The BRAF-inhibitor PLX4720 inhibits CXCL8 secretion in BRAFV600E mutated and normal thyroid cells: a further anti-cancer effect of BRAF-inhibitors

CXCL8 is a chemokine secreted by normal and thyroid cancer cells with proven tumor-promoting effects. The presence of BRAFV600E mutation is associated with a more aggressive clinical behavior and increased ability to secrete CXCL8 by papillary-thyroid-cancer cells. Aim of this study was to test the effect of the BRAF-inhibitor (PLX4720) on the basal and TNF-α-induced CXCL8 secretions in BRAFV600E mutated (BCPAP, 8305C, 8505C), in RET/PTC rearranged (TPC-1) thyroid-cancer-cell-lines and in normal-human-thyrocytes (NHT). Cells were incubated with increasing concentrations of PLX4720 alone or in combination with TNF-α for 24-hours. CXCL8 concentrations were measured in the cell supernatants. PLX4720 dose-dependently inhibited the basal and the TNF-α-induced CXCL8 secretions in BCPAP (F: 14.3, p < 0.0001 for basal and F: 12.29 p < 0.0001 for TNF-α), 8305C (F: 407.9 p < 0.0001 for basal and F: 5.76 p < 0.0001 for TNF-α) and 8505C (F:55.24 p < 0.0001 for basal and F: 42.85 p < 0.0001 for TNF-α). No effect was found in TPC-1 (F: 1.8, p = 0.134 for basal; F: 1.6, p = 0.178 for TNF-α). In NHT an inhibitory effect was found only at the highest concentration of PLX4720 (F: 13.13 p < 0.001 for basal and F: 2.5 p < 0.01 for TNF-α). Cell migration assays showed that PLX4720 reduced both basal and CXCL8-induced cell migration in BCPAP, 8305C, 8505C and NHT but not in TPC-1 cells. These results constitutes the first demonstration that PLX4720 is able to inhibit the secretion of CXCL8 in BRAFV600E mutated thyroid cancer cells indicating that, at least some, of the anti-tumor activities of PLX4720 could be exerted through a lowering of CXCL8 in the thyroid-cancer-microenvironment.

Thyroid cancer microenvironment is composed of a mixture of immune cells and soluble mediators, which are present within and surrounding primary thyroid tumors [13][14][15] . Among soluble mediators, specific chemokines not only attract different types of immune cell into the tumor site, but also produce pro-tumorigenic actions, including pro-angiogenetic, cyto-proliferative and pro-metastatic effects [13][14][15] . Thus, the composition of thyroid cancer microenvironment, and in particular the levels of CXCL8 are currently regarded as a key factor for driving tumor progression 16 . In this scenario, no previous study investigated the role of BRAFV600E oncogene in influencing the production of CXCL8 within thyroid cancer microenvironment. Our group recently demonstrated that the BCPAP thyroid cancer cell line, harboring the BRAFV600E mutation, does secrete high amounts of CXCL8 both in basal culture conditions and after incubation with TNFα 17 . CXCL8 has well-known pro-tumorigenic effects in several human cancers including the thyroid ones 16,18 and its targeting was shown to hamper tumor progression 18,19 . Indeed, in vitro and in vivo studies demonstrated that CXCL8 plays a crucial role in promoting epithelial-mesenchimal transition (EMT) and migration/metastatization of thyroid cancer cells 20,21 . Supporting these actions, the administration of recombinant CXCL8 in xenografted mice with PTC significantly increased mortality 19,22,23 while targeting of CXCL8 with an anti-CXCL8 monoclonal antibody significantly prolonged survival 19 . Previous attempts to inhibit CXCL8 secretion in normal and neoplastic thyroid cells were only partially effective owing to the presence of multiple intracellular pathways leading to CXCL8 secretion. Thus, lowering CXCL8 concentrations in thyroid cancer microenvironment requires specific strategies depending on the specific oncogenic background of neoplastic cells 22,24,25 .
Pharmacological compounds with BRAF kinase blocking activity were shown to inhibit the secretion of CXCL8 in melanoma cell lines harboring the BRAFV600E mutation 2 , but their effect in thyroid cancer cells remains to be investigated. Among them, the Plexxikon compound PLX4720 (7-azaindole derivative) is an orally administrable selective inhibitor of BRAFV600E with proven in vitro and in vivo therapeutic efficacy in melanoma models [26][27][28] . As for the thyroid, PLX4720 was shown to inhibit the proliferation of BRAFV600E mutated thyroid cancer cell lines in vitro 29 and to reduce tumor growth and distant metastases in a human thyroid cancer model in mice 30 . Whether this BRAF-inhibitor also exerts a CXCL8-lowering effect,in thyroid cancer cells remains to be investigated. Aim of the present study was to test the effect of PLX4720 on the secretion of CXCL8 in normal human thyroid (NHT) cells, in the BCPAP, 8305C and 8505C thyroid cancer cell line harboring the BRAF V600e mutation and in the TPC-1 thyroid cancer cell line bearing the RET/PTC re-arrangement. The potential effect of PLX4720 on the basal and CXCL8-induced thyroid cell migration was also investigated.

Results
Effects of treatment with PLX4720 in terms of inhibition of CXCL8 secretion in thyroid cancer cell lines and NHT. CXCL8 concentrations were assayed in the supernatants of, BCPAP, 8305C and 8505C in basal and TNF--stimulated condition. As previously reported, TNF-elicited a significant increase in the concentrations of CXCL8 in the supernatants of BCPAP 17,31 but also in 8305C and 8505C cells (data not shown). The treatment with PLX4720 significantly and in a dose-dependent manner, inhibited the secretion of CXCL8 in: BCPAP (ANOVA F: 14.3; p < 0.0001) (Fig. 1A), in 8305C (ANOVA F: 407.9; p < 0.0001) (Fig. 1B) and in 8505C (ANOVA F: 55.24; p < 0.0001) (Fig. 1C). Post Hoc analysis performed by Bonferroni demonstrated some differences in the strength of inhibition induced by PLX4720 in different cell types. Indeed, significant inhibition of the basal secretion of CXCL8 started from a 2 µM concentration of PLX4720 in BCPAP (p < 0.001 vs. basal) (Fig. 1A) and from a 0.1 µM concentration of PLX4720 in 8305C (Fig. 1B) and 8505C (Fig. 1C); (p < 0.001 vs. basal for both 8305C and 8505C).
www.nature.com/scientificreports www.nature.com/scientificreports/ Overall, the inhibition of CXCL8 secretion was paralleled by the inhibition of ERK phosphorylation, as assessed by Western blot, in all cell types with the exception of TPC-1 cells in which no CXCL8 inhibiting effect and no reduction of ERK phosphorylation was found (Supplemental Fig. S1).
Time course of PLX4720 inhibition of CXCL8 secretion. To assess whether the CXCL8-inhibiting effect of PLX4720 could vary over time, a time course of CXCL8 secretion following treatment with PLX4720 was performed. For this set of experiments those cells in which an inhibitory effect on CXCL8 secretion was exerted by PLX4720 (NHT, BCPAP, 8505C and 8305C) were treated with PLX4720 (0, 0.1, 1, 2, 5, 10 µM) for 24, 48 and 72 hours. A first finding was that, throughout the time course the basal levels of CXCL8 progressively increased in all cell types (although at a different magnitude) (ANOVAs: F = 13.39; p < 0.0001 for BCPAP; F = 6.76; p < 0.009 for 8305C; F = 11.09; p < 0.001 for 8505C; F = 2,7; p < 0.002 for NHT) ( Fig. 2A-D). The percentages of inhibition of the CXCL8 secretion were compared for each concentration of PLX4720 at 24-h, 48-h and 72-h in each cell type. Separated ANOVAs for each cell type were performed to assess possible differences in percentages of inhibition of CXCL8 secretion by PLX4720 at each concentration throughout the time course. As shown in Fig. 3(A-D), no significant change in the inhibitory power of a given concentration of PLX4720 as assessed at 24 h, 48 h and 72 h could be found in any cell type. These findings would suggest that the CXCL8-inhibitory effect of PLX4720 is maintained (at a similar strength of inhibition) up to 72 hours.

Migration Assays.
To assess whether the PLX4720-mediated inhibition of CXCL8 secretion would produce any biological consequence, cell migration assays were performed.
As a further confirmation of these findings, wound healing assay was performed (Supplemental Fig. S2).

Discussion
The results of the present study show that the selective BRAF-inhibitor, PLX4720, inhibits the basal and the TNFα-stimulated secretion of CXCL8 in BRAFV600E mutated thyroid cancer cell lines (BCPAP, 8305C and 8505C). The inhibitory effect of PLX4720 was observed also in primary cultures of NHT, even if only at the maximal concentration. On the other hand, PLX4720 did not produce any significant effect in terms of CXCL8 inhibition in TPC-1 cells harboring RET-PTC rearrangement. It should be highlighted that, the strength of the inhibitory effect of PLX4720 (both in terms of minimal concentration able to elicit a significant inhibitory effect and in terms of percentages of inhibition) was different in relation to the specific cell type. Indeed, the CXCL8-inhibiting effect of PLX4720 was stronger in 8305C and 8505C tumor cell lines as compared to BCPAP which would fit with the notion that not all tumors may show the same effects from a single targeted agent 32 . The finding that the inhibitory effect of PLX4720 on CXCL8 secretion was not limited to BRAFV600E mutated cells, but also occurred in NHT could be regarded as unexpected and deserves to be discussed.
The fact that the concentrations of CXCL8 in the supernatants of NHT cells treated with PLX4720 progressively increased up to 72 hours of culture and that the percentage of inhibition versus untreated NHT cells remained constant would not support the hypothesis of a cytotoxic effect exerted by PLX4720. The issue of a potential cytotoxic effect of PLX4720 was addressed by previous studies on both normal and thyroid cancer cells demonstrating lack of any cytotoxic effect 29,33 . In particular, Nucera et al. demonstrated that PLX4720 treatment (1 µM or 10 µM which corresponds to the concentrations used in the present study) did not lead to apoptosis as assessed by flow cytometric analysis 33 . Taken together these data indicate that the inhibition of CXCL8 in NHT cells is not due to a cytotoxic effect of PLX4720. A more likely explanation stems from the study by Tsai et al., who demonstrated that PLX4720 does inhibits wildtype BRAF kinase activity in several tumor cell lines, but only at higher concentrations than those effective on cells harboring a mutated BRAFV600E kinase 27 . These data provide an explanation for the high concentration of PLX4720 required to produce a significant inhibition of the CXCL8 in NHT cells. The inhibitory effect of PLX4720 on the secretion of CXCL8 by NHT cells should not be underestimated because these cells are by far the most abundant ones in a neoplastic thyroid, being their contribution to the global secreted amounts of CXCL8 in the tumor microenvironment not negligible 25 . www.nature.com/scientificreports www.nature.com/scientificreports/ A further aspect deserving to be commented is that the inhibitory power of PLX4720 (expressed as % of inhibition) for a given cell type and a given concentration of PLX4720 was maintained at a similar degree over a 72-h time course indicating lack of any rebound effect of the CXCL8 inhibiting effect. The existence of a rebound effect was previously demonstrated by Montero-Conde C et al., showing that the effects of BRAF inhibition in BRAF mutant thyroid cancer cells, were hampered (after 72 hours) by the release of negative feedback mechanisms 34 .
PLX4720 was previously demonstrated to exert several anticancer effects in both xenografts models of melanoma and thyroid cancer. In animal models of melanoma, treatment with PLX4720 leads to a reduction of tumor growth and induces a higher expression of apoptotic genes and cell cycle arrests in neoplastic cells 26,27 .
As far as thyroid tumors are concerned, treatment with PLX4720 lead to a reduction of proliferation, migration and invasion of thyroid cancer cells 29,33 . The incubation of BCPAP with rh-CXCL8 produced a significant increase of BCPAP migration (Post Hoc *p < 0.0001 vs. basal). The co-incubation with rh-CXCL8 and PLX4720 10 µM significantly inhibited the BCPAP migration induced by rh-CXCL8 (Post Hoc # p < 0.001 vs. CXCL8). Panel (B) in 8305C, the treatment with PLX4720 10 µM reduced the basal migration of 8305C (ANOVA F = 161.7, p < 0.0001; Post Hoc °p < 0.005 vs. basal). The incubation of 8305C with rh-CXCL8 produced a significant increase of 8305C migration (Post Hoc *p < 0.0001 vs. basal). The co-incubation with rh-CXCL8 and PLX4720 10 µM significantly inhibited the 8305C migration induced by rh-CXCL8 (Post Hoc # p < 0.0001 vs. CXCL8). Panel (C) in 8505C the treatment with PLX4720 10 µM reduced the basal migration of 8505C (ANOVA F = 21.6, p < 0.0001; Post Hoc °p < 0.005 vs. basal). The incubation of 8505C with rh-CXCL8 produced a significant increase of 8505C migration (Post Hoc *p < 0.0001 vs. basal). The co-incubation with rh-CXCL8 and PLX4720 10 µM significantly inhibited the 8505C migration induced by rh-CXCL8 (Post Hoc # p < 0.0001 vs. CXCL8). Panel (D) in NHT, the treatment with PLX4720 10 µM reduced the basal migration of NHT (ANOVA F = 25.3, p < 0.0001; Post Hoc °p < 0.005 vs. basal). The incubation of NHT with rh-CXCL8 produced a significant increase of NHT migration (Post Hoc *p < 0.05 vs. basal). The co-incubation with rh-CXCL8 and PLX4720 10 µM significantly inhibited the NHT migration induced by rh-CXCL8 (Post Hoc # p < 0.0001 vs. CXCL8). Panel (E) in TPC-1 the treatment with PLX4720 10 µM did not reduce the basal migration of TPC-1 (ANOVA F = 4.4, p < 0.01), Post Hoc NS vs. basal). The incubation of TPC-1 with rh-CXCL8 produced a significant increase of TPC-1 migration (Post Hoc *p < 0.05 vs. basal). The co-incubation with rh-CXCL8 and PLX4720 10 µM did not reduce the TPC-1 migration induced by rh-CXCL8 (Post Hoc # NS vs. CXCL8). Bar graphs show the corresponding analysis of migrated cells on the lower side of the transwell filter. Basal migration was conventionally estimated as 100%.
www.nature.com/scientificreports www.nature.com/scientificreports/ In this regards, the results of the present study suggest that at least some of the anti-cancer effect of PLX4720 might be exerted also through a reduction of CXCL8 within the tumor microenvironment. Indeed, even if CXCL8 inhibition might not be the only mechanism, the results of migration assays experiments clearly indicated that PLX4720 was able to reduce the migration of thyroid cells. In particular, this effect was evident only in those cells in which the treatment by PLX4720 also produced an inhibition of the CXCL8 secretion (BCPAP, 8305C, 8505C, NHT) but not in those cells in which PLX4720 did not inhibit CXCL8 secretion (TPC-1). These results highlight that the ability of PLX4720 in inhibiting CXCL8 secretion would in turn, result in potentially relevant biological consequences.
The results of the present study constitute the first demonstration of a CXCL8-inhibiting effect exerted by PLX4720 in thyroid cancer cells and should be regarded as potentially relevant. Indeed, CXCL8 is a chemokine with extensively described pro-tumorigenic effects which include influencing of tumor cell growth, angiogenesis, invasiveness and EMT 16,20,35,36 . Furthermore targeting/lowering of CXCL8 levels is known to produce beneficial effects in thyroid cancer. Interestingly, a previous in vivo study showed that patients with melanoma harboring the BRAFV600E mutation experience a significant decrease in the circulating levels of CXCL8 while treated with BRAF-inhibitors 2 .
In this view, the present study was specifically aimed at evaluating whether PLX4720 (already proven to have several anti cancer effects) was also able to inhibit CXCL8 secretion which could be regarded "per se" as beneficial against cancer. In conclusion, the here reported results, would suggest a further anti-cancer effect of PLX4720 which is manifested by its ability to reduce CXCL8 secretion. Although in vivo studies will be required to assess the therapeutic benefits deriving from the inhibition of CXCL8 secretion in thyroid tumor microenvironment, this could represent an alternative therapeutic strategy, at least for those minority of patients who are refractory to conventional therapies.

Materials and Methods
Primary cultures of NHT. Surgical specimens of normal human thyroid were obtained from the contralateral disease-free lobe of patients who underwent thyroidectomy for a solitary non-functioning nodule (n = 3). The study was approved by the Institutional Board of ICS-Maugeri. Before surgery, written informed consent to the study was obtained from all patients. All the experiments were performed in accordance with the relevant guidelines and regulations. Surgical specimens were minced and then incubated with collagenase type II (Sigma, Saint Louis, MO, USA) 5 mg/ml, in 5 ml of Coon's F12 medium, for 4 h at 37 °C as previously described 37 . Then, 10 ml of Coon's F12 medium were added, following which, cells were filtered, spun at 1000 × g for 10 min, washed with Coon's F12 medium, spun again, and finally re-suspended in complete medium containing 5% newborn calf serum and a mixture of six hormones including insulin (5 μg/ml), hydrocortisone (50 μg/ml), transferrin (5 μg/ml), somatostatin (10 ng/ml), gly-his-lysine (10 ng/ml) and bovine TSH (1 mU/ml).
Thyroid tumor cell lines BCPAP, TPC-1, 8305C and 8505C. Human thyroid cancer cell lines, BCPAP harboring the BRAF V600E mutation and TPC-1 bearing the RET/PTC rearrangement, were a gift of Prof. M. Santoro (Medical School, University "Federico II" of Naples, Naples, Italy). These cell lines had been previously tested and authenticated by DNA analysis. Cancer cells were propagated in Dulbecco's Modified Eagle Medium (DMEM) (Sigma, Saint Louis, MO, USA) supplemented with 10% fetal bovine serum (Sigma, Saint Louis, MO, USA), 2 mM L-glutamine and 100 U/ml penicillin/streptomycin (Sigma, Saint Louis, MO, USA) as previously described 37 . Human thyroid cancer cell lines, 8305C and 8505C harboring the BRAF V600E mutation were previously tested and authenticated by DNA analysis. 8305C cells were propagated in Mininum Essential Medium (MEM) (Sigma, Saint Louis, MO, USA) supplemented with 10% fetal bovine serum (Sigma, Saint Louis, MO, USA), 2 mM L-glutamine and 100 U/ml penicillin/streptomycin (Sigma, Saint Louis, MO, USA). 8505C cells were propagated in RPMI medium (Sigma, Saint Louis, MO, USA) supplemented with 10% fetal bovine serum (Sigma, Saint Louis, MO, USA), 2 mM L-glutamine and 100 U/ml penicillin/streptomycin (Sigma, Saint Louis, MO, USA). Cells were incubated with the chosen stimuli in serum-free medium.
Basal and TNF-α-induced CXCL8 secretion in BCPAP, 8305C, 8505C, TPC-1 thyroid cancer cell lines and in primary cultures of NHT in the presence or absence of increasing concentrations of PLX4720. For the CXCL8 secretion assays, 3000 cells were seeded into 96-well plates in complete medium.
Time course of PLX4720 inhibition of CXCL8 secretion in BCPAP, 8305C, 8505C, TPC-1 thyroid cancer cell lines and in primary cultures of NHT. For the CXCL8 secretion assays, 3000 cells were seeded into 96-well plates in complete medium. After adherence to the plastic surface, BCPAP, 8305C, 8505C and NHT cells were incubated for 24, 48 and 72 hours with or without PLX4720 0, 0.1, 1, 2, 5, 10 µM. All experiments were performed in triplicates.
ELISA for CXCL8. CXCL8 was measured in cell supernatants of BCPAP, TPC-1, 8305C, 8505C and NHT cells using commercially available kits (R&D Systems, Minneapolis, MN). The mean minimum detectable concentration of CXCL8 was 3.5 pg/ml. The intra-and inter-assay coefficients of variation were 3.4% and 6.8%, respectively. Samples were assayed in duplicates.