Fig 2: IGF-1R activation contributes to EMT and invasion of GDF15-overexpressing breast cancer cells(A) Total protein lysates were collected from BT474 stable empty vector control clone (pCMV) and GDF15 stable clones 2, 3, and 5 (C2, C3, and C5). Western blots were performed for p-Tyr1131 IGF-1R and total IGF-1R; actin was probed as loading control. Experiments were repeated 3 times; representative blots are shown. Quantification (shown beneath each band) was normalized to actin and performed using Odyssey Li-Cor imaging software. (B) BT474 GDF15 stable clone 2 (C2) cells were treated with normal mouse IgG control or 0.25 μg/mL alpha IR3 (aIR3) IGF-1R monoclonal antibody for 48 hours. Western blots of total protein lysates were performed for total IGF-1R, vimentin, FoxM1, and E-Cadherin. Bar graphs show quantification relative to actin loading control, and was performed using Odyssey Li-Cor imaging software. Error bars represent standard deviation between triplicates; experiments were performed at least 3 times. (C) BT474 stable empty vector control clone (pCMV) and BT474 GDF15 stable clone 5 (C5) cells were pre-treated with normal mouse IgG or 0.25 μg/mL alpha IR3 (aIR3) IGF-1R monoclonal antibody for 24 hours. Cells were then seeded in Matrigel-coated Boyden chambers in serum-free media plus control IgG or aIR3; 10% FBS was added to the well as a chemo-attractant. After 24 hours of invasion, cells in chambers were fixed and stained. Representative photos of invading cells are shown at 20× magnification. The total number of invading cells was counted in 10 random fields; the average number of invading cells is shown for triplicate cultures per cell line, *p<0.05.

Fig 3: GDF15 overexpression alters breast cancer cell cycle profile(A) Western blotting (above) and real-time PCR (graph) for GDF15 in BT474, JIMT1 and MDA-MB-231 (MDA231) breast cancer cell lines. PCR values reflect fold change in GDF15 transcript normalized to RPLPO housekeeping gene. Error bars represent standard deviation between triplicate samples; experiments were repeated at least 3 times. (B) Real-time PCR for GDF15 in BT474 pCMV stable empty vector control clone (pCMV) and GDF15 stable clones 2, 3, and 5 (C2, C3, and C5). Values reflect fold change in GDF15 transcript level normalized to RPLPO housekeeping gene. Error bars represent standard deviation between triplicate samples; experiments were repeated at least 3 times. (C-D) BT474 parental, pCMV empty vector control clone (pCMV), and GDF15 stable clones 2 and 3 (C2, C3) were fixed, stained with propidium iodide, and analyzed for DNA content by flow cytometry. The percentage of cells in each cell cycle phase is shown per cell line (C) (white, G0/G1; gray, S; black, G2/M). Error bars represent standard deviation between triplicate samples; experiments were repeated at least 3 times. Representative cell cycle histograms are shown per line (D).

Fig 4: Recombinant GDF15 is anti-proliferative and induces apoptosis in the TAM-Wd cells.A. Dose response to recombinant GDF-15 protein (1, 3, 10 ng/ml) in MCF-7, TAM-R and TAM-Wd cells. Cell number was assessed after 48 hrs post-treatment. The data shown represent the cell number relative to non-treated control cells. The dose-dependent anti-proliferative effect was statistically significant from 1 ng/ml across all cell-lines (*p<0.001). B. Both the addition of recombinant GDF15 and 5-Aza/E2 caused an increase in apoptotic cells as determined by flow cytometric analysis of M30 bound, TAM-Wd cells. Increases in apoptosis were significant for both treatments (*p<0.001), although 5-Aza/E2 induced a greater response.

Fig 5: FoxM1 promotes invasion and upregulation of MMP2 and MMP9 in GDF15-overexpressing breast cancer cells(A) BT474 stable empty vector control clone (pCMV) and GDF15 stable clone 5 (C5) were transfected with 100 nM control siRNA (siCtrl) or FoxM1 siRNA (siFoxM1) for 48 hours, and then plated in serum-free media in Matrigel-coated Boyden chambers. After 24 hours, cells were fixed and stained. Representative photos of invading cells are shown at 20× magnification. The total number of invading cells was counted in 10 random fields; the average number of invading cells is shown for triplicate cultures per cell line; student’s t-test, **p<0.005. (B) BT474 GDF15 stable clone 5 (C5) cells were transfected with 100 nM control siRNA (siCtrl) or FoxM1 siRNA (FoxM1) for 48 hours. Real-time PCR was performed for FoxM1, MMP2, and MMP9. Values reflect average fold in transcript normalized to internal control RPLPO. Error bars represent standard deviation between triplicate samples; experiments were repeated 3 times, **p<0.005.