VX-765 br Thus these data indicated that
Thus, these data indicated that CAFs can promote breast cancer cell invasion via integrin β3.
3.3. CAF-secreted IL32 serves as a mediator between CAFs and breast cancer cells
The interactions between tumour VX-765 and the tumour micro-environment are mediated via direct (stromal cells or deposited sub-stances) or indirect (secretion of proteins) mechanisms . The CM derived from CAFs promoted breast cancer cell invasion, suggesting
that some of the secreted factors may play a role in the cross-talk be-tween breast tumour cells and CAFs. Thus, we analysed the mRNA expression profiles of CAFs and NFs derived from breast tumour tissues. A set of cytokine genes was found to be dysregulated in CAFs. Among these, 41 cytokines were up-regulated, while 58 cytokines were down-regulated (Fig. 3A), as confirmed by qRT-PCR analysis of 10 randomly chosen dysregulated mRNAs (Fig. 3B). Integrin β3 is a cell surface re-ceptor that recognises its ligand through the RGD motif. To identify the potential ligands of integrin β3 among these dysregulated cytokines, we compared these cytokines with all RGD motif–containing proteins in a
(caption on next page)
Fig. 4. IL32 associates with transmembrane integrin β3 of breast cancer cells. (A) HEK293T cells transfected with an expression vector encoding integrin β3 were cultured in the FBS-free medium with recombinant wild-type IL32 (WT) or mutant IL32 (RGD motif mutated into RGE, Mut) for 3 h. Immunoprecipitation with western blotting was conducted with an antibody against IL32 or integrin β3. (B) BT549 cells were cultured in the FBS-free medium with recombinant wild-type IL32 (WT) or mutant IL32 (RGD motif mutated into RGE, Mut) for 3 h. Immunoprecipitation with western blotting was conducted as in (A). (C) BT549 or BT549/sh-β3 (shRNA-2) cells were cultured with the CM derived from CAFs for 3 h, and the binding of IL32 and integrin β3 was confirmed by immunoprecipitation with western blotting analysis. (D). BT549 cells were treated with recombinant wild-type IL32 (rIL32-WT) or mutant IL32 (rIL32-Mut) as in (A). Confocal microscopy showing that integrin β3 and IL32 co-localised at the membrane of BT549 cells stained with a Cy3-labelled anti-IL32 antibody and FITC-labelled anti–integrin β3 antibody. (E) BT549 cells (upper panel) or BT549/sh-β3 (shRNA-2) cells (middle panel) were treated with CM derived from CAFs or BT549 cells treated with CM from CAFs/sh-IL32 cells (lower panel) and stained with the antibodies described in (D). Co-localisation of transmembrane integrin β3 and IL32 was visualised by confocal mi-croscopy. Cell nuclei were counterstained with DAPI (scale bar, 5 μm).
public database (http://prosite.expasy.org/). Three candidate cytokines including IL32, CD70, and BMP1 were identified (Fig. 3C). After checking their expression by qRT-PCR (Fig. 3D), IL32 was found to be an abundant signalling protein with the RGD motif in CAFs, as also proved by an ELISA (Fig. 3E) and Chen's findings . Besides, the enhanced IL32 production in CAFs was next corroborated by measuring its expression in another 18 paired samples of primary CAFs and NFs from patients with breast cancer (Fig. 3F and G). Furthermore, treat-ment of BT549 cells with recombinant human IL32 (rIL32) enhanced the cell invasion ability in a dose-dependent manner (Fig. 3H).
Subsequently, we wanted to know whether IL32 interacts with in-tegrin β3 at the plasma membrane of breast cancer cells. Herein, the binding of IL32 to integrin β3 was firstly proved by immunoprecipita-tion with western blotting. HEK293T cells transfected with integrin β3 were subjected to immunoprecipitation with wild-type rIL32 or mutant rIL32 (RGD mutated to RGE). As depicted in Fig. 4A and B, wild-type rIL32 bound to ectopic integrin β3 in HEK293T cells (Fig. 4A) and to endogenous integrin β3 in BT549 cells (Fig. 4B). Substitution of the RGD motif with an inactive RGE motif in IL32 abrogated the ability of IL32 to bind to integrin β3, thereby confirming the RGD-dependent interaction of IL32 with integrin β3. Consistently with these findings, the IL32 secreted from CAFs also bound to integrin β3 in BT549 cells (Fig. 4C). By contrast, after the knockdown of integrin β3 in BT549 cells, the interaction between IL32 and integrin β3 was not de-tectable (Fig. 4C). In addition, immunofluorescent staining clearly in-dicated that wild-type but not mutant IL32 can co-localise with integrin β3 at the plasma membrane of HEK293T cells (Fig. 4D). Moreover, CAF-derived IL32 co-localised with integrin β3 at the plasma membrane of BT549 cells. shRNA-mediated silencing of integrin β3 in BT549 cells or silencing of IL32 in CAFs (Fig S2A and S2B) efficiently abrogated IL32 localisation at the tumour cell membrane (Fig. 4E). Accordingly, these data showed that IL32 specifically binds to integrin β3 at the breast cancer cell membrane, suggesting that IL32 serves as a mediator of the cross-talk between CAFs and breast cancer cells.