ETV4 and ETV5 drive synovial sarcoma through cell cycle and DUX4 embryonic pathway control
Joanna DeSalvo, Yuguang Ban, Luyuan Li, Xiaodian Sun, Zhijie Jiang, Darcy A. Kerr, Mahsa Khanlari, Maria Boulina, Mario R. Capecchi, Juha M. Partanen, Lin Chen, Tadashi Kondo, David M. Ornitz, Jonathan C. Trent, Josiane E. Eid
Joanna DeSalvo, Yuguang Ban, Luyuan Li, Xiaodian Sun, Zhijie Jiang, Darcy A. Kerr, Mahsa Khanlari, Maria Boulina, Mario R. Capecchi, Juha M. Partanen, Lin Chen, Tadashi Kondo, David M. Ornitz, Jonathan C. Trent, Josiane E. Eid
View: Text | PDF
Research Article Oncology

ETV4 and ETV5 drive synovial sarcoma through cell cycle and DUX4 embryonic pathway control

Abstract

Synovial sarcoma is an aggressive malignancy with no effective treatments for patients with metastasis. The synovial sarcoma fusion SS18-SSX, which recruits the SWI/SNF-BAF chromatin remodeling and polycomb repressive complexes, results in epigenetic activation of FGF receptor (FGFR) signaling. In genetic FGFR-knockout models, culture, and xenograft synovial sarcoma models treated with the FGFR inhibitor BGJ398, we show that FGFR1, FGFR2, and FGFR3 were crucial for tumor growth. Transcriptome analyses of BGJ398-treated cells and histological and expression analyses of mouse and human synovial sarcoma tumors revealed prevalent expression of two ETS factors and FGFR targets, ETV4 and ETV5. We further demonstrate that ETV4 and ETV5 acted as drivers of synovial sarcoma growth, most likely through control of the cell cycle. Upon ETV4 and ETV5 knockdown, we observed a striking upregulation of DUX4 and its transcriptional targets that activate the zygotic genome and drive the atrophy program in facioscapulohumeral dystrophy patients. In addition to demonstrating the importance of inhibiting all three FGFRs, the current findings reveal potential nodes of attack for the cancer with the discovery of ETV4 and ETV5 as appropriate biomarkers and molecular targets, and activation of the embryonic DUX4 pathway as a promising approach to block synovial sarcoma tumors.

Authors

Joanna DeSalvo, Yuguang Ban, Luyuan Li, Xiaodian Sun, Zhijie Jiang, Darcy A. Kerr, Mahsa Khanlari, Maria Boulina, Mario R. Capecchi, Juha M. Partanen, Lin Chen, Tadashi Kondo, David M. Ornitz, Jonathan C. Trent, Josiane E. Eid

×

Figure 10

ETV5 and E2F1 genomic binding in SS cells.

Options: View larger image (or click on image) Download as PowerPoint
ETV5 and E2F1 genomic binding in SS cells. (A, D, and G) Diagrams displa...
(A, D, and G) Diagrams display the sequence and location of the putative binding motifs (green) of ETV5 on the E2F1 gene, and of E2F1 on the CCNE2 and CHAF1B genes. The diagrams also delineate the regions selected for qPCR amplification and subcloning in the PGL3 reporter vector. (B, E, and H) Bar graphs show ETV5 binding to E2F1 (B) and E2F1 binding to CCNE2 and CHAF1B genes (E and H). Dots (12 per group in B and 6 per group in E and H) represent values from 2 independent ChIP-qPCR experiments. IgG served as background control. IgG binding and ETV5 or E2F1 binding were quantified as percentage of input chromatin. Error bars indicate SEM. P values (*P ≤ 0.00009; **P < 0.00001; ***P < 0.00001) compare ΔCt averages of ETV5 or E2F1 antibody versus IgG in naive cells and compare ΔCt averages of ETV5 or E2F1 binding in shETV5-expressing cells versus control (TRC2) cells. The PCR primers used are described in Supplemental Methods. (C, F, and I) Box plots demonstrate endogenous transcriptional activity measured by luminescence (RLU) and driven by the binding regions of ETV5 (PG-E2F1, C) and E2F1 (PG-CCNE2, F; and PG-CHAF1B, I), subcloned in PGL3. The PG-dlE2F1, PG-dlCCNE2, and PG-dlCHAF1B lanes show transcriptional activity of the 3 regions lacking the binding motifs. The deleted sequences are underlined on the respective diagrams. The 10 dots in each lane represent individual values from 2 independent assays performed in 5 replicates. Crossbars indicate the mean, and error bars indicate SEM. P values were measured by comparison of the average of each PG-dl group with that of the corresponding PG vector. *P < 0.00001. (J) Model for the ETV4/5 pathway in SS cells.