LncRNA PVT-1 promotes osteosarcoma cancer stem-like properties through direct interaction with TRIM28 and TSC2 ubiquitination

Abstract

Osteosarcoma, the most common pediatric bone tumor, is an aggressive heterogeneous malignancy defined by complex chromosomal aberrations. Overall survival rates remain at ~70%, but patients with chemoresistant or metastatic disease have extremely poor outcomes of <30%. A subgroup of tumors harbor amplification of chromosome 8q24.2 and increased expression of the oncogenic long noncoding RNA (lncRNA) Plasmacytoma Variant Translocation-1 (PVT-1), which is associated with an extremely poor clinical prognosis. This study demonstrates that PVT-1 is critical for osteosarcoma tumor-initiation potential. Chromatin Hybridization by RNA Purification analysis identified Tripartite-Motif Containing Family 28 (TRIM28) as a novel PVT-1 binding partner. Mechanistically, co-immunoprecipitation studies showed the PVT-1/TRIM28 complex binds and increases SUMOylation of phosphatidylinositol 3-kinase catalytic subunit type 3 (Vps34), which leads to enhanced ubiquitination and degradation of tumor suppressor complex 2 (TSC2), thus contributing to increased self-renewal and stem cell phenotypes. Furthermore, we identified that osteosarcoma cells with increased PVT-1 have enhanced sensitivity to the SUMOylation inhibitor, TAK-981. Altogether, this study elucidated a role for PVT-1 in the enhancement of cancer stem-like behaviors, including migration and invasion, in osteosarcoma, and identified the novel PVT-1/TRIM28 axis signaling cascade as a potential therapeutic target for osteosarcoma treatment.

Fig. 1. PVT-1 expression in osteosarcoma.

a Kaplan–Meier survival analysis from osteosarcoma patients at Texas Children’s Hospital that present with 8q24.2 copy number gain () (n = 30) to no gains () (n = 25). b Venn diagram depicting the differential gene expression profiles when comparing osteosarcoma patients with low vs. high PVT-1 and low vs. high MYC (left panel). Gene ontology (GO) pathway enrichment analysis defined signaling cascades dependent upon low vs. high PVT-1 or MYC. (right panel). c RNA expression profile of PVT-1 in human osteosarcoma cell lines, biopsies, and patient-derived xenografts compared to their corresponding control(s). d Spearman-correlation plot between PVT-1 copy number and PVT-1 transcript level (n = 11). e Kaplan–Meier event-free and overall survival of patients based on PVT-1 level. Error bars represent S.D.; a Student’s t test or one-way ANOVA, was used to calculate statistical significance values: *p < 0.05; **p < 0.001; ***p < 0.0001. Data represent three independent experiments.

Fig. 2. PVT-1 enhances osteosarcoma growth and metastatic potential.

a Cell viability assessment secondary to PVT-1 overexpression in HOS cells. b Cell viability assessment secondary knockdown of PVT-1 in TCCC-OS63. c Analysis of in vitro migration and invasion characterization for HOS-PVT1 o.e. cell lines by trans-well migration and invasion assay. d Migration and invasion analysis for TCCC-OS63-shPVT-1 cell line. Error bars represent S.D.; a Student’s t test or one-way ANOVA was used to calculate statistical significance values for all figures. *p < 0.05; **p < 0.001; ***p < 0.0001. Data represent three independent experiments.

Fig. 3. PVT-1 induces cancer stem-like phenotypes.

a RT-qPCR analysis for expression of stem cell genes in HOS PVT-1 Ctrl and PVT-1 o.e. cells. b Western blot analyzed the stem cell protein level in HOS PVT-1 Ctrl and PVT-1 o.e. cells. c Quantification of serial dilution secondary and tertiary sarcospheres from HOS PVT-1 Ctrl and PVT-1 o.e. cells. HOS PVT-1 Ctrl: 100 and 10 cell group developed no secondary sarcospheres. None detected (N.D.). Secondary sarcospheres were generated from the primary sarcosphere and secondary sarcospheres were used for tertiary sarcosphere production. d Expression level of PVT-1 in HOS adherent and HOS sarcospheres. *p < 0.05; **p < 0.001; ***p < 0.0001. Data represent three independent experiments.

Fig. 4. PVT-1 drives osteosarcoma cancer stem-like behavior through TSC2.

a RPPA analysis of HOS Ctrl and PVT-1 o.e. cells. b GSEA analysis of the RPPA data identified cellular pathways impacted by the enhancement of PVT-1. c Western blot analysis of TSC2 in HOS Ctrl and PVT-1 o.e. cells. d RT-qPCR of TSC2 in HOS Ctrl and PVT-1 o.e. cells. e–h Rescue experiments to understand the role of TSC2 as a downstream effector of PVT-1. Effects of TSC2 overexpression in HOS PVT-1 o.e. cell lines on: e Cell viability. f Migration and invasion assay. g Stem cell markers via western blot. h Serial-Dilution Sphere formation. None detected (N.D.). Error bars represent S.D. Statistical analysis was performed using linear regression of e and t-test or one-way ANOVA for the rest of the figures. *p < 0.05; **p < 0.001; ***p < 0.0001. Data represent three independent experiments.

Fig. 5. PVT-1 interacts with TRIM28 which induces cancer stem-like behaviors.

a ChIRP-mass spectrometry analysis to identify proteins directly bound to PVT-1. b ChIRP-Western blot to validate PVT-1/TRIM28 interaction. c TRIM28 RIP-qPCR analysis comparing enrichment of PVT-1 interaction in HOS PVT-1 Ctrl and PVT-1 o.e. cells. d Analysis of PVT-1/TRIM28 interaction in breast cancer cell lines with 8q24 non-amplification vs. 8q24 amplification. e–h HOS cells with PVT-1 and TRIM28 expression were modified to identify which factors are critical for PVT-1-induced tumorigenic and cancer stem cell-like behaviors and TSC2 inhibition. Cell lines transfected with PVT-1 o.e. or shTRIM28 are labeled with a “+” and if the corresponding control plasmid was inserted with a “−”. e metastatic behaviors: migration and invasion (f) sarcosphere production (g) qPCR analysis of stem cell markers (h) Western blot to assess for protein level of stem cell markers. Error bars represent S.D. Statistical analysis was performed using one-way ANOVA. *p < 0.05; **p < 0.001; ***p < 0.0001. Data represent three independent experiments.

Fig. 6. Mutation of PVT-1/TRIM28 interaction region negates PVT-1-induced behaviors.

a PVT-1 deletion constructs made from PVT-1 full length (FL) sequence. Deletion region denoted by (). b qPCR of two regions of PVT-1. One primer set was used to determine transfection efficiency in HOS cells of each construct and to ensure intact region was present (gray). A second primer set was used to identify the absence of the intended deletion region (black). c TRIM28 RIP-qPCR performed on HOS PVT-1 mutant and F.L. cells to identify the region of PVT-1 where TRIM28 interaction occurs. d–h Effects of deletion of PVT-1/TRIM28 interaction region in HOS cell line on: d Western blot analysis assessing the stem cell marker protein expression level. e TSC2 western blot (f) Cell viability (g) Migration and Invasion propensity (h) Sarcosphere assay. Error bars represent S.D. Statistical analysis was performed using one-way ANOVA. *p < 0.05; **p < 0.001; ***p < 0.0001. Data represent three independent experiments.

Fig. 7. PVT-1/TRIM28 axis is associated with TSC2 ubiquitination.

HOS cell lines transfected with PVT-1 o.e. or shTRIM28 are labeled with a “+” and if the corresponding control plasmid was inserted with a “−” a Western blot to detect TSC2 level dependent upon the perturbation of PVT-1 and/or TRIM28 in HOS cells. b HOS cells were transiently transfected with HA-Ubiquitin plasmid. 48 h post-transfection, endogenous TSC2 was immunoprecipitated from whole cell lysate and ubiquitinated TSC2 was detected by western blot using anti-HA antibody. c Western blot to determine presence of Vps34 co-immunoprecipitates: SUMO1, TRIM28, and TSC1 in HOS cells. d Cells were treated with MG132 at 10 uM for 5 h. Whole cell lysates of these cells were used for TSC1 co-immunoprecipitation. Immunoblots were analyzed for detection of TSC2 and Vps34. e CCK-8 assessment of HOS PVT-1 o.e. cell sensitivity to SUMOylation inhibitor. f Proposed model of PVT-1/TRIM28 signaling cascade which promotes tumorigenic and cancer-stem like phenotypes. Error bars represent S.D. Statistical analysis was performed using one-way ANOVA. *p < 0.05; **p < 0.001; ***p < 0.0001. Data represent three independent experiment.