Auranofin is a potent suppressor of osteosarcoma metastasis

Abstract

Osteosarcoma (OS) accounts for 56% of malignant bone cancers in children and adolescents. Patients with localized disease rarely develop metastasis; however, pulmonary metastasis occurs in approximately 50% of patients and leads to a 5-year survival rate of only 10-20%. Therefore, identifying the genes and pathways involved in metastasis, as new therapeutic targets, is crucial to improve long-term survival of OS patients. Novel markers that define metastatic OS were identified using comparative transcriptomic analyses of two highly metastatic (C1 and C6) and two poorly metastatic clonal variants (C4 and C5) isolated from the metastatic OS cell line, KHOS. Using this approach, we determined that the metastatic phenotype correlated with overexpression of thioredoxin reductase 2 (TXNRD2) or vascular endothelial growth factor (VEGF). Validation in patient biopsies confirmed TXNRD2 and VEGF targets were highly expressed in 29-42% of metastatic OS patient biopsies, with no detectable expression in non-malignant bone or samples from OS patients with localised disease. Auranofin (AF) was used to selectively target and inhibit thioredoxin reductase (TrxR). At low doses, AF was able to inhibit TrxR activity without a significant effect on cell viability whereas at higher doses, AF could induce ROS-dependent apoptosis. AF treatment, in vivo, significantly reduced the development of pulmonary metastasis and we provide evidence that this effect may be due to an AF-dependent increase in cellular ROS. Thus, TXNRD2 may represent a novel druggable target that could be deployed to reduce the development of fatal pulmonary metastases in patients with OS.

Keywords: auranofin; metastasis; osteosarcoma; oxidative stress; thioredoxin reductase.

Figure 1. Characterisation of highly and poorly metastatic OS clonal variants and validation of TXNRD2

A. Primary tumour growth of KHOS and clonal variants C1-C6 in an orthotopic mouse model of OS, measured as time in days for the primary tumour to reach 10 mm. B. Metastatic burden of KHOS and clonal variants, quantified as percentage area of metastasis/total area of lung in H&E stained lung tissue sections. C. Immunostaining of three representative biopsy sections from patients who developed lung metastasis, showing high expression of TrxR2, and three OS patient biopsies who did not develop metastatic disease, showing low expression of TrxR2. Magnification: 20X. Statistical analysis was performed using an unpaired t-test. *P < 0.05, **P < 0.01. Bars: SD.

Figure 2. Inhibition of TrxR and the in vitro metastatic phenotype in OS by auranofin

A. Inhibition of TrxR activity in the presence of AF was measured by the reduction rate of (5, 5′-dithiobis (2-nitrobenzoic) acid (DTNB) substrate in an NADPH dependent reaction over 30 min. B. AF cytotoxicity was measured by metabolic activity causing cleavage of MTS substrate to formazan, as an indication of cell viability of metastatic osteosarcoma cell lines. C. AF decreased proliferation after 24 h treatment, assessed by BrdU incorporation. D. Colony forming ability of AF treated cells over 14 days. Cells were stained with Coomassie and quantified by counting colonies with more than 20 cells. E. Migration of cells over 24 h through an 8 μM membrane. F. Invasion into Matrigel and through an 8 μM membrane, over 48 h. Quantification was performed by measuring Calcein uptake. Statistical analysis was performed using an unpaired t-test. *P < 0.05, **P < 0.01. Bars: SD.

Figure 3. Auranofin is a potent inhibitor of OS metastasis in vivo

A. Dose-dependent reduction of metastatic burden in mice injected orthotopically with KHOS, after treatment with AF. B. Reduction of metastatic burden in various metastatic OS cell lines, KHOS, KRIB and BTK143B, injected orthotopically, receiving 1 mg/kg AF, *P = 0.028; **P = 0.004. C. Representative H+E stained mouse lung sections of the experiment in B. Magnification = 10×. D. Orthotopic KHOS primary tumor volume over time in mice treated with 1 mg/kg AF. E. Orthotopic primary tumor growth, measured as number of days to reach 10 mm, of mice treated with various AF concentrations. F. Subcutaneous primary KHOS tumour growth, measured as tumour volume over time. G. Quantification of Ki67 immunostaining of tumours in (F). H. Quantification of cleaved caspase-3 immunostaining of subcutaneous tumors treated with vehicle or AF. Staining quantification was performed using Nikon NIS Elements software which uses image segmentation for measurement of colour. Statistical analysis was performed using an unpaired t-test. *P < 0.05, **P < 0.01. Bars: SD.

Figure 4. Auranofin treatment induces apoptosis in metastatic OS cells

A. Apoptotic cell death was determined in KHOS cells treated with AF for 12 h and measured by annexin-V and PI staining using flow cytometry. B. Apoptosis marker immunoblot of KHOS cells treated with 0, 1, 3 or 5 μM of AF for 12 h.

Figure 5. Oxidative stress is increased in OS cells treated with auranofin

A. Dose-dependent increase in ROS in KHOS cells treated with AF was determined by the oxidation of DCFDA over 15 mins, 37°C. B. Flow cytometry determined a shift in the cell population with an increase in the percentage of cells in the DCFDA-high population after 12 h AF treatment. C. An apoptosis array of KHOS cells treated with 1 μM AF for 12 h showing upregulation of oxidative stress response proteins and down-regulation of anti-apoptosis proteins. D. Increased mitochondrial membrane permeability occurring in cells treated with AF was measured using JC-1 dye incorporation over 15 mins at 37°C. Mitochondrial membrane permeability was visualised using flow cytometry. Functional (energised) mitochondria (red) and dispersed JC-1 dye (green), indicate a loss of mitochondrial membrane potential in AF-treated cells. Statistical analysis was done using an unpaired t-test. *P < 0.05, **P < 0.01. Bars: SD.