Tumor p38MAPK signaling enhances breast carcinoma vascularization and growth by promoting expression and deposition of pro-tumorigenic factors

Abstract

The breast carcinoma microenvironment strikingly influences cancer progression and response to therapy. Various cell types in the carcinoma microenvironment show significant activity of p38 mitogen-activated protein kinase (MAPK), although the role of p38MAPK in breast cancer progression is still poorly understood. The present study examined the contribution of tumor p38MAPK to breast carcinoma microenvironment and metastatic capacity. Inactivation of p38MAPK signaling in metastatic breast carcinoma cells was achieved by forced expression of the kinase-inactive mutant of p38/MAPK14 (a dominant-negative p38, dn-p38). Disruption of tumor p38MAPK signaling reduced growth and metastases of breast carcinoma xenografts. Importantly, dn-p38 markedly decreased tumor blood-vessel density and lumen sizes. Mechanistic studies revealed that p38 controls expression of pro-angiogenic extracellular factors such as matrix protein Fibronectin and cytokines VEGFA, IL8, and HBEGF. Tumor-associated fibroblasts enhanced tumor growth and vasculature as well as increased expression of the pro-angiogenic factors. These effects were blunted by dn-p38. Metadata analysis showed elevated expression of p38 target genes in breast cancers and this was an unfavorable marker of disease recurrence and poor-outcome. Thus, our study demonstrates that tumor p38MAPK signaling promotes breast carcinoma growth, invasive and metastatic capacities. Importantly, p38 enhances carcinoma vascularization by facilitating expression and deposition of pro-angiogenic factors. These results argue that p38MAPK is a valuable target for anticancer therapy affecting tumor vasculature. Anti-p38 drugs may provide new therapeutic strategies against breast cancer, including metastatic disease.

Keywords: angiogenesis; breast cancer; fibronectin; p38MAPK; tumor microenvironment.

Figure 1. p38MAPK contributes to breast carcinoma invasion and metastasis

(A) Immunoblotting of whole-cell lysates from breast cancer MDA-MB-231 cells transduced with empty-vector control (EGFP) or Flag-tagged p38MAPK-AGF (dn-p38). Cells were treated with 2 ng/mL TGF-β1 for the indicated times. (B) Invasion of MDA-MB-231 cells tested using Matrigel-covered transwells. Assays were done in triplicate and repeated at least twice. (C) Lung surface colonies of EGFP and dn-p38 MDA-MB-231 cells after tail-vein injection of tumor cells into female SCID mice, 6 mice/group). **, P<0.01

Figure 2. Tumor growth and lung metastasis of MDA-MB-231 cell xenografts in the orthotopic model

(A) Appearance of palpable tumors following injection of breast cancer MDA-MB-231 EGFP and dn-p38 cells into the mammary fat pad of SCID mice. (B) Tumor volume of the orthotopic xenografts of EGFP and dn-p38 MDA-MB-231 cells in female SCID mice, 6 mice/group. (C) Lung-surface colonies of tumor cells in the orthotopic xenograft model, 6 mice/group. (D) Microvessel density measured using CD31 staining of tumor sections in six fields for each tumor section (5 tumors per group) and presented as a mean number per field (0.2 mm²). **, P<0.01

Figure 3. Fibroblast-enhanced tumor growth and vasculature depend on tumor p38MAPK

(A) Graphs show tumor volumes at the endpoint of the xenograft study with empty-vector control (EGFP) and dn-p38 MDA-MB-231 breast carcinoma cells alone (Tu) or in combination with 208F fibroblasts (Tu+Fb). Images show tumors excised at the endpoint of the study. **, P<0.01. (B) Quantification of the microvessel density was done using CD31 staining of tumor sections in six fields for each tumor section (5 tumors per group) and presented as a mean number per field (0.2 mm²). (C) Blood-vessel diameters in tumor sections were measured at 400× magnification on slides immune-stained for CD31. **, P<0.01. (D) CD31 staining (blood vessels) of tumor xenograft sections of breast carcinoma MDA-MB-231 EGFP or dn-p38 cells alone (Tu) or in combination with fibroblasts (Tu+Fb); arrows mark vessels with large lumen. Enlarged images of boxed areas are shown in the right panels. (E) Staining of fibroblasts with an antibody to rPH, prolyl 4-hydroxylase. Images were taken at 400x magnification, bar size, 100μm.

Figure 4. Disruption of p38MAPK signaling does not reduce expression of MMP9 and ICAM1 by tumor cells

(A) Top panel shows gelatin zymography of 48-hour conditioned media from empty-vector control (EGFP) and dn-p38 MDA-MB-231 cells treated with 2 ng/mL TGF-β1 or 10 ng/mL TNF or their combination. Bottom panel shows immunoblotting of GAPDH, a loading control, in whole-cell lysates. (B) Immunoblot analysis of signaling markers in response to TNF in EGFP and dn-p38 MDA-MB-231 cells treated with 10 ng/mL TNF for the indicated times. (C) Immunoblotting of ICAM1 in whole-cell lysates from EGFP and dn-p38 MDA-MB-231 cells treated with 10 ng/mL TNF for the indicated times. Tubulin is a loading control. (D) Immunoblots of ICAM1 and GAPDH in lysates of MDA-MB-231 cells treated with 2 ng/mL TGF-β1 and 10 ng/mL TNF ± 5 μM SB202190, a p38MAPK inhibitor, for 24 hours.

Figure 5. p38MAPK signaling contributes to expression of Fibronectin in response to cytokines and tumor-fibroblast interactions

(A) Immunoblotting of Fibronectin and tubulin, a loading control, in whole-cell lysates from empty-vector control (EGFP) and dn-p38 MDA-MB-231 tumor cells treated with 2 ng/mL TGF-β1 or 10 ng/mL TNF or their combination for the indicated times. (B) Immunoblot analysis of Fibronectin and α-catenin, a loading control, in lysates of tumor cells treated with SB202190, a p38 inhibitor, and 2 ng/mL TGF-β1 for 24 hours. (C) Immunoblots of Fibronectin, p38, p-Smad2, p-RELA and GAPDH, a loading control, in lysates of MDA-MB-231 cells transfected with Scramble-control or siRNA to p38-alpha and then treated with 2 ng/mL TGF-β1 or 10 ng/mL TNF or their combination for the indicated times. (D-E) Immunoblotting of Fibronectin, phospho-HSP27 and tubulin, a loading control, in lysates from co-cultures (T+F) of MDA-MB-231 and A549 cancer cells (T) with 208F and WI-38 fibroblasts (F) for 72 hours. (F) Immunoblots of Fibronectin and GAPDH, a loading control, in lysates from co-cultures (T+F) of MDA-MB-231 EGFP or dn-p38 cells (T) and 208F fibroblasts (F) incubated for 48 hours.

Figure 6. Levels of p38MAPK targets are elevated in human breast cancer

(A) qRT-PCR analysis of pro-angiogenic cytokine mRNA levels in empty-vector control (EGFP) and dn-p38 MDA-MB-231 cells. (B) qRT-PCR of pro-angiogenic cytokine mRNA levels in EGFP and dn-p38 MDA-MB-231 cells individually cultured or in co-cultures with 208F fibroblasts for 48 hours. (C) Expression of Fibronectin (FN1) and VEGFA in the tumor and stromal compartments of human breast carcinomas obtained using data from the Richardson Breast Study (tumor), and the Finak Breast Study (stroma). (D) Kaplan-Meier survival estimation of the overall survival of breast cancer patients and Fibronectin levels using the Breast Cancer METABRIC dataset [35]. (E) A model of the p38MAPK role in the regulation of tumor angiogenesis in breast carcinomas. Crosstalk of tumor and fibroblast cells increases cytokine signaling via p38MAPK and TAK1/MAP3K7. p38MAPK promotes expression of Fibronectin, an extracellular matrix protein as well as pro-angiogenic cytokines including VEGFA, IL8 and HBEGF. TAK1 controls expression of MMP9 which releases matrix-bound VEGFA and activates IL8. Pro-angiogenic cytokines stimulate tumor vascularization thereby enhancing tumor growth. *, P<0.05, **, P<0.01.