The role of CCL21/CCR7 chemokine axis in breast cancer-induced lymphangiogenesis

Prior to the investigation of the role of CCL21/CCR7 pair in VEGF-C production, we have screened the constitutive expression of CCR7, CCL21, and VEGF-C in two well differentiated, luminal type (T47D, MCF-7) and two poorly differentiated basal type (Hs578t, MDA-MB-231) breast cancer cell lines (Additional file 1: Figure S1A and B). Based on these preliminary results, MDA-MB-231 breast carcinoma cell line – that is characterized by an invasive phenotype – was selected for its ability to express/secrete high levels of VEGF-C, which makes it adequate for use in a loss-of-function model. Conversely, for the gain-of-function approach, MCF-7 cell line was selected since expresses/secretes relatively low levels of VEGF-C. In this regard, CCR7 expression in MDA-MB-231 cells was knocked down with shRNA targeting CCR7 gene and the effectiveness of transfection was assessed by means of Western blot, real-time PCR, and quantitative real-time PCR (Figure 1A to C). Of note, low levels of CCR7 expression correlates with significant downregulations in VEGF-C protein and mRNA expressions (Figure 1D to F). To determine whether CCL21/CCR7 interaction regulates the secretion of lymphangiogenic factor VEGF-C, CCR7 shRNA and parental MDA-MB-231 cells were incubated with and without human CCL21/6Ckine (350 ng/ml) for 24 hours. While exogenous CCL21 has stimulated VEGF-C production by parental cells, the secretion level of VEGF-C from CCR7 shRNA transfected tumor cells has decreased significantly. The level was significantly lower compared to CCL21 treated or untreated parental cells or scrambled shRNA transfected cells (Figure 1G). Moreover, the proposed molecular mechanism responsible for the regulation of VEGF-C was analyzed in MCF-7 mock and MCF-7-CCL21-knocked in (KI) breast cancer cells and the efficacy of nucleotransfection is presented in Figure 1H and I. Quantitative PCR, Western blot, and ELISA confirmed that CCL21/CCR7 pair has the ability to regulate the expression of the lymphangiogenic factor VEGF-C in the analyzed MCF-7 breast cancer cells (Figure 1H - L). Overall, these findings suggest that CCL21/CCR7 pair has the potential to regulate VEGF-C expression/secretion in the analyzed breast cancer cell line models.

Prior studies have documented the role of phosphatidylinositol 3-kinase (PI3K) and its downstream mediator protein kinase B (AKT) in the survival and invasiveness of head and neck carcinoma cells through CCL21/CCR7 interaction [37,38]. Furthermore, extracellular signal-regulated kinase (ERK1/2) pathway is involved in cell cycle progression and survival of non-small cell lung carcinoma cells resulting from CCR7 activation [39,40].

To investigate whether AKT and ERK1/2 pathways are downstream activated by CCL21-CCR7 binding, the phosphorylation status of both pathways was assessed in MDA-MB-231 breast cancer cells. As expected, an increase in phosphorylation of AKT and ERK1/2 was observed for the entire duration of stimulation with CCL21/6Ckine (Figure 2A, B). To assess whether the phosphorylation of both proteins was dependent on the activation of CCR7, MDA-MB-231 cells were treated with various concentrations of CCR7 antibody (0, 5, 10, 20 μg/ml) before stimulation with CCL21/6Ckine. When CCR7 was blocked by its neutralizing antibody, the phosphorylation status of AKT and ERK1/2 was significantly decreased (Figure 2C, D). Similarly, the addition of PI3K/AKT and ERK1/2 specific inhibitors significantly reduced the effect of CCL21 on activation of both pathways (Figure 2E - G).

Finally, to test the functional roles of PI3K/AKT and/or ERK1/2 pathways in CCL21-induced VEGF-C secretion, MDA-MB-231 cells were pre-treated with various concentrations of PI3K inhibitor (LY294002 at 25, 50, and 75 μM), Akt-1/2 inhibitor (at 2.5, 5, and 10 μM), and ERK1/2 inhibitor (U0126 at 5, 10, and 15 μM) and then treated with CCL21/6Ckine (350 ng/ml) for 24 hours. Following that, the supernatants were collected and subjected to a VEGF-C ELISA. Both PI3K and AKT inhibitors yielded a significantly reduced VEGF-C secretion when compared with CCL21 treatment alone (Figure 2H, I), whereas no significant difference was observed for U0126 treatment (Figure 2J). As such, it can be affirmed that the activation of PI3K/AKT, but not ERK1/2 signaling pathway downstream of CCR7 is involved in the regulation of VEGF-C secretion.

Lymphatic vessel formation is a complex biological process that can be approached in vitro through separate assays, each attempting to mimic a different stage of the lymphangiogenic process such as: proliferation, migration, and formation of capillary-like tubes [41]. Within the limited scope of the current study, primary HMVECs-dLy were employed as an in vitro model of LECs and their CCR7 and CCL21 expressions were verified at the mRNA and protein levels. Agarose gel electrophoresis of real-time PCR products from HMVEC-dLy cells showed that both the chemokine receptor and ligand are expressed at the mRNA level (Figure 3A). On the other hand, lysates from cultured HMVEC-dLy cells were assayed by Western blot and CCR7 protein expression was detected at 45 kDa while only traces of CCL21 were observed (Figure 3B). Moreover, CCL21 protein secretion by HMVEC-dLy cells was quantified in 2D and 2D-matrix conditions (Matrigel was used because CCL21 is strongly matrix-binding). One important aspect to be emphasized is that CCL21 chemokine ligand is secreted as a small molecular weight protein that is readily immobilized within the extracellular matrix by binding to sulfated proteoglycans [42]. As such, there was no surprise that the determined bound CCL21 protein fraction was about twofold higher than the soluble fraction, but the overall CCL21 secretion by HMVEC-dLy was constitutively present at very low levels (Figure 3C). Therefore, serum-starved LECs were treated with various concentrations of exogenous CCL21/6Ckine (100–350 ng/ml) and their proliferation rates (BrdU uptake) increased significantly (Figure 3D). Conversely, when CCR7 activation was blocked with CCR7-neutralizing antibody (10 μg/ml), LECs proliferation significantly decreased to serum-free level (Figure 3E). LECs also responded to the chemotactic effect of CCL21 by increased migration (Figure 3F, G). Conversely, in the presence of various concentrations of CCR7-neutralizing antibody (5, 10, 20 μg/ml), CCL21-mediated LECs migration was significantly inhibited (Figure 3H). Finally, the extent of tube-like structures formed by LECs increased significantly with increasing concentrations of CCL21/6Ckine (100–350 ng/ml) (Figure 3I, J). By contrast, CCL21 mediated tubulogenesis was blocked by CCR7 neutralizing antibody (Figure 3K, L). To determine whether CCL21/CCR7 acts through the regulation of VEGFR-3 ligand, LECs were treated with CCL21 in the presence and absence of VEGFR-3 neutralizing antibody (1, 2.5, 5 μg/ml). Blocking VEGFR-3 significantly has reduced CCL21-induced LECs proliferation, migration, and tube formation (Figure 3M). Based on these results, it can be inferred that within the framework of the investigated in vitro model, CCL21/CCR7 pair has a direct lymphangiogenic potential through stimulation of the lymphangiogenic traits of LECs. Additionally, CCL21/CCR7 pair interacts in a ligand-independent manner with VEGFR-3 expressed by LECs to drive their lymphangiogenic response.

To validate our in vitro observations, Directed In Vivo Lymphangiogenic Assay (DIVLA) was conducted, as reported in the past [43]. In brief, the angioreactors containing basement membrane extract and breast cancer cells were implanted subcutaneously into the dorsal flanks of nude mice and lymphatic vessel formation was assessed through multiple approaches. Host lymphatic endothelial cells that invaded the angioreactors associated with different conditions were quantified with an immunofluorescence assay for LYVE1 and Prox1. The results revealed that the recruitment of LYVE1 and Prox1 labelled lymphatic endothelial cells was significantly higher in the angioreactors containing MCF-7-CCL21-KI cells when compared with their controls. Similarly, this recruitment was impaired when CCR7 was knocked down, as observed in the MDA-MB-231-CCR7-KD angioreactors compared to control, MDA-MB-231-mock angioreactors (Figure 4B). Relative LECs recruitment measured with quantitative real-time PCR for LYVE1 mRNA expression in cellular contents of the angioreactors also revealed a similar phenomenon: a significantly higher recruitment with MCF-7-CCL21-KI than with MCF-7-mock cells; and a significantly lower recruitment with MDA-MB-231-CCR7-KD than with MDA-MB-231 mock cells (Figure 4C). For a direct quantification of MVD and LVD, serial cryosections of angioreactors were performed and dual immunostaining was subsequently used to visualize lymphatics (LYVE1/Prox1/Podoplanin) and blood vessels (CD31) by means of “hot spot” method [34,43]. Higher MVD and LVD were observed in MCF-7-CCL21-KI sections when compared to control (MCF-7 mock). Similarly, lower densities were observed in MDA-MB-231-CCR7-KD sections than in MDA-MB-231 mock (Figure 4D – G, quantification 4I). Finally, the addition of recombinant VEGF-C to the angioreactors containing MDA-MB-231-CCR7-KD cells has rescued lymphatic vessel formation to the same level observed with control MDA-MB-231 (Figure 4H). To verify the in vivo expression of CCR7 by LECs, angioreactors containing only lymphangiogenic growth factors were implanted into nude mice and quantitative real-time PCR analysis was performed to analyze the content of the tubes (Additional file 2: Figure S2B). Thus, according to our in vivo results, CCL21/CCR7 pair is a positive regulator of lymphangiogenesis.

To characterize the role of CCL21/CCR7 in situ, we used a panel of breast cancer tissues (n = 105) collected from the primary tumor site. The demographic and clinical characterization of the samples is summarized in Additional file 3: Table S1. In brief, the majority of the tissues were classified as invasive mammary carcinomas characterized by various grades of differentiations: grade III (62%) and grade II (24%). At first, quantitative real-time PCR was employed to compare the expression profile of CCR7, CCL21, and VEGF-C in primary tumor and adjacent breast tissues (control). The stability in expression levels of two common housekeeping genes (GAPDH and ACTB) was verified across the entire panel of tissues. Of note, the level of CCR7 mRNA in tumor tissues was significantly higher than that in the control samples. A trend similar to CCR7 was observed for VEGF-C. However, CCL21 mRNA expression in tumors was not significantly different when compared to controls (Figure 5A). Moreover, it was found that CCR7 mRNA expression is positively correlated with lymph-vascular markers (LYVE1, podoplanin) and VEGF-C in breast cancer specimens while no correlation with endothelial marker CD31 was observed (Figure 5B – E). As for CCL21 expression, a weak correlation with lymph-vascular markers was observed (data not included). These results suggest a clinical association between CCR7 expression and lymph-vascular recruitment in primary breast tumor.