Transcriptomic Definition of Molecular Subgroups of Small Round Cell

Transcriptomic definition of molecular subgroups of small round cell sarcomas Sarah Watson, Virginie Perrin, Delphine Guillemot, Stéphanie Reynaud, Jean-Michel Coindre, Marie Karanian, Jean-Marc Guinebretière, Paul Fréneaux, Francois Le Loarer, Megane Bouvet, et al.

To cite this version:

Sarah Watson, Virginie Perrin, Delphine Guillemot, Stéphanie Reynaud, Jean-Michel Coindre, et al.. Transcriptomic definition of molecular subgroups of small round cell sarcomas. The Journal of pathology and bacteriology, John Wiley & Sons, 2018, 245 (1), pp.29-40. 10.1002/path.5053. inserm-02440510

HAL Id: inserm-02440510 https://www.hal.inserm.fr/inserm-02440510 Submitted on 15 Jan 2020

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Title: Transcriptomic definition of molecular subgroups of small round cell sarcomas

Running Title: Molecular classification of sarcoma subtypes

Authors Sarah Watson1,2, Virginie Perrin1,2, Delphine Guillemot3, Stephanie Reynaud3, Jean-Michel Coindre4,5, Marie Karanian6, Jean-Marc Guinebretière7, Paul Freneaux8, François Le Loarer4,5, Megane Bouvet3, Louise Galmiche-Rolland9,10, Frédérique Larousserie11, Elisabeth Longchampt12, Dominique Ranchere-Vince6, Gaelle Pierron3*, Olivier Delattre1,2,3,13*, Franck Tirode1,2,14* *Co-senior authors

Affiliations 1 INSERM U830, Laboratory of Genetics and Biology of Cancer, F-75005, Paris, France 2 Institut Curie, Paris Sciences et Lettres, F-75005, Paris, France 3 Institut Curie, Unité de génétique somatique, F-75005, Paris, France 4 Institut Bergonié, Department of Pathology, F-33000, Bordeaux, France. 5 Université Bordeaux 2, F-33000, Bordeaux, France. 6 Centre Leon Bérard, Department of Pathology, F-69008, Lyon, France. 7 Service de Pathologie, Hôpital René-Huguenin, Institut Curie, F-92210, Saint-Cloud, France. 8 Département de Biologie des Tumeurs, Institut Curie, Service d'anatomie pathologique, F- 75005, Paris, France 9 Service d’Anatomie Pathologique, Hôpital Necker Enfants malades, F-75015, Paris, France 10 Université Paris Descartes, F-75006, Paris, France 11 Service d'Anatomie Pathologique, Hôpital Cochin, F-75014, Paris, France 12 Service d'Anatomie et de Cytologie Pathologiques, Hôpital Foch, F-92151, Suresnes, France 13 Ligue Contre le Cancer, Equipe labellisée 14 Univ Lyon, Université Claude Bernard Lyon 1, INSERM 1052, CNRS 5286, Cancer Research Center of Lyon, Centre Léon Bérard, F-69008, Lyon, France

Corresponding authors: Franck Tirode, Ph.D. Biology of Rare Sarcomas Group Cancer Research Center of Lyon, INSERM U1052 Centre Léon Bérard, 28 Rue Laënnec, 69373 Lyon Cedex 08 Tel.: +33 4 69 85 61 45 Email: [email protected]

Olivier Delattre, M.D., Ph.D. INSERM U830 “Genetics and Biology of Cancers” Institut Curie Research Centre 26 rue d’Ulm, 75248 Paris cedex 05, France Tel.: +33 1 56 24 66 79 Fax.: +33 1 56 24 66 30 Email: [email protected] Conflict of Interest statement:

Authors have nothing to disclose Watson et al. 1

Word count: 3929

Abstract

Sarcoma represents a highly heterogeneous group of tumours. We report here the first unbiased and systematic search for gene fusions combined with unsupervised expression analysis of a series of 184 small round cell sarcomas. Fusion genes were detected in 59% percent of samples, with half of them being observed recurrently. We identified biologically homogeneous groups of tumours such as the CIC-fused (to DUX4, FOXO4 or NUTM1) and

BCOR-rearranged (BCOR-CCNB3, BCOR-MAML3, ZC3H7B-BCOR and BCOR internal duplication) tumour groups. VGLL2-fused tumours represented a more biologically and pathologically heterogeneous group. This study also refined the characteristics of some entities such as EWSR1-PATZ1 spindle cell sarcoma or FUS-NFATC2 bone tumours that are different from EWSR1-NFATC2 tumours and transcriptionally resemble CIC-fused tumour entities. We also describe a completely novel group of epithelioid and spindle-cell rhabdomyosarcomas characterized by EWSR1- or FUS-TFCP2 fusions. Finally, expression data identified some potentially new therapeutic targets or pathways.

Keywords

Sarcoma, Fusion genes, FET-TFCP2, FUS-NFATC2, EWSR1-PATZ1, VGLL2-NCOA2,

BCOR-Rearranged, CIC-Fused, RNAseq.

Introduction

Small round cell sarcoma is a heterogeneous group of tumours mostly affecting children and young adults and characterized by an overall poor prognosis [1]. They remain challenging for

Watson et al. 2

pathologists since those tumours share overlapping morphological and immunophenotypical features. The identification of specific fusion transcripts has considerably improved their diagnosis as many subtypes are characterized by a specific fusion gene, such as

EWSR1/FUS-ETS, SS18-SSX and PAX-FOXO1 fusions in Ewing sarcoma [2], synovial sarcoma [3] and aRMS [4], respectively, or the more recently described CIC-DUX4 [5],

EWSR1-NFATc2 [6] and BCOR-CCNB3 [7] in “Ewing-like” tumours. As a result, diagnostic evaluation of specific chromosome translocations by FISH or of the resulting fusion transcripts by RT-PCR has become an asset for the molecular assessment of small round cell sarcoma. However, a number of sarcomas remains uncharacterised, raising both diagnostic and therapeutic issues.

In the present study, we report RNA-sequencing of 184 small round cell sarcoma samples from three different cohorts. By applying several fusion algorithms and performing unsupervised clustering analysis, we were able to show that recurrent fusions define homogeneous groups of tumours based on pathological features and expression profile analyses, including FUS-NFATC2-positive, EWSR1-PATZ1-positive, CIC-fused or BCOR- rearranged tumours. We also identified a new epithelioid rhabdomyosarcoma group characterized by a EWSR1- or FUS-TFCP2 fusion gene. Finally, we propose genes and pathways specifically expressed in a variety of different entities that may serve as biomarkers or potential therapeutic targets.

Materials and Methods

Tumour samples. Tumour samples were chosen based on their pathological diagnosis either as unclassified sarcoma or as known sarcoma that did not carry pathognomonic genetic aberrations or suspicious for sarcoma with simple genetic or monomorphic features

Watson et al. 3

(supplementary material, Figure S1). Samples were used in accordance with the French

Biobank legislation.

Paired-end RNA sequencing. Total RNAs were isolated from crushed frozen tumours using a

Trizol reagent kit (Life technologies). Library constructions were performed following the

TruSeq Stranded mRNA LS protocol (Illumina). Sequencing were performed on either HiSeq

2500 (100nt paired-end) or NextSeq 500 (150nt paired-end) Illumina sequencing machines.

All fastq files were deposited on the European Genome-phenome Archive

(https://www.ebi.ac.uk/ega/studies/EGAS00001002189). Raw sequencing files of

SMARCA4-DTS and MRT control samples were previously deposited in the Sequence Read

Archive (#SRP052896).

Bioinformatics analyses. For fusion gene discovery, sequencing reads were injected into both the deFuse tools [8] and the FusionMap tool [9] with hg19 genome as reference. Only fusion transcripts supported by both tools with at least 2 split reads (and 3 spanning pairs for defuse) were considered. Gene expression values were extracted by Kallisto v0.42.5 [10] with GRCh38 release 79 genome annotation. Clustering, BGA, PCA and t-SNE were computed using R packages Cluster v2.0.3, made4 version 1.44.0, FactoMiner v1.31.4 and

Rtsne version 0.10, respectively. Gene ontology analyses were performed online

(https://david.ncifcrf.gov/) with DAVID v6.7 tool [11].

Fusion validation: PCR amplification of the fusion point using specifically designed primers and 50ng of cDNA were then performed with AmpliTaq Gold® DNA Polymerase with Buffer II

(Thermo Fisher Scientific, Waltham, MA USA 02451) prior to Sanger sequencing on an

Applied 3500XL Genetic Analyzer.

Watson et al. 4

Immunohistochemistry. Immunohistochemistry was performed using the following antibodies and methods: NUTM1 (C52B1, Cell Signaling, dilution 1:50) and ALK (5A4, Cell Signaling, dilution 1:50) IHC were performed on a Ventana BenchMark platform with Cell Conditioning

Solution 1 pre-treatment. NUTM1 was incubated for 56min and ALK for 80min prior to be revealed with the ultraView Universal DAB Detection Kit. ETV4 (clone 16, Santa Cruz

Biotechnology, dilution 1:15) IHC was performed overnight at 4°C, following a pre-heating at

98°C for 30 min in Trisbuffer pH9, and revealed with REAL EnVision kit (Dako, Glostrup,

Denmark).

Results

In molecular diagnostic routine, primitive small round or spindle cell sarcomas are systematically tested for the most common fusion genes using multiplexed Q-PCR

(supplementary material, Table S1). In the present study, we performed whole transcriptome sequencing to investigate its efficacy for proper tumour classification. Our investigation cohort was composed of 94 samples randomly selected from about 700 retrospective samples in which no common fusion gene could be identified by standard diagnostic investigation (see the general scheme of the analysis in supplementary material, Figure S1).

We first investigated the presence of gene fusions and the retained fusion candidates were subsequently validated and searched for within the remaining available samples (around 600 cases) using specific RT-PCR assay. The new cases thus identified were also RNA- sequenced and generated our follow-up cohort (12 samples). Basic clinical data, initial diagnosis and results of molecular analyses for all the 184 cases sequenced in this study are summarized in the supplementary material, Table S2. We next performed expression profile analyses that we compared using t-Distributed Stochastic Neighbour Embedding [12] and unsupervised clustering analyses (Figure 1A and B, supplementary material, Figure S2) to a Watson et al. 5

subset of 78 well defined cases composing our control cohort. For each group, differential expression analyses against each of all other tumour types were performed. Top variant genes are reported in the supplementary material, Table S3 together with ontology enrichment and gene set enrichment analyses (GSEA).

A fusion gene could be identified in almost 3/5 of the investigation cohort samples

We identified fusion genes in 55 out of the 94 tumours of the investigation cohort. Except for one case (SARC036) for which numerous fusions were found on chromosome 1, suggestive of chromothripsis, the mean number of fusion events detected per sample was 1.8. Forty samples carried a single fusion gene, while multiple (2 to 9) gene fusions were detected in 14 samples (supplementary material, Table S2). All fusion genes were subsequently confirmed by RT-PCR and Sanger sequencing.

Most of the RNAseq fusion-positive samples (26/55) expressed previously described fusion genes that were not included in our RT-PCR routine procedure, namely single cases of each of the following fusions: EWSR1-PBX1, ACTB-GLI1, PAX3-MAML3, COL1A1-PDGFB,

VGLL2-CITED2, FUS-NFATc2, BCOR-MAML3, ZC3H7B-BCOR, TPR-NTRK1, BRD3-

NUTM1, KIAA1549-BRAF; two cases with EWSR1-PATZ1, TPM3-NTRK1, EML4-ALK or

NAB2-STAT6 fusions; and three cases with VGLL2-NCOA2 or CIC-NUTM1 fusions. We also identified a variant of a SS18-SSX2 fusion with an atypical SS18 gene breakpoint. RT-PCR screening of these fusions in the initial cohort led to the identification of eleven additional cases: 3 EWSR1-PATZ1, 1 EML4-ALK, 2 FUS-NFATC2, 3 VGLL2-NCOA2 and 2 CIC-

NUTM1 (supplementary material, Table S2).

Novel fusions were detected in 29 samples (supplementary material, Table S2). Two cases presented variants of known fusions: UXT-TFE3, a variant of the ASPSCR1- or SFPQ-TFE3 fusions found in alveolar soft part sarcomas and renal cell carcinoma; and IKBKG-ALK, a Watson et al. 6

new variant among the numerous ALK fusions found in inflammatory myofibroblastic tumours. In two cases, we identified a completely new fusion between EWSR1 or FUS and

TFCP2. Eleven samples harboured previously undescribed fusions, involving genes relevant for tumorigenesis, but for which no additional sample could be identified during the RT-PCR screen. These fusions were then considered as private to their respective tumour. Finally, in the remaining 14 fusion-positive cases, we were unable to propose a definitive driver event due to our lack of knowledge on the implication in cancer of the fused genes.

Emerging VGLL2-NCOA2/CITED, CIC-fused and BCOR-rearranged sarcoma subtypes

A total of 7 samples harbouring a VGLL2 fusion with either NCOA2 (n=6) or CITED (n=1) were identified (supplementary material, Figure S3A) with two samples forming a primary/relapse couple (SARC070_Primary and SARC070_Relapse). These fusion genes characterized tumours occurring in very young children (below 5yo), as previously described

[13]. Centralized review of cases highlighted two subtypes with specific pathological aspects.

In three cases (SARC061, SARC065 and SARC070_primary), the tumours were composed of a large amount of fibrous stroma with scarce tumour cells (Figure 2A). Tumour cells presented neither atypia nor abnormal mitotic figures. They were negative for desmin, whereas few cells (below 1%) showed a nuclear positivity for myogenin. Fewer than 25% of cells were Ki67 positive. In contrast, in the four other cases (SARC070 at relapse, SARC085,

SARC088 and SARC102), the cellular mass was far denser with cells presenting numerous atypia and mitoses. Myogenin staining was strongly positive in around 10% of cells and desmin and Ki67 immuno-reactivity was displayed by over 30% of tumour cells (Figure 2B).

Consistently with this histological heterogeneity, t-SNE analysis and unsupervised clustering revealed a relatively disparate group of tumours (Figure 1A and supplementary material,

Figure S2A) which samples could further be separated according to the histology, when

Watson et al. 7

different clustering conditions were applied (supplementary material, Figure S2C-D). When considering these two histological subtypes separately it appeared that the “fibrous” subtype was enriched in immune/inflammatory response genes, while the “dense” subtype was enriched in cell cycle/proliferation genes (supplementary material, Table S3). Supervised group comparisons and between group analyses (BGA) indicated that VGLL2-fused tumours expressed numerous muscle-related genes as well as genes involved in extracellular matrix and in the epithelial-mesenchymal transition process. Despite clear positivity for muscle differentiation markers, none of the two subtypes clustered with rhabdomyosarcoma samples.

A total of five CIC-NUTM1 (supplementary material, Figure S3B) fusions were retrieved. All but one case were observed in young children, at various locations (supplementary material,

Table S2). Tumours harbouring CIC-NUTM1 fusion clustered tightly with the other CIC-

DUX4- or -FOXO4-positive samples (Figure 1). All these tumours overexpressed ETS family members of the PEA3 type [5,14,15] as well as a variety of secreted and matrix protein genes such as VGF, BMP2, glypican 3/4, NRG1, PTX3, pleiotrophin and spondins. GSEA revealed enrichment in genes of extracellular matrix but also in genes overlapping proliferation/response to drug/immune response categories (supplementary material, Table

S3).

The seven samples presenting a rearrangement of BCOR, including BCOR-MAML3 or

ZC3H7B-BCOR fusions [16] and internal tandem duplication (ITD) [17–19] (supplementary material, Figure S3C), grouped together with the BCOR-CCNB3-positive samples from the control cohort, describing a very-well defined cluster. These samples also shared some pathological features, as previously described [17]. Among the most significantly enriched gene ontologies for this BCOR-rearranged group were “developmental protein” and

Watson et al. 8

“homeobox” (supplementary material, Table S3) with a strong overexpression of HOX-A, -B,

-C and -D family genes as well as HMX1, PITX1, ALX4 or DLX1. More specifically, we observed very strong gene sets and ontology enrichments for the morphogenesis, development and differentiation of neurons and for the skeletal system development. Finally, different genes encoding membrane receptors including RET, FGFR2/3, EGFR, PDGFRA,

NTRK3, KIT or NGFR were also highly and constantly overexpressed in these tumours and may hence constitute actionable target genes.

Identification of a FUS-NFATC2 fusion gene in a new bone tumour entity transcriptionally distinct from EWSR1-NFATC2-positive tumours

We identified FUS-NFATC2 fusion genes in three tumours of the femur of adult patients

(median age 38.3yo), like the only case described in the literature (17). Tumours displayed areas composed of round tumour cells arranged in sheets or short fascicles. Tumour cells were embedded in a variably myxoid stroma (Figure 3A). Focal hemangiopericytic vascular network was present in all cases (Figure 3B). More distinctively, these tumours harboured focal myxohyaline foci raising suspicion of cartilaginous differentiation (Figure 3C). All tumours displayed brisk mitotic activity and necrosis. Altogether, these tumours were histologically reminiscent of CIC-fused sarcomas. Clustering analyses indicated that FUS-

NFATC2-positive tumours grouped together and were clearly distinct from all other FET- fused samples, including EWSR1-NFATC2-positive tumours (Figure 1). While the EWSR1-

NFATC2 tumours were strongly enriched in genes associated with inflammatory and immune responses (supplementary material, Tables S3 and S4), the FUS-NFATC2 tumours were enriched in proliferation and drug resistance signatures. In accordance with the potential areas of cartilaginous differentiation, genes involved in the extracellular matrix of cartilaginous tissues (like ACAN, COL9A2, MATN3, COMP or CILP2) or encoding secreted

Watson et al. 9

proteins (pleiotrophin, DKK3, ANGPTL2, SBSPON or WNT5B) were preferentially expressed in FUS-NFATC2-positive tumours.

EWSR1-PATZ1 fusion gene is found in a new tumour group unrelated to any other

EWSR1-fused tumour.

While only single cases of tumour carrying an EWSR1-PATZ1 fusion have been previously reported [21–23], we identified in our cohorts a total of 5 cases (Figure 4A and supplementary material, Figure S3D). All tumours were from soft tissues and occurred across a very broad age range (from 0.9 to 68.5yo). Centralized pathological review of three cases indicated three different morphological aspects (Figure 4B): i) A first tumour was composed of bundles of relatively monomorphic spindle cells with an eosinophilic cytoplasm and enlarged hyperchromatic nuclei; ii) A second case consisted of a diffuse proliferation with several vessels forming a “haemangioma-like” vasculature. Tumour cells were mostly round, and focally spindle, with scant cytoplasm and an enlarged nucleus containing one nucleolus; iii) The last case resided in a diffuse proliferation of spindle cells displaying an eosinophilic cytoplasm, oval nuclei with an irregular chromatin distribution with one or several nucleoli. Some nests of epithelioid cells with abundant, eosinophilic or clear, cytoplasm and atypical nuclei could be seen. Nevertheless, two features were observed consistently: a fibrous stroma and the presence of at least one component of spindle-shaped cells. All cases were negative for EMA and AE1/AE3 and inconsistent for PS100, cytoplasmic CD99 and vimentin staining. Ki67 was high (between 20% and 70%). In agreement with this heterogeneous histological presentation, the proposed diagnoses for EWSR1-PATZ1- positive tumours were quite variable including unclassified spindle cell sarcoma, myxoid liposarcoma, Ewing-like sarcoma, or unclassified malignant neuroectodermal tumours.

Nevertheless, all five tumours tightly clustered together and away from other EWSR1-fused

Watson et al. 10

tumours, indicative of a transcriptionally different entity. GSEA identified enrichment of genes correlated with SMARCA2 expression in prostate cancer (supplementary material, Table S3).

A new “epithelioid rhabdomyosarcoma” entity characterized by EWSR1/FUS-TFCP2 fusion

New fusions involving either EWSR1 (exon 5) or FUS (exon 6) and TFCP2 (exon2) were identified in three cases (Figure 5A and supplementary material, Figure S4A), linking part of the low complexity domains of EWSR1/FUS to the CP2 DNA binding and the SAM/pointed domains of TFCP2. The three tumour samples formed a discrete cluster away from any other

EWSR1/FUS-fused tumour samples (Figure 1 and supplementary material, Figure S2).

EWSR1/FUS-TFCP2-positive tumours arose in young adult females (age range 16-38yo) and developed in either the pelvic region, chest wall or the sphenoid bone. All three tumours were extremely aggressive, since patient survival did not exceed 5 months. Upon review, all three tumours were composed of an epithelioid proliferation arranged in small sheets or short fascicles. Tumour cells presented monotonous round nuclei with high grade features and prominent nucleoli (Figure 5B). They were associated with variable amounts of fibrous stroma with focal sclerosing areas that were present in all cases. The three tumours stained positive for desmin, MYOD1 and myogenin. Gene expression profile comparison confirmed a strong expression of MYOD1 and DES as well as an impressive overexpression of TERT and ALK (supplementary material, Figure S4B), the latter being heterogeneously expressed with both a cytoplasmic and membranous staining (Figure 5B). In silico functional analyses highlighted T cell immune response as well as keratin intermediate filament enrichment

(supplementary material, Table S3). Altogether the observation of positivity for both muscle and epithelial markers suggests that this EWSR1/FUS-TFCP2 fusion defines a new aggressive “epithelioid rhabdomyosarcoma” entity.

Watson et al. 11

Expression profiling reveals distinct transcriptomic patterns and potential biomarkers

Supervised group comparisons and between group analyses (BGA) highlighted genes specifically expressed in the molecularly defined entities (Figure 6; supplementary material,

Table S3). More specifically, crossing pairwise differential analysis and BGA analysis, we identified genes that were specifically expressed in each tumour type including VGLL2-fused tumours (LANCL2), CIC-fused tumours (ETV4), BCOR-rearranged (HES7), FUS-NFATC2- positive sarcomas (CD8B), EWSR1-PATZ1-positive tumours (GPR12) and FET-TFCP2- positive tumours (REQL), but also for almost all of the other tumour entities (supplementary material, Figure S5).

Discussion

We have used RNA sequencing to investigate unclassified small round or spindle cell sarcomas. We first confirm that sarcomas, like haematological malignancies [24], demonstrate a high incidence of gene fusions as compared to carcinomas. In this respect, it is noteworthy that the mesenchyme, from which sarcomas are derived, and haematopoiesis, from which arise leukaemias and lymphomas, both originate from the mesoderm. One hypothesis may rely on the activity of specific recombinases in mesodermal derived tissues.

In this respect we can mention the role of the RAG1 recombinase in the generation of lymphoma-specific translocations [25] and the recently suspected role of the PGBD5 recombinase in malignant rhabdoid tumours [26].

Combining fusion gene discovery and expression profiling, our analysis resulted in the delineation of biologically homogeneous groups of tumours. While the CIC-NUTM1 fusion was discovered as a new brain tumor entities [14], we show here that they are encompassed in an homogeneous CIC-fused group of tumors together with CIC-DUX4- and CIC-FOXO4- positive samples [27]. The overexpression of genes of the PEA3 type of the ETS family

Watson et al. 12

observed in all CIC-fused samples is known to be a consequence of a loss of function of CIC

[28], which was also involved in resistance to MAPK inhibitors [28,29] or in the promotion of metastasis [30]. Further analyses should confirm whether a dominant negative effect on wild type CIC is a consequence of CIC-fusions and if it may be correlated with the aggressiveness of CIC-fused tumors. The BCOR-rearranged group of tumors includes various BCOR-fusion genes and BCOR-ITD. BCOR-ITD were described in CCSK [19], in a new brain tumor entity CNS HGNET-BCOR [14] and in soft tissue undifferentiated round cell sarcoma of infancy sharing strong similarities with CCSK [17]. This homogeneous biological entity may also present clinical specificities depending on the type of genetic lesions. Indeed, while BCOR-CCNB3 was exclusively observed in bone, BCOR-ITD was only observed in soft tissues. BCOR rearrangements may lead to an abnormal activity of the non-conventional polycomb repressive BCOR (or PRC1.1) complex [31]. In this regard, the observation that

BCOR-rearrangements are associated with increased expression of genes correlated with

SMARCA2 expression (Figure 6A) may suggest that impairment of PRC1.1 activity lead to an increased activity of the antagonist SWI/SNF complex. Thanks to a unique collection, we found several samples carrying FUS-NFATC2 or EWSR1-PATZ1 fusions, which formed tight groups. FUS-NFATC2-positive tumours present different morphologies but with features rather reminiscent of CIC-fused tumours and are definitively different from EWSR1-NFATC2- positive tumours. EWSR1-PATZ1 tumours present relatively divergent cell morphologies, rendering their pathological identification challenging, though areas of spindle cells found in all three samples may guide pathologists during diagnosis. The identification of more cases is nevertheless mandatory to improve the characterization of these ultra-rare FET-fused sarcomas.

Watson et al. 13

We also describe here a new entity carrying an EWSR1- or FUS-TFCP2 fusion gene. TFCP2

(also known as LSF or LBP1) does not resemble any of the usual FET-fusion partners. It was first described as an activator of the late SV40 promoter and later found to bind globins and

HIV-1 promoters [32,33]. TFCP2 is ubiquitously expressed and plays determinant roles in lineage-specific gene expression or cell cycle regulation [34]. Potential involvement of

TFCP2 in oncogenesis has been suggested and may depend on tumor type: In hepatocellular carcinomas, TFCP2 is overexpressed [35] and its inhibition by small molecules leads to cell cycle arrest [36], whereas in skin melanoma TFCP2 is under- expressed, and its re-expression induces growth arrest [37]. Interestingly, in addition to the expression of MYOD1 (supplementary material, Table S3), a number of up-regulated genes in these FUS/EWSR1-TFCP2-positive tumours are involved in muscle biology (such as

Cholinergic Receptor Nicotinic subunit alpha1, delta and gamma or sarcoglycan alpha). This may suggest either a muscle cellular origin of these tumours or a muscle differentiation program triggered by the fusion protein. Consistently with the genes expressed, pathological review confirmed that TFCP2-rearranged tumours were an epithelioid variant of rhabdomyosarcoma, although the morphological spectrum of these tumours needs to be assessed on a larger scale. Despite a common muscle phenotype, these samples do not cluster together with other rhabdomyosarcomas. Considering potential therapeutic targets, it is noteworthy that ALK and TERT are highly expressed in these tumours. Moreover, recently developed small molecules impeding the DNA-binding activity of TFCP2 [36,38], which remains in the fusion protein, might also be effective in hindering EWSR1-TFCP2 binding and possibly in inhibiting the development of these very aggressive tumours.

Samples carrying VGLL2-NCOA2 or VGLL2-CITED2 fusion genes demonstrate some transcriptome heterogeneity related to two different histological subtypes: one presenting

Watson et al. 14

only few tumour cells surrounded by fibrous stroma, rather suggestive of relatively benign tumours, the other with a spindle cell rhabdomyosarcoma-like morphology similar to that reported [13] and exhibiting expression profiles enriched in cell cycle genes. Interestingly, one fibrous tumour and one rhabdomyosarcoma-like tumour represented the primary and the relapse tumours from the same patient, respectively. It is therefore likely that the rhabdomyosarcoma-like tumour represents a more aggressive evolution of the fibrous tumour. Further analyses including exome-sequencing of the two tumours may enable the identification of additional genetic mutations accounting for this evolution.

In addition to the fundamental role of expert surgical pathology, molecular analysis is now becoming an integral part of the diagnosis of small round cell tumours, in particular for resolving frequent diagnostic dilemmas. With the increase of sarcoma subclasses, it becomes difficult for the pathologist to ascertain a precise diagnosis in a reliable and cost effective way. In this respect, and depending on the expertise and technical availabilities at each pathology department, two molecular approaches may be proposed. One may rely on the design of a specific panel of genes, based on our selection of genes that are specifically expressed in each tumour group, tested in a simple assay (p.e., using Nanostring technology. Alternatively, thanks to the constant decrease of whole transcriptome costs and the availability of robust bioinformatics tools, we strongly believe that RNA-sequencing is a key approach. Indeed, our work indicates that the convergent information from gene fusion detection and expression profiles-based clustering is extremely powerful to identify biologically homogeneous groups of tumours. Hence, in the short term, RNA-sequencing is expected to constitute a mandatory methodology for an efficient molecular diagnosis of sarcoma, a requirement recently proposed in the GENSARC study [39]. Such a precise subgrouping of sarcomas is essential i) to investigate the clinical characteristics of each

Watson et al. 15

subgroup, particularly regarding the risk of evolution and response to current treatment options, ii) to identify new therapeutic opportunities, as potentially ALK overexpression in

EWSR1/FUS-TFCP2 sarcomas, iii) to construct relevant animal models to design robust pre- clinical studies, and iv) to design highly specific assays to monitor circulating cell and tumour

DNA during treatment. Further increasing the number of samples, within large international consortia, is essential to identify groups of tumours of sufficient size to enable robust conclusions and to allow the collection of otherwise “orphan” samples.

Acknowledgments

The authors wish to thank pathologists who provided tumour material: E. Angot, H. Antoine-

Poirel, S. Aubert, A. Babik, J.P. Barbet, C. Bastien, A.M. Bergemer-Fouquet, D. Berrebi, L.

Boccon-Gibod, C. Bossard, S. Boudjemaa, E. Cassagnau, J. Champigneulle, M.A.

Chrestian, A. Clemenson, S. Collardeau-Frachon, S. Corby, J.F. Cote, A. Coulomb-

Lhermine, A. Croue, C. Daniliuc, A. De Muret , A. Dhouibi, C. Douchet, F. Dujardin, J.M.

Dumollard, H. Duval, M. Fabre, C. Fernandez, S. Fraitag, F. Galateau-Salle, L. Gibault, A.

Gomez-Brouchet, C. Guettier, M.F. Heymann, J.F. Ikoli, J.F. Jazeron, C. Jeanne-Pasquier,

A. Jouvet, R. Kaci, M. Karanian-Philippe, O. Kerdraon, J. Klijanienko, C. Labit-Bouvier, M.

Lae, T. Lazure, F. Le Pessot , F. Lemoine, A. Liprandi, F. Llamas - Gutierrez, M.C. Machet,

A. Maran-Gonzalez, L. Marcellin, P. Marcorelles, C. Marin, A. Maues De Paula, C.A.

Maurage, A. Moreau, A. Neuville, H. Perrochia, J.M. Picquenot, C. Renard, V. Rigau, J.

Riviere, C. Rouleau, A. Rouquette, H.Sartelet, I. Serre, N. Stock, H. Szabo, P. Terrier, M.

Terrier-Lacombe, V. Thomas De Montpreville, J. Tran Van Nhieu, E. Uro-Coste, P. Validire, I.

Valo, V. Verkarre, J.M. Vignaud, M.O. Vilain, M.L. Wassef, D. Zachar, L. Zemoura and the tumorothèque Necker-Enfants Malades. We also thank Brigitte Manship for the careful

Watson et al. 16

reading of the manuscript. This work was supported by the Institut National de la Santé et de la Recherche Médicale, the Institut Curie, the Ligue National Contre Le Cancer, the Institut

National du Cancer and la Direction générale de l'offre de soins (INCa-DGOS_5716), the

European PROVABES (ERA- 649 NET TRANSCAN JTC-2011), ASSET (FP7-HEALTH-

2010-259348), and EEC (HEALTH-F2-2013-602856) projects. U830 is also indebted to the

Société Française des Cancers de l’Enfant, Enfants et Santé, Courir pour Mathieu, Dans les pas du Géant, La course de l’espoir du mont Valérien, Au nom d'Andréa, Association

Abigaël, Association Marabout de Ficelle, Les Bagouz à Manon, Les amis de Claire, the association Adam. SW was supported by a grant from the Fondation Nuovo-Soldati. High- throughput sequencing has been performed by the NGS platform of Institut Curie, supported by the grants ANR-10-EQPX-03 and ANR10-INBS-09-08 from the Agence Nationale de la

Recherche (investissements d’avenir) and by the Canceropôle Ile-de-France.

Author contributions

SW, GP, OD and FT designed the study and wrote the manuscript. SW, VP, DG, SR and MB performed all the experiments. SW, GP and FT, acquired and analysed the data. JMC, MK,

JMG, PF, DRV and FLL provided samples and were the reference senior pathologists. LGR,

FL and EL recruited patients and provided numerous samples and clinical information.

References 1 Antonescu C. Round cell sarcomas beyond Ewing: emerging entities. Histopathology 2014; 64: 26-37

2 Delattre O, Zucman J, Melot T, et al. The Ewing family of tumors – A subgroup of small-round-cell tumors defined by specific chimeric transcripts. N Engl J Med 1994; 331: 294-299

3 Clark J, Rocques PJ, Crew AJ, et al. Identification of novel genes, SYT and SSX, involved in the t(X;18)(p11.2;q11.2) translocation found in human synovial sarcoma. Nat Genet 1994; 7: 502-508

Watson et al. 17

4 Galili N, Davis RJ, Fredericks WJ, et al. Fusion of a fork head domain gene to PAX3 in the solid tumour alveolar rhabdomyosarcoma. Nat Genet 1993; 5: 230-235

5 Kawamura-Saito M, Yamazaki Y, Kaneko K, et al. Fusion between CIC and DUX4 up-regulates PEA3 family genes in Ewing-like sarcomas with t(4;19)(q35;q13) translocation. Hum Mol Genet 2006; 15: 2125-2137

6 Szuhai K, IJszenga M, de Jong D, et al. The NFATc2 gene is involved in a novel cloned translocation in a Ewing sarcoma variant that couples its function in immunology to oncology. Clin Cancer Res 2009; 15: 2259-2268

7 Pierron G, Tirode F, Lucchesi C, et al. A new subtype of bone sarcoma defined by BCOR-CCNB3 gene fusion. Nat Genet 2012; 44: 461-466

8 McPherson A, Hormozdiari F, Zayed A, et al. deFuse: An Algorithm for Gene Fusion Discovery in Tumor RNA-Seq Data. PLOS Comput Biol 2011; 7: e1001138

9 Ge H, Liu K, Juan T, et al. FusionMap: detecting fusion genes from next-generation sequencing data at base-pair resolution. Bioinforma Oxf Engl 2011; 27: 1922-1928

10 Bray NL, Pimentel H, Melsted P, et al. Near-optimal probabilistic RNA-seq quantification. Nat Biotechnol 2016; 34: 525-527

11 Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 2009; 37: 1-13

12 Van der Maaten L, Hinton G. Visualizing High-Dimensional Data Using t-SNE. J Mach Learn Res 2008; 9: 2579-2605

13 Alaggio R, Zhang L, Sung Y-S, et al. A Molecular Study of Pediatric Spindle and Sclerosing Rhabdomyosarcoma: Identification of Novel and Recurrent VGLL2-Related Fusions in Infantile Cases. Am J Surg Pathol 2016; 40: 224-235

14 Sturm D, Orr BA, Toprak UH, et al. New Brain Tumor Entities Emerge from Molecular Classification of CNS-PNETs. Cell 2016; 164: 1060-1072

15 Le Guellec S, Velasco V, Pérot G, et al. ETV4 is a useful marker for the diagnosis of CIC-rearranged undifferentiated round-cell sarcomas: a study of 127 cases including mimicking lesions. Mod Pathol 2016; 29: 1523-1531

16 Specht K, Zhang L, Sung Y-S, et al. Novel BCOR-MAML3 and ZC3H7B-BCOR Gene Fusions in Undifferentiated Small Blue Round Cell Sarcomas. Am J Surg Pathol January 2016

17 Kao Y-C, Sung Y-S, Zhang L, et al. Recurrent BCOR Internal Tandem Duplication and YWHAE- NUTM2B Fusions in Soft Tissue Undifferentiated Round Cell Sarcoma of Infancy: Overlapping Genetic Features With Clear Cell Sarcoma of Kidney. Am J Surg Pathol 2016; 40: 1009-1020

Watson et al. 18

18 Kao Y-C, Sung Y-S, Zhang L, et al. BCOR Overexpression Is a Highly Sensitive Marker in Round Cell Sarcomas With BCOR Genetic Abnormalities: Am J Surg Pathol 2016; 40: 1670-1678

19 Roy A, Kumar V, Zorman B, et al. Recurrent internal tandem duplications of BCOR in clear cell sarcoma of the kidney. Nat Commun 2015; 6: 8891

20 Brohl AS, Solomon DA, Chang W, et al. The genomic landscape of the Ewing sarcoma family of tumors reveals recurrent STAG2 mutation. PLoS Genet 2014; 10: e1004475

21 Mastrangelo T, Modena P, Tornielli S, et al. A novel zinc finger gene is fused to EWS in small round cell tumor. Oncogene 2000; 19: 3799-3804

22 Qaddoumi I, Orisme W, Wen J, et al. Genetic alterations in uncommon low-grade neuroepithelial tumors: BRAF, FGFR1, and MYB mutations occur at high frequency and align with morphology. Acta Neuropathol (Berl) 2016; 131: 833-845

23 Johnson A, Severson E, Gay L, et al. Comprehensive Genomic Profiling of 282 Pediatric Low‐ and High‐Grade Gliomas Reveals Genomic Drivers, Tumor Mutational Burden, and Hypermutation Signatures. The Oncologist September 2017: theoncologist.2017-0242

24 Lieber MR. Mechanisms of human lymphoid chromosomal translocations. Nat Rev Cancer 2016; 16: 387-398

25 Papaemmanuil E, Rapado I, Li Y, et al. RAG-mediated recombination is the predominant driver of oncogenic rearrangement in ETV6-RUNX1 acute lymphoblastic leukemia. Nat Genet 2014; 46: 116-125

26 Kentsis A, Eisenberg A, Blackford AN, et al. PGBD5 promotes site-specific oncogenic mutations in human tumors. Nat Genet 2017; 49: 1005-1014

27 Sugita S, Arai Y, Tonooka A, et al. A novel CIC-FOXO4 gene fusion in undifferentiated small round cell sarcoma: a genetically distinct variant of Ewing-like sarcoma. Am J Surg Pathol 2014; 38: 1571-1576

28 Dissanayake K, Toth R, Blakey J, et al. ERK/p90RSK/14-3-3 signalling has an impact on expression of PEA3 Ets transcription factors via the transcriptional repressor capicúa. Biochem J 2011; 433: 515-525

29 Wang B, Krall EB, Aguirre AJ, et al. ATXN1L, CIC, and ETS Transcription Factors Modulate Sensitivity to MAPK Pathway Inhibition. Cell Rep 2017; 18: 1543-1557

30 Okimoto RA, Breitenbuecher F, Olivas VR, et al. Inactivation of Capicua drives cancer metastasis. Nat Genet 2017; 49: 87-96

31 Gearhart MD, Corcoran CM, Wamstad JA, et al. Polycomb Group and SCF Ubiquitin Ligases Are Found in a Novel BCOR Complex That Is Recruited to BCL6 Targets. Mol Cell Biol 2006; 26: 6880- 6889

Watson et al. 19

32 Swendeman SL, Spielholz C, Jenkins NA, et al. Characterization of the genomic structure, chromosomal location, promoter, and development expression of the alpha-globin transcription factor CP2. J Biol Chem 1994; 269: 11663-11671

33 Yoon JB, Li G, Roeder RG. Characterization of a family of related cellular transcription factors which can modulate human immunodeficiency virus type 1 transcription in vitro. Mol Cell Biol 1994; 14: 1776-1785

34 Veljkovic J, Hansen U. Lineage-specific and ubiquitous biological roles of the mammalian transcription factor LSF. Gene 2004; 343: 23-40

35 Yoo BK, Emdad L, Gredler R, et al. Transcription factor Late SV40 Factor (LSF) functions as an oncogene in hepatocellular carcinoma. Proc Natl Acad Sci U S A 2010; 107: 8357-8362

36 Rajasekaran D, Siddiq A, Willoughby JLS, et al. Small molecule inhibitors of Late SV40 Factor (LSF) abrogate hepatocellular carcinoma (HCC): Evaluation using an endogenous HCC model. Oncotarget 2015; 6: 26266-26277

37 Goto Y, Yajima I, Kumasaka M, et al. Transcription factor LSF (TFCP2) inhibits melanoma growth. Oncotarget 2016; 7: 2379-2390

38 Grant TJ, Bishop JA, Christadore LM, et al. Antiproliferative small-molecule inhibitors of transcription factor LSF reveal oncogene addiction to LSF in hepatocellular carcinoma. Proc Natl Acad Sci 2012; 109: 4503-4508

39 Italiano A, Di Mauro I, Rapp J, et al. Clinical effect of molecular methods in sarcoma diagnosis (GENSARC): a prospective, multicentre, observational study. Lancet Oncol 2016; 17: 532-538

Figure Legend

Figure 1: Fusion genes define tumour groups.

Kallisto-extracted expression values were used for all genes having a CDS annotation in

Ensembl GRCh38p5. A. t-Distributed Stochastic Neighbour Embedding (t-SNE) analysis.

Tumour samples (coloured spheres) carrying identical or similar fusion genes are linked to the centroid of the group (small grey sphere). Fusion genes defining major groups are indicated. Unclassified sarcomas are not represented here for the sake of clarity but were taken into account in the analysis. An enlarged view of the centre of the figure is presented in Watson et al. 20

the supplementary material, Figure S2A. B. Hierarchical clustering. 1-pearson correlation and

Ward’s method were used as distance and clustering method, respectively; identified recurrent fusion genes are indicated. An enlarged view of TFE3-fused, CIC-fused and

BCOR-rearranged samples are presented below (see supplementary material, Figure S2B for the full annotated cluster).

Figure 2: Subtypes of VGLL2-NCOA2-positive samples

A. Representative images of fibrous tumour samples. SARC070_primary and SARC065 tumour samples contained abundant fibrous stroma, displayed low Ki67 and were mostly negative for myogenin and desmin. B. Representative images of dense tumour samples.

SARC070_relapse and SARC102 cells were packed and strongly positive for Ki67, with numerous cells being positive for myogenin and desmin. Scale bar: 100 µm

Figure 3: FUS-NFATC2-positive tumours

Morphology (hematoxylin/eosin staining) of FUS-NFATC2 tumours. A. Proliferation of round tumour cells embedded in a variable myxoid stroma. B. Round to spindle tumour cells arranged in sheets associated with haemangioperycitic vessels. C. Tumour cells were focally embedded within myxohyaline stroma reminiscent of cartilaginous differentiation. Scale bars:

100 µm

Figure 4: EWSR1-PATZ1-positive tumours

A. Schematic diagram of native EWSR1 and PATZ1 proteins indicating fusion point of

EWSR1-PATZ1 fusion gene. LC: Low complexity domain; RRM: RNA recognition motif;

RGG: Arg-Gly-Gly rich region; Zn: Zinc Finger, RanBP2-type; BTB/POZ: Broad-Complex,

Tramtrack and Bric a brac / POxvirus and Zinc finger; H: AT hook; C2H2: Cys2-His2 Zinc- finger. Exon number based on the indicated RefSeq accession number is indicated below Watson et al. 21

each protein scheme together with amino acid length scale. B. H & E staining of three

EWSR1-PATZ1-positive samples demonstrating a variety of morphologies (scale bar: 100

µm).

Figure 5: FET-TFCP2 fusion

A. Schematic diagram of native proteins and fusion proteins. EWSR1-TFCP2 fusion was identified in SARC049 while FUS-TFCP2 was identified in RNA009_16_062 and

RNA020_16_136 samples. LC: Low complexity domain; RRM: RNA recognition motif; RGG:

Arg-Gly-Gly rich region; Zn: Zinc Finger, RanBP2-type; SAM: Sterile alpha motif / pointed domain. Exon number based on the indicated RefSeq accession number is indicated below each protein scheme. Amino acid length scale is presented at the bottom. B. Morphological spectrum of FET-TFCP2 tumours. Hematoxylin/eosin staining (left panels) illustrating epithelioid (top), spindling (middle) and sclerosing (bottom) areas and immunohistochemistry

(right panels) for MYOD1 (top), desmin (middle), and ALK (bottom) seen in all cases are illustrated in EWSR1-TFCP2-positive sample SARC049. ALK staining was cytoplasmic- positive with nuclear and membrane exclusion. Morphological aspects were evocative of an epithelioid and spindled variant of rhabdomyosarcoma, with numerous mitotic figures. Scale bar: 40 µm

Figure 6: Expression profiles functional analyses

Heatmap representing the specificity of the top 100 genes (detailed in the supplementary material, Table S3) identified in the between group analysis for each 24 tumour groups containing at least two samples. Data were centred and scaled (in row direction) prior analysis. Colour key of the scaled expression values is shown.

Watson et al. 22

Supporting information:

Supplementary Figure S1: Constitution of the cohorts

Supplementary Figure S2: Details of the unsupervised analyses

Supplementary Figure S3: Details of VGLL2-NCOA2/CITED2, CIC-NUTM1, BCOR-ITD and EWSR1-PATZ1 fusion points.

Supplementary Figure S4: Sequence of the fusion points and genes expressed in FET- TFCP2-positive tumors

Supplementary Figure S5: Expression of the most specific gene for each tumor entity (with more than 2 samples)

Supplementary Table S1: List of the fusion genes routinely tested by RT-PCR at the Institut Curie Unité de Génétique Somatique.

Supplementary Table S2: Description of the samples from the three cohorts

Supplementary Table S3: Differentially expressed genes in the BGA and supervised analyses and Gene ontology / GSEA enrichment for all sarcomas subtypes

Supplementary Table S4: Differentially expressed genes and gene ontology enrichment for FUS-NFATC2- vs EWSR1-NFATC2-positive tumors

Watson et al. 23

A -60 EWSR1-NFATC2

FET-TFCP2

VGLL2-NCOA2/CITED2

EML4-ALK FET-DDIT3 EWSR1-PATZ1

FET-CREB3L1/2 NAB2-STAT6 BCOR-CCNB3 HEY1-NCOA2

SS18-SSX BCOR-MAML3 / ZC3H7B-BCOR BCOR-ITD

FUS-NFATC2 FET-NR4A3 2 0

Tsne EWSR1- CREB1/ATF1 BAF-deficient -60 CIC-DUX4 CIC-NUTM1 BRD3/4- NUTM1

0 FET-ETS 40 EWSR1-WT1

-60 0 60 Tsne 1 B

0.8 0.6

distance 0.4

0.2 ard's

W 0.0

ALK

ETS

SSX

NUT

WT1

ATF1

Fused

FOXO

ERMS AT/RT

STAT6

- TFCP2

NR4A3

CREB1/

NCOA2 -

PTATZ1

deficient

NFATC2

FET

SS18

CIC

EML4

PAX

Rearranged

FET

TFE3

BRD3/4

EWSR1

BAF

NAB2

FUS

VGLL2

HEY1

EWSR1

BCOR

ITD

TFE3

DUX4

BCOR

CCNB3

FOXO4

NUTM1

MAML3

CIC

UTX

BCOR

CIC

SFPQ

BCOR

ZC3H7B ASPSCR1

Watson et al. Figure 1 A SARC070 (primary) SARC065

HES HES Desmin

Ki67 Ki67 Myogenin

B SARC070 (relapse) SARC070

HES Ki67 Desmin SARC102

HES Ki67 Myogenin

Watson et al. Figure 2 A

Watson et al. Figure 3 A

LC RGG RRM RGG Zn RGG EWSR1 1 2 3 4 5 6 7 8 910 11 12 13 14 15 16 17 NM_005243

BTB/POZ H C2H2 PATZ1 1 2 3 4 5 NM_014323 0 100 200 300 400 500 600 687 1610 1620 1630 1640 1650 1660 1670 1680 1690 1700 * * * * * * * * * * * * * * * * * * * accagcctccagctgggctacatcgaccttcctcctccgaggctgggtgagaatgggctacccatctctgaagaccccgacggcccccgaaag T S L Q L G Y I D L P P P R L G E N G L P I S E D P D G P R K 320 330 340 NM_014323

G155_RNA045_17_001 G155_RNA045_17_001

SARC017 SARC041

Watson et al. Figure 4 A LC RGG RRM RGG Zn RGG EWSR1 1 2 3 4 5 6 7 8 910 11 12 13 14 15 16 17 NM_005243 LC RRM RGG Zn RGG FUS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 NM_004960 CP2CP2 transcription transcription factor factor SAM TFCP2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 NM_005653

SARC049: EWSR1-TFCP2 LC CP2 transcription factor SAM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 RNA009_16_062 & RNA020_16_136 : FUS-TFCP2 LC CP2 transcription factor SAM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

0 100 200 300 400 500 600 700

HES MYOD1

HES desmin

HES ALK

Watson et al. Figure 5

ALK

ETS

SSX

WT1

Fused

Fused Fused

Fused

FOXO

- PATZ1 -

- STAT6

TFCP2

NR4A3

NUTM1

cluster1 cluster2

NCOA2 NCOA2

NFATC2

FET

SS18

CIC EML4

rearranged

PAX

CREB3L1/2 CREB3L2/3

FET

TFE3

- -

EWSR1

CREB1/ATF1

NAB2

FUS

NTRK3 NTRK1

HEY1

EWSR1

BRD3/4

FET FET

sarcoma sarcoma

EWSR1

BCOR

rhabdomyosarcoma

EWSR1

Fused Fused Fused

- -

-3 -2 -1 0 1 2 3

Embryonal VGLL2 Scaled expression value (Log2) VGLL2

Watson et al. Figure 6 Sarcoma samples received for molecular diagnosis

Molecular investigations including QPCR on known fusion transcripts *

Identification of a pathognomonic No pathognomonic alteration identified alteration

Selection of 78 samples: 700+ retrospective samples without detectable fusion with good quality RNA Ewing Sarcoma (n=7), CIC-DUX4 sarcoma (n=6) Random selection Pool of samples without Rhabdomyosarcoma (n=6) of 94 cases detectable fusion BCOR-CCNB3 sarcoma (n=5) Clear cell sarcoma (n=5) RNA-sequencing Solitary fibrous tumor (n=5) EWSR1-NFATC2 Ewing-like sarcoma (n=4) Extraskeletal myxoid chondrosarcoma (n=4) Investigation cohort N=94 NUT midline carcinoma (n=4) Alveolar soft part sarcoma (n=3) Renal cell carcinoma (n=3) Desmoplastic small round cell tumor (n=3) Robust fusion gene No robust fusion gene Myxoid liposarcoma (n=3) detected in 55 cases detected in 39 cases Congenital fibrosarcoma (n=2) Low-grade fibromyxoid sarcoma (n=2) Mesenchymal chondrosarcoma (n=2) Single fusion gene Multiple fusion genes Synovial sarcoma (n=1) detected in 40 cases detected in 15 cases SMARCA4-DTS (n=4) (1 sample with Malignant rhabdoid tumor (n=7) chromothripsis) SCCOHT (n=1) PCR screening*** + CIC-FOXO sarcoma (n=1)** yes no

RNA-sequencing RNA-sequencing

Control cohort N=78 Follow-up cohort N=12

Supplementary Figure S1: Constitution of the cohorts

* See Supplementary table S1 ** : RNAseq raw fastq files kindly provided by Yasuhito Arai and Tadashi Hasegawa [27] *** Screened fusions were: BCOR-MAML3, CDKN2A-CCDC12, CIC-NUTM1, EML4-ALK, EWSR1-PATZ1, EWSR1- TFCP2, FUS-TFCP2, FUS-NFATC2, KMT2A-YAP1 / YAP1-KMT2A, MN1-TAF3, NF1- RHOT1, PARG-BMS1, VGLL2-NCOA2 and VGLL2-CITED2.

Watson et al., Supplementary Figure S1 A

PAX-FOXO VGLL2-Fused

ERMS FET-DDIT3 NTRK3-Fused

-20 FET-CREB3L1/2

NTRK1-Fused 0

EWSR1-CREB1/ATF1 Tsne

0 3

BRD3/4-NUTM1

0 -40 Tsne 1

Watson et al., Supplementary Figure S2

Ward's distance

0.0 0.2 0.4 0.6 0.8 B Fusion Gene SampleID ASPSCR1-TFE3 RNA010_16_052 ASPSCR1-TFE3 RNA011_16_065 ASPSCR1-TFE3 RNA011_16_066 UXT-TFE3 SARC066 ASPSCR1-TFE3 RNA010_16_053 ASPSCR1-TFE3 RNA011_16_067 SFPQ-TFE3 RNA010_16_054 EWSR1-ATF1 RNA001_16_002 EWSR1-ATF1 RNA004_16_011 EWSR1-CREB1 SARC055 EWSR1-CREB1 SARC069 EWSR1-ATF1 SARC058 RNA007_16_037 TAF15-NR4A3 RNA002_16_002 TAF15-NR4A3 RNA003_16_007 EWSR1-NR4A3 SARC046 EWSR1-NR4A3 RNA004_16_010 PAX3-MAML3 SARC053 SARC079 BRD4-NUTM1 SARC067 BRD3-NUTM1 SARC087 SARC011 SARC059 TANC1-DAPL1;FAM172A-THBS4;GABRA1-PSMB3 ;ARHGEF28-EFCAB3;ICK-BTBD9 SARC043 BRD4-NUTM1 RNA002_16_003 BRD4-NUTM1 RNA006_16_025 BRD4-NUTM1 RNA006_16_026 EML4-ALK RNA004_16_009 EML4-ALK SARC056 EML4-ALK SARC060 ETV6-NTRK3 RNA006_16_027 TPM3-NTRK1 RNA005_16_019 RNA003_16_003 IKBKG-ALK SARC042 SARC022 KMT2A-YAP1 SARC002 PDLIM5-USP8 SARC019 SARC024 TFE3-Fused SARC031 SARC073 EWSR1-CREB1/ATF1 FUS-DDIT3 SARC062 FET-NR4A3 EWSR1-PBX1 SARC003 BRD3/4-NUT SARC090 EWSR1-CREB3L1 SARC033 EML4-ALK FUS-CREB3L2 SARC047 RNA004_16_012 VGLL2-CITED2 SARC085 VGLL2-NCOA2 SARC102 VGLL2-Fused RNA004_16_013 VGLL2-NCOA2 SARC061 FET-TFCP2 VGLL2-NCOA2 SARC065 SARC070_(Primary) EWSR1-NFATC2 VGLL2-NCOA2 SARC070_(Relapse) FUS-DDIT3 RNA001_16_004 EWSR1-DDIT3 SARC005 PAX-FOXO ACTB-GLI1 SARC063 ERMS FUS-TFCP2 RNA009_16_062 EWSR1-TFCP2 SARC049 AT/RT FUS-TFCP2 RNA020_16_136 GNB1-PTGER3;RERE-CAMTA1;FAR1-ACSM5 SARC006 BCOR-Rearranged SARC016 SEC14L1-CYP39A1 SARC012 MAPKAPK5-ACAD10 SARC021 CREB3L2-AUTS2;TAX1BP1-RARB;PACS1-SEPT9 SARC071 EWSR1-NFATC2 RNA002_16_004 EWSR1-NFATC2 RNA034_16_245 SS18-SSX HEY1-NCOA2 EWSR1-NFATC2 RNA041_16_300 EWSR1-PTATZ1 EWSR1-NFATC2 RNA041_16_303 NAB2-STAT6 CDKN2A-CCDC12 RNA003_16_005 SARC098 FET-ETS SARC010 BAF-deficient SARC091 SARC020 FUS-NFATC2 SARC089 EWSR1-WT1 SARC072 SARC095 CIC-Fused RNA007_16_034 PVT1-KIAA0125 SARC039 SARC035

Watson et al., Supplementary Figure S2

Ward's distance

0.0 0.2 0.4 0.6 0.8

B (continued) Fusion Gene SampleID PAX3-FOXO1 RNA001_16_005 PAX7-FOXO1 RNA006_16_032 RNA006_16_030 PAX3-FOXO1 RNA006_16_031 SARC036 PPFIBP1-TUBA1B RNA005_16_020 PDLIM5-LOC285419 RNA006_16_029 SARC086 RASSF3-XPOT;IGF2AS-LSP1 RNA007_16_040 STAG1-ATF7 SARC025 RNA007_16_036 SARC097 SARC075 SMARCA4_MRT02 SMARCB1_MRT05 SMARCB1_MRT04 TPM3-NTRK1 RNA009_16_064 SARC007 Chr1-chromothripsis SARC027 CCNT2-KYNU;RGL1-CCDC149;SMARCD1-KNTC1 SARC082 KPNA6-EPHB2;XPR1-IGSF21;TNFRSF14-ACOT7;CAMSAP1L1-CSRP1;EPB41-RBM34 SARC015 SARC096 BCOR-CCNB3 RNA003_16_001 BCOR-CCNB3 RNA004_16_015 BCOR-CCNB3 RNA004_16_014 BCOR-CCNB3 RNA004_16_016 BCOR-CCNB3 RNA003_16_002 BCOR_ITD SARC074 BCOR_ITD SARC030 BCOR_ITD SARC078 BCOR_ITD SARC068 BCOR-MAML3 SARC048 BCOR_ITD SARC050 ZC3H7B-BCOR RNA009_16_058 EPC2-PHC2;ETNK2-IKBKE SARC028 TRPS1-RIMBP2;RIMBP2-TRPS1 SARC032 MN1-TAF3 RNA003_16_004 FAM133B-DRG1;LATS1-SNX9;HPS4;LOC96610;HADH-TBC1D24;TRIM4;CALN1;LARGE;TPST1 SARC077 SARC001 NF1-RHOT1 SARC009 AKAP13-ABHD2 SARC023 SARC057 ETV6-NTRK3 RNA006_16_028 KIAA1549-BRAF SARC040 MEMO1-BCL11A1 SARC076 H19-COL1A RNA005_16_018 SARC083 COL1A-PDGFB RNA009_16_063 SARC099 TPR-NTRK1 SARC100 SARC004 PVT1-ATP10B SARC081 SS18-SSX2 RNA001_16_007 TFE3-Fused SS18-SSX2 SARC034 HEY1-NCOA2 SARC029 EWSR1-CREB1/ATF1 HEY1-NCOA2 SARC045 EWSR1-PATZ1 SARC017 FET-NR4A3 EWSR1-PATZ1 SARC026 EWSR1-PATZ1 RNA045_17_001 BRD3/4-NUT EWSR1-PATZ1 SARC041 EML4-ALK EWSR1-PATZ1 RNA034_16_244 NAB2-STAT6 RNA002_16_005 NAB2-STAT6 RNA005_16_023 NAB2-STAT6 RNA007_16_039 NAB2-STAT6 SARC044 NAB2-STAT6 SARC037 VGLL2-Fused NAB2-STAT6 RNA007_16_038 NAB2-STAT6 SARC052 FET-TFCP2 EWSR1-FLI1 RNA001_16_001 EWSR1-FLI1 RNA001_16_006 EWSR1-FLI1 RNA005_16_017 EWSR1-NFATC2 EWSR1-FLI1 RNA005_16_021 EWSR1-FLI1 RNA001_16_008 FUS-ERG SARC064 FUS-ERG SARC080 PAX-FOXO SMARCA4_MRT01 RNA005_16_024 ERMS SARC018 AT/RT SARC093 SARC008 SARC014 SMARCB1_MRT01 BCOR-Rearranged SMARCB1_MRT02 SMARCB1_MRT03 FUS-NFATC2 RNA020_16_134 FUS-NFATC2 RNA022_16_147 FUS-NFATC2 SARC094 SARC054 SARC092 SS18-SSX SARC088 HEY1-NCOA2 EWSR1-WT1 RNA001_16_003 EWSR1-PTATZ1 EWSR1-WT1 RNA002_16_008 NAB2-STAT6 EWSR1-WT1 RNA002_16_007 CIC-DUX4 EWN1069T FET-ETS CIC-DUX4 SARC051 CIC-DUX4 SARC013 BAF-deficient CIC-DUX4 SARC038 CIC-NUTM1 SARC084 FUS-NFATC2 CIC-DUX4 EWN1072T CIC-FOXO4 4SXT1 EWSR1-WT1 CIC-DUX4 INI42 CIC-NUTM1 RNA012_16_073 CIC-Fused CIC-NUTM1 RNA003_16_008 CIC-NUTM1 SARC101 CIC-NUTM1 RNA012_16_074

Watson et al., Supplementary Figure S2 TopHat/Cufflinks expression

5 4 3 distance 2 1 Watson et al., Supplementary Figure S2 Supplementary al., et Watson ward's 0 Tumor Type Age Gender Tumor site Tumor_tissue Technology Sequencing_Run Cohort ) PATZ1 PATZ1 - - Primary SARC093 SARC014 SARC018 SARC008 SARC099 SARC016 SARC036 SARC086 SARC097 SARC075 SARC054 SARC088 SARC091 SARC092 SARC098 SARC001 SARC057 SARC011 SARC004 SARC083 SARC007 SARC096 SARC079 SARC020 SARC089 SARC035 SARC073 SARC022 SARC010 SARC024 SARC031 SARC090 SARC059 SARC072 SARC095 INI42_CIC.DUX4 RNA007_16_037 RNA005_16_024 RNA007_16_036 RNA004_16_012 RNA004_16_013 RNA006_16_030 RNA003_16_003 RNA007_16_034 SMARCA4_MRT02 SMARCB1_MRT05 SMARCB1_MRT04 SMARCA4_MRT01 SMARCB1_MRT01 SMARCB1_MRT02 SMARCB1_MRT03 SARC070 SARC070 ( 4SXT1_CIC.FOXO4 SARC052B2.STAT6 SARC037B2.STAT6 SARC044B2.STAT6 SARC064_FUS.ERG SARC080_FUS.ERG SARC013_CIC.DUX4 SARC051_CIC.DUX4 SARC038_CIC.DUX4 SARC066_UXT.TFE3 EWN1069_CIC.DUX4 EWN1072_CIC.DUX4 SARC056_EML4.ALK SARC060_EML4.ALK SARC074_BCOR_ITD SARC050_BCOR_ITD SARC068_BCOR_ITD SARC078_BCOR_ITD SARC030_BCOR_ITD SARC063_ACTB.GLI1 SARC042_IKBKG.ALK SARC062_FUS.DDIT3 SARC034_SS18.SSX2 SARC101_CIC.NUTM1 SARC084_CIC.NUTM1 SARC100_TPR.NTRK1 SARC009_NF1.RHOT1 SARC025_STAG1.ATF7 SARC002_KMT2A.YAP1 SARC094_FUS.NFATC2 SARC058_EWSR1.ATF1 SARC053_PAX3.MAML3 SARC029_HEY1.NCOA2 SARC045_HEY1.NCOA2 SARC003_EWSR1.PBX1 SARC067_BRD4.NUTM1 SARC087_BRD3.NUTM1 SARC081_PVT1.ATP10B SARC019_PDLIM5.USP8 SARC048_BCOR.MAML3 SARC047_FUS.CREB3L2 SARC005_EWSR1.DDIT3 SARC065_VGLL2.NCOA2 SARC061_VGLL2.NCOA2 SARC102_VGLL2.NCOA2 SARC017_EWSR1.PATZ1 SARC026_EWSR1.PATZ1 SARC041_EWSR1.PATZ1 SARC085_VGLL2.CITED2 SARC049_EWSR1.TFCP2 SARC046_EWSR1.NR4A3 SARC069_EWSR1.CREB1 SARC055_EWSR1.CREB1 SARC039_PVT1.KIAA0125 SARC023_AKAP13.ABHD2 RNA007_16_038B2.STAT6 RNA007_16_039B2.STAT6 RNA002_16_005B2.STAT6 RNA005_16_023B2.STAT6 SARC040_KIAA1549.BRAF RNA004_16_009_EML4.ALK RNA003_16_004_MN1.TAF3 SARC076_MEMO1.BCL11A1 SARC033_EWSR1.CREB3L1 RNA001_16_004_FUS.DDIT3 RNA001_16_007_SS18.SSX2 RNA005_16_018_H19.COL1A RNA012_16_073_CIC.NUTM1 RNA012_16_074_CIC.NUTM1 RNA003_16_008_CIC.NUTM1 RNA009_16_062_FUS.TFCP2 RNA020_16_136_FUS.TFCP2 RNA010_16_054_SFPQ.TFE3 SARC012_SEC14L1.CYP39A1 SARC027_Chr1.chromothripsis RNA005_16_021_EWSR1.FLI1 RNA005_16_017_EWSR1.FLI1 RNA001_16_003_EWSR1.WT1 RNA002_16_008_EWSR1.WT1 RNA002_16_007_EWSR1.WT1 RNA006_16_028_ETV6.NTRK3 RNA006_16_027_ETV6.NTRK3 RNA020_16_134_FUS.NFATC2 RNA022_16_147_FUS.NFATC2 SARC021_MAPKAPK5.ACAD10 RNA001_16_005_PAX3.FOXO1 RNA006_16_032_PAX7.FOXO1 RNA006_16_031_PAX3.FOXO1 RNA009_16_064_TPM3.NTRK1 RNA005_16_019_TPM3.NTRK1 RNA001_16_002_EWSR1.ATF1 RNA004_16_011_EWSR1.ATF1 RNA002_16_003_BRD4.NUTM1 RNA006_16_026_BRD4.NUTM1 RNA006_16_025_BRD4.NUTM1 RNA003_16_001_BCOR.CCNB3 RNA004_16_015_BCOR.CCNB3 RNA004_16_014_BCOR.CCNB3 RNA003_16_002_BCOR.CCNB3 RNA004_16_016_BCOR.CCNB3 RNA003_16_007_TAF15.NR4A3 RNA002_16_002_TAF15.NR4A3 RNA001_16_001_EWSR1.FLI1.1 RNA001_16_006_EWSR1.FLI1.2 RNA001_16_008_EWSR1.FLI1.1 RNA009_16_058_ZC3H7B.BCOR RNA045_17_001_EWSR1 RNA034_16_244_EWSR1 RNA009_16_063_COL1A.PDGFB RNA004_16_010_EWSR1.NR4A3 RNA011_16_067_ASPSCR1.TFE3 RNA010_16_052_ASPSCR1.TFE3 RNA011_16_065_ASPSCR1.TFE3 RNA011_16_066_ASPSCR1.TFE3 RNA010_16_053_ASPSCR1.TFE3 RNA034_16_245_EWSR1.NFATC2 RNA041_16_300_EWSR1.NFATC2 RNA002_16_004_EWSR1.NFATC2 RNA041_16_303_EWSR1.NFATC2 SARC070 SARC070 (Relapse)_VGLL2.NCOA2 RNA005_16_020_PPFIBP1.TUBA1B RNA003_16_005_CDKN2A.CCDC12 SARC028_EPC2.PHC2;ETNK2.IKBKE RNA006_16_029_PDLIM5.LOC285419 SARC032_TRPS1.RIMBP2;RIMBP2.TRPS1 RNA007_16_040_RASSF3.XPOT;IGF2AS.LSP1 SARC006_GNB1.PTGER3;RERE.CAMTA1;FAR1.ACSM5 SARC071_CREB3L2.AUTS2;TAX1BP1.RARB;PACS1.SEPT9 SARC082_CCNT2.KYNU;RGL1.CCDC149;SMARCD1.KNTC1

SARC043_TANC1.DAPL1;FAM172A.THBS4;GABRA1.PSMB3 ;ARHGEF28.EFCAB3;ICK.BTBD9 SARC015_KPNA6.EPHB2;XPR1.IGSF21;TNFRSF14.ACOT7;CAMSAP1L1.CSRP1;EPB41.RBM34 SARC077_FAM133B.DRG1;LATS1.SNX9;HPS4;LOC96610;HADH.TBC1D24;TRIM4;CALN1;LARGE;TPST1 C

Sequencing_Run

Tumor_tissue

Tumor Technology

D Tumor Gender

Cohort ward's distance

Type

Age

0 1 2 3 4 5 site

SARC066_UXT.TFE3 RNA010_16_052_ASPSCR1.TFE3 RNA011_16_065_ASPSCR1.TFE3 RNA011_16_066_ASPSCR1.TFE3 RNA010_16_053_ASPSCR1.TFE3 RNA011_16_067_ASPSCR1.TFE3 RNA010_16_054_SFPQ.TFE3 SARC053_PAX3.MAML3 SARC079 SARC058_EWSR1.ATF1 RNA001_16_002_EWSR1.ATF1 RNA004_16_011_EWSR1.ATF1 SARC033_EWSR1.CREB3L1 SARC047_FUS.CREB3L2 SARC002_KMT2A.YAP1 SARC003_EWSR1.PBX1 SARC069_EWSR1.CREB1 SARC055_EWSR1.CREB1 SARC090 RNA007_16_037 RNA002_16_007_EWSR1.WT1 SARC087_BRD3.NUTM1 SARC020 SARC102_VGLL2.NCOA2 SARC089 SARC072 SARC095 SARC091 SARC070 (Relapse)_VGLL2.NCOA2 SARC085_VGLL2.CITED2 SARC088 SARC067_BRD4.NUTM1 RNA006_16_026_BRD4.NUTM1 RNA006_16_025_BRD4.NUTM1 RNA002_16_003_BRD4.NUTM1 SARC011 SARC015_KPNA6.EPHB2;XPR1.IGSF21;TNFRSF14.ACOT7;CAMSAP1L1.CSRP1;EPB41.RBM34 SARC057 SARC059 SARC043_TANC1.DAPL1;FAM172A.THBS4;GABRA1.PSMB3 ;ARHGEF28.EFCAB3;ICK.BTBD9 SARC056_EML4.ALK SARC060_EML4.ALK RNA004_16_009_EML4.ALK SARC042_IKBKG.ALK SARC062_FUS.DDIT3 SARC073 SARC022 SARC010 SARC019_PDLIM5.USP8 SARC024 SARC031 RNA004_16_012 RNA004_16_013 SARC065_VGLL2.NCOA2 SARC061_VGLL2.NCOA2 SARC070 (Primary) SARC049_EWSR1.TFCP2 RNA009_16_062_FUS.TFCP2 RNA020_16_136_FUS.TFCP2 SARC012_SEC14L1.CYP39A1 SARC016 SARC006_GNB1.PTGER3;RERE.CAMTA1;FAR1.ACSM5 SARC092 SARC021_MAPKAPK5.ACAD10 SARC071_CREB3L2.AUTS2;TAX1BP1.RARB;PACS1.SEPT9 RNA034_16_245_EWSR1.NFATC2 RNA041_16_300_EWSR1.NFATC2 RNA002_16_004_EWSR1.NFATC2 RNA041_16_303_EWSR1.NFATC2 RNA006_16_027_ETV6.NTRK3 RNA005_16_019_TPM3.NTRK1 Kalllisto SARC063_ACTB.GLI1 RNA003_16_003 SARC035 RNA007_16_034 SARC039_PVT1.KIAA0125 RNA009_16_063_COL1A.PDGFB RNA006_16_028_ETV6.NTRK3 SARC040_KIAA1549.BRAF SARC076_MEMO1.BCL11A1 SARC083 RNA005_16_018_H19.COL1A expression RNA003_16_004_MN1.TAF3 SARC099 SARC100_TPR.NTRK1 SARC017_EWSR1-PATZ1 RNA045_17_001_EWSR1-PATZ1 SARC026_EWSR1-PATZ1 RNA034_16_244_EWSR1-PATZ1 SARC041_EWSR1-PATZ1 SARC029_HEY1.NCOA2 SARC045_HEY1.NCOA2 SARC032_TRPS1.RIMBP2;RIMBP2.TRPS1 SARC034_SS18.SSX2 RNA001_16_007_SS18.SSX2 SARC004 SARC081_PVT1.ATP10B SARC009_NF1.RHOT1 SARC001 SARC023_AKAP13.ABHD2 SARC077_FAM133B.DRG1;LATS1.SNX9;HPS4;LOC96610;HADH.TBC1D24;TRIM4;CALN1;LARGE;TPST1 SARC028_EPC2.PHC2;ETNK2.IKBKE SARC046_EWSR1.NR4A3 RNA003_16_007_TAF15.NR4A3 RNA002_16_002_TAF15.NR4A3 RNA004_16_010_EWSR1.NR4A3 SARC005_EWSR1.DDIT3 RNA001_16_004_FUS.DDIT3 SARC098 RNA003_16_005_CDKN2A.CCDC12 SARC052B2.STAT6 RNA007_16_038B2.STAT6 SARC037B2.STAT6 SARC044B2.STAT6 RNA007_16_039B2.STAT6 RNA002_16_005B2.STAT6 RNA005_16_023B2.STAT6 RNA001_16_005_PAX3.FOXO1 RNA006_16_032_PAX7.FOXO1 RNA006_16_030 RNA006_16_031_PAX3.FOXO1 SARC036 RNA005_16_020_PPFIBP1.TUBA1B RNA006_16_029_PDLIM5.LOC285419 SARC086 SARC025_STAG1.ATF7 RNA007_16_040_RASSF3.XPOT;IGF2AS.LSP1 SARC097 RNA007_16_036 SMARCA4_MRT02 SMARCB1_MRT04 SMARCB1_MRT05 SARC096 RNA009_16_064_TPM3.NTRK1 RNA001_16_003_EWSR1.WT1 RNA002_16_008_EWSR1.WT1 SARC094_FUS.NFATC2 RNA020_16_134_FUS.NFATC2 RNA022_16_147_FUS.NFATC2 SARC007 SARC027_Chr1.chromothripsis SARC082_CCNT2.KYNU;RGL1.CCDC149;SMARCD1.KNTC1 SARC054 SARC075 SMARCA4_MRT01 SARC093 SARC014 SARC018 SARC008 RNA005_16_024 SMARCB1_MRT02 SMARCB1_MRT03 SMARCB1_MRT01 SARC064_FUS.ERG SARC080_FUS.ERG RNA001_16_001_EWSR1.FLI1.1 RNA001_16_006_EWSR1.FLI1.2 RNA005_16_021_EWSR1.FLI1 RNA001_16_008_EWSR1.FLI1.1 RNA005_16_017_EWSR1.FLI1 SARC013_CIC.DUX4 EWN1069_CIC.DUX4 SARC051_CIC.DUX4 SARC084_CIC.NUTM1 EWN1072_CIC.DUX4 SARC038_CIC.DUX4 4SXT1_CIC.FOXO4 INI42_CIC.DUX4 RNA012_16_073_CIC.NUTM1 SARC101_CIC.NUTM1 RNA012_16_074_CIC.NUTM1 RNA003_16_008_CIC.NUTM1 RNA003_16_001_BCOR.CCNB3 RNA004_16_015_BCOR.CCNB3 RNA004_16_014_BCOR.CCNB3 RNA003_16_002_BCOR.CCNB3 RNA004_16_016_BCOR.CCNB3 SARC074_BCOR_ITD SARC068_BCOR_ITD SARC078_BCOR_ITD SARC030_BCOR_ITD SARC050_BCOR_ITD SARC048_BCOR.MAML3 RNA009_16_058_ZC3H7B.BCOR Watson et al., Supplementary Figure S2 Legend

Tumor Type Age Sequencing Run <3 Alveolar soft part sarcoma <10 HiSeq.RunB127 ARMS <25 NextSeq.Run46 AT/RT >25 NextSeq.Run49 BCOR-Rearranged sarcoma NextSeq.Run1 Biphenotypic sinonasal sarcoma Tumor site NextSeq.Run41 CIC-Fused sarcoma Head and neck NextSeq.Run34 Clear cell sarcoma Trunk Renal cell carcinoma Distal limbs NextSeq.Run35 Dermatofibrosarcoma protuberans Proximal limbs HiSeq.RunB102 DSCRT HiSeq.RunB137 EML4-ALK sarcoma Gender HiSeq.RunB68 ERMS M HiSeq.RunB72 Ewing sarcoma F NextSeq.Run48 EWSR1-NFATc2 sarcoma EWSR1-PATZ1 HiSeq.RunB135 Extraskeletal myxoid chondrosarcoma Tumor tissue HiSeq.RunB70 FET-TFCP2 soft tissue HiSeq.RunA119 FIC bone HiSeq.RunA53 Inflammatory myofibroblastic tumor brain tissue EXT.HASEGAWA Lipofibromatosis-like Neural Tumor HiSeq.RunB143 Technology Low-grade fibromyxoid sarcoma NextSeq.Run54 MRT Extracerebral HiSeq PE100 NextSeq.Run29 Mesenchymal chondrosarcoma NextSeq PE150 Myoepithelioma NextSeq.Run38 HiSeq.RunA125 Myxoid Liposarcoma Cohort NUT midline carcinoma NextSeq.Run118 INVESTIGATION COHORT Pericytoma CONTROL COHORT NextSeq.Run142 SCCOHT SCREENED SAMPLES HiSeq.RunA304 SMARCA4-DTS NextSeq.Run73 Solitary fibrous tumor NextSeq.Run81 Synovial Sarcoma Unclassified Sarcoma NextSeq.Run43 VGLL2-Fused NextSeq.Run45

Supplementary Figure S2: Detailed unsupervised analyses.

A) Zoom on the center region of the TSNE analysis (Figure 1A) with another angle, demonstrating dispersion of EWSR1-CREB3L1/2, NTRK1-, NTRK3- and VGLL2-fused tumors, and at a lesser extend EWSR1-CREB1/ATF1 and FET-DDIT3-fused tumors, as compared to the more defined EWSR1-NR4A3, EWSR1-WT1 or EWSR1-PATZ1 groups. B). Details of the clustering analysis (identical to Figure 1B) showing all the fusion genes identified. C&D) Expression profiles were generated using two different methods: in C) sequencing reads were aligned with TopHat2 and expression profiles were extracted using Cufflinks2 tool. In D) expression profiles were extracted with Kallisto tools that does not require prior alignment. In both cases, hierarchical clustering using the ten percent most variant genes based on interquartile range, proved to be quite similar. The main difference with Figure 1 concerned the VGLL2-fused samples that split in two or more groups. No clustering bias due to either patient ‘s clinical data (age, tumor site, gender or tissue type) or experimental condition (technology used or sequencing run) could be observed.

Watson et al., Supplementary Figure S2 A VGLL2 exon 2 NCOA2 exon 13

CCAAGGAGCTCTGGGCCCTGGCGAGCTTTTAATAACCCACGACCAGGGCAA P R S S G P W R A F N N P R P G Q SARC061 SARC070 SARC088 SARC102 VGLL2 exon 2 NCOA2 exon 14

CCAAGGAGCTCTGGGCCCTGGCGAGGAATGATTGGTAACAGTGCTTCTCGG P R S S G P W R G M I G N S A S R SARC065

VGLL2 exon 3 NCOA2 exon 14

GGTCTGGGCCTCAGCGTGGACTCAGGAATGATTGGTAACAGTGCTTCTCGG G L G L S V D S G M I G N S A S R

B SARC084 & RNA012_16_074 SARC101

CIC exon 18 NUTM1 exon 3 CIC exon 18 NUTM1 exon 6

GACATCTTCACCTTTGACCGTACAGCATCTGCATTGCCGGGACCGGATATG GACATCTTCACCTTTGACCGTACAGTGTACATTCCGAAGAAGGCAGCCTCC D I F T F D R T A S A L P G P D M D I F T F D R T V Y I P K K A A S

RNA003_16_008 RNA012_16_073

CIC exon 17 NUTM1 exon 6 CIC exon 20 NUTM1 exon 6

CGCAAGAAGAGGAAGAACTCCACGGTGTACATTCCGAAGAAGGCAGCCTCC CCCAGCCCCGCAGGGGGCCCAGACCACGGACCTCCTGCTCCTGAGGCACCC R K K R K N S T V Y I P K K A A S P S P A G G P D H G P P A P E A P

WT CSKDLEAFNPESKELLDLVEFTNEIQTLLGSSVEWLHPSDLASDNYW* SARC068 CSKDLEAFNPESKELLDLVEFTNEIQTLLGSSVSLEAFNPESKELLDLVEFTNEIQTLLGSSVEWLHPSDLASDNYW* SARC050 CSKDLEAFNPESKELLDLVEFTNEIQTLLGSSVFMEAFNPESKELLDLVEFTNEIQTLLGSSVEWLHPSDLASDNYW* SARC078 CSKDLEAFNPESKELLDLVEFTNEIQTLLGSSVEWLHPSDLASDKELLDLVEFTNEIQTLLGSSVEWLHPSDLASDNYW* SARC074 CSKDLEAFNPESKELLDLVEFTNEIQTLLGSSVEWLHPSDLASDELLDLVEFTNEIQTLLGSSVEWLHPSDLASDNYW* SARC030 CSKDLEAFNPESKELLDLVEFTNEIQTLLGSSVEWLHPSDLASDNYWLDLVEFTNEIQTLLGSSVEWLHPSDLASDNYW*

Watson et al., Supplementary Figure S3 D

RNA034_16_244

EWSR1 exon 8 EWSR1 intron 8 (1689nt from exon8) PATZ exon 1 EWSR1 exon 9 (NM_001163287)

CGGGGAGGAGGACGCGGTGGAATGGGTCCCACTTTTCATTATGCTGCCGGGGGCCCCCGAAAGAGGAGCCGG R G G G R G G M G P T F H Y A A G G P R K R S R

SARC041

EWSR1 exon 8 EWSR1 intron 8 PATZ exon 1 (936nt from exon8)

CGGGGAGGAGGACGCGGTGGAATGGGAATGAACTTCAAAATTAAAGACGGCCCCCGAAAGAGG R G G G R G G M G M N F K I K D G P R K R

SARC026

EWSR1 exon 8 PATZ exon 1

CGGGGAGGAGGACGCGGTGGAATGGGAAGTGACCCCGACGGCCCCCGAAAG R G G G R G G M G S D P D G P R K

RNA045_17_001

EWSR1 exon 8 PATZ exon 1

CGGGGAGGAGGACGCGGTGGAATGGGCCTCCAGCTGGGCTACATCGACCTT R G G G R G G M G L Q L G Y I D L

SARC017

EWSR1 exon 8 PATZ exon 1

GCTGGGTGAGAATGGGCTACCCATC CGGGGAGGAGGACGCGGTGGAATGGG L G E N G L P I R G G G R G G M G CCTTCCTCCTCCGAGGCTGGGTGAG L P P P R L G E

Supplementary Figure S3: Details of VGLL2-NCOA2/CITED2 (A), CIC-NUTM1 (B), BCOR-ITD (C) and EWSR1-PATZ1 (D) fusion points. Sanger sequences of the fusion points are given in A, B and D together with amino acid sequences of the translated fusion genes. Amino acid sequences of the internal tandem duplications in exon 15 of BCOR are shown for BCOR-ITD-positive tumors (C).

Watson et al., Supplementary Figure S3 A SARC049 RNA020_16_136 & RNA009_16_062

EWSR1 exon 5 TFCP2 exon 2 FUS exon 6 TFCP2 exon 2

CCAGCAGCCACTGCACCTACAAGTGATGTCCTTGCATTGCCCATTTTTAAG CGTGGAGGCAGAGGTGGCATGGGTGATGTCCTTGCATTGCCCATTTTTAAG P A A T A P T S D V L A L P I F K R G G R G G M G D V L A L P I F K

B **

8 ENSG00000171094 (ALK) 8 ENSG00000164362 (TERT) **

6 6

Log2(TPM+2)

Log2(TPM+2) in

in 4 4

2 2

Intensities

(5)

(2) (3)

(3)

(7)

(3)

(5) (7)

(4)

(5)

(4)

(2)

(4)

(5)

(3) (2)

(3)

(7)

(3)

(7) (5)

(4)

(5)

(4)

(2)

(4)

60)

(12)

(60)

(12)

(

(12)

NT (3) NT

FIC (2) FIC

tumor

ERMS (2) ERMS (4) ARMS

ARMS (4) ARMS (2) ERMS

TFCP2 (3) TFCP2

Fused

DSRCT (3) DSRCT

TFCP2 (3) TFCP2

LPF

Fused

DSRCT (3) DSRCT

LPF

sarcoma

Sarcoma

sarcoma

Sarcoma

sarcoma

Deficient

sarcoma

carcinoma

sarcoma

Deficient

sarcoma

carcinoma

FET

CIC

FET

cell CIC

fibrous

cell

ALK

fibrous

ALK

cell

BAF

Fused Fused 2 (4) cluster Fused 1 (3) cluster

BAF

Fused Fused 1 (3) cluster Fused 2 ( cluster

PATZ1 PATZ1

- -

PATZ1 PATZ1

- -

chondrosarcoma

chondrosarcoma -

chondrosarcoma

soft part part soft

- chondrosarcoma

NFATC2 NFATC2

soft part softpart

Synovial Synovial

midline

NFATC2 NFATC2

Clear

- Synovial

midline

Clear

EML4

ETS Ewing Ewing ETS

Myxoid (3) Liposarcoma Myxoid

EML4

Renal

ETS Ewing Ewing ETS

Myxoid (3) Liposarcoma Myxoid

Solitary

Renal

Solitary

Rearranged

Unclassified

Rearranged

Unclassified

VGLL2 VGLL2 FUS

NUT NUT

VGLL2 VGLL2 FUS

NUT NUT

EWSR1

FET

EWSR1

FET

Alveolar

EWSR1

Alveolar

grade fibromyxoid fibromyxoid grade

EWSR1

grade fibromyxoid fibromyxoid grade

BCOR

Mesenchymal

Low

extraskeletal myxoid extraskeletal extraskeletal myxoid extraskeletal

Supplementary Figure S4: Sequence of the fusion points and genes expressed in FET-TFCP2-positive tumors A. Sanger sequencing of EWSR1-TFCP2-positive sample (SARC049) and FUS-TFCP2-positive samples (RNA020_16_136, similar to RNA009_16_062) B. Boxplots for ALK and TERT gene expression level as log2(transcript per million + 2) across all tumor samples demonstrate strong signals in FET-TFCP2 tumor samples. Number of samples for each boxplot is indicated under brackets. ** : Welsh t-test p-value < 0.01.

Watson et al., Supplementary Figure S4 A *** B ** ** 10 7 ENSG00000132434 (LANCL2) ENSG00000175832 (ETV4)

8 6

inLog2(TPM+2) inLog2(TPM+2) 4

4 Intensities

Intensities 2 3

C D n.s. *** 8 ENSG00000183337 (BCOR) 8 ENSG00000179111 (HES7)

6 6

inLog2(TPM+2) 4 inLog2(TPM+2) 4

Intensities 2 Intensities 2

E 8 F 8 ENSG00000162624 (LHX8) * ENSG00000172116 (CD8B) * ** *

6 6

4 4

inLog2(FPKM+2) inLog2(TPM+2)

2 2

Intensities Intensities

G 8 ENSG00000132975 (GPR12) *** H 12 ENSG00000004700 (RECQL) ***

10 6

inLog2(TPM+2) 4 inLog2(TPM+2) 6

Intensities 2 Intensities 2

Watson et al., Supplementary Figure S5 12 I ENSG00000158022 (TRIM63) *** 10 J *** 8 ENSG00000165553 (NGB) 8

inLog2(TPM+2) inLog2(TPM+2) 4

Intensities Intensities 2 2

*** K 10 ENSG00000179761 (PIPOX) *** L 8 ENSG00000151715 (TMEM45B) 8

inLog2(TPM+2) inLog2(TPM+2) 4

4 Intensities 2 Intensities 2

*** M *** N 12 ENSG00000197696 (NMB) 10 ENSG00000188992 (LIPI)

10 8

6 inLog2(TPM+2)

Intensities Intensities

2 2

O * 10 10 ENSG00000181092 (ADIPOQ) P ENSG00000119866 (BCL11A) ** 8 8

6 6

inLog2(TPM+2) inLog2(TPM+2)

4 4 Intensities

2 Intensities 2

Watson et al., Supplementary Figure S5 Q R 10 ENSG00000189362 (NEMP2) 8 ENSG00000112164 (GLP1R) * 8 *** 6

6 inLog2(TPM+2) inLog2(TPM+2) 4

4 Intensities Intensities 2 2

Supplementary Figure S5: Expression of the most specific gene for each tumor entity (with more than 2 samples)

A. Expression level of LANCL2 gene throughout all samples demonstrating its specificity for VGLL2-Fused tumors B. Expression level of ETV4 gene throughout all samples demonstrating its specificity for CIC-Fused tumors C. Expression level of BCOR gene throughout all samples demonstrating its lack of specificity for BCOR-rearranged tumors D. Expression level of HES7 gene throughout all samples demonstrating its specificity for BCOR-rearranged tumors E. Expression level of LHX8 gene throughout all samples demonstrating its specificity for EML4-ALK-positive tumors F. Expression level of CD8B gene throughout all samples demonstrating its specificity for FUS-NFATC2-positive tumors G. Expression level of GPR12 gene throughout all samples demonstrating its specificity for EWSR1-PATZ1-positive tumors H. Expression level of RECQL gene throughout all samples demonstrating its specificity for FET-TFCP2-positive tumors I. Expression level of TRIM63 gene throughout all samples demonstrating its specificity for TFE3-fused sarcoma J. Expression level of NGB gene throughout all samples demonstrating its specificity for EWSR1-WT1 desmoplastic small round cell tumors K. Expression level of PIPOX gene throughout all samples demonstrating its specificity for PAX-FOXO alveolar rhabdomyosarcoma L. Expression level of TMEM45 gene throughout all samples demonstrating its specificity for EWSR1-NFATC2-positive tumors M. Expression level of NMB gene throughout all samples demonstrating its specificity for FET-NR4A3 extraskeletal myxoid chondrosarcoma N. Expression level of LIPI gene throughout all samples demonstrating its specificity for FET-ETS Ewing sarcoma O. Expression level of ADIPOQ gene throughout all samples demonstrating its specificity FET-DDIT3 myxoid liposarcoma P. Expression level of BCL11A gene throughout all samples demonstrating its specificity for NUTM1-BRD3/4 NUT-midline carcinoma Q. Expression level of NEMP2 gene throughout all samples demonstrating its specificity for NAB2-STAT6 Solitary Fibrous Tumors R. Expression level of GLP1R gene throughout all samples demonstrating its specificity for FET-CREB1/ATF1 clear cell sarcoma Number of tumors by groups are indicated under brackets. *, ** and ***: Welsh t-test p-value < 0.05, 0.01 and 10-4, respectively, n.s.: not significant

Watson et al., Supplementary Figure S5