TRIM22 inhibits osteosarcoma progression through destabilizing NRF2 and thus activation of ROS/AMPK/mTOR/autophagy signaling

Abstract

Osteosarcoma (OS) is a malignant bone tumor that mainly occurs in adolescents. It is accompanied by a high rate of lung metastasis, and high mortality. Recent studies have suggested the important roles of tripartite motif-containing (TRIM) family proteins in regulating various substrates and signaling pathways in different tumors. However, the detailed functional role of TRIM family proteins in the progression of OS is still unknown and requires further investigations. In this study, we found that tripartite motif-containing 22 (TRIM22) was downregulated in OS tissues and was hence associated with better prognosis. In vitro and in vivo functional analysis demonstrated that TRIM22 inhibits proliferation and metastasis of OS cells. Nuclear factor erythroid 2-related factor 2 (NRF2), a redox regulator, was identified as a novel target for TRIM22. TRIM22 interacts with and accelerates the degradation of NRF2 by inducing its ubiquitination dependent on its E3 ligase activity but independent of Kelch-like ECH-associated protein 1 (KEAP1). Further, a series of gain- and loss-of-function experiments showed that knockdown or overexpression of NRF2 reversed the functions of knockdown or overexpression of TRIM22 in OS. Mechanistically, TRIM22 inhibited OS progression through NRF2-mediated intracellular reactive oxygen species (ROS) imbalance. ROS production was significantly promoted and mitochondrial potential was remarkably inhibited when overexpressing TRIM22, thus activating AMPK/mTOR signaling. Moreover, TRIM22 was also found to inhibit Warburg effect in OS cells. Autophagy activation was found in OS cells which were overexpressed TRIM22, thus leading to autophagic cell death. Treatment with N-Acetylcysteine (NAC), a ROS scavenger or the autophagy inhibitor 3-Methyladenine (3-MA) abolished the decreased malignant phenotypes in TRIM22 overexpressing OS cells. In conclusion, our study indicated that TRIM22 inhibits OS progression by promoting proteasomal degradation of NRF2 independent of KEAP1, thereby activating ROS/AMPK/mTOR/Autophagy signaling that leads to autophagic cell death in OS. Therefore, our findings indicated that targeting TRIM22/NRF2 could be a promising therapeutic target for treating OS.

Keywords: Autophagic cell death; Osteosarcoma; ROS; TRIM22/NRF2; Warburg effect.

Graphical abstract

Fig. 1

TRIM22 is downregulated in OS and correlated with a better prognosis. (a) The expression of TRIM22 in tumor tissues and normal non-tumor tissues from GSE12865, GSE28424, GSE33382 and GSE36001 (unpaired two-tailed Student's t-test). (b) The expression of TRIM22 from the Human Cancer Metastasis Database (unpaired two-tailed Student's t-test, http://hcmdb.i-sanger.com/index, HCMDB). (c) Kaplan-Meier analysis of the overall survival and metastasis-free survival from an online database (http://hgserver1.amc.nl/cgi-bin/r2/main/cgi). (d) The correlations between TRIM22 expression and overall survival in the GEPIA database (http://gepia.cancer-pku.cn/index.html). (e) Representative IHC staining of TRIM22 in OS tissues and paired non-tumor tissues. Scale bar, 100 μm. (f) Expression of TRIM22 by IHC staining of OS tissues. Scale bar, 100 μm. (g) The expression of TRIM22 was determined in twelve pairs of OS tissues and matched non-tumor tissues by Western blot. (h) The expression of TRIM22 was examined in forty pairs of OS tissues and matched non-tumor tissues by qRT-PCR (n = 40, ***P<0.001, unpaired two-tailed Student's t-test). (i) The expression of TRIM22 in OS cell lines and hFOB 1.19 cell line was evaluated by Western blot. IHC = immunohistochemistry.

Fig. 2

Inhibitory effects of TRIM22 on proliferation of OS in vitro and in vivo. (a) Western blot analysis of TRIM22 expression after transfection in HOS and 143B cells. (b) CCK-8 assays in HOS and 143B cells with TRIM22 knockdown or overexpression, respectively (n = 3, ***P<0.001, two-way ANOVA with post hoc test). (c, d) EdU assays and colony formation assays in HOS and 143B cells with TRIM22 knockdown or overexpression, respectively and quantification of the EdU assay and colony formation assay (n = 3, **P<0.01, unpaired two-tailed Student's t-test). Scale bar, 200 μm. (e, f) Flow cytometric analysis and quantification of apoptosis in HOS and 143B cells with TRIM22 depletion or overexpression, respectively (n = 3, **P<0.01, unpaired two-tailed Student's t-test). (g) Representative images of xenograft tumors after subcutaneous injection of OS cells with TRIM22 knockdown or overexpression and corresponding controls 28d after inoculation. (h) Time course of growth and tumor weight of OS xenografts (n = 6, ***P<0.001, two-way ANOVA with post hoc test (tumor volume) and unpaired two-tailed Student's t-test (average tumor weight)). (i) IHC staining of Ki67 of tumor sections. Scale bar, 100 μm. (j) IHC staining of cleaved caspase 3 of tumor sections. Scale bar, 100 μm. CCK-8 = cell counting kit-8; EdU = 5-ethynyl-2′-deoxyuridine; IHC = immunohistochemistry.

Fig. 3

Inhibitory effects of TRIM22 on metastasis of OS in vitro and in vivo. (a, b) Effects of TRIM22 on invasion in OS cells in vitro using transwell invasion assay and quantification of transwell invasion assay (n = 3, *P<0.05, **P<0.01, unpaired two-tailed Student's t-test). Scale bar, 200 μm. (c, d) Representative images of 3D tumor spheroid cell-invasion assay in transfected HOS and 143B cells and quantification of 3D tumor spheroid cell-invasion assay (n = 3, **P<0.01, ***P<0.001, unpaired two-tailed Student's t-test). Scale bar, 500 μm. (e, f) Migration plots of OS cells in a gradient of FBS. The starting point of every track was normalized to the position x = 0 and y = 0. Mean velocities of OS cells and mean chemotaxis indices of OS cells were quantified (n = 3, *P<0.05, **P<0.01, unpaired two-tailed Student's t-test). (g) Representative images of lung metastasis models in nude mice after tail injection of HOS and 143B cells with TRIM22 knockdown or overexpression, respectively. (h) H & E staining of pulmonary metastatic nodules. Scale bar, 500 μm. (i) Quantification of lung metastatic foci (n = 6, **P<0.01, unpaired two-tailed Student's t-test). FBS = fetal bovine serum; H & E = hematoxylin-eosin.

Fig. 4

TRIM22 binds to and degrades NRF2. (a) IP/MS analysis indicated that NRF2 interacts with TRIM22. (b) Detection of endogenous protein interactions between TRIM22 and NRF2 in OS cells lysates. (c) Detection of exogenous protein interactions between TRIM22 and NRF2 in HEK 293T cells. Flag-tagged TRIM22 and Myc-tagged NRF2 plasmids were transfected into HEK 293T cells. (d) NRF2 mRNA levels in OS cells in indicated groups (n = 3, ns, no significance, unpaired two-tailed Student's t-test). (e) Western blot analysis of NRF2 and HO-1 protein levels in transfected OS cells. (f) Evaluation of NRF2 protein levels in 143B cells overexpressed with TRIM22 with or without proteasome inhibitor MG132 treatment. (g) Increasing amounts of Flag-tagged TRIM22 (WT or C15/18A mutant) were transfected and the expression levels of NRF2 and Flag-tagged TRIM22 were detected by Western blot in 143B cells. (h) Detection of NRF2 protein levels in shNC and shTRIM22 transfected HOS cells in the presence of CHX (10 μg/ml) for indicated time point (n = 3, ***P<0.001, two-way ANOVA with post hoc test). (i) Detection of NRF2 protein levels in 143B cells with overexpression of TRIM22 or corresponding control in the presence of CHX (10 μg/ml) for the indicated time points (n = 3, **P<0.01, two-way ANOVA with post hoc test). IP/MS = Immunoprecipitation coupled with mass spectrometry; CHX = cycloheximide.

Fig. 5

Effects of TRIM22 on ubiquitination of NRF2. (a) Evaluation of endogenous NRF2 ubiquitination in HOS cells transfected with shNC or shTRIM22. (b) Evaluation of endogenous NRF2 ubiquitination in 143B cells with overexpression of TRIM22 or corresponding control. (c) Evaluation of exogenous NRF2 ubiquitination in HEK 293T cells co-transfected with Flag-tagged TRIM22 (WT) or Flag-tagged TRIM22 (C15/18A), HA-tagged Ub and Myc-tagged NRF2. (d) HOS cells were co-transfected with shTRIM22, shDUB3 as well as HA-Ub, and then the NRF2 ubiquitylation linkage was assessed. (e) 143B cells transfected with Flag-TRIM22, Myc-DUB3 as well as Ub, and then NRF2 ubiquitylation linkage was assessed.

Fig. 6

TRIM22 promotes KEAP1-independent NRF2 proteasomal degradation. (a) Myc-tagged NRF2 was expressed in HEK 293T Tet-On Flag-TRIM22 cells with doxycycline (Dox) inducible expression of Flag-TRIM22. Cells were exposed to Dox as indicated. Cell lysates were evaluated by immunoblotting with anti-Myc and anti-Flag. (b–c) Plasmids encoding NRF2 variant deficient for KEAP1 binding including E82G or ΔETGE were transfected and cell lysates were evaluated by immunoblotting with anti-Myc and anti-Flag. (d) Quantification of Myc-NRF2 expression level in a-c (n = 3, ns, no significance, two-way ANOVA with post hoc test). (e) Cells were exposed to a vehicle or the NRF2 inducer sulforaphane for 6h prior to Dox administration and cell lysates were evaluated by immunoblotting with anti-Myc and anti-Flag. (f) Quantification of Myc-NRF2 expression level in e (n = 3, ns, no significance, two-way ANOVA with post hoc test). Dox = doxycycline.

Fig. 7

TRIM22 inhibits OS cells proliferation and metastasis through interacting with and degrading NRF2 in vivo. (a) Representative images of tumors and quantification of tumor volumes and tumor weights in nude mice bearing HOS cells in indicated groups (n = 6, ***P<0.001, two-way ANOVA with post hoc test (tumor volume) and one-way ANOVA with post hoc test (average tumor weight)). (b) IHC staining of Ki67 and cleaved caspase 3 of tumor sections in indicated groups. Scale bar, 100 μm. (c) Representative images of tumors and quantification of tumor volumes and tumor weights in nude mice bearing 143B cells in indicated groups (n = 6, ***P<0.001, two-way ANOVA with post hoc test (tumor volume) and one-way ANOVA with post hoc test (average tumor weight)). (d) IHC staining of Ki67 and cleaved caspase 3 of tumor sections in indicated groups. Scale bar, 100 μm. (e, f) H & E staining and quantification of pulmonary metastatic nodules in indicated groups (n = 6, *P<0.05, **P<0.01, ***P<0.001, one-way ANOVA with post hoc test). Scale bar, 500 μm. IHC = immunohistochemistry; H & E = hematoxylin-eosin.

Fig. 8

Effects of TRIM22 on ROS production, mitochondrial potential and Warburg effect in OS cells. (a) Heatmap of DEGs of HOS cells transfected with shNC and shTRIM22. (b) GSEA was used to identify the distribution of genes in the ROS-related, AMPK signaling-related, mTOR signaling-related and autophagy-related pathway gene set of HOS cells transfected with shNC and shTRIM22. (c, d) Mitochondrial potential was assessed and quantified by JC-1 staining in HOS and 143B cells (n = 3, ***P<0.001, unpaired two-tailed Student's t-test). Scale bar, 100 μm. (e, f) Detection and quantification of ROS levels in HOS and 143B cells by flow cytometry (n = 3, **P<0.01, ***P<0.001, unpaired two-tailed Student's t-test). (g, h) Evaluation of GSH/GSSG ratio and GPX activity in HOS and 143B cells (n = 3, *P<0.05, **P<0.01, ***P<0.001, unpaired two-tailed Student's t-test). (i) Western blot analysis of AMPK/mTOR signaling and LC3 as well as p62 protein levels in HOS and 143B cells. (j, k) Detection of ECAR and OCR in HOS and 143B cells (n = 3, two-way ANOVA with post hoc test). DEGs = differentially expressed genes; GSEA = gene set enrichment analysis; ROS = reactive oxygen species; GSH = reduced glutathione; GSSG = oxidized glutathione; GPX = glutathione peroxidase; ECAR = extracellular acidification rate; OCR = O₂ consumption rate.

Fig. 9

Effects of TRIM22 on autophagy in OS cells. (a) HOS cells transfected with GFP-mRFP-LC3 lentivirus were monitored, and the cellular puncta were evaluated. Scale bar, 20 μm (n = 3, **P<0.01, one-way ANOVA with post hoc test). (b) TEM was applied to determine the autophagic microstructure of transfected HOS cells (n = 3, **P<0.01, unpaired two-tailed Student's t-test). Scale bar, 2 μm. (c) 143B cells transfected with GFP-mRFP-LC3 lentivirus were monitored, and the cellular puncta were evaluated (n = 3, ***P<0.001, one-way ANOVA with post hoc test). Scale bar, 20 μm. (d) TEM was applied to determine the autophagic microstructure of transfected 143B cells. Scale bar, 2 μm (n = 3, ***P<0.001, unpaired two-tailed Student's t-test). (e) Potential underlying mechanism by which TRIM22/NRF2 axis inhibits OS progression through ROS/AMPK/mTOR/Autophagy signaling.