- Email: [email protected]
Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures
YGYNO-976169; No. of pages: 6; 4C: Gynecologic Oncology xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures Yeheli Ravid a, Malka Formanski a, Yoav Smith b, Reuven Reich a,⁎,1, Ben Davidson c,d,⁎⁎ a
Institute of Drug Research, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem 91120, Israel Genomic Data Analysis Unit, Hadassah Medical School, The Hebrew University of Jerusalem, Jerusalem 91120, Israel Department of Pathology, Norwegian Radium Hospital, Oslo University Hospital, N-0310 Oslo, Norway d The Medical Faculty, University of Oslo, N-0316 Oslo, Norway b c
H I G H L I G H T S • microRNA array analysis differentiates uterine endometrial stromal sarcoma from leiomyosarcoma. • microRNAs are predominantly underexpressed in metastatic compared to primary leiomyosarcoma. • Frizzled-6 silencing suppresses invasion in leiomyosarcoma cells in vitro.
a r t i c l e
i n f o
Article history: Received 14 October 2015 Received in revised form 1 January 2016 Accepted 2 January 2016 Available online xxxx Keywords: Endometrial stromal sarcoma Leiomyosarcoma Differential diagnosis Tumor progression MicroRNA
a b s t r a c t Objective. To compare the microRNA (miRNA) profiles of uterine endometrial stromal sarcoma (ESS) and leiomyosarcoma (LMS), and to compare the miRNA signatures of primary and metastatic uterine LMS. Methods. Eight primary LMS, 9 primary ESS and 8 metastatic LMS were analyzed for miRNA profiles using TaqMan Human miRNA Array Cards. Findings for 20 differentially expressed miRNAs were validated in a series of 44 uterine sarcomas (9 primary uterine ESS, 17 primary uterine LMS, 18 metastatic LMS) using qPCR. Frizzled-6 protein expression was analyzed in 30 LMS (15 primary, 15 metastases). Frizzled-6 was silenced in SK-LMS-1 uterine LMS cells using siRNA and the effect on invasion, wound healing and matrix metalloproteinase-2 (MMP2) activity was assessed. Results. Ninety-four miRNAs were significantly differentially expressed in ESS and LMS, of which 76 were overexpressed in ESS and 18 overexpressed in LMS. Forty-nine miRNAs were differentially expressed in primary and metastatic LMS, of which 45 were overexpressed in primary LMS and 4 in metastases. Differential expression was confirmed for 10/20 miRNA analyzed using qPCR. Frizzled-6 silencing in SK-LMS-1 cells significantly inhibited cellular invasion, wound healing and MMP-2 activity. Conclusions. Differential miRNA signatures of ESS and LMS provide novel data regarding transcriptional regulation in these cancers, based on which new potential diagnostic markers, prognostic biomarkers and therapeutic targets may be explored. Differences in miRNA profiles of primary and metastatic LMS may improve our understanding of disease progression in this aggressive malignancy. © 2016 Elsevier Inc. All rights reserved.
1. Introduction
⁎ Correspondence to: R. Reich, Institute of Drug Research, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem 91120, Israel. ⁎⁎ Correspondence to: B. Davidson, Department of Pathology, Norwegian Radium Hospital, Ullernchausseen 70, Montebello, N-0310 Oslo, Norway. E-mail addresses: [email protected] (R. Reich), [email protected] (B. Davidson). 1 R.R. is affiliated with the David R. Bloom Center for Pharmacy and the Adolf and Klara Brettler Center for Research in Molecular Pharmacology and Therapeutics at The Hebrew University of Jerusalem, Israel.
Uterine sarcomas are rare tumors, comprising 7% of all soft tissue sarcomas and about 3% of uterine malignancies [1–3]. Leiomyosarcoma (LMS) and endometrial stromal sarcoma (ESS) are the most common histological types [2,3]. Adenosarcoma and carcinosarcoma are both mixed epithelial–mesenchymal tumors, but only the former have a true sarcomatous component, whereas carcinosarcomas are currently regarded as metaplastic carcinomas. ESS has been previously classified as low-grade or high-grade, later regarded as a single entity, and recently re-divided into low-grade and high-grade categories, although the latter group constitutes rare tumors [4].
http://dx.doi.org/10.1016/j.ygyno.2016.01.001 0090-8258/© 2016 Elsevier Inc. All rights reserved.
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
2
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
MicroRNAs (miRNAs) are small (19–25 nucleotides), non-coding RNAs that post-transcriptionally regulate gene expression [5]. miRNAs are synthesized as a long double-stranded precursor called pri-miRNA by DNA polymerase II in the nucleus. Pri-miRNA is cleaved at specific sites by the RNAse Drosha inside the nucleus, producing a pre-miRNA which is exported to the cytoplasm by the exportin 5 protein, where it is processed by Dicer into mature miRNA. Mature miRNAs are subsequently activated through binding to the RNA-induced silencing complex (RISC) [6,7]. Through the RISC, miRNAs can regulate their targets, mediating translational repression or degradation. The sequence at the 5′ end of the mature miRNA, called the “seed region”, binds its complementary sequence within the 3′ untranslated regions (UTR) of the target mRNA [8]. Perfect or near-perfect complementarity between the miRNA and its mRNA target results in mRNA degradation, whereas lesser complementarity leads to translational inhibition. Data regarding the miRNA profile of uterine sarcomas is limited to date. An inverse association was reported for the miRNA Let-7 and its target high-mobility group AT-hook-2 (HMGA2) in uterine LMS [9]. miRNA profiles which differentiate uterine LMS from leiomyoma [10] or from different types of leiomyoma as well as smooth muscle tumors of uncertain malignant potential (STUMP) [11] have been described. In another report, the miRNA profiles of uterine sarcomas and carcinosarcomas were compared to those of patient-matched benign tissue [12]. Our group has previously reported on gene expression profiles which differentiate LMS from ESS [13] and primary LMS from metastatic LMS [14]. Similar data with respect to miRNA profiles are unavailable to date to the best of our knowledge. The present study compared the miRNA profiles of primary ESS, primary LMS and metastatic LMS. 2. Material and methods 2.1. Patients and material Specimens consisted of 44 uterine sarcomas, including 9 primary uterine ESS, 17 primary uterine LMS and 18 metastatic LMS, submitted for routine diagnostic purposes to the Department of Pathology at the Norwegian Radium Hospital during the period 1993–2009. Tumors were snap-frozen and kept at −70 °C. Frozen sections were evaluated for the presence of a N 80% tumor component and absence of necrosis. Diagnoses were established by an experienced gynecologic pathologist (BD) based on morphology and immunohistochemistry. The series studied consisted of the same 35 tumors analyzed in our previous reports [13,14], to which 9 new tumors were added (2 high-grade ESS, 4 primary LMS, 3 metastatic LMS). As previously detailed [14], the majority of primary and metastatic LMS were not patient-matched. Clinicopathologic data are detailed in Table 1. The study was approved by the Regional Committee for Medical Research Ethics in Norway.
array real-time PCR was run on the ABI PRISM 7900 System using the TaqMan Universal PCR Master Mix (Applied Biosystems). Raw miRNA array data were analyzed by using the RQ manager and Data Assist software of the ABI system (Applied Biosystems). Normalized CT (ΔCT) was calculated by comparing each miRNA value to the geometric mean of three endogenous controls, RNU 44, RNU 46 and small nuclear U6 RNA. One assay not related to human miRNA (ath-miR159a) was included as negative control. Eight primary and 8 metastatic LMS, as well as all 9 ESS were analyzed for miRNA profile and served as training group of our study. For validation, the entire material (17 primary LMS, 18 metastatic LMS, 9 primary ESS) was analyzed. Conversion of miRNA and mRNA into cDNA and quantitative PCR (qPCR) detection of miRNAs was carried out according to the manufacturer's protocols using the miScript Reverse Transcription Kit and miScript SYBR Green PCR Kit (Qiagen GmbH). qRT-PCR was performed on Mx3000P® QPCR System (Stratagene, La Jolla CA). Bioinformatics tools were used to identify their predicted targets and the molecular networks they may affect.
2.4. Western blot analysis Thirty LMS (15 primary, 15 metastatic) were analyzed using Western blotting. Tumors were lysed in lysis buffer (1% NP40, 20 mM Tris–HCl (pH 7.5), 137 mM NaCl, 0.5 mM EDTA, 10% glycerol, 1 mM sodium orthovanadate, 1 × protease inhibitors cocktail and 0.1% SDS (Sigma, St. Louis MO). Fifteen micrograms of protein from each sample were loaded under reducing conditions on 8% SDS-PAGE gel. Following electrophoresis, the proteins were transferred to PVDF transfer membranes (Millipore, Bedford MA). Membranes were blocked in Tris Buffered Saline with Tween 20 (TBST; 10 mM Tris–HCl [pH 8.0], 150 mM NaCl and 0.1% Tween 20) containing 5% non-fat dry milk (Nestlé, Vevey, Switzerland) for 1 h at room temperature. Membranes were then incubated with a rabbit monoclonal antibody against frizzled-6 (Abcam; cat. # 128916) overnight in 4 °C, washed with TBST and incubated with secondary antibody (Peroxidase-conjugated AffinPure goat anti-mouse, 1:20000 dilution, Jackson Immuno-Research Laboratories, Inc., West Grove, PA). After washing with TBST, bands were visualized by enhanced chemiluminescence (ECL; Pierce, Rockford IL), according to the manufacturer's instructions. Membranes were stripped, blocked and then re-incubated with anti-GAPDH antibody (Sigma). The developed X-ray films were scanned for quantification and analysis was performed on a Macintosh computer using the public domain NIH Image program (http://rsb.info.nih.gov/nih-image/). Loading was corrected by calculating the ratio between frizzled-6 and GAPDH values and was expressed as arbitrary units (AU).
2.5. Frizzled-6 silencing using siRNA 2.2. Isolation of total RNA and miRNA Total RNA, including miRNA, was isolated from the uterine biopsies using the miRVana miRNA Isolation Kit (Ambion-Applied Biosystems, Austin TX). RNA concentration and integrity were determined using a NanoDrop-1000 spectrophotometer (Thermo Scientific, Wilmington DE). RNA samples were stored at −80 °C before TaqMan miRNA array studies. 2.3. TaqMan miRNA arrays cards miRNA expression profiling was performed using the TaqMan Human miRNA Array Card A (Applied Biosystems) containing 381 mature human miRNAs in miRBase 10.1 (http://microRNA.sanger.ac.uk). Briefly, miRNA was reverse-transcribed to cDNA using the Megaplex™ RT Human Primers Pool and the TaqMan MiRNA Reverse Transcription Kit (Applied Biosystems). Quantitative 384-well TaqMan low-density
SK-LMS-1 uterine LMS cells were obtained from ATCC (Manassas, VA) and cultured according to ATCC instructions. Cells were used for silencing studies at sub-confluent conditions. Gene silencing-Frizzled-6 siRNAs (3 sequences) were purchased from Sigma-Aldrich (St. Louis MO). Silencing efficiency was estimated for each of the sequences and for the combination of all three. Cellular invasion was measured using Boyden chamber assay with Matrigel-coated PVPF filters with 8 μm pores as previously described [15]. Zymography was performed as previously detailed [15]. Scratch assay (“wound healing” assay) was performed as follows: 24 h prior to the experiment, cells transfected with siFRZ6 construct were seeded to create a confluent monolayer. Cell monolayers were scraped with a pipette tip to create approximately similarly sized scratches. Cells were incubated for 24 h, and images acquired at 0 and 24 h. Scratch closure was analyzed by comparing images obtained at these time points.
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
3
Table 1 Clinicopathologic data for patients with primary ESS, primary LMS and metastatic LMS.a Case
Diagnosis
Age
Anatomic site
Tumor diameter (cm)
Atypia
Mitotic rate (10 HPF)
Necrosis
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28d 29 30e 31 32 33 34d 35 36f 37f 38 39 40 41 42g 43 44h
LG-ESS LG-ESS HG-ESSb HG-ESSb LG-ESS LG-ESS LG-ESS LG-ESS LG-ESS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS
51 60 57 70 70 70 34 33 58 44 70 49 49 45 59 70 59 49 45 63 72 49 57 60 69 61 55 55 56 56 31 49 47 55 50 45 45 46 59 42 32 46 61 63
Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Intestine Intestine Pelvis Abdomen Intestine Bone Pelvis and chest wall Pelvis Omentum Pelvis Retroperitoneum Lung Retroperitoneum Abdomen Lung Abdomen Abdomen Lung
8 NA 9 18 NA 4 2.5 10 11 3 14 22 6 7 13 N20 5 3 17 6 10 NA 4 15 21 5 17 17 17 17 9 9 8 17 10 7 7 10 6 NA 15 10 10 6
Low Low Severe Moderate Moderate Severec Low Low Low Moderate Severe Moderate Severe Severe Severe Moderate Severe Severe Moderate Severe Moderate Severe Moderate Severe Severe Severe Severe Severe Moderate Moderate Severe Severe Severe Severe Severe Severe Severe Severe Severe Severe Moderate Severe Severe Severe
b5 b5 30 25 9 b5 15 b5 b5 1 N20 15 4 6 16 4 20 26 32 7 b5 16 20 13 35 20 20 20 32 32 50 2 1 20 16 6 6 12 30 2 40 12 10 7
No Yes Yes No No No Yes No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Abbreviations: LG-ESS = low-grade endometrial stromal sarcoma; HG-ESS = high-grade endometrial stromal sarcoma; and NA = not available. a Age and tumor characteristics for patients with metastases are for the primary tumor. b Diagnosis based on morphology. c Focal finding; low-grade atypia in the majority of tumor. d Same patient as #27. e Same patient as #29. f Same patient as #14. g Same patient as #38. h Same patient as #20.
3. Results Forty-nine miRNAs were found to be differentially expressed between primary and. metastatic LMS, of which 4 were overexpressed and 45 underexpressed in the metastatic lesions. Comparative analysis of primary LMS and ESS identified 94 miRNAs that were significantly differentially expressed between these two sarcoma types, including 76 overexpressed in ESS and 18 overexpressed in LMS (p b 0.05; Supplementary Table 1 and Supplementary Figs. 1-A, -B). Among the 20 miRNAs found to be most significantly differentially expressed in primary ESS vs. LMS or in primary vs. metastatic LMS, 10 were confirmed to be differentially expressed in validation experiments. Comparison of primary LMS and ESS showed overexpression of 7 miRNAs (mir-15b, mir-21, mir-23b, mir-25, mir-145, mir-148b and mir-195) in ESS. Comparison of primary and metastatic LMS lesions showed lower mir-15a and mir-92a levels and higher mir-31 levels in primary LMS (Table 2).
Analysis using the Onto-Tool-Pathway-Express suite identified 5 pathways involving the genes downregulated by the 9 miRNA underexpressed in primary LMS compared to ESS or metastatic LMS, including the mitogen-activated protein kinase (MAPK) signaling pathway, Wnt signaling pathway, Focal adhesion, the mTOR signaling pathway and the transforming growth factor-β (TGF-β) signaling pathway. Two pathways identified for mir-31 were the MAPK signaling pathway and the Wnt signaling pathway. Analysis of the gene expression array data of the same samples from our previous study [14] identified involvement of the TGF-β and Wnt pathways. Six different genes that are part of the Wnt signaling pathway are controlled by miRNAs 15a, 31 and 92a in LMS — FZD4, FZD6, FZD8, FZD10, LRP5 and LRP6. Among these genes, FZD6 was found to be significantly higher in metastatic compared to primary LMS (p = 0.025; Fig. 1-A). Western blot analysis revealed the opposite results, as
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
4
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
Fig. 1. Differential expression of Frizzled-6 in primary and metastatic LMS. A: FZD6 is overexpressed in metastatic compared to primary LMS. B: Frizzled-6 protein is overexpressed in primary compared to metastatic LMS.
Frizzled-6 protein expression was higher in primary compared to metastatic LMS (p = 0.031; Fig. 1-B). Since Frizzled-6 is hypothesized to be involved in regulation of cellular invasion, we tested the effect of Frizzled-6 silencing on the invasiveness of LMS cells. Fig. 2-A shows the silencing efficiency obtained for each sequence and combination of all three siRNAs used. Frizzled-6 silencing using siRNA significantly inhibited cellular invasion (Fig. 2-B). In wound closure assay, control cells closed 55% of the scratched area, while si-FZD6 cells closed only 34% (Fig. 2-C). MMP-2 activity in siFZD6 cells was reduced to b 40% compared to control cells (Fig. 2-D). 4. Discussion ESS and LMS are both rare sarcomas affecting the uterine corpus. However, they differ considerably at the genotypic and phenotypic levels, as well as their clinical behavior. Our understanding of
Table 2 Validated differentially expressed miRNAs in primary ESS, primary LMS and metastatic LMS. miRNA
Primary vs. metastatic LMS
Primary LMS vs. ESS
p-Value
Mir-15a Mir-31 Mir-92a Mir-15b Mir-21 Mir-23b Mir-25 Mir-145 Mir-148b Mir-195
Overexpressed in metastases Overexpressed in primary Overexpressed in metastases – – – – – – –
– – – Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS
0.025 0.0384 0.0384 0.0026 0.0043 0.0026 0.0336 0.0041 0.0233 0.0041
differences in the molecular drivers of each of these cancers is limited by the paucity of comparative molecular analyses of ESS and LMS. Beyond scientific interest, such studies may improve and expand the very limited panel available for differentiating these tumors in routine surgical pathology, which in our opinion does not include a single reliable ESS marker to date, despite claims that CD10 has such role. While ESS may be differentiated from LMS based solely on morphology in many cases, ESS with smooth muscle phenotype has been described [16,17] and may pose greater diagnostic difficulty. Molecular analyses and studies at the protein level comparing ESS and LMS may further pinpoint the predictive and prognostic role certain biomarkers may have in one, but not the other cancer, as recently observed by our group regarding hormone receptors [18]. The present study provides the first analysis of miRNA profiles in ESS vs. LMS. In addition, we compared the miRNA profiles of primary and metastatic LMS, in the hope of gaining knowledge regarding the molecular changes undergone by LMS during metastasis. Predictably, as in our gene expression array analyses [13,14], differences were more pronounced in the ESS vs. LMS analysis, these being two different cancers, compared to the more subtle differences between primary and metastatic LMS. Based on our literature search, none of the 10 miRNAs identified and validated as differentially expressed in ESS vs. LMS or primary vs. metastatic LMS has been previously studied in these tumors. Several of these miRNAs, including miR-15, miR-25, miR-92 and miR-148b, have not been studied in any sarcoma. Among these, miR-15 is of the most obvious interest for further research. The miRNA cluster miR-15a/miR16-1 at chromosome 13q is the target of a frequent deletion in chronic lymphoblastic leukemia, and downregulation of these miRNAs removes their repression of target molecules driving cell proliferation, including Bcl-2, cyclin D1 and Mcl-1, in different cancers [19]. The remaining differentially expressed miRs have been the focus of analyses of other, non-genital sarcomas. miR-21 is an onco-miR expressed in multiple cancers, including carcinomas of various origin, hematological cancers and glioblastoma, and is implicated in resistance to multiple chemotherapeutic agents through regulation of PTEN, PTCD4, Bcl-2 and MSH2 [20]. Part of the biological role of miR-21 in cancer may be related to its complex modulating effect on the immune system [21]. Several studies have focused on the role of miR-21 in sarcoma. miR-21 was shown to be overexpressed in malignant compared to benign Phyllodes tumors of the breast, and its upregulation is associated with the acquisition of a myofibroblastic phenotype [22]. miR-21 is overexpressed in osteosarcoma compared to normal bone and mediates migration and invasion in MG-63 osteosarcoma cells through repression of the negative matrix metalloproteinase (MMP) regulator RECK [23]. Higher miR-21 and lower miR-199a and miR-143 levels were observed in the serum of osteosarcoma patients compared to controls and suggested to be biomarkers of this disease [24]. Of note, miR-21 levels were higher in ESS than in LMS in the present study, despite the generally more aggressive behavior of the latter. miR-23b, another miRNA overexpressed in ESS in our study, has not been studied in sarcomas to date. However, miR-23a has been reported to be a tumor suppressor in osteosarcoma, its effect mediated via suppression of the transcription factor RUNX2 and the cytokine CXCL12 [25]. miR-31, which was overexpressed in primary compared to metastatic LMS in our series, was reported to have intermediate expression level in osteosarcoma compared to osteoblasts and normal bone, with highest levels in osteoblasts [26]. miR-145 is the most extensively studied miRNA in sarcoma among the differentially-expressed miRs identified in the present study. It was identified as the miR most strongly negatively related to the EWS-FLI1 fusion product in Ewing sarcoma, with reciprocal repression of the two molecules [27]. miR-145 downregulation in chondrosarcoma abrogates its negative regulation of SOX9, which in turn increases the
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
5
Fig. 2. Frizzled-6 regulates invasion in LMS cells. A: silencing efficiency for each FZD6 siRNA sequence and combination of all three siRNAs used. B: FZD6 silencing by siRNA significantly inhibited cellular invasion. C: Wound closure assay. Control cells closed 55% of the scratched area, while si-FRZ6 cells closed only 34% within 24 h. D: Zymography. MMP-2 activity in si-FRZ6 cells is reduced to b40% compared to control cells.
levels of the transcription factor ETV5, leading to MMP-2 expression and bone resorption [28]. miR-145 was further shown to be downregulated in liposarcoma compared to normal fat in two studies [29,30] and overexpression of miR-145 and miR-451 in liposarcoma cell lines reduced proliferation and induced apoptosis [29]. Finally, miR-145 was identified in Epstein–Barr virus (EBV)-associated smooth muscle tumors developing post-transplantation, as well as uterine leiomyomas, and therefore postulated to be a marker of smooth muscle differentiation unrelated to EBV infection [31]. Notwithstanding, in the present study its expression was higher in ESS than in LMS, more in agreement with the above publications [27–30] associating its loss with malignancy and clinically aggressive sarcomas. miR-195 was reported to be overexpressed in the serum of osteosarcoma patients compared to controls [32]. In the above report by Namløs et al., miR-195 levels were intermediate in osteosarcoma compared to
osteoblasts and normal bone, as were those of miR-31, though with highest levels in normal bone [26]. Fricke and co-workers found higher levels of miR-195-5p in blood from synovial sarcoma patients compared to both healthy controls and patients with other sarcomas. Results were validated for the comparative analysis with controls for 7 miRNAs using qRT-PCR, though miR-195-5p was not overexpressed in synovial sarcoma compared to LMS (n = 5; organ of origin not specified) [33]. As we did not have an ESS cell line at our lab, we chose to focus on the Wnt pathway, regulated by several of the miRNAs found to be differentially expressed in primary vs. metastatic LMS. Wnt proteins are secreted glycoproteins that bind to receptors of the Frizzled family and several co-receptors. In humans, 19 members of the Wnt family and 10 Frizzled receptors are known. Based on the different biological readouts, Wnt ligands and Frizzled receptors are divided into canonical and non-canonical pathways. The canonical pathway involves the
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
6
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
accumulation of β-catenin and its function as transcription factor together with TCF/LEF upon its transport into the nucleus. The noncanonical pathway involves the activation of rho, cdc42 and JNK. These pathways regulate central cancer-associated functions, including cell growth, migration, invasion and metastasis [34]. Differential expression of Frizzled-6 was found in primary and metastatic LMS, with opposite findings at the mRNA and protein level. The discrepancy between mRNA expression and protein expression of this receptor may be due to ceRNA regulation, though further research is needed to elucidate this finding. In vitro experiments with LMS cells further showed that FZD6 inhibition by siRNA inhibits cellular invasion, wound closure and MMP-2 activity, suggesting a tumor-promoting role in this cancer. In conclusion, ESS and LMS have unique miRNA signatures, with differences for primary vs. metastatic LMS being of lesser degree. The differences in miRNA profiles between LMS and ESS are likely to affect cellular regulatory pathways and result in unique phenotype for each of these cancers. These in turn may form the basis for identifying new diagnostic and therapeutic candidates in each tumor. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2016.01.001. Financial acknowledgment This work was supported by a grant from the National Sarcoma Foundation at the Norwegian Radium Hospital. Conflict of interest statement The authors declare that they have no competing interests.
References [1] J.R. Toro, L.B. Travis, H.J. Wu, et al., Incidence patterns of soft tissue sarcomas, regardless of primary site, in the surveillance, epidemiology and end results program, 1978–2001: an analysis of 26,758 cases, Int. J. Cancer 119 (2006) 2922–2930. [2] E. D'Angelo, J. Prat, Uterine sarcomas: a review, Gynecol. Oncol. 116 (2010) 131–139. [3] V.M. Abeler, O. Røyne, S. Thoresen, et al., Uterine sarcomas in Norway. A histopathological and prognostic survey of a total population from 1970 to 2000 including 419 patients, Histopathology 54 (2009) 355–364. [4] R.J. Kurman, M.L. Carcangiu, C.S. Herrington, et al., WHO Classification of Tumours of Female Reproductive Organs, IARC, Lyon, France, 2014. [5] B.R. Cullen, Transcription and processing of human microRNA precursors, Mol. Cell 16 (2004) 861–865. [6] J. Permuth-Wey, Z. Chen, Y.Y. Tsai, et al., Ovarian Cancer Association Consortium (OCAC).microRNA processing and binding site polymorphisms are not replicated in the Ovarian Cancer Association Consortium, Cancer Epidemiol. Biomarkers Prev. 20 (2011) 1793–1797. [7] A.J. Pratt, I.J. MacRae, The RNA-induced silencing complex: a versatile gene-silencing machine, J. Biol. Chem. 284 (2009) 17897–17901. [8] A. Grimson, K.K. Farh, W.K. Johnston, P. Garrett-Engele, L.P. Lim, D.P. Bartel, MicroRNA targeting specificity in mammals: determinants beyond seed pairing, Mol. Cell 27 (2007) 91–105.
[9] G. Shi, M.A. Perle, K. Mittal, H. Chen, X. Zou, M. Narita, E. Hernando, P. Lee, J.J. Wei, Let-7 repression leads to HMGA2 overexpression in uterine leiomyosarcoma, J. Cell. Mol. Med. 13 (2009) 3898–3905. [10] L.S. Danielson, S. Menendez, C.S. Attolini, et al., A differentiation-based microRNA signature identifies leiomyosarcoma as a mesenchymal stem cell-related malignancy, Am. J. Pathol. 177 (2010) 908–917. [11] Q. Zhang, J. Ubago, L. Li, et al., Molecular analyses of 6 different types of uterine smooth muscle tumors: emphasis in atypical leiomyoma, Cancer 120 (2014) 3165–3177. [12] M. Kowalewska, E. Bakula-Zalewska, M. Chechlinska, et al., microRNAs in uterine sarcomas and mixed epithelial–mesenchymal uterine tumors: a preliminary report, Tumour Biol. 34 (2013) 2153–2160. [13] B. Davidson, V.M. Abeler, E. Hellesylt, et al., Gene expression signatures differentiate uterine endometrial stromal sarcoma from leiomyosarcoma, Gynecol. Oncol. 128 (2013) 349–355. [14] B. Davidson, V.M. Abeler, M. Førsund, et al., Gene expression signatures of primary and metastatic uterine leiomyosarcoma, Hum. Pathol. 45 (2014) 691–700. [15] R. Reich, E.W. Thompson, Y. Iwamoto, et al., Inhibition of plasminogen activator, serine proteinases and collagenase IV prevents the invasion of basement membrane by metastatic cells, Cancer Res. 48 (1988) 3307–3312. [16] F. Moinfar, M. Azodi, F.A. Tavassoli, Uterine sarcomas, Pathology 39 (2007) 55–71. [17] E. Oliva, Cellular mesenchymal tumors of the uterus: a review emphasizing recent observations, Int. J. Gynecol. Pathol. 33 (2014) 374–384. [18] B. Davidson, M.L. Kjæreng, M. Førsund, H.E. Danielsen, G.B. Kristensen, V.M. Abeler, Progesterone receptor expression is an independent prognosticator in FIGO stage I uterine leiomyosarcoma, Am. J. Clin. Pathol. (2015) (in press). [19] Y. Pekarsky, C.M. Croce, Role of miR-15/16 in CLL, Cell Death Differ. 22 (2015) 6–11. [20] L. Hong, Y. Han, Y. Zhang, et al., MicroRNA-21: a therapeutic target for reversing drug resistance in cancer, Expert Opin. Ther. Targets 17 (2013) 1073–1080. [21] F.J. Sheedy, Turning 21: induction of miR-21 as a key switch in the inflammatory response, Front. Immunol. 6 (2015) 19. [22] C. Gong, Y. Nie, S. Qu, et al., miR-21 induces myofibroblast differentiation and promotes the malignant progression of breast phyllodes tumors, Cancer Res. 74 (2014) 4341–4352. [23] W. Ziyan, Y. Shuhua, W. Xiufang, L. Xiaoyun, MicroRNA-21 is involved in osteosarcoma cell invasion and migration, Med. Oncol. 28 (2011) 1469–1474. [24] L. Ouyang, P. Liu, S. Yang, S. Ye, W. Xu, X. Liu, A three-plasma miRNA signature serves as novel biomarkers for osteosarcoma, Med. Oncol. 30 (2013) 340. [25] Y. He, C. Meng, Z. Shao, H. Wang, S. Yang, MiR-23a functions as a tumor suppressor in osteosarcoma, Cell. Physiol. Biochem. 34 (2014) 1485–1496. [26] H.M. Namløs, L.A. Meza-Zepeda, T. Barøy, et al., Modulation of the osteosarcoma expression phenotype by microRNAs, PLoS ONE 7 (2012), e48086. [27] J. Ban, G. Jug, P. Mestdagh, et al., Hsa-mir-145 is the top EWS-FLI1-repressed microRNA involved in a positive feedback loop in Ewing's sarcoma, Oncogene 30 (2011) 2173–2180. [28] I.W. Mak, S. Singh, R. Turcotte, M. Ghert, The epigenetic regulation of SOX9 by miR145 in human chondrosarcoma, J. Cell. Biochem. 116 (2015) 37–44. [29] C.M. Gits, P.F. van Kuijk, M.B. Jonkers, et al., MicroRNA expression profiles distinguish liposarcoma subtypes and implicate miR-145 and miR-451 as tumor suppressors, Int. J. Cancer 135 (2014) 348–361. [30] S. Ugras, E. Brill, A. Jacobsen, et al., Small RNA sequencing and functional characterization reveals MicroRNA-143 tumor suppressor activity in liposarcoma, Cancer Res. 71 (2011) 5659–5669. [31] D. Jonigk, N. Izykowski, L. Maegel, et al., MicroRNA expression in Epstein–Barr virusassociated post-transplant smooth muscle tumours is related to leiomyomatous phenotype, Clin. Sarcoma Res. 3 (2013) 9. [32] H. Cai, H. Zhao, J. Tang, H. Wu, Serum miR-195 is a diagnostic and prognostic marker for osteosarcoma, J. Surg. Res. 194 (2015) 505–510. [33] A. Fricke, P.V. Ullrich, J. Heinz, et al., Identification of a blood-borne miRNA signature of synovial sarcoma, Mol. Cancer 14 (2015) 151. [34] K. Ueno, H. Hirata, Y. Hinoda, R. Dahiya, Frizzled homolog proteins, microRNAs and Wnt signaling in cancer, Int. J. Cancer 132 (2013) 1731–1740.
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
Contents lists available at ScienceDirect
Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno
Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures Yeheli Ravid a, Malka Formanski a, Yoav Smith b, Reuven Reich a,⁎,1, Ben Davidson c,d,⁎⁎ a
Institute of Drug Research, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem 91120, Israel Genomic Data Analysis Unit, Hadassah Medical School, The Hebrew University of Jerusalem, Jerusalem 91120, Israel Department of Pathology, Norwegian Radium Hospital, Oslo University Hospital, N-0310 Oslo, Norway d The Medical Faculty, University of Oslo, N-0316 Oslo, Norway b c
H I G H L I G H T S • microRNA array analysis differentiates uterine endometrial stromal sarcoma from leiomyosarcoma. • microRNAs are predominantly underexpressed in metastatic compared to primary leiomyosarcoma. • Frizzled-6 silencing suppresses invasion in leiomyosarcoma cells in vitro.
a r t i c l e
i n f o
Article history: Received 14 October 2015 Received in revised form 1 January 2016 Accepted 2 January 2016 Available online xxxx Keywords: Endometrial stromal sarcoma Leiomyosarcoma Differential diagnosis Tumor progression MicroRNA
a b s t r a c t Objective. To compare the microRNA (miRNA) profiles of uterine endometrial stromal sarcoma (ESS) and leiomyosarcoma (LMS), and to compare the miRNA signatures of primary and metastatic uterine LMS. Methods. Eight primary LMS, 9 primary ESS and 8 metastatic LMS were analyzed for miRNA profiles using TaqMan Human miRNA Array Cards. Findings for 20 differentially expressed miRNAs were validated in a series of 44 uterine sarcomas (9 primary uterine ESS, 17 primary uterine LMS, 18 metastatic LMS) using qPCR. Frizzled-6 protein expression was analyzed in 30 LMS (15 primary, 15 metastases). Frizzled-6 was silenced in SK-LMS-1 uterine LMS cells using siRNA and the effect on invasion, wound healing and matrix metalloproteinase-2 (MMP2) activity was assessed. Results. Ninety-four miRNAs were significantly differentially expressed in ESS and LMS, of which 76 were overexpressed in ESS and 18 overexpressed in LMS. Forty-nine miRNAs were differentially expressed in primary and metastatic LMS, of which 45 were overexpressed in primary LMS and 4 in metastases. Differential expression was confirmed for 10/20 miRNA analyzed using qPCR. Frizzled-6 silencing in SK-LMS-1 cells significantly inhibited cellular invasion, wound healing and MMP-2 activity. Conclusions. Differential miRNA signatures of ESS and LMS provide novel data regarding transcriptional regulation in these cancers, based on which new potential diagnostic markers, prognostic biomarkers and therapeutic targets may be explored. Differences in miRNA profiles of primary and metastatic LMS may improve our understanding of disease progression in this aggressive malignancy. © 2016 Elsevier Inc. All rights reserved.
1. Introduction
⁎ Correspondence to: R. Reich, Institute of Drug Research, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, Jerusalem 91120, Israel. ⁎⁎ Correspondence to: B. Davidson, Department of Pathology, Norwegian Radium Hospital, Ullernchausseen 70, Montebello, N-0310 Oslo, Norway. E-mail addresses: [email protected] (R. Reich), [email protected] (B. Davidson). 1 R.R. is affiliated with the David R. Bloom Center for Pharmacy and the Adolf and Klara Brettler Center for Research in Molecular Pharmacology and Therapeutics at The Hebrew University of Jerusalem, Israel.
Uterine sarcomas are rare tumors, comprising 7% of all soft tissue sarcomas and about 3% of uterine malignancies [1–3]. Leiomyosarcoma (LMS) and endometrial stromal sarcoma (ESS) are the most common histological types [2,3]. Adenosarcoma and carcinosarcoma are both mixed epithelial–mesenchymal tumors, but only the former have a true sarcomatous component, whereas carcinosarcomas are currently regarded as metaplastic carcinomas. ESS has been previously classified as low-grade or high-grade, later regarded as a single entity, and recently re-divided into low-grade and high-grade categories, although the latter group constitutes rare tumors [4].
http://dx.doi.org/10.1016/j.ygyno.2016.01.001 0090-8258/© 2016 Elsevier Inc. All rights reserved.
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
2
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
MicroRNAs (miRNAs) are small (19–25 nucleotides), non-coding RNAs that post-transcriptionally regulate gene expression [5]. miRNAs are synthesized as a long double-stranded precursor called pri-miRNA by DNA polymerase II in the nucleus. Pri-miRNA is cleaved at specific sites by the RNAse Drosha inside the nucleus, producing a pre-miRNA which is exported to the cytoplasm by the exportin 5 protein, where it is processed by Dicer into mature miRNA. Mature miRNAs are subsequently activated through binding to the RNA-induced silencing complex (RISC) [6,7]. Through the RISC, miRNAs can regulate their targets, mediating translational repression or degradation. The sequence at the 5′ end of the mature miRNA, called the “seed region”, binds its complementary sequence within the 3′ untranslated regions (UTR) of the target mRNA [8]. Perfect or near-perfect complementarity between the miRNA and its mRNA target results in mRNA degradation, whereas lesser complementarity leads to translational inhibition. Data regarding the miRNA profile of uterine sarcomas is limited to date. An inverse association was reported for the miRNA Let-7 and its target high-mobility group AT-hook-2 (HMGA2) in uterine LMS [9]. miRNA profiles which differentiate uterine LMS from leiomyoma [10] or from different types of leiomyoma as well as smooth muscle tumors of uncertain malignant potential (STUMP) [11] have been described. In another report, the miRNA profiles of uterine sarcomas and carcinosarcomas were compared to those of patient-matched benign tissue [12]. Our group has previously reported on gene expression profiles which differentiate LMS from ESS [13] and primary LMS from metastatic LMS [14]. Similar data with respect to miRNA profiles are unavailable to date to the best of our knowledge. The present study compared the miRNA profiles of primary ESS, primary LMS and metastatic LMS. 2. Material and methods 2.1. Patients and material Specimens consisted of 44 uterine sarcomas, including 9 primary uterine ESS, 17 primary uterine LMS and 18 metastatic LMS, submitted for routine diagnostic purposes to the Department of Pathology at the Norwegian Radium Hospital during the period 1993–2009. Tumors were snap-frozen and kept at −70 °C. Frozen sections were evaluated for the presence of a N 80% tumor component and absence of necrosis. Diagnoses were established by an experienced gynecologic pathologist (BD) based on morphology and immunohistochemistry. The series studied consisted of the same 35 tumors analyzed in our previous reports [13,14], to which 9 new tumors were added (2 high-grade ESS, 4 primary LMS, 3 metastatic LMS). As previously detailed [14], the majority of primary and metastatic LMS were not patient-matched. Clinicopathologic data are detailed in Table 1. The study was approved by the Regional Committee for Medical Research Ethics in Norway.
array real-time PCR was run on the ABI PRISM 7900 System using the TaqMan Universal PCR Master Mix (Applied Biosystems). Raw miRNA array data were analyzed by using the RQ manager and Data Assist software of the ABI system (Applied Biosystems). Normalized CT (ΔCT) was calculated by comparing each miRNA value to the geometric mean of three endogenous controls, RNU 44, RNU 46 and small nuclear U6 RNA. One assay not related to human miRNA (ath-miR159a) was included as negative control. Eight primary and 8 metastatic LMS, as well as all 9 ESS were analyzed for miRNA profile and served as training group of our study. For validation, the entire material (17 primary LMS, 18 metastatic LMS, 9 primary ESS) was analyzed. Conversion of miRNA and mRNA into cDNA and quantitative PCR (qPCR) detection of miRNAs was carried out according to the manufacturer's protocols using the miScript Reverse Transcription Kit and miScript SYBR Green PCR Kit (Qiagen GmbH). qRT-PCR was performed on Mx3000P® QPCR System (Stratagene, La Jolla CA). Bioinformatics tools were used to identify their predicted targets and the molecular networks they may affect.
2.4. Western blot analysis Thirty LMS (15 primary, 15 metastatic) were analyzed using Western blotting. Tumors were lysed in lysis buffer (1% NP40, 20 mM Tris–HCl (pH 7.5), 137 mM NaCl, 0.5 mM EDTA, 10% glycerol, 1 mM sodium orthovanadate, 1 × protease inhibitors cocktail and 0.1% SDS (Sigma, St. Louis MO). Fifteen micrograms of protein from each sample were loaded under reducing conditions on 8% SDS-PAGE gel. Following electrophoresis, the proteins were transferred to PVDF transfer membranes (Millipore, Bedford MA). Membranes were blocked in Tris Buffered Saline with Tween 20 (TBST; 10 mM Tris–HCl [pH 8.0], 150 mM NaCl and 0.1% Tween 20) containing 5% non-fat dry milk (Nestlé, Vevey, Switzerland) for 1 h at room temperature. Membranes were then incubated with a rabbit monoclonal antibody against frizzled-6 (Abcam; cat. # 128916) overnight in 4 °C, washed with TBST and incubated with secondary antibody (Peroxidase-conjugated AffinPure goat anti-mouse, 1:20000 dilution, Jackson Immuno-Research Laboratories, Inc., West Grove, PA). After washing with TBST, bands were visualized by enhanced chemiluminescence (ECL; Pierce, Rockford IL), according to the manufacturer's instructions. Membranes were stripped, blocked and then re-incubated with anti-GAPDH antibody (Sigma). The developed X-ray films were scanned for quantification and analysis was performed on a Macintosh computer using the public domain NIH Image program (http://rsb.info.nih.gov/nih-image/). Loading was corrected by calculating the ratio between frizzled-6 and GAPDH values and was expressed as arbitrary units (AU).
2.5. Frizzled-6 silencing using siRNA 2.2. Isolation of total RNA and miRNA Total RNA, including miRNA, was isolated from the uterine biopsies using the miRVana miRNA Isolation Kit (Ambion-Applied Biosystems, Austin TX). RNA concentration and integrity were determined using a NanoDrop-1000 spectrophotometer (Thermo Scientific, Wilmington DE). RNA samples were stored at −80 °C before TaqMan miRNA array studies. 2.3. TaqMan miRNA arrays cards miRNA expression profiling was performed using the TaqMan Human miRNA Array Card A (Applied Biosystems) containing 381 mature human miRNAs in miRBase 10.1 (http://microRNA.sanger.ac.uk). Briefly, miRNA was reverse-transcribed to cDNA using the Megaplex™ RT Human Primers Pool and the TaqMan MiRNA Reverse Transcription Kit (Applied Biosystems). Quantitative 384-well TaqMan low-density
SK-LMS-1 uterine LMS cells were obtained from ATCC (Manassas, VA) and cultured according to ATCC instructions. Cells were used for silencing studies at sub-confluent conditions. Gene silencing-Frizzled-6 siRNAs (3 sequences) were purchased from Sigma-Aldrich (St. Louis MO). Silencing efficiency was estimated for each of the sequences and for the combination of all three. Cellular invasion was measured using Boyden chamber assay with Matrigel-coated PVPF filters with 8 μm pores as previously described [15]. Zymography was performed as previously detailed [15]. Scratch assay (“wound healing” assay) was performed as follows: 24 h prior to the experiment, cells transfected with siFRZ6 construct were seeded to create a confluent monolayer. Cell monolayers were scraped with a pipette tip to create approximately similarly sized scratches. Cells were incubated for 24 h, and images acquired at 0 and 24 h. Scratch closure was analyzed by comparing images obtained at these time points.
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
3
Table 1 Clinicopathologic data for patients with primary ESS, primary LMS and metastatic LMS.a Case
Diagnosis
Age
Anatomic site
Tumor diameter (cm)
Atypia
Mitotic rate (10 HPF)
Necrosis
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28d 29 30e 31 32 33 34d 35 36f 37f 38 39 40 41 42g 43 44h
LG-ESS LG-ESS HG-ESSb HG-ESSb LG-ESS LG-ESS LG-ESS LG-ESS LG-ESS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS LMS
51 60 57 70 70 70 34 33 58 44 70 49 49 45 59 70 59 49 45 63 72 49 57 60 69 61 55 55 56 56 31 49 47 55 50 45 45 46 59 42 32 46 61 63
Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Uterus Intestine Intestine Pelvis Abdomen Intestine Bone Pelvis and chest wall Pelvis Omentum Pelvis Retroperitoneum Lung Retroperitoneum Abdomen Lung Abdomen Abdomen Lung
8 NA 9 18 NA 4 2.5 10 11 3 14 22 6 7 13 N20 5 3 17 6 10 NA 4 15 21 5 17 17 17 17 9 9 8 17 10 7 7 10 6 NA 15 10 10 6
Low Low Severe Moderate Moderate Severec Low Low Low Moderate Severe Moderate Severe Severe Severe Moderate Severe Severe Moderate Severe Moderate Severe Moderate Severe Severe Severe Severe Severe Moderate Moderate Severe Severe Severe Severe Severe Severe Severe Severe Severe Severe Moderate Severe Severe Severe
b5 b5 30 25 9 b5 15 b5 b5 1 N20 15 4 6 16 4 20 26 32 7 b5 16 20 13 35 20 20 20 32 32 50 2 1 20 16 6 6 12 30 2 40 12 10 7
No Yes Yes No No No Yes No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Abbreviations: LG-ESS = low-grade endometrial stromal sarcoma; HG-ESS = high-grade endometrial stromal sarcoma; and NA = not available. a Age and tumor characteristics for patients with metastases are for the primary tumor. b Diagnosis based on morphology. c Focal finding; low-grade atypia in the majority of tumor. d Same patient as #27. e Same patient as #29. f Same patient as #14. g Same patient as #38. h Same patient as #20.
3. Results Forty-nine miRNAs were found to be differentially expressed between primary and. metastatic LMS, of which 4 were overexpressed and 45 underexpressed in the metastatic lesions. Comparative analysis of primary LMS and ESS identified 94 miRNAs that were significantly differentially expressed between these two sarcoma types, including 76 overexpressed in ESS and 18 overexpressed in LMS (p b 0.05; Supplementary Table 1 and Supplementary Figs. 1-A, -B). Among the 20 miRNAs found to be most significantly differentially expressed in primary ESS vs. LMS or in primary vs. metastatic LMS, 10 were confirmed to be differentially expressed in validation experiments. Comparison of primary LMS and ESS showed overexpression of 7 miRNAs (mir-15b, mir-21, mir-23b, mir-25, mir-145, mir-148b and mir-195) in ESS. Comparison of primary and metastatic LMS lesions showed lower mir-15a and mir-92a levels and higher mir-31 levels in primary LMS (Table 2).
Analysis using the Onto-Tool-Pathway-Express suite identified 5 pathways involving the genes downregulated by the 9 miRNA underexpressed in primary LMS compared to ESS or metastatic LMS, including the mitogen-activated protein kinase (MAPK) signaling pathway, Wnt signaling pathway, Focal adhesion, the mTOR signaling pathway and the transforming growth factor-β (TGF-β) signaling pathway. Two pathways identified for mir-31 were the MAPK signaling pathway and the Wnt signaling pathway. Analysis of the gene expression array data of the same samples from our previous study [14] identified involvement of the TGF-β and Wnt pathways. Six different genes that are part of the Wnt signaling pathway are controlled by miRNAs 15a, 31 and 92a in LMS — FZD4, FZD6, FZD8, FZD10, LRP5 and LRP6. Among these genes, FZD6 was found to be significantly higher in metastatic compared to primary LMS (p = 0.025; Fig. 1-A). Western blot analysis revealed the opposite results, as
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
4
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
Fig. 1. Differential expression of Frizzled-6 in primary and metastatic LMS. A: FZD6 is overexpressed in metastatic compared to primary LMS. B: Frizzled-6 protein is overexpressed in primary compared to metastatic LMS.
Frizzled-6 protein expression was higher in primary compared to metastatic LMS (p = 0.031; Fig. 1-B). Since Frizzled-6 is hypothesized to be involved in regulation of cellular invasion, we tested the effect of Frizzled-6 silencing on the invasiveness of LMS cells. Fig. 2-A shows the silencing efficiency obtained for each sequence and combination of all three siRNAs used. Frizzled-6 silencing using siRNA significantly inhibited cellular invasion (Fig. 2-B). In wound closure assay, control cells closed 55% of the scratched area, while si-FZD6 cells closed only 34% (Fig. 2-C). MMP-2 activity in siFZD6 cells was reduced to b 40% compared to control cells (Fig. 2-D). 4. Discussion ESS and LMS are both rare sarcomas affecting the uterine corpus. However, they differ considerably at the genotypic and phenotypic levels, as well as their clinical behavior. Our understanding of
Table 2 Validated differentially expressed miRNAs in primary ESS, primary LMS and metastatic LMS. miRNA
Primary vs. metastatic LMS
Primary LMS vs. ESS
p-Value
Mir-15a Mir-31 Mir-92a Mir-15b Mir-21 Mir-23b Mir-25 Mir-145 Mir-148b Mir-195
Overexpressed in metastases Overexpressed in primary Overexpressed in metastases – – – – – – –
– – – Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS Overexpressed in ESS
0.025 0.0384 0.0384 0.0026 0.0043 0.0026 0.0336 0.0041 0.0233 0.0041
differences in the molecular drivers of each of these cancers is limited by the paucity of comparative molecular analyses of ESS and LMS. Beyond scientific interest, such studies may improve and expand the very limited panel available for differentiating these tumors in routine surgical pathology, which in our opinion does not include a single reliable ESS marker to date, despite claims that CD10 has such role. While ESS may be differentiated from LMS based solely on morphology in many cases, ESS with smooth muscle phenotype has been described [16,17] and may pose greater diagnostic difficulty. Molecular analyses and studies at the protein level comparing ESS and LMS may further pinpoint the predictive and prognostic role certain biomarkers may have in one, but not the other cancer, as recently observed by our group regarding hormone receptors [18]. The present study provides the first analysis of miRNA profiles in ESS vs. LMS. In addition, we compared the miRNA profiles of primary and metastatic LMS, in the hope of gaining knowledge regarding the molecular changes undergone by LMS during metastasis. Predictably, as in our gene expression array analyses [13,14], differences were more pronounced in the ESS vs. LMS analysis, these being two different cancers, compared to the more subtle differences between primary and metastatic LMS. Based on our literature search, none of the 10 miRNAs identified and validated as differentially expressed in ESS vs. LMS or primary vs. metastatic LMS has been previously studied in these tumors. Several of these miRNAs, including miR-15, miR-25, miR-92 and miR-148b, have not been studied in any sarcoma. Among these, miR-15 is of the most obvious interest for further research. The miRNA cluster miR-15a/miR16-1 at chromosome 13q is the target of a frequent deletion in chronic lymphoblastic leukemia, and downregulation of these miRNAs removes their repression of target molecules driving cell proliferation, including Bcl-2, cyclin D1 and Mcl-1, in different cancers [19]. The remaining differentially expressed miRs have been the focus of analyses of other, non-genital sarcomas. miR-21 is an onco-miR expressed in multiple cancers, including carcinomas of various origin, hematological cancers and glioblastoma, and is implicated in resistance to multiple chemotherapeutic agents through regulation of PTEN, PTCD4, Bcl-2 and MSH2 [20]. Part of the biological role of miR-21 in cancer may be related to its complex modulating effect on the immune system [21]. Several studies have focused on the role of miR-21 in sarcoma. miR-21 was shown to be overexpressed in malignant compared to benign Phyllodes tumors of the breast, and its upregulation is associated with the acquisition of a myofibroblastic phenotype [22]. miR-21 is overexpressed in osteosarcoma compared to normal bone and mediates migration and invasion in MG-63 osteosarcoma cells through repression of the negative matrix metalloproteinase (MMP) regulator RECK [23]. Higher miR-21 and lower miR-199a and miR-143 levels were observed in the serum of osteosarcoma patients compared to controls and suggested to be biomarkers of this disease [24]. Of note, miR-21 levels were higher in ESS than in LMS in the present study, despite the generally more aggressive behavior of the latter. miR-23b, another miRNA overexpressed in ESS in our study, has not been studied in sarcomas to date. However, miR-23a has been reported to be a tumor suppressor in osteosarcoma, its effect mediated via suppression of the transcription factor RUNX2 and the cytokine CXCL12 [25]. miR-31, which was overexpressed in primary compared to metastatic LMS in our series, was reported to have intermediate expression level in osteosarcoma compared to osteoblasts and normal bone, with highest levels in osteoblasts [26]. miR-145 is the most extensively studied miRNA in sarcoma among the differentially-expressed miRs identified in the present study. It was identified as the miR most strongly negatively related to the EWS-FLI1 fusion product in Ewing sarcoma, with reciprocal repression of the two molecules [27]. miR-145 downregulation in chondrosarcoma abrogates its negative regulation of SOX9, which in turn increases the
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
5
Fig. 2. Frizzled-6 regulates invasion in LMS cells. A: silencing efficiency for each FZD6 siRNA sequence and combination of all three siRNAs used. B: FZD6 silencing by siRNA significantly inhibited cellular invasion. C: Wound closure assay. Control cells closed 55% of the scratched area, while si-FRZ6 cells closed only 34% within 24 h. D: Zymography. MMP-2 activity in si-FRZ6 cells is reduced to b40% compared to control cells.
levels of the transcription factor ETV5, leading to MMP-2 expression and bone resorption [28]. miR-145 was further shown to be downregulated in liposarcoma compared to normal fat in two studies [29,30] and overexpression of miR-145 and miR-451 in liposarcoma cell lines reduced proliferation and induced apoptosis [29]. Finally, miR-145 was identified in Epstein–Barr virus (EBV)-associated smooth muscle tumors developing post-transplantation, as well as uterine leiomyomas, and therefore postulated to be a marker of smooth muscle differentiation unrelated to EBV infection [31]. Notwithstanding, in the present study its expression was higher in ESS than in LMS, more in agreement with the above publications [27–30] associating its loss with malignancy and clinically aggressive sarcomas. miR-195 was reported to be overexpressed in the serum of osteosarcoma patients compared to controls [32]. In the above report by Namløs et al., miR-195 levels were intermediate in osteosarcoma compared to
osteoblasts and normal bone, as were those of miR-31, though with highest levels in normal bone [26]. Fricke and co-workers found higher levels of miR-195-5p in blood from synovial sarcoma patients compared to both healthy controls and patients with other sarcomas. Results were validated for the comparative analysis with controls for 7 miRNAs using qRT-PCR, though miR-195-5p was not overexpressed in synovial sarcoma compared to LMS (n = 5; organ of origin not specified) [33]. As we did not have an ESS cell line at our lab, we chose to focus on the Wnt pathway, regulated by several of the miRNAs found to be differentially expressed in primary vs. metastatic LMS. Wnt proteins are secreted glycoproteins that bind to receptors of the Frizzled family and several co-receptors. In humans, 19 members of the Wnt family and 10 Frizzled receptors are known. Based on the different biological readouts, Wnt ligands and Frizzled receptors are divided into canonical and non-canonical pathways. The canonical pathway involves the
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001
6
Y. Ravid et al. / Gynecologic Oncology xxx (2016) xxx–xxx
accumulation of β-catenin and its function as transcription factor together with TCF/LEF upon its transport into the nucleus. The noncanonical pathway involves the activation of rho, cdc42 and JNK. These pathways regulate central cancer-associated functions, including cell growth, migration, invasion and metastasis [34]. Differential expression of Frizzled-6 was found in primary and metastatic LMS, with opposite findings at the mRNA and protein level. The discrepancy between mRNA expression and protein expression of this receptor may be due to ceRNA regulation, though further research is needed to elucidate this finding. In vitro experiments with LMS cells further showed that FZD6 inhibition by siRNA inhibits cellular invasion, wound closure and MMP-2 activity, suggesting a tumor-promoting role in this cancer. In conclusion, ESS and LMS have unique miRNA signatures, with differences for primary vs. metastatic LMS being of lesser degree. The differences in miRNA profiles between LMS and ESS are likely to affect cellular regulatory pathways and result in unique phenotype for each of these cancers. These in turn may form the basis for identifying new diagnostic and therapeutic candidates in each tumor. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2016.01.001. Financial acknowledgment This work was supported by a grant from the National Sarcoma Foundation at the Norwegian Radium Hospital. Conflict of interest statement The authors declare that they have no competing interests.
References [1] J.R. Toro, L.B. Travis, H.J. Wu, et al., Incidence patterns of soft tissue sarcomas, regardless of primary site, in the surveillance, epidemiology and end results program, 1978–2001: an analysis of 26,758 cases, Int. J. Cancer 119 (2006) 2922–2930. [2] E. D'Angelo, J. Prat, Uterine sarcomas: a review, Gynecol. Oncol. 116 (2010) 131–139. [3] V.M. Abeler, O. Røyne, S. Thoresen, et al., Uterine sarcomas in Norway. A histopathological and prognostic survey of a total population from 1970 to 2000 including 419 patients, Histopathology 54 (2009) 355–364. [4] R.J. Kurman, M.L. Carcangiu, C.S. Herrington, et al., WHO Classification of Tumours of Female Reproductive Organs, IARC, Lyon, France, 2014. [5] B.R. Cullen, Transcription and processing of human microRNA precursors, Mol. Cell 16 (2004) 861–865. [6] J. Permuth-Wey, Z. Chen, Y.Y. Tsai, et al., Ovarian Cancer Association Consortium (OCAC).microRNA processing and binding site polymorphisms are not replicated in the Ovarian Cancer Association Consortium, Cancer Epidemiol. Biomarkers Prev. 20 (2011) 1793–1797. [7] A.J. Pratt, I.J. MacRae, The RNA-induced silencing complex: a versatile gene-silencing machine, J. Biol. Chem. 284 (2009) 17897–17901. [8] A. Grimson, K.K. Farh, W.K. Johnston, P. Garrett-Engele, L.P. Lim, D.P. Bartel, MicroRNA targeting specificity in mammals: determinants beyond seed pairing, Mol. Cell 27 (2007) 91–105.
[9] G. Shi, M.A. Perle, K. Mittal, H. Chen, X. Zou, M. Narita, E. Hernando, P. Lee, J.J. Wei, Let-7 repression leads to HMGA2 overexpression in uterine leiomyosarcoma, J. Cell. Mol. Med. 13 (2009) 3898–3905. [10] L.S. Danielson, S. Menendez, C.S. Attolini, et al., A differentiation-based microRNA signature identifies leiomyosarcoma as a mesenchymal stem cell-related malignancy, Am. J. Pathol. 177 (2010) 908–917. [11] Q. Zhang, J. Ubago, L. Li, et al., Molecular analyses of 6 different types of uterine smooth muscle tumors: emphasis in atypical leiomyoma, Cancer 120 (2014) 3165–3177. [12] M. Kowalewska, E. Bakula-Zalewska, M. Chechlinska, et al., microRNAs in uterine sarcomas and mixed epithelial–mesenchymal uterine tumors: a preliminary report, Tumour Biol. 34 (2013) 2153–2160. [13] B. Davidson, V.M. Abeler, E. Hellesylt, et al., Gene expression signatures differentiate uterine endometrial stromal sarcoma from leiomyosarcoma, Gynecol. Oncol. 128 (2013) 349–355. [14] B. Davidson, V.M. Abeler, M. Førsund, et al., Gene expression signatures of primary and metastatic uterine leiomyosarcoma, Hum. Pathol. 45 (2014) 691–700. [15] R. Reich, E.W. Thompson, Y. Iwamoto, et al., Inhibition of plasminogen activator, serine proteinases and collagenase IV prevents the invasion of basement membrane by metastatic cells, Cancer Res. 48 (1988) 3307–3312. [16] F. Moinfar, M. Azodi, F.A. Tavassoli, Uterine sarcomas, Pathology 39 (2007) 55–71. [17] E. Oliva, Cellular mesenchymal tumors of the uterus: a review emphasizing recent observations, Int. J. Gynecol. Pathol. 33 (2014) 374–384. [18] B. Davidson, M.L. Kjæreng, M. Førsund, H.E. Danielsen, G.B. Kristensen, V.M. Abeler, Progesterone receptor expression is an independent prognosticator in FIGO stage I uterine leiomyosarcoma, Am. J. Clin. Pathol. (2015) (in press). [19] Y. Pekarsky, C.M. Croce, Role of miR-15/16 in CLL, Cell Death Differ. 22 (2015) 6–11. [20] L. Hong, Y. Han, Y. Zhang, et al., MicroRNA-21: a therapeutic target for reversing drug resistance in cancer, Expert Opin. Ther. Targets 17 (2013) 1073–1080. [21] F.J. Sheedy, Turning 21: induction of miR-21 as a key switch in the inflammatory response, Front. Immunol. 6 (2015) 19. [22] C. Gong, Y. Nie, S. Qu, et al., miR-21 induces myofibroblast differentiation and promotes the malignant progression of breast phyllodes tumors, Cancer Res. 74 (2014) 4341–4352. [23] W. Ziyan, Y. Shuhua, W. Xiufang, L. Xiaoyun, MicroRNA-21 is involved in osteosarcoma cell invasion and migration, Med. Oncol. 28 (2011) 1469–1474. [24] L. Ouyang, P. Liu, S. Yang, S. Ye, W. Xu, X. Liu, A three-plasma miRNA signature serves as novel biomarkers for osteosarcoma, Med. Oncol. 30 (2013) 340. [25] Y. He, C. Meng, Z. Shao, H. Wang, S. Yang, MiR-23a functions as a tumor suppressor in osteosarcoma, Cell. Physiol. Biochem. 34 (2014) 1485–1496. [26] H.M. Namløs, L.A. Meza-Zepeda, T. Barøy, et al., Modulation of the osteosarcoma expression phenotype by microRNAs, PLoS ONE 7 (2012), e48086. [27] J. Ban, G. Jug, P. Mestdagh, et al., Hsa-mir-145 is the top EWS-FLI1-repressed microRNA involved in a positive feedback loop in Ewing's sarcoma, Oncogene 30 (2011) 2173–2180. [28] I.W. Mak, S. Singh, R. Turcotte, M. Ghert, The epigenetic regulation of SOX9 by miR145 in human chondrosarcoma, J. Cell. Biochem. 116 (2015) 37–44. [29] C.M. Gits, P.F. van Kuijk, M.B. Jonkers, et al., MicroRNA expression profiles distinguish liposarcoma subtypes and implicate miR-145 and miR-451 as tumor suppressors, Int. J. Cancer 135 (2014) 348–361. [30] S. Ugras, E. Brill, A. Jacobsen, et al., Small RNA sequencing and functional characterization reveals MicroRNA-143 tumor suppressor activity in liposarcoma, Cancer Res. 71 (2011) 5659–5669. [31] D. Jonigk, N. Izykowski, L. Maegel, et al., MicroRNA expression in Epstein–Barr virusassociated post-transplant smooth muscle tumours is related to leiomyomatous phenotype, Clin. Sarcoma Res. 3 (2013) 9. [32] H. Cai, H. Zhao, J. Tang, H. Wu, Serum miR-195 is a diagnostic and prognostic marker for osteosarcoma, J. Surg. Res. 194 (2015) 505–510. [33] A. Fricke, P.V. Ullrich, J. Heinz, et al., Identification of a blood-borne miRNA signature of synovial sarcoma, Mol. Cancer 14 (2015) 151. [34] K. Ueno, H. Hirata, Y. Hinoda, R. Dahiya, Frizzled homolog proteins, microRNAs and Wnt signaling in cancer, Int. J. Cancer 132 (2013) 1731–1740.
Please cite this article as: Y. Ravid, et al., Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures, Gynecol Oncol (2016), http://dx.doi.org/10.1016/j.ygyno.2016.01.001