Melanoma is an aggressive form of skin cancer characterized by poor prognosis in the late stages. However, in recent years, outcomes of melanoma patients have greatly improved due to new therapies specifically targeting oncogenic driver mutations or immune checkpoints [1
]. Oncogenic mutations in BRAF are present in up to 50% of melanomas, the most frequent being a valine to glutamic acid at position 600 (V600E), which constitutively activates the BRAF/MEK/ERK-signaling pathway, transmitting constant cell growth signals [2
The BRAF inhibitors (BRAFi) vemurafenib and dabrafenib, the MEK inhibitor (MEKi) trametinib, and combinations of BRAFi and MEKi (dabrafenib + trametinib, vemurafenib + cobimetinib, encorafenib + binimetinib) are presently approved as first line therapies for patients with unresectable or metastatic melanoma harboring BRAFV600E
(vemurafenib, dabrafenib) or BRAFV600E/K
(trametinib, combined therapies) [3
]. Unfortunately, although most patients experience a remarkable initial response to BRAFi and MEKi, the long-term efficacy of therapy is limited by the development of secondary drug resistance in the majority of patients [6
Several mechanisms are considered responsible for acquired resistance to targeted therapy, including secondary mutations, bypass signaling and activation of other compensatory downstream effectors, modifications of the tumor microenvironment (TME), and cross talk with the immune system [7
]. In particular, melanoma is characterized by a remarkable metabolic plasticity that leads to the development of resistance mechanisms through metabolic interactions between tumor cells and the TME [9
]. Moreover, it has been demonstrated that treatment with BRAFi/MEKi can induce TME modifications through autocrine and paracrine effects [10
]. In fact, melanoma is an immunogenic cancer that escapes immune surveillance through the production of specific cytokines and growth factors in the TME [11
Dendritic cells (DCs) are professional antigen-presenting cells physiologically present in tissues, and they exert a pivotal role in immune surveillance through the regulation of both innate and adaptive immunity. Appropriate stimuli—such as pathogen-associated molecular patterns (PAMPS) able to stimulate Toll-like receptors or cytokines—cause the switch from immature to mature DCs with a phenotype characterized by increased expression of both stimulatory and co-stimulatory molecules, the acquisition of activation markers, and the active secretion of cytokines and/or chemokines [12
]. DCs are a critical component of antitumor immunity, being potent inducers of T cell responses. On the other hand, defects in DC maturation and function have also been described in several types of cancer, including melanoma [13
], suggesting that, under certain circumstances, DCs can contribute to immune suppression and tumor progression [14
]. A better knowledge of whether and how the secretory pathways of melanoma cells with acquired resistance to BRAFi lead to modifications of the TME and affect DCs is essential for developing new therapeutic approaches aimed at increasing drug sensitivity and overcoming the emergence of secondary resistance.
In this study, the capacity of conditioned media from vemurafenib-resistant melanoma cells to modulate or interfere with DC activation was studied. Moreover, proteins released in the medium by melanoma cells showing acquired resistance were identified to better clarify their interplay with the TME.
2. Materials and Methods
2.1. Chemicals and Reagents
Roswell Park Memorial Institute medium (RPMI-1640), phosphate-buffered saline without Ca++ and Mg++ (PBS), glutamine, penicillin (10,000 UI/mL), and streptomycin (10,000 μg/mL) were from Eurobio Laboratoires (Le Ulis Cedex, France). Fetal calf serum (FCS) was from HyClone (South Logan, UT, USA). All solvents were purchased from Mallinckrodt Baker (Milan, Italy). Dimethyl sulfoxide (DMSO), lipopolysaccharide (LPS) of Escherichia coli, and all other reagents were from Sigma Chemicals (St. Louis, MO, USA). Vemurafenib (from Selleck Chemicals, Houston, TX, USA) and bortezomib (from Santa Cruz Biotechnology, Santa Cruz, CA, USA) were dissolved in DMSO. Matrigel (MG) was from Becton Dickinson Bioscience (Franklin Lakes, NJ, USA). Granulocyte-macrophages colony stimulating factor (GM-CSF) was from Sandoz (Basel, Switzerland), and recombinant human interleukin (IL)-4 was purchased from R&D Systems (Minneapolis, MN, USA).
2.2. Cell Culture and Generation of Vemurafenib-Resistant Cell Lines
The BRAF-mutant (V600E) human melanoma cell line SK-MEL-28 [15
] was obtained from the American Type Culture Collection (ATCC). Cells were cultured in RPMI-1640, supplemented with 10% FCS, 0.05% L-glutamine, penicillin (100 U/mL), and streptomycin (100 μg/mL) and maintained at 37 °C in a 5% CO2
humidified atmosphere. Vemurafenib-resistant (VR) variants (namely VR2 and VR3) were derived from the original parental cell line according to a published procedure [16
]. Briefly, human melanoma cells were initially treated with 20 µM vemurafenib and then cultured in complete medium containing 5 µM vemurafenib for at least 3 months before they were used for the subsequent studies. Several resistant subcultures were obtained, and these cells were further propagated in growth medium containing 2 µM vemurafenib. Resistant clones were cultured for one cell cycle in the absence of vemurafenib before each experiment.
Evaluation of cell growth in the presence of vemurafenib was achieved by a colori-metric sulphorhodamine B (SRB) assay (cell density determination based on cellular pro-tein content), as previously described [17
]. Briefly, melanoma cells (4 × 103
cells/well) were seeded and grown in 96-well plates for 24 h and then exposed to different concentrations of vemurafenib for 72 h. Subsequently, cells were fixed with 50% trichloroacetic acid and stained with 0.2% SRB solution. Cell proliferation was determined by spectrophotometric quantification (540 nm wavelength).
2.3. Expression Levels of Cytokines/Chemokines in Conditioned Medium
Cytokine/chemokine quantification in cell cultures was achieved by xMAP technology through a Luminex platform (Bio-Rad Laboratories, Hercules, CA, USA) equipped with a magnetic washer workstation according to the manufacturer’s protocol. Melanoma cell lines were cultured in complete medium for 24 h. Then cells were washed in PBS, and fresh complete medium was applied. The conditioned medium (CM) was collected 72 h later and stored at −20 °C until needed. Samples were analyzed (using a Bio-Plex Pro human cytokine multiplex assay) for IL-1β, IL-4, IL-6, IL-8, IL-10, IL-12, tumor necrosis factor(TNF)-α, interferon(IFN)-γ, Eotaxin/CCL11, basic fibroblast growth factor (bFGF), GM-CSF, granulocyte-colony stimulating factor (G-CSF), monocyte chemoattractant protein-1 (MCP-1/CCL2), RANTES/CCL5 (Regulated on Activation, Normal T-expressed and Secreted), macrophage inflammatory protein 1 alpha (MIP-1α/CCL3), MIP-1β/CCL4, IFN-γ-inducible protein 10 (IP-10/CXCL10), and vascular endothelial growth factor (VEGF). The quantification was carried out with a Bio-Plex array reader (Bio-Plex 200 System) and Bio-Plex Manager (Version 6.1 Bio-Rad Laboratories, Hercules, CA, USA) software.
2.4. Generation of Dendritic Cells and Swap Experiments
Human DCs were differentiated from monocytes isolated from human peripheral blood mononuclear cells of healthy donors according to published methods [18
]. Briefly, cells were isolated by Ficoll density gradient, and then monocytes were positively sorted using anti-CD14-labeled magnetic beads (MACS, Miltenyi Biotech, Germany) according to the manufacturer’s instructions. The sorted cells, obtained from two different healthy donors, were cultured for 5 days in complete medium supplemented with 25 ng/mL GM-CSF and 1000 U/mL IL-4 at a cell density of 4 × 105
cells/mL in 6-well plates (3 mL/well) to induce their differentiation into DCs.
For medium swap experiments at the 5th day of culture, DCs were collected, washed, and plated overnight at 5 × 105 cells/mL in 24-well plates (1 mL/well) in complete medium containing melanoma CM previously analyzed for cytokines/chemokines content (20% vol/vol). To control maturation, 0.2 µg/mL LPS was added overnight. When required, melanoma CM was added overnight together with LPS.
2.5. FACS Analysis
DCs were collected and stained for 30 min at 4 °C with the following mAbs: anti-human HLA class I, HLA-DR class II, CD80, CD83, and CD86 or appropriate isotype controls (all from BD Pharmingen, San Diego, CA, USA) according to the manufacturer’s instructions. Then, stained cells were analyzed using a Beckman Coulter Gallios flow cytometer equipped with three lasers and Kaluza Software (Beckman Coulter), acquiring 2 × 104 events gated according to DC forward and size scatters.
2.6. Collection of Cell Line Secretome, Mass Spectrometry, and Protein Identification
Melanoma cells (SK-MEL-28 and SK-MEL-28-VR2) were cultured in standard conditions until they reached 70% confluence. Culture medium was then removed, and cells were washed three times with PBS to eliminate residual FCS. To collect the secretome, cells were incubated with serum-free medium for 24 h at 37 °C. For each cell type, serum-free conditioned media (SF-CM) from three different flasks were pooled and centrifuged to remove cellular debris. SF-CM were then concentrated (10-fold) using MWCO of 3 kDa Ultra centrifugal filters (Amicon Millipore, Billerica, MA, USA) at 4 °C.
To optimize their proteomic identification, proteins (10 µg) from concentrated SF-CM were denatured and electrophoretically separated as described [19
]. Briefly, the total gel lanes were cut, and proteins were reduced, alkylated, and digested overnight with bovine trypsin sequencing grade (Roche Applied Science, Monza, Italy). The peptide mixtures were then analyzed by nano-reversed-phase liquid chromatography tandem mass spectrometry (RP-LC-MS/MS), and proteins were identified as described [19
2.7. Bioinformatics Analysis
Differentially expressed proteins identified by proteomic analysis were further analyzed by the Database for Annotation, Visualization, and Integrated Discovery (version 6.8, DAVID) software (http://david.abcc.ncifcrf.gov/
]. DAVID functional annotation cluster analysis was performed on the list of proteins of SF-CM from SK-MEL-28 and SK-MEL-28-VR2 identified by LC-MS/MS analysis. Terms with a p
-value ≤ 0.05 were selected for DAVID analysis. The gene ontology (GO) terms of cellular component (GOTERM_CC_FAT), molecular function (GOTERM_MF_DIRECT), biological processes (GOTERM_BP_DIRECT), and UP_KEYWORDS in the functional categories section in DAVID were used. To obtain GO (significant q
-value threshold level of <0.05) of differentially expressed proteins, we also used the GOnet database (https://tools.dice-database.org/GOnet/
]. Proteins were categorized according to GO molecular function. Functional protein–protein interactions were also analyzed using STRING software, version 10.5 (http://string-db.org
). The interaction networks were obtained on the basis of confidence scores (threshold score 0.4 with no more than 5 interactors) as described [23
2.8. Evaluation of Melanoma Cell Sensitivity to Bortezomib, Adhesion Assay, and Determination of CD147/Basigin and MMP-2
For proliferation studies, cells (8 × 104) were seeded and grown for 24 h in 6-well plates. Thereafter, cells were treated with bortezomib (0, 10, 20, and 40 nM) for 24 h. Cells were then harvested and counted with a Neubauer modified chamber.
The adhesion assay was performed on 24-well plates coated with MG (50 μg/well). After MG polymerization, cells were seeded at a density of 1 × 106 cells/mL, followed by incubation at 37 °C for 1 h. The adherent cells were detached with trypsin/EDTA and counted. Attachment to MG was expressed as the percentage of cells adhered, and the percentage of parental cell line was taken as 100%.
Surface expression of CD147/basigin (or extracellular matrix metalloproteinase inducer, EMMPRIN) in melanoma cells was determined by flow cytometry using a FITC-conjugated anti-CD147 antibody (BD Pharmingen, kindly provided by Dr. Elvira Pelosi, Istituto Superiore di Sanità, Rome, Italy), as previously described [20
]. CD147/basigin and matrix metalloproteinase 2 (MMP-2) in CM were quantified using a human magnetic Luminex assay (R&D Systems, Minneapolis, MN, USA) according to manufacturer’s instructions.
Plasma levels of CD147/basigin and MMP-2 were determined (as described in the previous paragraph) also in 5 patients with BRAFV600-mutant metastatic cutaneous melanoma treated with either vemurafenib or vemurafenib plus cobimetinib at Istituto Dermopatico dell’Immacolata, IDI-IRCCS; from such patients, peripheral blood samples were collected both before the start of therapy and at disease progression. Baseline evaluation included medical history, physical examination, and radiologic tumor assessment with computer tomography (CT) or positron emission tomography scans. Vemurafenib (Zelboraf) was given at a dose of 960 mg/bid, and vemurafenib plus cobimetinib (Cotellic) at a dose of 960 mg/bid and 60 mg/qd, respectively, for three weeks with one week of break. All patients underwent physical examination and assessment of biochemical parameters monthly, whereas tumor response was determined with CT every three months. Tumor response was classified according to RECIST 1.1 criteria. Time-to-treatment-failure (TTF) was defined as the time from the start of therapy to the first observation of disease progression per RECIST 1.1. The study was conducted in accordance with good clinical practice guidelines and the Declaration of Helsinki. The study was approved by the IDI-IRCCS Ethics Committee (ID #407/1, 2013 and #407/2, 2016), and a written informed consent was obtained from the patients.
2.10. Plasma Preparation
Blood was collected into BD vacutainer tubes (#367704, BD Biosciences, Plymouth, UK) and centrifuged at 1200× g for 10 min at 4 °C. Plasma was collected and centrifuged again at 1200× g for 10 min at 4 C°, aliquoted, and stored at −80 °C until use.
2.11. Statistical Analysis
Results are expressed as means of three independent experiments ± standard deviations (SDs). The statistical significance of differences was determined by two-tailed t-tests; the significance threshold was set at p ≤ 0.01. The analysis of the plasma CD147/basigin and MMP-2 expression levels was carried out with Mann–Whitney tests; the significance threshold was set at p < 0.05.
Understanding the interaction between tumor cells and the TME is essential to identifying new therapeutic strategies. Indeed, the cross talk between tumor cells and the TME contributes to tumor progression, metastasis, and response to therapy [29
]. In this regard, it has been clearly demonstrated that the TME is able to support proliferation and BRAFi resistance in melanoma [30
]. Soluble mediators play a key role in the cross talk between tumor cells and TME. Therefore, to get further insight into the role of melanoma cell secretome in BRAFi resistance, we first evaluated the secretion of a set of cytokines/chemokines—known to be implicated in the modulation of immune responses and/or tumor cell proliferation and invasiveness—in melanoma cells sensitive or with acquired resistance to vemurafenib and assessed whether CM from these cells could affect DC maturation. We then carried out a multiplex analysis of CM of vemurafenib-sensitive and vemurafenib -resistant cells to identify additional soluble mediators potentially involved in BRAFi resistance.
As compared with parental cells, both vemurafenib-resistant cell clones showed in-creased secretion of IL-10, considered an important autocrine growth factor for malignant melanoma [32
] and found to be upregulated in vemurafenib-resistant cells [33
] of VEGF, a potent angiogenic factor also involved in BRAFi-resistance of melanoma cells [34
], and of IL-1β and IL-8, cytokines that promote inflammation, tumorigenesis, and invasiveness [35
]. Notably, increased expression of IL-8 has been found to be associated with multidrug resistance in breast cancer cells, sunitinib resistance in renal cell carcinoma, and RO4929097 (a γ-secretase inhibitor) resistance in NSCLC cells [36
]. Upregulation of IL-8 secretion and MMP-2 activity was also reported by Sandri et al. [37
] in melanoma cells with acquired resistance to vemurafenib, and our data are consistent with those findings and with the knowledge that MMP-2 activity is promoted by IL-8 expression [38
]. On the other hand, a recent study by Hartman et al. [39
] analyzing six different cell lines with acquired resistance to vemurafenib and their drug-sensitive counterparts evidenced a de-crease of IL-8 expression in four resistant cell lines and no changes in the other two cell lines. In the same study, a heterogenous phenotype of vemurafenib-resistant cells was also observed for MMP-2 expression, which was decreased in three cell lines, upregulated in one cell line, and not affected in the remaining two cell lines. Melanoma is a heterogenous tumor, and previous studies have shown that the transcriptional heterogeneity of melanoma cells affects their initial response to BRAFi and the development of resistance [40
]. It is possible to hypothesize that the genetic background and the transcriptional heterogeneity at baseline of the melanoma cell lines used in the different studies to investigate mechanisms underlying acquired resistance to vemurafenib might be responsible for the diverse modulation of IL-8 and MMP-2 expression observed in the drug-resistant sublines. In the VR2 clone, we also detected enhanced secretion of the chemokines MCP-1, MIP-1α, MIP-1β, RANTES, and eotaxin, which play a key role in monocytes and DC chemotaxis [41
]. Notably, MCP-1, also considered as autocrine growth factor for melanoma and a metastasis-inducer, is produced by BRAF-resistant melanoma cells, and its plasma level was shown to increase in melanoma patients during BRAFi treatments [42
The presence of cytokines and chemokines able to induce DC maturation [41
] prompted us to study the effects of melanoma CM on human DCs in order to extend the knowledge of the possible functional consequences of increased cytokine and chemokine secretion by vemurafenib-resistant melanoma cells. CM derived from the drug-resistant cells induced DC maturation, as was shown in the results of both phenotypic and cytokine/chemokine secretion analyses. Indeed, DCs exposed to CM from resistant cells showed increased expression of activation markers and released higher levels of pro-inflammatory factors, as was reported for IL-1β [35
], than DCs treated with CMs from parental cells. Notably, IL-6 and MCP-1 were secreted at concentrations even higher than those observed in LPS-stimulated DCs. The interplay between melanoma and stromal cells in the TME, including DCs, is mainly supported by presence of IL-6, IL-10, and VEGF [13
]. IL-6 promotes tumor growth by inhibition of apoptosis and induction of tumor angiogenesis. Increased serum concentration of IL-6 has been correlated with a worse prognosis in patients with melanoma, even if the specific biological functions of IL-6 in progression of melanoma are unknown [45
]. It is noteworthy that the stimulation of DCs with CM from vemurafenib-resistant cells induced a significant release of IL-10, an immunosuppressive cytokine whose expression correlates with melanoma progression and metastasis [46
]. It is well demonstrated, indeed, that melanoma alters DC in a pro-tumorigenic way and in a VEGF-dependent manner too [47
]. Interestingly, the presence in melanoma CM of cytokines with antitumor activity, such as IL-12 and IFN-γ, has a potential role in DC maturation [49
]. Accumulating evidence suggests that TME-located DCs, with intermediate mature states and expressing high levels of pro-inflammatory signals, might be considered as facilitators in cancer progression [50
]. Our observations indicate for the first time that vemurafenib-resistant melanoma cells can modify DC maturation in order to benefit from their activation and subsequent cytokine production. Therefore, the delicate balance in the TME of cytokines levels may contribute, finally, to BRAFi resistance. Melanoma CM did not interfere with the upregulation of surface antigens on LPS-stimulated DCs (data not shown), suggesting their inability to interfere with DC under strong stimulation. In this study we investigated the effects of melanoma CM on DCs derived from monocytes of healthy donors. Of course, it is possible that in melanoma patients other factors may interfere with DC maturation, including the prevalence of other DC populations, such as the plasmacytoid or the CD34+
derived DCs, as well as a disabled maturation of monocytes into DC.
The mass spectrometry analysis identified 22 and 48 specific proteins in SK-MEL-28 and SK-MEL-28-VR2 cells, respectively. Enrichment analysis of DAVID (regarding the involved biological processes) highlighted that resistant cells suggest a higher activation of oxidative metabolism, in agreement with other studies [52
]. Moreover, STRING analysis suggested a possible increase in the activity of the proteasome pathway in those cells. The proteasome is a large multi-subunit complex present in nucleus and cytoplasm that controls degradation of intracellular proteins. Due to this activity, the proteasome is strictly involved in many regulatory pathways within the tumor cell, including proliferation and apoptosis [24
]. It should be noted that the presence of typical intracellular proteins (i.e., proteasome subunits) in CM was expected [54
], since proteasome subunits occurrence in extracellular space and in plasma has been reported [55
]. The proteasome role in BRAF-mutant cells has been widely explored. Recent studies have revealed that BRAF mutation enhances proteasome capacity and resistance to proteasome inhibitors in myeloma patients [56
]. On the other hand, Zecchin and colleagues clearly demonstrated that proteasome inhibitors possess a significant selectivity toward BRAFV600E
-mutant colorectal cancer cells as a consequence of persistent BRAF signaling and a nononcogenic addiction to the proteasome function in those cells [57
]. Furthermore, the combination of bortezomib and vemurafenib was shown to produce synergistic antitumor effects in thyroid cancer, both in vitro and in a xenograft model [58
]. In the present study, we show that our vemurafenib-resistant cells were more sensitive to bortezomib than parental cells. Based on the results of STRING analysis, we can hypothesize that the development of resistance to BRAFi renders the cells more dependent on proteasome function for proliferation and/or survival, leading to increased susceptibility to proteasome-targeting drugs. Further studies are however required to define the therapeutic potential of proteasome inhibitors in BRAFi-resistant melanomas since these agents have also been shown to have both immunosuppressive and immunostimulatory effects [59
Bioinformatic analysis also indicated that vemurafenib-resistant cells expressed proteins related to cell adhesion and cell–cell and cell–matrix interactions. For example, we detected (Table 1
) the presence of thrombospondin-1, which is involved in a melanoma epithelial-to-mesenchymal transition (EMT)-like process [60
], or to protein FAM3C, and possibly related to EMT, tumor progression, and metastasis [61
]. We also detected neuropilin-2, whose expression is known to be associated with melanoma progression [62
]. However, we focused our attention on CD147/basigin since its expression was previously found to be upregulated in vemurafenib resistant cells [63
]. CD147/basigin is a transmembrane protein member of the immunoglobulin superfamily that can shed from the cell membrane via an MMPs-dependent cleavage [64
]. This soluble CD147 acts as a paracrine molecule able to stimulate the production of MMPs, with a consequent increase in the invasiveness of cancer cells [65
]. The role of CD147/basigin in promoting melanoma proliferation, angiogenesis, progression, and metastasis is well documented [66
]. In particular, downregulation of CD147/basigin induces apoptosis in melanoma cells [67
] and impairs VEGF production [68
]. Moreover, this molecule promotes tumor cell invasiveness by regulating MMP expression, including MMP-2, in neighboring fibroblasts or cancer cells [69
], and previous studies have demonstrated that metalloproteases’ expression and enzymatic activity play an important role in determining an aggressive phenotype of melanoma cells [63
]. As demonstrated by proteomic analyses and multiplex assays, VR2 and VR3 cells secreted a higher level of CD147/basigin than parental cells, a finding consistent with the enhanced secretion of VEGF and MMP-2 detected in the resistant cells [70
] as well as with their predicted increased oxidative metabolism [71
]. In this study, we demonstrated that in patients who developed resistance to vemurafenib or vemurafenib plus cobimetinib, plasma levels of CD147/basigin at disease progression were significantly higher than those detected before the start of therapy. These results, although preliminary and confirmed only in a small number of patients, suggest a possible contribution of CD147/basigin upregulation in BRAFi/MEKi resistance. Interestingly, several inhibitors [72
] or an antibody [73
] targeting CD147/basigin have been suggested as potential therapeutic agents.