Recruitment of circulating breast cancer cells is stimulated by radiotherapy

Abstract

Radiotherapy (RT) is a localized therapy that is highly effective in killing primary tumor cells located within the field of the radiation beam. We present evidence that irradiation of breast tumors can attract migrating breast cancer cells. Granulocyte-macrophage colony stimulating factor (GM-CSF) produced by tumor cells in response to radiation stimulates the recruitment of migrating tumor cells to irradiated tumors, suggesting a mechanism of tumor recurrence after radiation facilitated by transit of unirradiated, viable circulating tumor cells to irradiated tumors. Data supporting this hypothesis are presented through in vitro invasion assays and in vivo orthotopic models of breast cancer. Our work provides a mechanism for tumor recurrence in which RT attracts cells outside the radiation field to migrate to the site of treatment.

Figure 1. Supernatant from irradiated cells promotes cell invasion, but not cell proliferation, in both murine and human cell lines

(A) Transwell migration assays using supernatant (SN) from a murine breast cancer cell line (4T1) collected 2 days after irradiation (IR) to a dose of 0 (non-IR) or 20Gy (IR) (B) Transwell migration assays using SN from human lung (A549), melanoma (A375) and breast cancer cells (MD-MB-231) collected 7 days after IR to a dose of 0 (non-IR) or 20Gy (IR). (C and D) Transwell migration assays using SN from 4T1 and MDA-MB-231 cells treated with increasing doses of radiation (0, 5, 10 and 20Gy). (E and F) Proliferation of 4T1 and MDA-MB-231 cells grown with SN from control and IR cells, respectively. Scale bars represent 100µm. (*p<0.05, **p<0.01, ***p<0.001).

Figure 2. Seeding of tumors by migrating tumor cells is enhanced by irradiation in vivo in both murine and human tumor models

(A) Schematic of the donor-recipient experimental protocol. (B) Ex vivo bioluminescence images (BLI) of 20Gy irradiated (IR) and unirradiated (non-IR) 4T1 recipient tumors from 2 representative mice from each group. Quantification of the ex vivo BLI images from 22 IR 4T1 recipient tumors and 20 non-IR 4T1 recipient tumors. Values are represented on a logarithmic scale. (C) Ex vivo BLI images of IR and non-IR MDA-MB-231 recipient tumors from 2 representative mice from each group. Quantification of the images in from 5 IR MDA-MB-231 recipient tumors and 5 non-IR MDA-MB-231 recipient tumors. (D) Ex vivo BLI of non-IR and 20Gy IR 4T1 recipient tumors 10 days after intravenous injection of 4T1-luci cells. Quantification of the images in from 5 IR 4T1 recipient tumors and 5 non-IR 4T1 recipient tumors. (E) Quantification of photons released from ex vivo BLI of 4T1 recipient tumors IR or non-IR after 0, 5, and 10 days (n= 5 mice per group) (F) Quantification of ex vivo BLI of 4T1 recipient tumors 10 days after treatment with a variety of radiation doses (0, 2, 10, 15 and 20Gy) (n=5 mice per group) (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0005).

Figure 3. GM-CSF is overexpressed after irradiation in both human and murine cell lines

(A) Membranes from cytokine arrays probed with SN from 4T1 cells IR to doses of 0, 5, 10 or 20Gy and its quantification. (B) Western blot (WB) for GM-CSF and IL-8 from SN from 4T1 cells IR to doses of 0, 5, 10 and 20Gy and its quantification relative to total protein loaded (Ponceau). (C) GM-CSF WB from 4T1 IR SN with lower doses (0,1,2,3,4,5,10,15,20Gy). Quantification is relative to the total protein loaded. (D) WB analysis for GM-CSF and IL-8 in the SN of MDA-MB-231 cells treated with different doses (0, 5, 10 and 20Gy) and its quantification relative to total protein loaded. (E) ELISA analysis of GM-CSF in SN of 4T1 and MDA-MB-231 IR cells with different doses (0, 5, 10, 20Gy) (F) ELISA analysis of GM-CSF in the serum of mice bearing IR and control orthotopic 4T1 tumors. See also Suppl. Fig 2 and Suppl. Fig 3.

Figure 4. GM-CSF promotes cell invasion in vitro and tumor seeding in vivo

(A) Western blot analysis of GM-CSF in 4T1 and MDA-MB-231 cells after transfection with an shRNA targeting GM-CSF. (B) Transwell migration assay of 4T1 WT cells toward SN collected from 4T1 WT, 4T1 shRNA GFP, and 4T1 shRNA GM-CSF clones (C) Transwell migration assay using GM-CSF KD 4T1 cells and 4T1 shRNA GFP cells. SN from control or IR 4T1 WT cells was used as chemoattractant. GM-CSF neutralizing antibody or recombinant GM-CSF protein was added to IR and control SN. (D) Transwell migration assay using GM-CSF KD MDA-MB-231 cells and MDA-MB-231 shRNA GFP cells. SN from IR MDA-MB-231 WT cells was used as chemoatractant and GM-CSF neutralizing antibody was added. (E) Ex vivo BLI images and corresponding quantification for recipient tumors IR to a dose of 0 or 20Gy from WT, shGFP-expressing, or shGM-CSF-expressing 4T1 cells. (F) Tumor growth curves from 4T1 shGFP expressing and 4T1 shGM-CSF clones (#19 and #22) after 0 or 10Gy IR. (G) Ex vivo BLI images and corresponding quantification of recipient tumors IR to a dose of 0 or 10Gy, after subsequent intravenous delivery of 4T1-luci cells, from WT, shGFP-expressing, or shGM-CSF-expressing 4T1 cells. (H) Clonogenic survival of 4T1 shRNA GFP and 4T1 GM-CSF KD clones after RT. (I) A model of baseline and radiation-induced tumor reseeding. Scale bars represent 100µm. (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0005).