Silencing vimentin expression decreases pulmonary metastases in a pre-diabetic mouse model of mammary tumor progression

Abstract

Increased breast cancer risk and mortality has been associated with obesity and type 2 diabetes (T2D). Hyperinsulinemia, a key factor in obesity, pre-diabetes and T2D, has been associated with decreased breast cancer survival. In this study, a mouse model of pre-diabetes (MKR mouse) was used to investigate the mechanisms through which endogenous hyperinsulinemia promotes mammary tumor metastases. The MKR mice developed larger primary tumors and greater number of pulmonary metastases compared with wild-type (WT) mice after injection with c-Myc/Vegf overexpressing MVT-1 cells. Analysis of the primary tumors showed significant increase in vimentin protein expression in the MKR mice compared with WT. We hypothesized that vimentin was an important mediator in the effect of hyperinsulinemia on breast cancer metastasis. Lentiviral short hairpin RNA knockdown of vimentin led to a significant decrease in invasion of the MVT-1 cells and abrogated the increase in cell invasion in response to insulin. In the pre-diabetic MKR mouse, vimentin knockdown led to a decrease in pulmonary metastases. In vitro, we found that insulin increased pAKT, prevented caspase 3 activation, and increased vimentin. Inhibiting the phosphatidylinositol 3 kinase/AKT pathway, using NVP-BKM120, increased active caspase 3 and decreased vimentin levels. This study is the first to show that vimentin has an important role in tumor metastasis in vivo in the setting of pre-diabetes and endogenous hyperinsulinemia. Vimentin targeting may be an important therapeutic strategy to reduce metastases in patients with obesity, pre-diabetes or T2D.

Figure 1

Tumors from hyperinsulinemic mice have increased Vimentin protein expression. Wild type (WT) and hyperinsulinemic (MKR) were injected with MVT-1 cells (c-Myc/Vegf overexpressing cell line) into the 4^th mammary fat pad (5–12 mice per group). (A.) Mammary tumor volume was measured twice weekly with calipers. (B.) Number of surface pulmonary macrometastases in both WT and MKR mice. (C.) Representative blots showing protein extracted from tumor tissue and analyzed by Western blot for Vimentin expression (both bands). β-Actin antibody used as loading control. (D.) Densitometry of western blot (* p<0.05). All graphs represent mean per group and error bars are SEM.

Figure 2

Vimentin expression is enhanced in tumors from MKR mice compared to WT controls. (A.) Tumor sections from wild type and MKR mice analyzed by immunofluorescence for Vimentin (red) expression. Nuclei stained with DAPI (blue). (B.) Quantification of Vimentin expression n=5 per group, with 6 sections per tumor. * p<0.05. Graphs represent mean per group and error bars are SEM. (C.) MVT-1 cells were GFP tagged and injected into wild type and MKR mice. Representative tumor sections show co-expression Vimentin and GFP by immunofluorescence. Nuclei were stained with DAPI (blue).

Figure 3

Knockdown (KD) of Vimentin expression in MVT-1 cells. (A) Analysis of Vimentin protein expression from MVT-1 cells stimulated with 10nM insulin (n=3) or PBS control (n=3) for 48 hours, repeated 4 times. β-Actin was used as a loading control. Representative western blot of experiments. Densitometry of western blot from 4 experiments (* p<0.05). (B.) Lentiviral shRNA for Vimentin and control sequence was used to generate MVT-1 cells with a knockdown of Vimentin protein expression. Cells were grown, protein was harvested and protein lysates were run to analyze two different clones for Vimentin KD. Densitometry analysis of Vimentin protein expression (both bands) of the parental, control, Vimentin 317673 clone, and Vimentin 317676 clone (n=2, repeated 3 times). (C.) mRNA was extracted and gene expression levels were analyzed using quantitative Real Time PCR. (D.) 400,000 MVT-1 Control, Vimentin 317673 clone, and Vimentin 317676 clone cells were plated into 0.8μm inserts coated with matrigel in Complete Medium. The number of cells that invaded through the matrigel coated insert after 24 hours was quantified by crystal violet staining, repeated 3 times. * p<0.05. All graphs represent mean per group and error bars are SEM.

Figure 4

Insulin Signaling preserved in Vimentin Knockdown Cells. (A.) Cells were serum starved overnight and stimulated with 10nM insulin (Ins) for 15 min. Protein was extracted from the cells and analyzed by Western blotting. (n=3 per condition, repeated two times) (B.) Densitometry of pAKT/AKT. * p<0.05. Graphs represent mean per group and error bars are SEM.

Figure 5

Silencing Vimentin decreased cell invasion but not proliferation or migration. (A.) Cell counting kit 8 (CCK-8) was used for determination of cell proliferation between MVT-1 control (Ctrl) and MVT-1 Vimentin knockdown (VimKD) cells. 5,000 cells were plated in Complete Medium an analyzed for cell proliferation at 24, 48 and 72 hours, repeated twice. (B.) 150,000 MVT-1 Ctrl and MVT-1 VimKD cells were plated in 0.8μm inserts. The number of cells that migrated through the insert after 24 hours in Complete Medium was quantified by crystal violet staining, repeated three times. (C.) 400,000 MVT-1 Ctrl and MVT-1 VimKD cells were plated into 0.8μm inserts coated with matrigel. Cells were stimulated with PBS or 10nM Insulin for 24hours in Serum Free Medium. The number of cells that invaded through the matrigel-coated insert after 24 hours was quantified by crystal violet staining, repeated 3 times.

Figure 6

Knockdown of Vimentin leads to decreased number of pulmonary metastases in the hyperinsulinemic mice. Wild type (10 per group) and MKR (13 per group) mice were injected with 100,000 of MVT-1 Control (Ctrl) and 100,000 of MVT-1 Vimentin KD (VimKD) into the 4^th mammary fat pad. (A.) Tumor volumes were measured with calipers. Significant difference between the tumor sizes of MKR Ctrl and WT Ctrl group was observed. * p<0.05. (B.) Numbers of surface pulmonary macrometastases in WT and MKR mice. * p<0.05, as indicated. (C.) Tumor tissue was extracted and analyzed for Western Blot for Vimentin antibody. Representative blot is shown. Densitometry of Vimentin of representative blot n=3 WT Ctrl, n=4 WT VimKD, n=3 MKR Ctrl, and n=4 MKR VimKD. β-Actin antibody used as a loading control. Graphs represent mean per group and error bars are SEM.

Figure 7

Inhibition of PI3K/AKT signaling downregulates Vimentin expression. Analysis of the effects of a PI3K inhibitor on the expression of Vimentin was analyzed. MVT-1 cells were serum starved overnight and pretreated for 1 hour with either vehicle or 500nM of NVP-BKM120 (BKM120). Subsequently, cells were stimulated with 10nM of insulin for 48 hours (n=3 per condition, repeated 3 times). Protein was extracted from the cells and analyzed by Western blotting for (A) Vimentin expression and for (B) phosphorylated AKT and total AKT. β-Actin was used as a loading control. (B, D) Densitometry of western blots A and C, respectively (* p<0.05). Graphs represent mean per group and error bars are SEM.

Figure 8

Insulin prevents the activation of Caspase 3. MVT-1 cells were serum starved overnight and pretreated for 1 hour with either vehicle or 500nM of NVP-BKM120 (BKM120). Subsequently, cells were stimulated with 10nM of insulin for 48 hours (n=3 per condition, repeated 3 times). (A.) Western blotting for cleaved Caspase 3 and total Caspase 3 expression. (B.) Densitometry of cleaved Caspase 3/total Capsase 3 western blot (* p<0.05). (C.) Western blotting for cleaved PARP and total PARP expression. (D.) Densitometry of cleaved PARP/total PARP western blot (* p<0.05). β-Actin was used as a loading control. All graphs represent mean per group and error bars are SEM.