Inhibition of Age-Related Therapy Resistance in Melanoma by Rosiglitazone-Mediated Induction of Klotho

Abstract

Purpose: Aging is a poor prognostic factor for melanoma. We have shown that melanoma cells in an aged microenvironment are more resistant to targeted therapy than identical cells in a young microenvironment. This is dependent on age-related secreted factors. Klotho is an age-related protein whose serum levels decrease dramatically by age 40. Most studies on klotho in cancer have focused on the expression of klotho in the tumor cell. We have shown that exogenous klotho inhibits internalization and signaling of Wnt5A, which drives melanoma metastasis and resistance to targeted therapy. We investigate here whether increasing klotho in the aged microenvironment could be an effective strategy for the treatment of melanoma.Experimental Design: PPARγ increases klotho levels and is increased by glitazones. Using rosiglitazone, we queried the effects of rosiglitazone on Klotho/Wnt5A cross-talk, in vitro and in vivo, and the implications of that for targeted therapy in young versus aged animals.Results: We show that rosiglitazone increases klotho and decreases Wnt5A in tumor cells, reducing the burden of both BRAF inhibitor-sensitive and BRAF inhibitor-resistant tumors in aged, but not young mice. However, when used in combination with PLX4720, tumor burden was reduced in both young and aged mice, even in resistant tumors.Conclusions: Using glitazones as adjuvant therapy for melanoma may provide a new treatment strategy for older melanoma patients who have developed resistance to vemurafenib. As klotho has been shown to play a role in other cancers too, our results may have wide relevance for multiple tumor types. Clin Cancer Res; 23(12); 3181-90. ©2017 AACR.

Figure 1. Loss of klotho in the aged microenvironment affects klotho and Wnt5A expression in melanoma cells

A, Immunohistochemical analysis of klotho in young and aged human skin (60X objective, 1X digital zoom). B, qRT-PCR of klotho mRNA in young (N=6) and aged (N=4) human fibroblasts (two-tailed unpaired t-test, p=0.04). C, Western analysis of klotho in the media of young and aged human fibroblasts. D, Immunofluorescent analysis of klotho (green), HMB45 (red) and Dapi (blue) expression in 3D organotypic skin reconstructs built with young or aged fibroblasts (63X objective). E, Western analysis of Wnt5A in melanoma cells exposed to DMEM +/− serum, media from young fibroblasts or from aged fibroblasts. HSP90 was used as loading control. F, Immunofluorescent analysis of Wnt5A (green), HMB45 (red) and Dapi (blue) expression in 3D organotypic skin reconstructs built with young or aged fibroblasts (63X objective). G, Western analysis of Wnt5A in melanoma cells exposed to media from young fibroblasts or young fibroblasts after klotho knock-down or aged fibroblasts +/− recombinant klotho. HSP90 was used as loading control. H, Immunofluorescent analysis of klotho (green) expression in 3D organotypic skin reconstructs built with young fibroblasts infected with either a control shRNA lentivirus or a sh_klotho lentivirus (63X objective). HMB45 (red) was used as melanoma cell marker and Dapi (blue) was used to stain nuclei. I, Immunofluorescent analysis of Wnt5A (green) expression in 3D organotypic skin reconstructs built with young fibroblasts infected with either a control shRNA lentivirus or a sh_klotho lentivirus (63X objective).

Figure 2. Wnt5A affects klotho expression via CAMKII mediated inhibition of PPARγ

A, Schematic of Wnt5A signaling showing activation of CAMKII, inhibition of PPARγ, and decrease of klotho transcription. Klotho is a sialidase and cleaves Wnt5A from syndecans at the cell surface, as we have previously published. B, Western analysis of active (phospho) and total CAMKII in melanoma cells treated with indicated dose of recombinant Wnt5A for 16h. C, Immunofluorescent analysis of PPARγ and klotho in melanoma cells treated with rWnt5A (63X objective). D,qRT-PCR analysis of klotho expression in aged fibroblasts treated with the CAMKII inhibitor KN93 (10μM, 48h) (two-tailed unpaired t-test, ***=p<0.001). E, Immunofluorescent analysis of PPARγ in KN93-treated aged fibroblasts (63X objective). F, Immunofluorescent analysis of PPARγ expression in melanoma cells after KN93 treatment (10μM, 48h) (63X objective). G, Immunofluorescent analysis of Klotho expression in melanoma cells after KN93 treatment (10μM, 48h) (63X objective). H, Immunofluorescent analysis of PPARγ expression in aged fibroblasts after treatment with PPARγ agonist Rosiglitazone (Roz) (63X objective). I, qRT-PCR analysis of klotho expression in aged fibroblasts treated with rosiglitazone (10μM, 48h) (two-tailed unpaired t-test, ***=p<0.001). J, Tumor growth curves of Yumm1.7 murine melanoma cells in aged C57/BL6 mice, in the presence of recombinant klotho (n=3), rosiglitazone (n=5), or vehicle control (n=4). Tumor burden was significantly less in mice treated with recombinant mouse klotho or rosiglitazone. Repeated measures analysis of variance (Anova) was calculated between samples. Holm Sidak correction was done, ***=p=0.0002 (Day-20). K, Analysis of Klotho and Wnt5A levels in mouse tumors by immunohistochemistry analysis after treatment with rosiglitazone or vehicle control (40X objective). Data represented as +/− SEM.

Figure 3. Rosiglitazone-mediated inhibition of phospho-Erk can inhibit growth of BRAF inhibitor resistant tumors

A, PO₄-Erk, Erk, and Wnt5A protein expression examined by western blot analysis in BRAF sensitive (Yumm1.7, WM983B, WM793) and resistant subclones (Yumm1.7_BR, WM983B_BR, WM793_BR). B, Western blot analysis for PO₄-Erk and Erk protein expression in WM983b_BR cells after treatment with rosiglitazone (10μM) for the indicated time period. C, Expression of PO₄-Erk, Erk and Wnt5A following treatment with rosiglitazone in WM983b_BR cells transfected with Ctrl or Klotho siRNA. D, PO₄-Erk, Erk, and Wnt5A protein expression examined by western blot analysis in WM983B-BR cells in the presence or absence of PLX4720 in conditioned media of aged fibroblasts treated with +/− rosiglitazone. E, Operetta® analysis of mito-SOX intensity (measurement of mitochondrial ROS) in WM983B_BR cells in the presence or absence of PLX4720 in conditioned media of aged fibroblasts treated with vehicle i.e. DMSO:PBS (1:10) or rosiglitazone (Anova p<0.0001; two tailed unpaired t-test, ***=p<0.001). C=control; P=plx4720; R= rosiglitazone; R+P= Rosiglitazone + plx4720. F, Tumor growth of BRAFi resistant Yumm1.7BR cells in aged (52 weeks) C57/BL6mice, +/− PLX4720 and +/− rosiglitazone. N=5 mice/group. Repeated measures analysis of variance (Anova) was calculated between samples, p<0.0001. Holm Sidak correction was performed, ***=p<0.005 (Day-18). G, Tumor growth of BRAFi resistant Yumm1.7BR cells in young (8 weeks) C57/BL6mice, +/− PLX4720 and +/− rosiglitazone (N=7 mice/group. Repeated measures analysis of variance (Anova) was calculated between samples, p<0.0001. Multiple comparision with Holm Sidak correction was done. H, Klotho levels in aged (left two panels) or young (right two panels) mouse serum (western blot) or tumors (immunohistochemistry) after treatment with rosiglitazone or vehicle control (40X objective).