thus creating space into which the tumor cells may proliferate (42). Acidosis promotes tumor invasion by degradation of the extracellular matrix through the discharge of proteolytic enzymes (43). Lowered pH levels also inhibit natural killer cell activity and the cytolytic activity of cytotoxic T-lymphocytes, therefore causing a diminished immune response to tumor antigens (44). This way, acidosis facilitates tumor progression and the development of invasive phenotypes. Furthermore, acidosis and hypoxia have been associated with resistance to therapeutic strategies such as multi-drug resistance and poor prognosis (45).

The present study investigated the quantitative expression of 24 different miRNAs secreted by EC cells in vitro in response to the exogenic stimuli hypoxia and acidosis. The panel of 24 specific miRNA types (let-7a, let-7b, let-7d, let-7i, miR-10a, −10b, −15a, −15b, −17, −19b, −20a, −20b, −21, −22, −26.1, 27a, −29c, −30b, −92a, −125b, −128-1, −135b, −200c, −222) was collocated based on a targeted pre-selection process focusing on EC-associated circulating miRNAs with tumor biological, diagnostic and therapeutic relevance. An initial screening expression analysis attested that all 24 secreted miRNAs were stably detectable in the supernatant of the selected three EC in vitro models (Ishikawa, EFE 184, AN3-CA; see Materials and methods for details). The current data on the tumor biological relevance of the selected miRNA types is summarized in Table I (1,25,26,28–30,41,46–75).

Table I. Tumor-biological relevance of pre-selected miRNA types.

Table I.

Tumor-biological relevance of pre-selected miRNA types.

Pre-selected miR	Putative target	Reported function	References
Let-7	BAX	Enhanced survival and proliferation of cancer cells (promoting EC)	Zhang et al (70)
Let-7a	Aurora B	Inhibition of EC growth	Liu et al (57)
Let-7b	High mobility group AT-hook2	Inhibition of aggressive phenotypes	Romero-Perez et al (64)
Let-7d	LIN28, C-MYC, K-RAS, HMGA2 and IMP-1	Tumor suppressor or oncogene	Kolenda et al (54)
Let-7i	RAS, HMGA2, c-Myc, CDC25A, CDK6 and cyclin D2	Tumor suppressor	Yang et al (94)
miR-10a	USF2, HOXA1, HOXD10, HOXB1, HOXB3, RB1CC1	Invasion, metastasis (promoting EC)	Dai et al (50)
miR-10b	TBX5m DYRK1A, PTEN	Oncomir, Migration, invasion, proliferation	Kim et al (53)
miR-15a	Bck, MCL1, CCND1, WNT3	Tumor suppressor	Aqeilan et al, Bonci et al (46,47)
miR-15b	VEGF, CCND1, CCNE1	Tumor suppressor, antiangiogenic	Zhao et al (71)
miR-17	VEGF-A, NOR-1, GALNT3	Antiangiogenic	Doebele et al, Hua et al, Lu et al, Ramon et al (41,51,63,95)
miR-19b	PTEN, TGFβ	Oncogene component of mir-17-92-Cluster	Doebele et al, Lu et al, Ramon et al, Fuziwara et al (51,63,73,95)
miR-20a	VEGF-A	Reduces angiogenesis in EC	Doebele et al, Lu et al, Ramon et al (51,63,95)
miR-20b	MMP-2	Tumor suppressor	Park et al (59)
miR-21	Maspin, Pdcd4, PTEN	Tumorigenesis Cell proliferation (promoting EC)	Qin et al, Torres et al (30,62)
miR-22	Cyclin D1, MMP2, MM9 Erα	Tumor suppressor	Li et al, Wang et al (56,69)
miR-26a-1	CCNE1, ERα, FGF9, MTDH, EZH2, MCL-1	Antiproliferation, proapoptotic, metastasis inhibition, growth inhibition	Chen et al (49)
miR-27a	FOXO1 BAX	Myometrial invasion Inhibition of apoptosis, increased survival of cancer cells (promoting EC)	Zhang et al, Mozos et al, Myatt et al (70,22,23)
miR-29c	SPARCS (in EC) PI3K, CD42	Tumor suppressor	Castilla et al, Park et al (24,25)
miR-30b	SIX1	Inhibits migration and invasion	Zhao et al (72)
miR-92a	PTEN/PI3K/Akt/mTOR	Member of miR-17-92 cluster, promotes carcinogenesis in endometrial epithelium, oncogene	Torres et al (67)
miR-125b	TP53INP1	Proliferation and migration of cancer cells (promoting EC)	Jiang et al, Shang et al (28,29)
	ERBB2	Inhibition of cancer cell invasion (protecting against EC)
miR-128-1	BMI1, E2F3	Tumor suppressor	Shan et al (65)
miR-135b	VEGF A, HIF1 A, HIFAN	Tumor promoting Upregulated in EC	Tsukamoto et al (68)
miR-200c	TIMP2 RD7 ZEBs, VEGF-A, FLT1, IKKβ, KLF9, FBLN5	Promoting EC: Metastasis of cancer cells Induction of cell survival and proliferation Reduction of apoptosis Transition into cancerous states Protective:	Li et al, Park et al, Snowdon et al (32,33,34)
	FN1, MSN, NTRK2, LEPR, ARHGAP19	Downregulation promotes EMT phenotype and aggressive behavior of cells
miR-222	c-kit CD117, PBX3	Antiangiogenic activity	Ramon et al, Poliseno et al, Yanokura et al (14,35,36)

Materials and methods

Cell culture conditions and treatments

Established endometrial cancer cell lines Ishikawa i) (type I, well differentiated, ER and PR positive) (76). AN3-CA ii) (type II, poorly differentiated, ER negative) and EFE-184 iii) (type I) were incubated in humidified atmosphere at 37°C and 5% CO2 in i) RPMI-1640 or (ii and iii) DMEM/F12 medium supplemented with 10% newborn calf serum (Gibco™, Thermo Fisher Scientific, Karlsruhe, Germany), 1% HEPES buffer (Gibco) and 100 U/ml penicillin/streptomycin (Sigma-Aldrich®, Taufkirchen, Germany).

For the induction of stress conditions hypoxia and acidosis cells were subcultured in 25 cm2 cell culture flasks and grown for 18 h. Hypoxia was induced by incubating cells in hypoxic chamber (<3% CO2) for a period of 18 h overnight. For acidosis exposure, cells were treated with 0.2% lactic acid in cell culture media and incubated for 48 h at 37°C. Cells incubated under standard conditions in parallel served as controls. Following treatment, 2 ml of culture media (supernatant) were harvested and immediately processed. All experiments were performed in triplicates.

Total RNA isolation

Total RNAs from supernatant of cultured cells were isolated by using the innuPREP Micro RNA Kit (Analytic Jena AG, Jena, Germany) according to the manufacturer's protocol. Assessment of RNA quantity was done by UV-spectrometry (NanoDrop ND1000 (PEQLAB, Erlangen, Germany). Until further processing RNA samples were stored at −80°C.

Reverse transcription (RT) and quantitative PCR

The reaction mixture of RT consists of 4 µl Maxima RT-buffer (Thermo), 1 µl 5 µM poly(T) adaptor primer (Biomers, Konstanz, Germany), 1 µl 5 mM dNTPs (Jena Bioscience, Jena, Germany), 0.25 µl Maxima reverse transcriptase (Thermo), 0.25 µl SUPERase in RNase inhibitor (Thermo) and 500 ng of the total RNA sample. The reaction was carried out at 42°C for 30 min and by 85°C for 10 min in a Nexus Thermal Cycler (Eppendorf, Hamburg, Germany). Until further analysis, processed cDNA was stored at −20°C.

Relative expression levels of specific microRNAs were assessed by quantitative PCR applying a SYBR Green assay in a duplicate analysis. cDNA (1 µl) with concentration of 5 ng/µl was mixed up with 9 µl master mix containing 1 µl 10 Xq PCR buffer, 0.5 µl 5 mM dNTPs (Jena Bioscience), 0.5 µl miRNA specific qPCR primer (Biomers), 0.5 µl SYBR Green (Roche Diagnostics GmbH, Mannheim Germany), 0.05 µl HotStart Taq (Jena Bioscience) and 6.54 µl nuclease-free water (Analytic Jena). Primers consisted of a specific primer pair for each miRNA type and for the housekeeping genes.

Quantitative PCRs were performed on a LightCycler96 (Roche Diagnostics GmbH) at 95°C for 2 min, followed by 40 cycles at 95°C, for 5 sec and for 30 sec at 60°C.

Data were analyzed in Microsoft Excel using ∆Ct method based on reference value (geometric mean of housekeeping genes, RNU48, miR-26b, miR-16 and miR-103 defined via ‘BestKeeper’ software tool (77) to determine single microRNA expression levels.

Pre-selection of miRNA specimen

A comprehensive survey was performed applying Pubmed interface for original research (http://www.ncbi.nlm.nih.gov/) and the miRTarBase platform (http://mirtarbase.mbc.nctu.edu.tw/) (78), specific miRNA information registry, to determine potential study-relevant miRNA types. These sources were screened for relevant specimen in regard to ‘endometrial cancer’, typical endometrial cancer target genes and pathways as well as literature. A methodological pre-screening of >40 miRNA types was performed, which was based on the acquired information on potential miRNA candidates, to evaluate the experimental in vitro setting focusing on secreted miRNA specimen in cell culture media. The pre-set of miRNAs was filtered to collocate the final set of miRNAs which were reliably detectable in culture media. The hereby identified miRNA types and their functional implications are summarized in Table I.

Statistical analysis

Observed miRNA expression levels were graphically visualized in box plots, showing median and quartiles in cell lines Ishikawa, EFE-184 and AN3-CA under hypoxia, acidosis and normoxic conditions. The influence of treatment on miRNA expression levels of the different BC cell lines was investigated using a linear model with factor treatment (hypoxia and acidosis compared to normoxic conditions) and cell line, including an interaction term.

Results

In the present study we investigated the regulatory effect of hypoxia and acidosis on the expression levels of 24 different EC associated microRNA types (let-7a, let-7b, let-7d, let-7i, miR-10a, −10b, −15a, −15b, −17, −19b, −20a, −20b, −21, −22, −26-1, 27a, −29c, −30b, −92a, −125b, −128-1, −135b, −200c, −222) that are actively secreted into the tissue culture supernatant by EC cells. The characteristics of the investigated EC cell lines Ishikawa, EFE-184 and AN3-CA are summarized in Table II.

Table II. Molecular classification of investigated endometrial cancer cell lines.

Table II.

Molecular classification of investigated endometrial cancer cell lines.

Cell line	Classification	Immunoprofile	Primary tumor	Origin
Ishikawa	Type I	ER+, PR+,	Endometrial adenocarcinoma	Uterus
EFE-184	Type I	ER+, PR−	Endometrium carcinoma	Ascitic fluid
AN3-CA	Type II	ER−, PR+	Endometrial adenocarcinoma	Lymph node metastasis

[i] ER, estrogen receptor; PR, progesterone receptor.

All 24 examined microRNAs were stably detectable in the supernatants of all three in vitro models. In order to examine potential regulatory effects triggered by typical microenvironmental alterations occurring in solid tumors during tumor growth, progression and metastasis on miRNA expression (79–82) hypoxia and acidosis were induced in the three cell lines. Expression levels of all EC associated miRNA types were determined under normoxic, hypoxic and acidotic conditions as mean ∆Ct values of the distinct miRNA type normalized against the geometric mean of the four housekeepers RNU48, miR-16, miR-26b and miR-103.

Under standard culture conditions each of the EC cell lines displayed a unique miRNA expression profile (Fig. 1). In response to hypoxia and acidosis, singular marked alterations in secreted miRNA expression levels were identified predominantly in Ishikawa cells.

Figure 1. Expression profiles of secreted miRNAs in Ishikawa, EFE-184 and AN3-CA cells under normoxic (control) conditions. miRNA-expressions were determined in triplicates by real-time quantitative PCR and calculated using ∆Ct method based on reference value (geometric mean of housekeeping genes, RNU48, miR-26b, miR-16 and miR-103) to determine single microRNA's expression levels.

Hypoxia triggered regulatory effects on the expression levels of miR-15a, miR-20a, miR-20b, miR-21 and miR-128-1. In Ishikawa cells, multivariable regression analysis identified a significant hypoxia-dependent downregulation of miR-15a, miR-20a, miR-20b and miR-128-1 expression. In detail, the expression level of miR-15a was decreased by a mean value of 0.020 (CI: −0.038 - −0.002; p=0.048) (Table III, Fig. 2), of miR-20a by 3.577 (CI: −6.663 - −0.491; p=0.036) (Table IV, Fig. 2), miR-20b by 0.340 (−0.544 - −0.136; p=0.004) (Table V, Fig. 2) and of miR-128-1 by 0.073 (−0.136 - −0.011; p=0.034) (Table VI, Fig. 2). In EFE-184 and AN3-CA cells a comparable hypoxia-driven downregulatory effect on miRNA expression could not be detected. In EFE-184 cells a hypoxia-induced upregulation of miR-21 expression was observed. The expression levels of secreted miR-21 levels in hypoxia-conditioned EFE-184 cells were increased by a mean value of 2.103 (CI: 0.300 - 3.907; p=0.035) (Table VII, Fig. 2). A hypoxia-dependent upregulatory effect on miRNA expression of AN3-CA and Ishikawa cells was not observed.

Figure 2. Box plot diagram of expression levels of significantly altered secreted miRNAs in response to hypoxia. Expression levels of secreted miR-15a, miR20a, miR-20b, miR-21 and miR-128-1 in endometrial cancer cell lines are displayed under varying treatment conditions. Levels of secreted miRNA expression were determined in AN3-CA, EFE-184 and Ishikawa cells under control conditions (N), extracellular acidosis (Ac) and hypoxia (Hx). Box plots demonstrate median (thick black line), lower and upper quantile range (box lines), and standard deviation range (dashed lines bounded by horizontal lines). Significant expression level alterations (*p≤0.05) of control conditions versus treatment options. Based on triplicate experiments, real-time quantitative PCR.

Table III. Two-way analysis of variance for levels of secreted miR-15a under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

Table III.

Two-way analysis of variance for levels of secreted miR-15a under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

	Estimate	Standard error	95% CI	p-value
Intercept	0.047	0.007	0.034, 0.060	0.000
Control
EFE-184	−0.033	0.009	−0.052, −0.015	0.002
Ishikawa	0.023	0.009	−0.005, 0.032	0.174
Acidosis
AN3-CA	−0.007	0.009	−0.025, 0.012	0.489
EFE-184	0.000	0.009	−0.018, 0.018	1.000
Ishikawa	−0.007	0.009	−0.025, 0.012	0.489
Hypoxia
AN3-CA	0.000	0.009	−0.018, 0.018	1.000
EFE-184	0.003	0.009	−0.015, 0.022	0.728
Ishikawa	−0.020	0.009	−0.038, −0.002	0.048

[i] Intercept is calculated on basis of AN3-CA control.

Table IV. Two-way analysis of variance for levels of secreted miR-20a under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

Table IV.

Two-way analysis of variance for levels of secreted miR-20a under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

	Estimate	Standard error	95% CI	p-value
Intercept	6.573	1.113	4.391, 8.756	0.000
Control
EFE-184	−1.657	1.575	−4.743, 1.429	0.307
Ishikawa	5.797	1.575	2.711, 8.883	0.002
Acidosis
AN3-CA	−1.453	1.575	−4.539, 1.633	0.368
EFE-184	−0.660	1.575	−3.746, 2.426	0.680
Ishikawa	−1.993	1.575	−5.079, 1.093	0.222
Hypoxia
AN3-CA	−0.177	1.575	−3.263, 2.909	0.912
EFE-184	0.007	1.575	−3.079, 3.093	0.997
Ishikawa	−3.577	1.575	−6.663, −0.491	0.036

[i] Intercept is calculated on basis of AN3-CA control.

Table V. Two-way analysis of variance for levels of secreted miR-20b under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

Table V.

Two-way analysis of variance for levels of secreted miR-20b under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

	Estimate	Standard error	95% CI	p-value
Intercept	0.487	0.073	0.343, 0.631	0.000
Control
EFE-184	−0.097	0.104	−0.300, 0.107	0.365
Ishikawa	0.527	0.104	0.323, 0.730	0.000
Acidosis
AN3-CA	−0.083	0.104	−0.287, 0.120	0.433
EFE-184	−0.097	0.104	−0.300, 0.107	0.365
Ishikawa	−0.207	0.104	−0.410, −0.003	0.062
Hypoxia
AN3-CA	−0.003	0.104	−0.207, 0.200	0.975
EFE-184	−0.010	0.104	−0.214, 0.194	0.924
Ishikawa	−0.340	0.104	−0.544, −0.136	0.004

[i] Intercept is calculated on basis of AN3-CA control.

Table VI. Two-way analysis of variance for levels of secreted miR-128-1 under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

Table VI.

Two-way analysis of variance for levels of secreted miR-128-1 under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

	Estimate	Standard error	95% CI	p-value
Intercept	0.230	0.023	0.186, 0.274	0.000
Control
EFE-184	−0.110	0.032	−0.173, −0.047	0.003
Ishikawa	0.090	0.032	0.027, 0.153	0.011
Acidosis
AN3-CA	−0.017	0.032	−0.079, 0.046	0.608
EFE-184	−0.007	0.032	−0.069, 0.056	0.837
Ishikawa	−0.033	0.032	−0.096, 0.029	0.310
Hypoxia
AN3-CA	0.000	0.032	−0.063, 0.063	1.000
EFE-184	−0.010	0.032	−0.073, 0.053	0.757
Ishikawa	−0.073	0.032	−0.136, −0.011	0.034

[i] Intercept is calculated on basis of AN3-CA control.

Table VII. Two-way analysis of variance for levels of secreted miR-21 under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

Table VII.

Two-way analysis of variance for levels of secreted miR-21 under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

	Estimate	Standard error	95% CI	p-value
Intercept	2.853	0.651	1.578 4.128	0.000
Control
EFE-184	11.013	0.920	9.210 12.817	0.000
Ishikawa	−1.253	0.920	−3.057 0.550	0.190
Acidosis
AN3-CA	−0.377	0.920	−2.180 1.427	0.687
EFE-184	−0.167	0.920	−1.970 1.637	0.858
Ishikawa	0.510	0.920	−1.293 2.313	0.586
Hypoxia
AN3-CA	−0.177	0.920	−1.980 1.627	0.850
EFE-184	2.103	0.920	0.300 3.907	0.035
Ishikawa	1.330	0.920	−0.473 3.133	0.165

[i] Intercept is calculated on basis of AN3-CA control.

In response to acidosis, single marked alterations in secreted expression for let-7a, miR-22 and miR-125b were observed. In Ishikawa cells, ∆Ct expression levels of let-7a were increased by a mean value of 1.980 (−0.442 - 3.518; p=0.021) (Table VIII, Fig. 3) and ∆Ct expression levels of miR-22 were decreased by a mean value of 0.023 (−0.044 - −0.003; p=0.040) (Table IX, Fig. 3). In AN3-CA cells a specific acidosis-dependent upregulation was observed for miR-125b. In detail, a mean increase of secreted miR-125b ∆Ct level by 1.007 (CI: 0.512 - 1.501; p=0.001) became evident (Table X, Fig. 3). No significant acidosis-dependent miRNA expression level alterations were detected in EFE-184 cells.

Figure 3. Box plot diagram of expression levels of significantly altered secreted miRNAs in response to acidosis. Expression levels of secreted let-7a, miR22 and miR-125b in endometrial cancer cell lines are displayed under varying treatment conditions. Levels of secreted miRNA expression were determined in AN3-CA, EFE-184 and Ishikawa cells under control conditions (N), extracellular acidosis (Ac) and hypoxia (Hx). Box plots demonstrate median (thick black line), lower and upper quantile range (box lines), and standard deviation range (dashed lines bounded by horizontal lines). Significant expression level alterations (*p≤0.05) of control conditions versus treatment options. Based on triplicate experiments, real-time quantitative PCR.

Table VIII. Two-way analysis of variance for levels of secreted let7a under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

Table VIII.

Two-way analysis of variance for levels of secreted let7a under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

	Estimate	Standard error	95% CI	p-value
Intercept	5.243	0.555	4.156, 6.331	0.000
Control
EFE-184	−2.020	0.785	−3.558, −0.482	0.019
Ishikawa	−0.917	0.785	−2.454, 0.621	0.258
Acidosis
AN3-CA	−0.737	0.785	−0.801, 2.274	0.360
EFE-184	1.180	0.785	−0.358, 2.718	0.150
Ishikawa	1.980	0.785	0.442, 3.518	0.021
Hypoxia
AN3-CA	−1.183	0.785	−2.721, 0.354	0.149
EFE-184	0.427	0.785	−1.111, 1.964	0.593
Ishikawa	1.140	0.785	−0.398, 2.678	0.163

[i] Intercept is calculated on basis of AN3-CA control.

Table IX. Two-way analysis of variance for levels of secreted miR-22 under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

Table IX.

Two-way analysis of variance for levels of secreted miR-22 under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

	Estimate	Standard error	95% CI	p-value
Intercept	0.053	0.007	0.039, 0.068	0.000
Control
EFE-184	0.020	0.011	−0.001, 0.041	0.074
Ishikawa	0.033	0.011	0.013, 0.054	0.005
Acidosis
AN3-CA	−0.003	0.011	−0.024, 0.017	0.755
EFE-184	0.007	0.011	−0.014, 0.027	0.535
Ishikawa	−0.023	0.011	−0.044, −0.003	0.040
Hypoxia
AN3-CA	0.003	0.011	−0.017, 0.024	0.755
EFE-184	−0.003	0.011	−0.024, 0.017	0.755
Ishikawa	−0.010	0.011	−0.031, 0.011	0.355

[i] Intercept is calculated on basis of AN3-CA control.

Table X. Two-way analysis of variance for levels of secreted miR-125b under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

Table X.

Two-way analysis of variance for levels of secreted miR-125b under control, acidotic and hypoxic conditions in three different endometrial cancer cell lines.

	Estimate	Standard error	95% CI	p-value
Intercept	3.893	0.178	3.544, 4.243	0.000
Control
EFE-184	−3.830	0.252	−4.324, −3.336	0.000
Ishikawa	−3.783	0.252	−4.278, −3.289	0.000
Acidosis
AN3-CA	1.007	0.252	0.512, 1.501	0.001
EFE-184	0.027	0.252	−0.468, 0.521	0.917
Ishikawa	0.000	0.252	−0.494, 0.494	1.000
Hypoxia
AN3-CA	0.153	0.252	−0.341, 0.648	0.551
EFE-184	0.003	0.252	−0.491, 0.498	0.990
Ishikawa	−0.010	0.252	−0.504, 0.484	0.969

[i] Intercept is calculated on basis of AN3-CA control.

No marked hypoxia- or acidosis-driven alterations in expression levels of let-7b, let-7d, let-7i, miR-10a, −10b, −15b, −17, −19b, −26-1, 27a, −29c, −30b, −92a, −135b, −200c and 222 were observed in the investigated EC cell lines (data not shown).

Discussion

Recent and ongoing studies account for a continually growing number of miRNA types with regulatory influence on EC tumorigenesis. While some of these miRNAs play an important role as tumor suppressors, expression of others is associated with promotion of tumor growth, invasion and metastasis (oncogenic miRNAs, oncomiRs). In the present study, we analyzed the expression of a panel of secreted EC associated microRNAs with impact on angiogenesis, proliferation, invasion, migration and metastasis including let-7a, let-7b, let-7d, let-7i, miR-10a, −10b, −15a, −15b, −17, −19b, −20a, −20b, −21, −22, −26.1, 27a, −29c, −30b, −92a, −125b, −128-1, −135b, −200c, −222. We were able to reliably detect all 24 secreted miRNA types in the supernatants of the three analyzed EC cell lines. Based on these EC in vitro models a quantitative analysis of secreted miRNA expression levels was employed to detect potential alterations under the typical tumor biological epiphenomena hypoxia and acidosis.

Our data clearly account for a hypoxia- and acidosis-dependent regulatory effect in the expression levels of eight specific secreted miRNAs (let-7a, miR-15a, miR-20a, miR-20b, miR-21, miR-22, miR-125b and miR-128-1) among the investigated miRNA types. The observed alterations in miRNA expression levels were not consistent in all three EC cell lines. Significant miRNA deregulation due to hypoxia or acidosis was predominantly observed in Ishikawa cells. In EFE-184 and AN3-CA cells a regulatory effect was observed solely for one specific miRNA type. Hypoxia resulted in downregulation of miR-15a, miR-20a, miR-20b and miR-128-1 in Ishikawa cells and upregulation of miR-21 in EFE-184 cells.

miR-15a is considered to trigger tumor suppressive effects by inhibiting several target oncogenes, like BCL2, MCL1, CCND1 and WNT3A (46,47). Thus, cells react with diminished proliferation, an increase in apoptosis rates and suppressed tumorigenicity both in vitro and in vivo. In chronic lymphocytic lymphoma (CLL), prostate carcinomas and pituitary adenomas, downregulation of miR-15a expression has been observed (46). In the present study, hypoxic conditions resulted in decreased miR15a expression in Ishikawa cells. This finding accounts for a reduced inhibitory signaling function on angiogenesis in type I EC cells under hypoxic stress and is supported by the results of Hua et al reported that in a tumor cell model VEGF expression is repressed, besides others, by miRNAs (miR15b, miR16) from the miR15-16 clustered family (41).

The miR-17-92 cluster, which contains six mature miRNAs (miR-17, −18a, −19a, −19b, −20a and −92a), conducts a complex role in angiogenetic signaling cascades (65) In the present analysis, we evaluated four specific miRNAs of this cluster: miR-17, miR-19b, miR-20a and miR-92a. Anti-angiogenetic activity of miR-20a by targeting VEGF-A has been reported (41,65). Inconsistent results have been reported for the role of miR-17 so far. While some investigatory approaches hypothesize an anti-angiogenetic activity (41,51), other studies account for either pro-angiogenetic activity (83,84) or no association with VEGF-A expression (65). Our data identified a significant hypoxia-driven expression reduction of miR20a in Ishikawa cells. For miR-17, miR-19b and miR-92a no significant changes in expression levels were observed. Our results are in line with the findings of Ramon et al and support the proposition that miR-20a, but not miR-17 and miR-19b, displays an anti-angiogenetic activity, which is reduced under hypoxic stress conditions in EC tumor cells (65).

MiR-20b has a tumor suppressive function by impeding MMP-2 expression leading to cell cycle arrest (61) as well as regulative function on oxygen balance of cells. Lei et al described that tumor cells are able to adapt to alterations of oxygen concentration by miR-20b triggered regulation of HIF-1α and VEGF. In their study inhibition of miR-20b increased protein levels of VEGF and HIF-1α in normoxic tumor cells, an increase of miR-20b expression in hypoxic tumor cells resulted in reduced protein levels of VEGF and HIF-1α (85). Park et al showed that miR-20b expression was downregulated in bladder cancer and an upregulation of this miRNA type lead to inhibition of proliferation, migration and invasion (61). In gastric cancer, downregulated miR-20b expression was associated with hypoxia-induced chemoresistance. In the present study we observed reduced miR-20b expression levels in Ishikawa cells under hypoxic conditions. These data are in line with aforementioned reports of Lei et al as well as of Danza et al in gastric cancer (85,86).

Under hypoxic conditions expression levels of miR-128-1 were also found reduced in Ishikawa cells. miR-128-1 is supposed to act as a tumor suppressor inhibiting tumor cell proliferation, invasion and self-renewal by direct targeting of BMI1 and E2F3 in glioblastomas (67). Furthermore, it has been associated with inhibition of the mTOR signaling pathway in glioblastomas. Decreased miR-128-1 expression levels in glioblastomas could be strongly associated with poor survival (87). The observed reduced expression levels of miR-128-1 in Ishikawa cells under hypoxic conditions may also account for a regulative function of miR-128-1 in EC.

MiR-21 expression levels were increased in EFE-184 cells under oxygen deprivation. This observation is in line with previous studies showing that miR-21 is an active participant in EEC malignant transformation by abrogating the regulatory function of PTEN, which has been found to be upregulated in EEC (24,64). Furthermore, a regulating role of miR-21 on the expression of the tumor suppressor gene Pdcd4 (programmed cell death-4) and Maspin, which is strongly associated with angiogenesis in EC, has been indicated (88). Whether Maspin acts as tumor suppressor or tumor promoter, seems to be complex and dependent on the tissue specific environment (89). In EC high Maspin expression has been associated with higher FIGO stage, lymph node involvement and the depth of myometrial invasion as well as poor prognosis (90). Additionally a correlation between high miR-21 expression and increased Maspin expression has been reported in EC (89).

Acidosis resulted in divergent miRNA expression patterns in the analyzed EC cell lines. While let-7a expression was found to be upregulated, miR-22 expression levels were downregulated under acidotic conditions in Ishikawa cells. Let-7a acts as tumor suppressor due to inhibition of EC growth by targeting and downregulating Aurora B (1, 21, 24). In EC let-7a expression levels are usually downregulated (58). miR-22 exerts tumor suppressive functions by influencing Cyclin D1, MMP2, PTEN and others (71). Li et al reported an inhibitory function of miR-22 expression on proliferation and invasion in estrogen receptor α-positive endometrial endometrioid carcinoma cells (57).

In AN3-CA cells miR-125b was significantly upregulated under acidotic conditions. These data are in line with previous findings showing an upregulated miR-125b expression in type II EC cells. Proliferation and migration of type II EC cancer cells was promoted by miR-125b in this study by targeting the tumor suppressor gene TP53INP1 (53).

In summary, our data clearly identify a hypoxia-dependent downregulation of secreted miRNAs with tumor suppressive and anti-angiogenetic function (miR-15a, miR-20a, miR-20b, miR-128-1) as well as upregulation of secreted miRNAs with tumor and angiogenesis promoting function (miR-21) in type I EC cell lines (Ishikawa, EFE-184). In contrast, in the analyzed type II EC cell line (AN3-CA) hypoxia did not show any significant impact on the expression pattern of the analyzed EC related miRNAs.

Acidosis triggered upregulation of the tumor promoting miRNA miR-125b in AN3-CA cells (type II EC). In Ishikawa cells (type I EC) the tumor promoting or tumor suppressive impact of acidosis is not clear yet since miRNAs with tumor suppressive function were found altered in divergent directions, both up- (let-7a) and down- (miR-22) regulated.

In conclusion, our current findings emphasize the functional importance of secreted miRNAs in the immediate response of EC cells to exogenic stress situations such as the typical tumor epiphenomena hypoxia and acidosis. Focusing on the specific potential of secreted, thus circulating miRNA molecules, that also conduct an inter-cellular (hormone-like) signaling function, alterations in expression levels not only influence intracellular gene expression and signaling cascades, but also transfer the induction of (tumor)biological cellular changes to adjacent cells (91,92) Thus, the fate of whole tissue areas may undergo transformation in respect to metabolic pathways, cell cycle control and neo-angiogenesis, that support malignant progression. To rule out the possibility of in vitro/in vivo differences of the analyzed secreted miRNA expression levels as well as to support the transferability of the observed results to in vivo settings, the data of the present in vitro study on EC cancer cell lines should be confirmed by corresponding in vivo analyses based on tumor tissue specimen and/or body fluids (e.g. blood, urine) of EC patients.

Continuous elucidation of miRNA functions augments the potential of these regulatory nucleic acid elements to either be implemented as useful biomarkers in disease/malignancy diagnosis, prognosis and therapy monitoring, as well as novel therapeutic targets (‘AntagomiRs’) in clinical cancer management (1,21,93). The potential applicability of circulating miRNAs as non-invasive biomarkers for diagnosis (32) or as therapeutic targets (94,95) in EC is currently subject of various studies. In this context this in vitro model may serve as initial step to elucidate the influence of microenvironmental changes on expression profiles of circulating microRNAs in EC.

References