Comprehensive DSRCT multi-omics analyses unveil CACNA2D2 as a diagnostic hallmark and super-enhancer-driven EWSR1::WT1 signature gene

Abstract

Desmoplastic small round cell tumor (DSRCT) is an aggressive cancer that predominantly affects adolescents and young adults, typically developing at sites lined by mesothelium [1, 2]. DSRCT is genetically defined by a chromosomal translocation that fuses the N-terminus of EWS RNA binding protein 1 (EWSR1) to the C-terminus of Wilms tumor protein (WT1), forming EWSR1::WT1 [3]. This fusion encodes a potent transcription factor and is the only known driver of oncogenic transformation in DSRCT [4]. The lack of a comprehensive understanding of DSRCT biology parallels its dismal survival rate (5%-20%) [1]. These challenges are exacerbated by the absence of clinical trials, the limited systematic collection and analysis of DSRCT biomaterial [1], and the notable lack of specific diagnostic markers, necessitating resource-intensive molecular testing for an accurate diagnosis.

Here we first focused on identifying promising candidates for validation as single, fast, and reliable diagnostic DSRCT markers. For this, we performed differential gene expression (DEG) analysis on datasets comprising patient samples from 32 DSRCT and 20 morphological mimics, identifying 23 genes overexpressed in DSRCT (log₂ fold change (log₂FC) > 2.5; adjusted P-value (Padj) < 0.01; Figure 1A, Supplementary Figure S1A). Secondly, we analyzed EWSR1::WT1 binding sites derived from chromatin immunoprecipitation followed by sequencing (ChIP-seq) data [5] obtained from the JN-DSRCT-1 cell line, identifying 2,065 genomic loci likely regulated by EWSR1::WT1 (Figure 1A). Third, we established JN-DSRCT-1 and SK-DSRCT2 cell lines expressing doxycycline (DOX)-inducible short hairpin RNA (shRNA)-mediated EWSR1::WT1 knockdown (KD) (Supplementary Figure S1B). Differential protein expression (DEP) analysis of these cells identified 104 proteins consistently regulated across both cell lines (log₂FC > 1.0 and Padj < 0.01; Figure 1A, Supplementary Table S1). The intersection of these analyses revealed calcium voltage-gated channel auxiliary subunit alpha2delta 2 (CACNA2D2) and IQ motif containing G (IQCG) as potential DSRCT biomarkers (Figure 1A). CACNA2D2 was selected for validation due to its significantly higher expression in DSRCTs compared to IQCG (P < 0.001; Figure 1A). Indeed, DSRCT exhibited the highest expression of CACNA2D2 among all studied morphological mimics and normal tissues (P < 0.001; Supplementary Figures S1C-D). Further ChIP-seq data and motif analyses of EWSR1::WT1 binding coordinates and histone marks in JN-DSRCT-1 and four DSRCT patient samples [5, 6] suggested a direct regulatory role of EWSR1::WT1 through an enhancer interaction at the CACNA2D2 locus (Figure 1B). Notably, KD of EWSR1::WT1 in JN-DSRCT-1 resulted in a loss of the EWSR1::WT1 signal and Histone H3 lysine 27 acetylation (H3K27ac) enhancer marks at the CACNA2D2 locus (Figure 1B). Additionally, chromatin interaction data [6] revealed 19 loops connecting the EWSR1::WT1 binding site to the transcription start site of CACNA2D2, which were depleted upon KD of EWSR1::WT1 (Figure 1C). Super enhancer (SE) analysis further demonstrated that the EWSR1::WT1-bound enhancer exhibited a characteristic SE H3K27ac profile in JN-DSRCT-1, which was lost upon EWSR1::WT1 KD (Figure 1D, Supplementary Table S2).

Post-transcriptional and post-translational KD of EWSR1::WT1 in three DSRCT cell line models expressing different EWSR1::WT1 isoforms (Supplementary Figure S2A) resulted in a significant reduction in CACNA2D2 expression (Figures 1E–F, Supplementary Figure S1B, Supplementary Figures S2B–F). Additionally, ChIP-seq data derived from MeT-5A mesothelial cells [6] – the potential cell of origin of DSRCT [7, 8] – ectopically expressing different EWSR1::WT1 isoforms (-KTS, +KTS, or -KTS/+KTS) suggested direct regulation, as evidenced by the co-occurrence of H3K27ac signals and signals for V5- or HA-tagged EWSR1::WT1 isoforms at the CACNA2D2 enhancer region (Supplementary Figure S2G). Notably, MeT-5A cells transfected with a control vector showed no substantial signal at this locus (Supplementary Figure S2G). Publicly available RNA-sequencing (RNA-seq) data from MeT-5A cells [6] expressing different EWSR1::WT1 isoforms showed that CACNA2D2 was differentially expressed in the presence of EWSR1::WT1 (4.1 ≤ log₂FC ≤ 5.9, Padj < 0.001) (Supplementary Figure S2H). Finally, quantitative polymerase chain reaction (qPCR) analysis of MeT-5A cells stably expressing a DOX-inducible ectopic EWSR1::WT1 expression cassette confirmed that upon EWSR1::WT1 induction, CACNA2D2 was significantly and highly overexpressed (Supplementary Figure S2I). Taken together, these results emphasize that EWSR1::WT1 is sufficient to drive CACNA2D2 expression. SE analysis of MeT-5A-derived data strikingly showed that the CACNA2D2 enhancer bound by EWSR1::WT1 became a SE upon ectopic expression of EWSR1::WT1^{− KTS + KTS} (Supplementary Figure S2J).

To explore whether CACNA2D2 could serve as a surrogate indicator of oncogenic EWSR1::WT1 transformation, we defined a CACNA2D2 gene set and gene signature by performing a correlation analysis of gene expression data from 32 DSRCT patient samples (Supplementary Figure S3A, Supplementary Tables S3-S4). Next, an EWSR1::WT1 signature was computed by performing a combined DEG analysis of newly generated in vivo and in vitro [4] material derived from three DSRCT cell lines upon EWSR1::WT1 KD (Supplementary Figure S3A, Supplementary Table S4). Notably, fast gene set enrichment analysis (fGSEA) of the CACNA2D2 gene set demonstrated a highly significant (Padj < 0.001) and strong positive enrichment for the EWSR1::WT1 signature (normalized enrichment score, NES_EWSR1::WT1 = 3.6). Moreover, single sample gene set enrichment analysis (ssGSEA) of expression data from 32 DSRCT patient samples confirmed that the EWSR1::WT1 signature significantly correlated with that of CACNA2D2 (r = 0.75), highlighting a transcriptional interconnection between CACNA2D2 and EWSR1::WT1 in situ (Figure 1G). These observations were further supported by single-cell (sc)-derived signatures from orthotopically-generated tumors using two DSRCT cell lines with DOX-inducible KD of EWSR1::WT1 at primary (n = 221) and metastatic (n = 221) locations (Figure 1G, Supplementary Table S4). Indeed, ssGSEA of our single-cell data showed highly significant correlation between the NES of our generated EWSR1::WT1 and CACNA2D2 signatures (Figure 1G), regardless of tumor location, implying that CACNA2D2-associated genes are also characteristic features of metastasized DSRCT cells (Supplementary Figure S3B).

To delineate the specificity of the interaction between CACNA2D2 and EWSR1::WT1 in DSRCT, we performed ssGSEA using our EWSR1::WT1 and CACNA2D2 signatures on expression data from 20 DSRCT morphological mimics (Figure 1H). Here, non-DSRCT cancer entities showed significantly lower NES and correlation strength for all signatures compared to DSRCT (Supplementary Figures S3C-D). These results further emphasized the high specificity of the CACNA2D2 and EWSR1::WT1 interplay in DSRCT. Moreover, both bulk- and sc-derived CACNA2D2 signatures precisely distinguished DSRCT cell clusters from non-tumor cells in single-cell RNA-sequencing (scRNA-seq) data from four DSRCT patients (n = 11 samples) [9] (Figure 1I, Supplementary Figure S3E). Concordantly, all predicted normal cell types within these tumors exhibited low enrichment of both CACNA2D2 signatures (Supplementary Figures S3F-G).

Further, dimensional reduction of CACNA2D2-associated CpG sites in 24 DSRCT patient samples, compared with 192 samples from 13 morphological mimics [10] revealed distinct clustering of all DSRCT samples, which was unique to CACNA2D2 compared to other described EWSR1::WT1-regulated genes or IQCG (Figure 1A, Supplementary Figures S4A-B). Additionally, these CACNA2D2-associated CpG sites exhibited significant (P < 0.001) and specific hypomethylation in DSRCT patient samples, collectively suggesting that the CACNA2D2-associated methylation signature is a distinct and specific feature of DSRCT (Supplementary Figure S4C).

To assess the diagnostic utility of CACNA2D2, we optimized a staining protocol for DSRCT cell line xenografts, achieving consistent and robust membranous or cytoplasmatic staining, even uncovering micrometastases (Figure 1J, Supplementary Figure S4D).

Finally, we assembled the largest collection of fresh-frozen and paraffin-embedded DSRCT patient samples analyzed to date (n = 61), comprising primary, metastatic, and post-treatment samples, and supplemented it with 249 patient samples from 18 different DSRCT morphological mimics (Supplementary Table S5). CACNA2D2 immunoreactivity was evaluated using a modified Immune Reactive Score (IRS) (Supplementary Material and Methods). Excitingly, DSRCT tumor sections exhibited the highest IRS for CACNA2D2 (IRS_mean = 10.5, 6 ≤ IRS_DSRCT ≤ 12, P < 0.001) (Supplementary Figure S4E-F), with specificity reaching 98% when applying a cutoff of IRS > 1 (Figure 1K-M, Supplementary Figure S4E). Indeed, even samples derived from CIC- and BCOR-rearranged sarcomas, as well as fusion-positive alveolar rhabdomyosarcomas, showed negligible mean protein expression compared to DSRCT (IRS_CIC = 0.21, IRS_BCOR = 0, IRS_fp-ARMS = 0.56). Furthermore, 100% sensitivity was achieved when applying an IRS cutoff of ≤ 6, implying that DSRCT samples consistently displayed strong staining for CACNA2D2 (Figure 1M). Thus, we recommend a single CACNA2D2 staining for clinically and histologically compatible DSRCT differential diagnosis. If IRS_CACNA2D2 ≤ 1, the diagnosis should be reconsidered or re-evaluated using molecular diagnostic procedures (such as fluorescence in situ hybridization, qRT-PCR, or next-generation sequencing), if available (Figure 1N). Conversely, if IRS_CACNA2D2 > 1, a diagnosis of DSRCT may be established. Also, CACNA2D2 staining may be used to rule out DSRCT within the broad spectrum of small-round-blue-cell tumors, potentially offering extensive diagnostic utility.

Finally, the high, specific, and homogenous membranous expression of CACNA2D2 in DSRCT, combined with the highly specific antibody described here, makes CACNA2D2 an ideal candidate for targeted therapeutic approaches, including drug delivery using antibody-drug conjugates or CAR-T cell therapy. Future studies should investigate the precise role of CACNA2D2 in DSRCT biology, with a focus on its potential contributions in tumor cell fitness, differentiation, and tumorigenic potential.

In conclusion, here we developed an extensive toolset for DSRCT research (Supplementary Figure S4G), a validated blueprint for how such resources could be harnessed in other cancer entities, and identified CACNA2D2 as a singular, powerful DSRCT biomarker.

Florian Henning Geyer, Florencia Cidre-Aranaz, and Thomas Georg Phillip Grünewald conceived the study. Florian Henning Geyer and Florencia Cidre-Aranaz wrote the paper and drafted all figures and tables. Florian Henning Geyer carried out all in vitro and in vivo experiments and performed all bioinformatic and statistical analyses. Florian Henning Geyer, Alina Ritter, and Thomas Georg Phillip Grünewald performed immunohistochemical evaluation and scoring of tumor samples and TMAs. Florencia Cidre-Aranaz, Roland Imle, and Ana Banito performed and/or coordinated in vivo experiments. Olivier Delattre provided microarray expression data. Seneca Kinn-Gurzo performed in vitro experiments on BER cell lines. Tobias Faehling and Clémence Henon performed single-cell bioinformatic analyses. Karim Aljakouch and Azhar Orynbek performed MassSpec and analyzed MassSpec data. Alina Ritter, Jing Li, Endrit Vinca, Laura Romero-Perez, Martin Sill, and Shunya Ohmura contributed to experimental procedures. Wolfgang Hartmann and Benjamin Friedrich Berthold Mayer provided clinical and/or histological guidance. Enrique De Álava, Juan Díaz-Martín, Stefanie Bertram, Sophie Postel-Vilnay, Martin Ebinger, Monika Sparber-Sauer, Daniel Baumhoer, Carine Ngo, David Horst, Yvonne Versleijen-Jonkers, Armin Jarosch, Sabine Stegmaier, and Thomas Kirchner provided clinical samples. Patrick Joseph Grohar, Thomas Georg Phillip Grünewald, and Jeroen Krijgsveld provided laboratory infrastructure. Florencia Cidre-Aranaz and Thomas Georg Phillip Grünewald supervised the study and data analysis. All authors read and approved the final manuscript.

The authors declare no competing interests.

The laboratory of Thomas Georg Phillip Grünewald is supported by grants from the Matthias-Lackas Foundation, the Dr. Leopold und Carmen Ellinger Foundation, the European Research Council (ERC CoG 2023 #101122595), the Deutsche Forschungsgemeinschaft (DFG 458891500), the German Cancer Aid (DKH-70112257, DKH-7011411, DKH-70114278, DKH-70115315), the Dr. Rolf M. Schwiete foundation, the SMARCB1 association, the Ministry of Education and Research (BMBF; SMART-CARE and HEROES-AYA), and the Barbara and Wilfried Mohr foundation. The research team of Florencia Cidre-Aranaz was supported by the German Cancer Aid (DHK-70114111), and the Dr. Rolf M. Schwiete Stiftung (2020-028 and 2022-31). In addition, this work was delivered as part of the PROTECT team supported by the Cancer Grand Challenges partnership funded by Cancer Research UK, the National Cancer Institute, the Scientific Foundation of the Spanish Association Against Cancer And KiKa (Children Cancer Free Foundation). Florian Henning Geyer, Tobias Faehling, Endrit Vinca, and Alina Ritter were supported by the German Academic Scholarship Foundation. In addition, Endrit Vinca was supported by scholarships from the Heinrich F.C. Behr foundation and the Rudolf and Brigitte Zenner foundation, Tobias Faehling by the Heinrich F.C. Behr foundation, and Florian Henning Geyer and Alina Ritter are supported by the German Cancer Aid through the ‘Mildred-Scheel-Doctoral Program’ (DKH-70114866). This project is co-funded by the European Union (ERC, CANCER-HARAKIRI, 101122595). All views and opinions expressed are however those of the authors only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them.

In vivo experiments were approved by the government of North Baden and conducted in accordance with ARRIVE guidelines and recommendations of the European Community (86/609/EEC) and UKCCCR (guidelines for the welfare and use of animals in cancer research). Open slides or tissue-microarrays from human formalin-fixed, paraffin-embedded or cryopreserved tissue samples were retrieved from the archives of the Institute of Pathology of the LMU Munich, the Charité Berlin, The Biobank of the Hospital Universitario Virgen del Rocío of Seville, the Hospital Gustave Roussy (Villejuif), the Bone Tumor Reference Center at the University of Basel, the University of Essen, the Cooperative Weichteilsarkom Studiengruppe (CWS) study center, the Klinikum Stuttgart (ethics committee from the Medical Faculty of the Eberhard-Karls University and University Hospital of Tübingen, approval no. 207/2022BO2), the Radboud University Medical Center, the Pathology Institute of the LMU Munich (approval no. 550-16 UE), and the University of Heidelberg (approval no. S-211/2021).