Wiskott-Aldrich syndrome protein regulates leukocyte-dependent breast cancer metastasis

Abstract

A paracrine interaction between epidermal growth factor (EGF)-secreting tumor-associated macrophages (TAMs) and colony-stimulating factor 1 (CSF-1)-secreting breast carcinoma cells promotes invasion and metastasis. Here, we show that mice deficient in the hematopoietic-cell-specific Wiskott-Aldrich syndrome protein (WASp) are unable to support TAM-dependent carcinoma cell invasion and metastasis in both orthotopic and transgenic models of mammary tumorigenesis. Motility and invasion defects of tumor cells were recapitulated ex vivo upon coculture with WASp(-/-) macrophages. Mechanistically, WASp is required for macrophages to migrate toward CSF-1-producing carcinoma cells, as well as for the release of EGF through metalloprotease-dependent shedding of EGF from the cell surface of macrophages. Our findings suggest that WASp acts to support both the migration of TAMs and the production of EGF, which in concert promote breast tumor metastasis.

Figure 1. WASp is required for cell elongation and invasion of both breast carcinoma cells and macrophages

MTLn3GFP cells cultured alone or with WASp^+/+, WASp^−/− or WASp^−/− rescued (WASp-Re) BMMs for 16 hr. (A) Representative phase-contrast images of the indicated conditions. MTLn3 cells are denoted by a C in co-culture images. Scale bar, 10 μm. (B,C) Quantification of average elongation of (B) MTLn3GFP and (C) the indicated macrophage types. n ≥ 4 independent experiments. * p < 0.05, ** p < 0.01 compared to control cells. Quantification of the fraction of MTLn3 (D) or BMMs (E) invading above 20 μm in the absence (−) or presence of WASp^+/+ or WASp^−/− BMMs from 3 independent mice. Error bars signify SEM from 5 independent fields. **** p < 0.0001 compared to MTLn3 cultured conditions with WASp^+/+ BMMs. See also Figure S1 and Movies S1 and S2.

Figure 2. WASp is required for TAM-dependent breast carcinoma cell metastasis but not primary tumor growth in a rat orthotopic breast tumor model

(A) Primary MTLn3-Cerulean tumor growth was determined by average day of tumor detection after injection, age of primary tumors and volume of primary tumors upon harvest. (B) Representative fluorescent images of spontaneous lung metastases. Scale bar, 10 μm. (C) Quantification of spontaneous lung metastases; number of metastatic nodules (< 5 cells in cluster; left graph) or total carcinoma cell number from the whole lung (right graph). n= 9 WASp^+/− and 11 WASp^−/− tumor bearing mice. * p < 0.05 ** p < 0.01 by two-tailed Student’s t-test and Mann-Whitney significance test. Error bars signify SEM. See also Figure S2.

Figure 3. WASp is required for TAM-dependent carcinoma cell motility, invasion and intravasation in vivo

(A) The number of Texas-red dextran positive TAMs from WASp^+/− and WASp^−/− mice was quantified 24 hours after Texas-red dextran injection. Images were acquired at 5 μm steps starting from the tumor cortex to a final depth of 100 μm using intravital two-photon microscopy. (B) Quantification of the average number of motile MTLn3 cells per field determined by intravital imaging. n > 3 mice per genotype. * p < 0.05 using Mann-Whitney’s significance test. See also Movies S3 and S4. (C) Total number of invasive cells collected into microneedles containing either 25 nM EGF or buffer. n=4 mice per genotype. ** p < 0.01 Student’s T-test compared to buffer containing needles in WASp^+/− mice. (D) Tumor cell blood burden in WASp^+/− or WASp^−/− mice bearing MTLn3-derived breast tumor following 3 weeks of tumor growth. Graphs show average number of cells (left) and colonies (right) per mL of blood. n > 4 mice per genotype. Error bars signify SEM. * p < 0.05.

Figure 4. WASp is required for EGF release

(A) Quantification of in vitro 3D invasion using MTLn3GFP cells cultured alone (−) or with conditioned media (CM) from WASp^+/+ or WASp^−/− BMMs. (B) Quantification of 3D invasion of MTLn3GFP cells in the absence (−) or presence of WASp^+/+ conditioned media with either a non-specific IgG or two different EGF neutralizing antibodies (αEGF Ab1 or Ab2). (C) Addition of exogenous EGF (5nM) to WASp^−/− BMM conditioned media restores invasion of MTLn3GFP cells. n≥3 experiments per condition. (D) Confocal sections of control sh and shWASp RAW/LR5 cells expressing APEGF-sfGFP. Unpermeabilized fixed cells were stained with an antibody against extracellular AP. Scale bar, 10 μm. (E) Equal expression levels of APEGF-sfGFP (arrow) expressed by the control sh and shWASp cells, detected with an antibody against GFP. (F) Conditioned media from the indicated APEGF-sfGFP expressing cells under the conditions shown were collected and assayed for shed AP activity. (G) Quantification of MTLn3GFP invasion with CM collected from WASp^−/− BMM infected with either a control vector or a vector expressing recombinant soluble EGF. Error bars signify SEM. * p < 0.05, ** p<0.01 and ***p<0.001. See also Figures S3 and S4.