C291 Davidson thematic crop

Molecular differentiators of uterine leiomyosarcoma and endometrial stromal sarcoma

Leiomyosarcoma and endometrial stromal sarcoma are the most common types of uterine sarcoma, a group of rare and clinically aggressive mesenchymal cancers. These two sarcomas may have overlapping clinical presentation, morphology and protein expression profiles, making their diagnosis occasionally difficult. This article discusses molecular approaches that may be applied to the diagnosis of these two cancers and may generate data expanding our therapeutic options and patient outcome.

by Professor Ben Davidson

The majority of cancers affecting the uterine corpus are carcinomas, i.e. tumours of epithelial origin. Uterine sarcomas, tumours that are of mesenchymal origin, are a group of rare and clinically aggressive tumours constituting 7 % of all soft tissue sarcomas and 3 % of malignant uterine tumours [1, 2]. The most common entities within this group are leiomyosarcoma (LMS) and endometrial stromal sarcoma (ESS) [2, 3]. Although LMS and ESS are readily diagnosed based on morphology and a limited immunohistochemistry (IHC) panel in many cases, some tumours may pose diagnostic difficulty, and currently used antibodies are not 100 % sensitive or specific [4]. Improved understanding of the molecular make-up of these tumours may lead to more accurate diagnosis and better understanding of their biology, eventually improving our ability to design targeted therapy approaches with the objective of improving patient outcome.

The genetic make-up of ESS and LMS
Low-grade ESS, the more common type of ESS, is characterized by several gene rearrangements creating fusion genes, of which the first described was fusion of the zinc finger gene 1 JAZF1, located at 7p15, and JJAZ1, also termed SUZ12, at 17q21 through a 7;17-translocation. Other fusions in low-grade ESS include the one between JAZF1 and the PHD finger protein 1 gene (PHF1) in 6p21, as well as between PHF1 and enhancer of polycomb homologue 1 (EPC1) gene at 10p11 and the MYST/Esa1 associated factor 6 gene (MEAF6) at 1p34. X chromosome rearrangements include fusion of the open reading frame CXorf67 and the BCL-6 interacting corepressor (BCOR) gene, both at Xp11, with the MBT domain-containing protein 1 gene (MBTD1) at 17q21 and with the zinc finger CCCH-type containing 7B gene (ZC3H7B) at 22q13, respectively.

High-grade ESS is characterized by a fusion between the tyrosine 3/tryptophan 5 monooxygenase gene (YWHAE) gene at 17p13 and the NUT family member gene (NUTM2; previously known as FAM22) at 10q22, creating YWHAE-NUTM fusion through a 10;17-translocation (reviewed by Davidson and Micci, invited review submitted to Expert Rev Mol Diagn). These alterations were recently confirmed by analysis of the ESS transcriptome and/or whole-exome sequencing, including the application of next generation sequencing [5–7].

The body of data with respect to the molecular characteristics of LMS is more limited. An observation found in several studies is the presence of exon 2 mutations in the mediator complex subunit 12 (MED12) gene on chromosome band Xq13.1 in some LMS. MED12 protein forms complex with MED13, cyclin-dependent kinase 8 (CDK8), and cyclin C, termed the CDK8 submodule of the Mediator, the mediator being a large multiprotein complex regulating transcription [8]. Though less frequent in LMS compared to leiomyomas, the benign counterpart of LMS, this finding appears to be absent in other malignant soft tissue sarcomas, and is rare in carcinomas, and is thus potentially relevant in the diagnostic setting (reviewed by Croce & Chibon [9]).

RNA sequencing of 99 LMS, of which 49 were uterine, identified 3 distinct molecular subtypes. Leiomodin (LMOD1) and ADP-ribosylation factor-like 4C (ARL4C) were found to be markers for type I and II tumours, respectively, and the latter group was associated with poor prognosis when located in the uterus [10].

Comparative molecular analysis of ESS and LMS
Our group performed two studies of uterine LMS and ESS with the aim of identifying novel biomarkers that may expand the arsenal of markers currently used in diagnosing these tumours, as well as improving our understanding of their unique biology.

In the first study, the gene expression profiles of 7 ESS and 13 LMS were compared using the HumanRef-8 BeadChip from Illumina. We identified 549 unique probes that were significantly differentially expressed in the two tumour entities, of which 336 and 213 were overexpressed in ESS and LMS, respectively. Genes found to be overexpressed in ESS included CCND2, ECEL1, ITM2A, NPW, SLC7A10, EFNB3, PLAG1 and GCGR, whereas genes overexpressed in LMS included FABP3, TAGLN, CDKN2A, JPH2, GEM, NAV2 and RAB23. qPCR analysis confirmed these differences for 14 of 16 genes selected for validation. Five protein products were selected for validation by IHC, including the LMS markers fatty acid binding protein (FABP3), transgelin (TAGLN) and neuron navigator 2 (NAV2) and the ESS markers cyclin D2 (CCND2) and integral membrane protein 2A (ITM2A). All were found to be significantly differentially expressed in LMS vs ESS (Fig. 1) [11]. Data for FABP3, TAGLN, NAV2 and CCND2 were recently confirmed in a large (approx. 350 tumours) uterine sarcoma series [Davidson et al., manuscript submitted].

Recently, we compared the microRNA (miRNA) profiles of primary ESS (n=9), primary LMS (n=8) and metastatic LMS (n=8) using Taqman Human miRNA Array Cards. Ninety-four miRNAs were significantly differentially expressed in ESS vs LMS, of which 76 and 18 were overexpressed in ESS and LMS, respectively. Forty-nine miRNAs were differentially expressed in primary and metastatic LMS, among which 45 and 4 were overexpressed in primary and metastatic LMS, respectively. Twenty miRNAs found to be most significantly differentially expressed in primary ESS vs LMS or in primary vs metastatic LMS were further studied in a validation series of 44 tumours using qPCR. Of these, 10 were confirmed to be differentially expressed in these groups, including overexpression of 7 miRNAs (mir-15b, mir-21, mir-23b, mir-25, mir-145, mir-148b and mir-195) in ESS compared to primary LMS. The remaining 3 differentially expressed miRNAs were in comparative analysis of primary and metastatic LMS (lower mir-15a and mir-92a levels and higher mir-31 levels in primary LMS). Differentially expressed miRNA regulated the mitogen-activated protein kinase (MAPK) signaling pathway, Wnt signaling, focal adhesion, the mTOR signaling pathway and the transforming growth factor-β (TGF-β) signaling pathway. As Wnt signaling pathway genes are controlled by miRNAs 15a, 31 and 92a in LMS, we looked at the biological role of Frizzled-6 in LMS cells and found that Frizzled-6 silencing by siRNA significantly inhibited cellular invasion, wound closure and matrix metalloproteinase (MMP-2) activity [12]

Conclusion and future perspectives
Recent years have brought about considerable progress in our understanding of the molecular events occurring in ESS and LMS. Our studies and data from other groups may aid in the diagnosis and classification of these cancers, hopefully providing rationale for targeted therapy. Uterine sarcomas express different cancer-related molecules that may be targeted (reviewed by Cuppens et al. [13]). Anti-hormonal treatment is used in patients with hormone receptor-positive tumours, and expression of progesterone receptor was recently shown to be a prognostic marker in stage I LMS [14]. In two studies, targeting of mTOR, Aurora kinases and other mitotic checkpoint regulators has been suggested as therapeutic modality in LMS [15,16]. Additional studies are likely to identify new relevant targets in the future, hopefully improving the outcome of uterine sarcoma patients.

The work of Dr Davidson is supported by the National Sarcoma Foundation at the Norwegian Radium Hospital.

1. Toro JR, Travis LB, Wu HJ, Zhu K, Fletcher CD, Devesa SS. 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 2006; 119: 2922–2930.
2. D’Angelo E, Prat J. Uterine sarcomas: a review. Gynecol Oncol. 2010; 116: 131–139.
3. Kurman RJ, Carcangiu ML, Herrington CS, Young RH (Eds.). WHO classification of tumours of female reproductive organs. IARC 2014.
4. Abeler VM, Nenodovic M. Diagnostic immunohistochemistry in uterine sarcomas: a study of 397 cases. Int J Gynecol Pathol. 2011; 30: 236–243.
5. Micci F, Gorunova L, Agostini A, Johannessen LE, Brunetti M, Davidson B, Heim S, Panagopoulos I. Cytogenetic and molecular profile of endometrial stromal sarcoma. Genes Chromosomes Cancer 2016; 55: 834–846.
6. Choi YJ, Jung SH, Kim MS, Baek IP, Rhee JK, Lee SH, Hur SY, Kim TM, Chung YJ, Lee SH. Genomic landscape of endometrial stromal sarcoma of uterus. Oncotarget 2015; 6: 33319–33328.
7. Li X, Anand M, Haimes JD, Manoj N, Berlin AM, Kudlow BA, Nucci MR, Ng TL, Stewart CJ, Lee CH. The application of next-generation sequencing-based molecular diagnostics in endometrial stromal sarcoma. Histopathology 2016; 69: 551–559.
8. Clark AD, Oldenbroek M, Boyer TG. Mediator kinase module and human tumorigenesis. Crit Rev Biochem Mol Biol. 2015; 50: 393–426.
9. Croce S, Chibon F. MED12 and uterine smooth muscle oncogenesis: state of the art and perspectives. Eur J Cancer 2015; 51: 1603–1610.
10. Guo X, Jo VY, Mills AM, Zhu SX, Lee CH, Espinosa I, Nucci MR, Varma S, Forgó E, Hastie T, Anderson S, Ganjoo K, Beck AH, West RB, Fletcher CD, van de Rijn M. Clinically relevant molecular subtypes in leiomyosarcoma. Clin Cancer Res. 2015; 21: 3501–3511.
11. Davidson B, Abeler VM, Hellesylt E, Holth A, Shih IeM, Skeie-Jensen T, Chen L, Yang Y, Wang TL. Gene expression signatures differentiate uterine endometrial stromal sarcoma from leiomyosarcoma. Gynecol Oncol. 2013; 128: 349–355.
12. Ravid Y, Formanski M, Smith Y, Reich R, Davidson B. Uterine leiomyosarcoma and endometrial stromal sarcoma have unique miRNA signatures. Gynecol Oncol. 2016; 140: 512–517.
13. Cuppens T, Tuyaerts S, Amant F. Potential therapeutic targets in uterine sarcomas. Sarcoma 2015; 2015: 243298.
14. Davidson B, Kjæreng ML, Førsund M, Danielsen HE, Kristensen GB, Abeler VM.. Progesterone receptor expression is an independent prognosticator in FIGO stage I uterine leiomyosarcoma. Am J Clin Pathol. 2016; 145: 449–458. 
15. Brewer Savannah KJ, Demicco EG, Lusby K, Ghadimi MP, Belousov R, Young E, Zhang Y, Huang KL, Lazar AJ, Hunt KK, Pollock RE, Creighton CJ, Anderson ML, Lev D.. Dual targeting of mTOR and aurora-A kinase for the treatment of uterine leiomyosarcoma. Clin Cancer Res. 2012; 18: 4633–4645.
16. Shan W, Akinfenwa PY, Savannah KB, Kolomeyevskaya N, Laucirica R, Thomas DG, Odunsi K, Creighton CJ, Lev DC, Anderson ML. A small-molecule inhibitor targeting the mitotic spindle checkpoint impairs the growth of uterine leiomyosarcoma. Clin Cancer Res. 2012; 18: 3352–3365.

The author
Ben Davidson1,2 MD, PhD
1Department of Pathology, Norwegian Radium Hospital, Oslo University Hospital, N-0310 Oslo, Norway
2University of Oslo, Faculty of Medicine, Institute of Clinical Medicine, N-0316 Oslo, Norway

*Corresponding author
E-mail: bend@medisin.uio.no