Autocrine PDGFR signaling promotes mammary cancer metastasis

Abstract

Metastasis is the major cause of cancer morbidity, but strategies for direct interference with invasion processes are lacking. Dedifferentiated, late-stage tumor cells secrete multiple factors that represent attractive targets for therapeutic intervention. Here we show that metastatic potential of oncogenic mammary epithelial cells requires an autocrine PDGF/PDGFR loop, which is established as a consequence of TGF-beta-induced epithelial-mesenchymal transition (EMT), a faithful in vitro correlate of metastasis. The cooperation of autocrine PDGFR signaling with oncogenic Ras hyperactivates PI3K and is required for survival during EMT. Autocrine PDGFR signaling also contributes to maintenance of EMT, possibly through activation of STAT1 and other distinct pathways. Inhibition of PDGFR signaling interfered with EMT and caused apoptosis in murine and human mammary carcinoma cell lines. Consequently, overexpression of a dominant-negative PDGFR or application of the established cancer drug STI571 interfered with experimental metastasis in mice. Similarly, in mouse mammary tumor virus-Neu (MMTV-Neu) transgenic mice, TGF-beta enhanced metastasis of mammary tumors, induced EMT, and elevated PDGFR signaling. Finally, expression of PDGFRalpha and -beta correlated with invasive behavior in human mammary carcinomas. Thus, autocrine PDGFR signaling plays an essential role during cancer progression, suggesting a novel application of STI571 to therapeutically interfere with metastasis.

Figure 1. EMT-specific upregulation of PDGFR pathway genes generates an autocrine PDGF/PDGFR loop.

(A) Regulation of PDGFR pathway genes in various cell pairs based on EpH4 mouse mammary epithelial cells that undergo TGF-β–induced alterations in epithelial plasticity (see ref. 29). Two cell pairs underwent EMT in response to TGF-β (yellow), 2 underwent reversible scattering (green), and 2 cell pairs served as oncogene controls (black). Shown are the oncogenes cooperating with TGF-β, the predominant signaling pathways activated, and the fold regulation of 5 PDGF pathway genes (red: upregulation; pink: no significant regulation, blue: downregulation). (B) Concentrated, serum-free supernatants from EpRas and EpRasXT cells were tested on huPDGFR-A/FDCP-1 cells for mitogenic activity ([³H]thymidine incorporation). Specificity of the mitogenic response for PDGF was verified by addition of saturating amounts of a neutralizing α–PDGF-A/B antibody.

Figure 2. Increased metastasis induced by RTK plus TGF-β signaling in a transgenic mammary carcinoma model correlates with EMT and invasion.

(A) Spontaneous formation of lung metastases in transgenic mice bearing a normal (MMTV-Neu) or constitutively active (MMTV-C.A.-Neu) EGFR 2 (HER2) under the control of the MMTV promoter (40) before or after breeding MMTV-TGF-β1 transgenic mice (Neu/TGF-β, C.A.-Neu/TGF-β). Thirty to 36 mice per group were analyzed for metastasis formation, and differences in numbers of metastases between the different groups were significant (P < 0.01). (B) Primary tumors from MMTV-Neu mice or MMTV-Neu × MMTV-TGF-β1 mice were subjected to cryosectioning and histological staining. (C) Cryosections from the above-described tumors were stained with antibodies against vimentin (red, left panels) or tenascin C (TN-C; green) plus CD31 to indicate endothelial cells (red; see Methods). DNA (nuclei) stained with DAPI (blue). MMTV-Neu tumors expressed tenascin C only in CD31-positive endothelial cells from tumor blood vessels, while the MMTV-Neu × MMTV-TGF-β1 tumors clearly expressed tenascin C in the tumor cells themselves. Original magnification, ×40 (B and C). (D) Tumors from MMTV-Neu or MMTV-Neu × MMTV-TGF-β mice were processed for mRNA extraction and real-time PCR, using oligonucleotide primers for PDGF-A, PDGF-B, and the PDGFR target gene JE/MCP-1. The amount of mRNA obtained for double transgenic tumors was higher in all cases (P < 0.05).

Figure 3. PDGFRα and PDGFRβ expression is specifically upregulated in human late stage mammary tumors.

(A and B) Shown are typical examples of intraductal mammary carcinomas, in which no staining for PDGFRα (A) or PDGFRβ (B) could be detected. (C and D) Typical examples of invasive mammary carcinomas that show intense staining for PDGFRα (C) or PDGFRβ (D). Original magnification, ×40 (A–D).

Figure 4. Interference with autocrine PDGF signaling in EpRas cells prevents EMT by causing apoptosis.

(A) EpRas and EpRasXT cells were treated with PDGF, VEGF (control, VEGFR pathway not altered in EpRasXT cells), PDGF-neutralizing antibodies (α-PDGF-Ab), 2 nonimmune control antibodies (Contr. Ab 1 and 2), the tumor drug STI571, or a specific PDGFR tyrosine kinase inhibitor (PDGFR inh; see Methods). Levels of p-AKT were determined by Western blot analysis. For normalization, blots were stripped and reprobed for total AKT. Signals were quantified by densitometry, normalized to levels of untreated EpRas cells, and shown as histograms. Data from 3 independent experiments are represented as mean ± SD. (B) EpRas cells were seeded into collagen gels and induced or not induced to undergo EMT by addition of TGF-β, in the presence or absence of neutralizing PDGF antibodies (α-PDGF), no immune antibodies (Contr. Ab), or STI571. Cultures were photographed after 7 days. (C) Similar collagen cultures were subjected to in situ TUNEL staining (green) and counterstaining for DNA (red). Lumina of polarized epithelial structures (yellow arrows), spindle-shaped mesenchymal cells (white arrows), and TUNEL-positive nuclei (green arrows) are indicated. Original magnification, ×10 (B) and ×40 (C). (D) Quantification of the data in B (>300 cells from several gel structures were evaluated for TUNEL and DAPI staining).

Figure 5. Interference with autocrine PDGF signaling in mesenchymal EpRasXT cells causes reversal of EMT rather than apoptosis.

(A) EpRasXT cells were allowed to grow in collagen gels for 4 days and treated when indicated with the specific PDGFR tyrosine kinase inhibitor (see Figure 4) for another 5 days. Gels were stained for E-cadherin (green, top panels) and mesenchymal markers calgranulin A (red, top panels), vimentin (green, bottom panels), and CD68 (red, bottom panels). Insets: EpRas control structures stained with the same antibodies. (B) EpRasXT and EpRas cells (insets) were inhibitor-treated and stained for E-cadherin/F-actin (top panels) or vimentin/F-actin (bottom panels), before (left) or after treatment with STI571 (middle) or the specific PI3K inhibitor LY294.002 (right). Original magnification, ×60 (A) and ×100 (B).

Figure 6. dnP expressed in EpRas cells prevents EMT but not tumor growth.

(A) EpRas clones infected with retroviral vectors expressing dnP-GFP (EpRas-dnP) or GFP alone (EpRas) were analyzed for expression of exogenous dnP by Western blot analysis. (B) EpRas and EpRas-dnP cells were seeded into collagen gels and treated as indicated. Results of brightfield microscopy (top panels) and staining with α–E-cadherin, α-vimentin, and α–F-actin antibodies (bottom panels) are shown. Original magnification, ×10 (B, top panels) and ×60 (B, bottom panels). (C) Cells as analyzed in A and B were injected into the fat pads of nude mice (6 mice per cell type; 2 injection sites per mouse; 2 × 10⁵ cells per injection site) and total tumor weight determined after 2–3 weeks.

Figure 7. dnP and STI571 (Gleevec) prevent metastasis of EpRas cells in a cell-autonomous fashion.

(A) EpRas cells expressing empty GFP vector (EpRas) or dnP (EpRas-dnP) were injected into the tail veins of nude mice (5 × 10⁵ cells per mouse). Photographs of respective lungs are shown. (B) Mean numbers of lung metastases per lung (3 lungs per cell type) were quantitated by serial sectioning. (C and D) Mixtures of EpRas cells (no GFP) and GFP-expressing EpRas (43% GFP⁺ cells) or EpRas-GFP-dnP cells (37% GFP⁺ cells) were injected (5 × 10⁵ cells) into the tail veins of nude mice. After 3 weeks, cells were recultivated from individual lungs and analyzed for GFP by FACS. One day after injection, 1 mouse per group was analyzed for injected cells initially reaching the lungs, yielding 8% (EpRas/EpRas-GFP) and 12% (EpRas/EpRas-GFP-dnP) GFP⁺ cells. (C) FACS profiles of lung cell cultures from representative mice of the 2 groups. (D) Percentage of GFP⁺ cells for individual mice. (E) Nude mice were treated or not treated for 6 days with STI571 and tail vein–injected with EpRas cells or CT26 cells (5 × 10⁵ cells per animal) 1 day after start of STI571 treatment. All mice were sacrificed when controls were moribund (~3 weeks), and lungs were photographed. Lung metastases were quantified either by determining mean total lung weights from a total of 6–9 lungs each (F; mean weight ± SD) or by serial sectioning to determine mean numbers of metastases per lung (G; mean ± SD from 3 lungs total).