Introduction

Melanoma is one of the most aggressive cancers, whose incidence is predicted to rise globally by 2040 in both sexes (63.5% in males and 49.5% in females) and the number of estimated deaths reveals a similar trend1.

Melanoma commonly arises from a proliferative benign lesions in which melanocytes have lost cell-cycle control, evaded the restrictive effects of tumor suppressor mechanisms, and bypassed the senescent state. The multistep sequence of genetic mutations that leads to the acquisition of a frankly malignant phenotype in melanoma depends on multiple intrinsic and extrinsic factors, among others genetic predisposition and exposure to ultraviolet (UV) radiation-induced DNA damage2. The genetic classification of melanoma identify four different subtypes according to specific mutations on genes involved in the activation of the mitogen activated protein kinese/extracellular signal-regulated kinase (MAPK/ERK) pathway. This classification inculdes three groups of melanoma lesions in which cancer cells harbor v-Raf murine sarcoma viral oncogene homolog B1 (BRAF), neuroblastoma RAS viral (v-ras) oncogene homolog (NRAS), and neurofibromin 1 (NF1) mutations, and a fourth group lacking mutations in these genes. Among the four different groups of melanoma, BRAF-mutated melanoma accounts for about 50% of the cases3,4.

Recently, a series of recommendations for cutaneous melanoma diagnosis and treatment have been proposed by a multidisciplanary consensus of experts. The consensus highlights the critical importance of preventing local recurrence and emphasizes that monotherapy is probably inadequate for patients with advanced and high-risk clinical stages5. According to this consensus, the surgical excision of the whole lesion with a safety margin (1–3 mm) represents the standard of care for cutaneous primary melanoma patients. Furthermore, sentinel lymph node biopsy should be performed in patients with tumor thickness ≥ 1.0 mm to ensure accurate staging and guide treatment decisions.

Radiotherapy of the primary tumor is rarely indicated. Adjuvant therapies with Immune Checkpoint Inhibitors (ICI, i.e. ipilimumab plus nivolumab) or a combination of BRAF and Mitogen-activated protein kinase kinase (MEK) inhibitors, have been proposed in completely resected stage IIB-IV melanoma. In addition, for patients with clinically detectable, resectable macroscopic disease, neoadjuvant therapy with ICI, may be administered prior to complete surgical resection. Next, adjuvant treatment, tailored according to the pathological response and BRAF mutation status, may be offered. Treatment for patients with disease recurrence, irresectable stage III/IV disease, and stage IV melanoma is even more complex. Indeed, several factors must be considered, including the type of prior neoadjuvant therapy, the timing of recurrence, and the choice of the first-line therapy. Above all, BRAF mutation status remains a critical determinant in treatment selection. Therefore, improving existing therapies, innovating new therapeutic drugs, and investigating new combination regimens is greatly needed6,7,8,9.

There is increasing evidence that, as in other cancers, aggressiveness in melanoma is associated to ECM stiffness. Indeed, retrospective analyses have high-lighted that mechanical stress may increase the risk of acral melanoma10,11. Furthermore, the presence of a stiff gradient in the ECM surrounding cancer cells facilitates the process of negative durotaxis, which induces melanoma to switch from non-invasive Radial Growth Phase (RGP) to Vertical Growth Phase (VGP) and allow cancer cells to invade and disseminate12. However, the intricate interplay between melanoma aggressiveness and ECM stiffness remains a captivating and contentious subject, since certain studies have underscored that, rather than ECM rigidity, it is the plastic adaptation of cancer to a dynamic microenvironment that facilitates cancer progression13.

ECM stiffening and fibrotic tumor microenvironment (TME) are closely related to response to therapy in different ways. The transition from the RGP to the VGP, accompained by the loss of E-cadherin and the increase in N-cadherin expression, leads to the strengthening of the cross-talk between melanoma cells and fibroblasts. A stiffed ECM promotes mechanotransduction through the nuclear accumulation of Yes-associated protein 1 (YAP) which correlates with increased proliferation and resistance to therapy14. In addition, MAPK-targeted therapies, in BRAF mutated melanoma cells, promotes the phenotypic reprogramming of tumor cells towards a myofibroblast-like phenotype which confers drug resistance15,16. Moreover, a fibrotic TME, characterized by high interstitial pressure and hypoperfusion, leads to an hypoxic and acidic tumor microenvironment that promotes tumor progression and metastastatic diffusion17,18,19,20. Finally, a fibrotic TME may impair an efficient diffusion of chemotherapeutic agents within the tumor mass. Thus the melanoma cells/fibroblasts crosstalk in a fibrotic TME may represent a novel target for the development of effective treatment approaches in melanoma21.

In this scenario we explored the effect of pirfenidone (5-methyl-1-phenyl-2- (1H)-pyridone, PFD), an anti-fibrotic agent, on two different BRAF mutated melanoma cell lines to shed light on the opportunity to repurpose the use of this drug in melanoma. PFD belongs to the class of small molecules and thus is potentially capable of interacting with multiple targets. Indeed, it has long been recognized for its anti-inflammatory, antioxidant, and anti-fibrotic properties. The multi-target capability of PFD represents the strength of this small molecule, especially in the context of a rapidly progressing, debilitating, and life-threatening disease such as idiopathic pulmonary fibrosis (IPF), which, even under treatment, can worsen quickly with acute exacerbations. Indeed, in 2008, PFD became the first drug approved for the treatment of IPF in Japan, in 2011 in Europe and in 2014 in the U.S. as ESBRIET®22,23.

PFD showed inhibitory effect on myofibroblast differentiation, collagen production and Transforming Growth Factor β1 (TGFβ1) signaling, contributing to the slowing of disease progression of IPF.

The main mechanism of PFD action has been identified as the down-regulation of TGFβ1 signaling. However, PFD exerts its beneficial effect on IPF through other collateral mechanisms such as suppressing cytokine release and reactive oxygen species (ROS) production. The role of PFD in cytokine release, particullary tumor necrosis factor-α (TNF-α) and interleukin-6 (IL-6), has been extensively investigated in preclinical model of intersitial lung diseases24,25. Moreover, the key role of PFD on oxidative stress inhibition has been investigated directly in preclinical model of pulmunary fibrosis26,27,28.

In view of its antifibrotic effect, PFD in combination with gemcitabine, suppresses in vivo desmoplasia by inhibiting stellate cells which are responsible for the extensive fibrosis associated with aggressive pancreatic cancer29. Furthermore, in a preclinical model of orthotopic breast cancer, PFD induces remodeling of the tumor ECM and enhances therapeutic treatment by downregulating cancer-associated desmoplasia30. Interestingly, because of the role in TGFβ in epithelial-to-mesenchymal transition (EMT), it has been demonstrated that PFD inhibits EMT in lung cancer through the downregulation of TGFβ signaling31,32. Recently, PFD demonstrated inhibitory effects on the TGFβ signaling pathway in colorectal cancer (CRC) models, both in vitro and in vivo, especially when combined with 5-fluorouracil( 5-FU) 33. In addition to that PFD has been evaluated in association with vitamin D to reduce fibrosis in a preclinical model of drug induced cancer fibrosis34. Additional investigations have shown that PFD may act in synergism with current therapeutic treatment of cancer cells. Indeed, it has been demonstrated that PFD sensitizes non-small cell lung cancer (NSCLC) cells to paclitaxel by inhibiting cell proliferation highlight that PFD may act targeting other pathways in addition to the TGFβ signaling35. Finally, PFD showed encouraging beneficial effects in vivo in a preclinical model of mesothelioma through the inhibition of ERK and Protein Kinase B (AKT) pathways36. In this context, particular importance is given to the clinical observation by Miura and collaborators, who reported a reduced risk of lung cancer development in patients with IPF receiving PFD treatment compared to untreated patients37. The common background of the aforementioned studies is the inhibitory ability of PFD on cancer-associated fibroblast, which in turn impacts on cancer as a consequence of the TME cross-talk.

Our aim was to investigate the direct antitumoral effect of PFD on melanoma cells to decipher the mechanisms involved in its beneficial effect. To this end we explored the intriguing network of PFD targets exploiting different experimental approaches and focusing our attention to the less explored ability of PFD to inhibit furin convertase enzyme activity.

Furin is a proprotein convertase enzyme localized mainly in the Golgi apparatus of mammalian cells. This enzyme plays a fundamental role in the maturation of a wide range of intracellular pro-proteins by the cleavage at a specific consensus amino acid motif (Arg-X-(Lys/Arg)-Arg). Furin mediates essential steps in the maturation of albumin, complement component C3, von Willebrand factor, matrix-metalloproteases (MMPs), growth factors and growth factor receptors from their respective precursors, and thus furin is involved in the fine-tuning of homeostasis, and embryogenesis38. Moreover, furin exerts also a critical role in the pathogenicity of viruses and bacteria. Indeed, very recently, it has been demostrated that furin cleaves SARS-CoV-2 spike protein enambling viral entry into lung cells39.

TGFβ is one of the most important mediator processed by furin. It has been demonstrated that the acquisition of a more aggressive phenotype by cancer cells may be partially due to the overexpression of furin convertase by cancer cells40. In addition, furin has been shown to play a critical role in melanoma progression41. Notably, furin activates mutated oncogenes (such as BRAF), enhancing their oncogenic activity42. Therefore, furin convertase became a candidate for the development of innovative anticancer therapy43.

Existing and effective inhibitors of furin activity contain an imidazole or thiazolidine ring, such as rosiglitazone and thiazolidinediones. Currently, significant efforts are emplyed in the development of novel small-molecule inhibitors with greater specificity, lower molecular weight and improved cellular permeability44,45.

Recent findings suggest that PFD, along with other small-molecule inhibitors, may directly interact with furin active site via its pyridone ring. Indeed, Burghardt and collaborators reported that PFD inhibits furin at concentrations ranging from 5 to 8 mM, which are comparable to those used in preclinical in vitro studies.46,47.

Here we present for the first time the effect of PFD in melanoma cells. Firstly, exploiting in silico studies, we found that PFD efficiently interact with the active site of furin. Next we reported an antiproliferative effect of PFD on BRAF mutated melanoma cells. Interestingly, the inhibition on cell growth and colony formation was similar to that of vemurafenib, although no synergistic effect was found. The antioxidant activity of PFD on melanoma cells positively correlates with the inhibition of proliferation and with the increase in p21 expression. Moreover, we explored the impact of PFD on melanoma cell aggressiveness under the light of its furin-inhibitor properties. Indeed, we found that PFD inhibits MMP-dependent invasiveness, and TGFβ1 release. Globally, we suggest that the wide spectrum of antitumoral effects of PFD on melanoma cells depends on the inhibition of furin activity.

Methods and materials

Chemicals

Chemicals were purchased from MedChemExpress, New Jersey, USA. PFD (#HY-B0673/CS-2905) was diluted in DMSO at a 1.5 M concentration and used for melanoma cell treatment at 2.7 mM and 8.1 mM concentrations, corresponding to a range of 0.2–0.5% of DMSO that did not significantly affect cell viability. Vemurafenib (#HY-12057) was diluted in DMSO at a 50 mM concentration and used for melanoma cell treatment at 1 µM, corresponding to 0.002% of DMSO. BOS318 (#HY-147140) was diluted in DMSO at a 50 mM concentration and used at 5 nM concentration (less than 0.001% of DMSO). In each experiments melanoma cells exposed to DMSO 0.2% were used as reference control (not-trated cells, NT).

PFD docking studies

We evaluated the ability of PFD to intecact with furin catalytic domain (PDB: 7QXZ) using in Molecular Docking Analysis. We identified and validated the coordinates of the active site of Fur based on the binding site location of the following co-crystallographic ligands previously reported as highly potent.

50 run modes of localized molecular docking were carried out for PFD with a box size of 15 Å × 15 Å × 15 Å and a center on ASP154 with γ-Carbon coordinates X = 45.134, Y = − 31.172, Z = − 15.157 using Autodocktools 1.5.6, and AutoDock4.

The analysis yielded 50 results, grouped into 4 clusters. Cluster 1 contains 44 calculated poses, which suggests that this is the most favorable conformation in the binding site. The analysis of the best pose revealed that the carbonyl group (C=O) is positioned to form hydrogen bonds with the side chains of HIS194 and TRP254. The central nitrogen heteroatom in the PFD molecule forms hydrogen bonds with the carboxyl group of ASP154. The hydrophobic portion, consisting of the central benzene ring and the lateral benzene ring, is nestled within hydrophobic regions of furin.

Cell cultures

The human melanoma cell line M21 was kindly provided by Dr. Antony Montgomery (The Scripps Research Institute, La Jolla, CA), and A375M6 were obtained from ATCC), both cell lines carry the BRAFV600E mutation48. Melanoma cells were cultured in DMEM 4.5 g/l glucose (#ECM0728, Euroclone, Milan, Italy), supplemented with 2 mM L-glutamine (#ECB3000D, Euroclone) and with 10% Fetal Bovine Serum (#ECS0180L; Euroclone). Cells in exponential growth were washed with 1X phosphate-buffered saline (containing 8.0 g/L NaCl, 0.2 g/L KCl, 1.42 g/L Na2HPO4, and 0.24 g/L KH2PO4, at pH 7.4, next PBS), harvested using a trypsin-ethylenediaminetetraacetic acid (EDTA) solution (#ECB3052, EuroClone), and propagated every three days. Melanoma cell cultures at a 60–70% confluence were used for the experiments. Cells exposed to vehicle (DMSO) were used as control and designated as not-treated (NT). Alternatively, cells were exposed to PFD to final concentrations of 2.7 mM (equal to 0.5 mg/ml) and 8.1 mM (equal to 1.5 mg/ml). The treatments were performed in complete standard medium (10% FCS DMEM) for 24 h or 48 h in accordance with in vitro protocols used in previous studies46,49.

MTT assay

Melanoma cell viability and proliferation were assessed using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide reduction assay (MTT) (M5655 Sigma Aldrich, Milan, Italy), as described in50. Briefly, melanoma cells (2 × 103) were plated into 96-multiwell plates in 100 µL of complete medium. After 24 h adhesion, cells were treated with vehicle (not-treated), 2.7 mM, 8.1 mM of PFD or with vemurafenib (1 µM). Vemurafenib treatment concentration was choosed accordingly to previous research investingation conducted in our laboratory51. After 48 h exposure, the medium was replaced with 100 µL MTT (0.5 mg/ml) reagent in each well, and plates were incubated at 37 °C. After 2 h, MTT was removed, and the blue MTT–formazan product was solubilized with dimethyl sulfoxide (DMSO, Sigma Aldrich). The absorbance of the formazan solution was read at 595 nm using the microplate reader (Bio-Rad, Milan, Italy). Inhibition of cell proliferation (IC50) was calculated by GraphPad Prism Software (San Diego, CA, USA).

Colony formation assay

5 × 102 cells were seeded in six-well plate in 2 ml complete medium in standard conditions. After 24 h adhesion, cells were treated once with PFD alone (2.7 mM and 8.1 mM concentrations) or vemurafenib (1 µM). Cultures were left untreated until the end of the incubation. After 14 days of cell growth, developed colonies were counted upon 20 min-fixation in 4% paraformaldehyde at 4 °C and 30 min-staining with crystal violet solution at room temperature. The clonogenic efficacy was evaluated by measuring the absorbance at 595 nm of the clone populations obtained in each treatment using the microplate reader (Bio-Rad, Milan, Italy).

Mitosox

The Mitosox green assay is a fluorescent method used to detect reactive oxygen species (ROS) by cytofluorimetric analysis, BD FACSCanto (BD Biosciences, Franklin Lakes, New Jersey, USA). 4.0 × 105 cells were plated in 2 mL of complete DMEM per well in a 6-well plate. After 24 h adhesion, treatments were added and cells were incubated for 6 h. At the end of the incubation, cell cultures were washed with PBS, gently detached using a trypsin–EDTA solution, and resuspended in complete medium. After centrifugation (1500 rpm for 5 min), the cells were washed in 500 uL of complete phosphate-buffered saline (CPBS, containing 0.7 mM CaCl2, 0.5 mM MgCl2, 0.1% glucose and PBS; pH 7.5, pre-warmed to 37 °C), centrifuged again, and resuspended in 500 μL of a Mitosox Green solution (Thermo Fisher, Waltham, MA, USA; M36005, work solution 0.5 µM in CPBS). Cell suspensions were incubated at 37 °C for 40 min in the dark with gentle agitation (300 rpm). Subsequently, 500 μL of CPBS were added in each sample, cells were centrifuged and resuspended in 500 μL of fresh CPBS for FACS analysis. Data were analyzed using FlowJo™ Software (version X.X, BD Biosciences). Briefly, the homogeneity of the cell populations was determined using Side Scatter (SSC) and Forward Scatter (FSC), and populations of interest were selected for each treatment. Next, within the different gated populations, the percentage of FITC-positive expressing cells were compared between samples.

Western blotting analysis

1 -2 × 106 melanoma cells were seeded in p100 plastic dishes and allowed to grow in standard conditions to reach 50–70% of cell confluence. Next, melanoma cell cultures were exposed for 24 h to different concentration of PFD or to vehicle. At the end of the incubation media were rapidly removed and melanoma cell monolayers were washed with ice cold PBS containing 1 mM Na4VO3, and lysed in about 100 μl of cell RIPA lysis buffer (Cat. No. 20–188; Merk Millipore, Vimodrone, MI, Italy), containing phenylmethylsulfonyl fluoride (PMSF) (Cat. No. 10837091001, Sigma-Aldrich), sodium orthovanadate (Cat. No. 567540 Sigma-Aldrich) and protease inhibitor mini tablets (Cat. No. A32955, Life Technologies, Monza, Italy) as previously described50. In some experiments melanoma cells were exposed for 24 h to BOS318 furine convertase inhibitor (5 nM) and/or PFD. Aliquots of supernatants containing equal amounts of protein (50–100 μg) in Bolt LDS Sample Buffer (Cat. No. B0007, Life Technologies) were separated on Bolt® Bis–Tris Plus gels 4–12% precast polyacrylamide gels (Cat. No. NW04122BOX; Life Technologies). Fractionated proteins were transferred from the gel to a polyvinylidene fluoride (PVDF) nitrocellulose membrane using a Bio-Rad system. Blots were blocked for 5 min, at room temperature, with EveryBlot Blocking Buffer (Cat. No. 12010020, Bio-Rad) and the membrane was probed at 4 °C overnight with primary antibodies diluted in a solution of 1:1 Immobilon® Block -Fluorescent Blocker (Cat. No. WBAVDFL01, Merk Millipore) /T-PBS buffer. The primary antibodies were: rabbit anti-p21 (Cat. No. 1201, 1:1000, Cell signaling Technology, Danvers, MA, USA), rabbit anti-TGFβ1 (Cat. No. orb11468, 1:500, Biorbyt, Durham, North Carolina, USA), rabbit anti-pNRF2 (Biorbyt orb1724698, 1:500), and mouse anti-vinculin (Cat. No. sc-73614, 1:1000, Santa Cruz, Biotechnology, Inc, USA). Membranes were washed in T-PBS buffer, and incubated for 1 h at room temperature with goat anti-rabbit IgG Alexa Flour 750 antibody (Cat. No. A21039; Invitrogen, Monza, Italy) or with goat anti-mouse IgG Alexa Fluor 680 antibody (Cat. No. A21057; Invitrogen, Monza, Italy), and then visualized by an Odyssey Infrared Imaging System (LI-COR® Bioscience). Mouse anti-vinculin monoclonal antibody (Cat. No, Sigma-Aldrich) was used to assess equal amount of protein loaded in each lane.

Cytofluorimetric evaluation of cell death

Melanoma cells were plated (4.0 × 105 cells/well) in 6-well plates, and allowed to grow in standard conditions to reach 60–70% of cell confluence. Next, melanoma cell cultures were exposed for 24 h to vehicle, to different concentration of PFD and/or BOS318 furine convertase inhibitor (5 nM). After 24 h incubation, cell death was determined by flow cytometer analysis using allophycocyanin (APC)-conjugated or Fluorescein Isothiocyanate (FITC)-conjugated recombinant Annexin V (from Immunotools GmbH #31490016 and # 31490013, Germany) and Propidium Iodide PI (from Sigma-Aldrich) according to the manufacturer’s protocol. Briefly, cells were harvested with Accutase (Eurolone, Pero, Italy), collected in flow cytometer tubes (1 × 105 cells/tube), washed in PBS, and incubated 15 min at 4 °C in the dark with 100 µl Annexin Binding buffer (100 mM HEPES, 140 mM NaCl, 25 mM CaCl2, pH 7.4), 1 µl of 100 µg/ml PI working solution, and 2.5–5 µl APC or FITC-conjugated Annexin V. Each sample was added with Annexin Binding Buffer to reach 500 µl volume/tube. Samples were then analyzed at BD FACSCanto (BD Biosciences, Franklin Lakes, New Jersey, USA). Cellular distribution depending on Annexin V and/or PI positivity allowed the measure of the percentage of viable cells (Annexin V and PI negative cells), early apoptosis (Annexin V positive and PI negative cells), late apoptosis (Annexin V and PI positive cells), and necrosis (Annexin V negative and PI positive cells).

Invasion assay

Melanoma cell invasiveness was determined in vitro on Geltrex™ -precoated polycarbonate inserts (Sarstedt 83.3932.800, 8 µm pore size, 0.3 cm2 growth surface). Briefly, 50 µL of 15 μg/mL of Geltrex™ solution were seeded on each insert and allowed to dry in sterile conditions. 5 × 104 cells were seeded in the upper compartment of the insert in their 24 h growth conditioned medium, and treated with PFD during migration at 2.7 mM and 8.1 mM concentrations. Cells were free to invade for 5 h at 37 °C in 10% CO2 in air. After incubation, the non-invading cells on the upper surface of the insert were wiped off mechanically with a cotton swab, while migrated cells on the lower side of the inserts were fixed in 4% paraformaldehyde for 20 min, and then stained with 20% of crystal violet (C6158, Sigma-Aldrich). Photographs of randomly chosen fields were taken, and migrated cells were counted in 4 field/photograph.

Statistical analysis

Statistical analysis between two groups was performed using unpaired Student’s t-tests. In the case of comparative analysis of three or more groups, one-way and two-way analysis of variance (ANOVA) was performed followed by Tukey’s post hoc test. Statistical analysis was presented only for significant values; where no significance was found, no symbols were depicted. Band intensities in Western blot analysis were quantified using computer-based ImageJ software. Each band was quantified and the corresponding histogram was constructed by normalizing the density of each band to that of vinculin and compared to not-treated cells (NT). The Western blot investigation was analyzed using Student’s t-test. Statistical analysis was performed using GraphPad Prism Software (San Diego, CA, USA). Data are expressed as mean ± S.E.M. were obtained from at least three independent experiments (n = 3) and P values ***p < 0.0001 and *p < 0.05 were considered significant.

Results

PFD interaction with furin convertase enzyme

We explored the effect of PFD on melanoma cells evaluating the role of PFD on the inhibition of furin, a subtilisin-like proprotein convertase enzyme, indeed it has been recently described that PFD interact directely with furin24.

Using Molecular Docking Analysis we evaluated the ability of PFD to intecact with furin catalytic domain (PDB: 7QXZ). The analysis yielded 50 results, grouped into 4 clusters. Cluster 1 contains 44 calculated poses, which suggests that this is the most favorable conformation in the binding site. The analysis of the best pose revealed that the carbonyl group (C=O) is positioned to form hydrogen bonds with the side chains of HIS194 and TRP254. The central nitrogen heteroatom in the PFD molecule forms hydrogen bonds with the carboxyl group of ASP154. The hydrophobic portion, consisting of the central benzene ring and the lateral benzene ring, is nestled within hydrophobic regions of furin (Fig. 1).

Fig. 1
figure 1

A Interaction between PFD and furin (PDB: 7QXZ) is illustrated. B Key amino acid residues represented by blue carbon sticks and the PFD carbons depicted in black. Hydrogen bonds are shown as dotted blue lines.

Full size image

PFD inhibits the viability and clonogenic ability of BRAF mutated melanoma cells

Human melanoma cells A375M6, and M21 were exposed to PFD at two different concentrations (2.7 mM and 8.1 mM) or to vemurafenib (at 1 µM concentration) or to their combination. PFD treatments were chosen accordingly with the corresponding concentration used for treatment of IPF patients. The MTT colorimetric assay was used for measuring cellular metabolic activity as indicator for cell viability, and proliferation, rather than for cytotoxicity. We found that in A375M6 cells, 2.7 mM treatment with PFD significantly reduced cell proliferation compared to untreated cells (P ≤ 0.05). Interestingly, 8.1 mM treatment with PFD significantly reduced cell proliferation (P < 0.0001) compared to untreated cells. Moreover, the 8.1 mM PFD treatment was more efficient compared to vemurafenib alone (P ≤ 0.05). However, after the exposure of A375M6 to the combination of PFD and vemurafenib, we did not observed a significant cooperative or synergistic effect (Fig. 1A, left panel). Slightly different results were found in M21 cells. In this cell line, the proliferation of melanoma cells exposed to PFD 2.7 mM was similar to that of untreated cells. M21 cell proliferation was reduced compared to untreated cells (***P < 0.001) after the exposure to 8.1 mM PFD. Interestingly, no statistically significant difference in cell proliferation was found comparing M21 cells treated with 8.1 mM PFD to M21 cells treated with vemurafenib alone. Similarly to the findings in A375M6 cells, also in M21 cells no significant cooperative or synergistic effect was observed following the treatment with drug combination (Fig. 2A, right panel). Interestingly, the exposure of Human Umbilical Vein Endothelial Cells (HUVEC) to PFD reveal a moderate and not significant decrease of cell proliferation (Fig. 1S).

Fig. 2
figure 2

Effect of PFD on melanoma cells viability and clonogenic ability. A375M6, and M21 human melanoma cells were concurrently exposed to 2.7 mM, 8.1 mM of PFD, to vemurafenib (1 µM) or to their combination. (A) MTT assays were performed at 48 h and results are represented as a mean ± SD from at least three independent experiments. (B) Clonogenic ability of A375M6, and M21 human melanoma cells exposed to 2.7 mM, 8.1 mM of PFD, to vemurafenib (1 µM) or to their combination. Statistical analysis was performed using Graph Pad Prism 4 software and one way ANOVA test. Statistical significance was assumed for p values ≤ 0.05: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 compared to not-treated cells.

Full size image

To further evaluate the inhibitory effect of PFD on cell proliferation, we evaluated the melanoma cells clonogenic ability. This assessment contributes to determine whether cancer cell populations retain the ability to survive in constrained environmental conditions, indicating the presence and manteinance of a cancer stem cell population.

Our investigations confirmed that PFD, at both concentrations, decreased the number of colonies in A375M6 (Fig. 2B, left panel) and M21 cells (right panel). The reduction of colonies obtained after the exposure of A375M6 melanoma cells to PFD, at both concentrations, was similar to that obtained after the exposure of melanoma cells to vemurafenib. The reduction of colonies obtained after the exposure of M21 melanoma cells to PFD was dose dependent, and only the higher dose of PFD revealed an inhibition similar to that obtained after the exposure of M21 melanoma cells to vemurafenib. Similarly to the findings using MTT assay, no significant cooperative or synergistic effect was observed following the treatment with drug combination. A moderate, though not statistically significant, increase in clonogenic ability was observed in cells treated with the combination of PFD 8.1 mM and vemurafenib, compared to each treatment alone (Fig. 2B).

Effect of PFD on ROS and cell cycle/apoptosis modulation

Next, we investigated which was the mechanism involved in PFD inhibition of cell proliferation. PFD is recognized as an antioxidant agent. It is well-known that in cancer cells sustained levels of ROS is a key feature of cancer cell survival and aggressivenes. Indeed, in cancer cells intracellular ROS are physiological byproducts of many different cellular processes52. Thus, we explored the effect of PFD on the release of mitochondrial ROS. We found that the 6 h exposure of M21 melanoma cells to PFD significantly reduced intracellular ROS. The amount of ROS reduction was similar with no significant difference between the two treatments (Fig. 3A,B). This reduction was associated with a dose–response reduction of Nrf2 phosphorylation in M21 cells (Fig. 3C,D) but not in A375M6 cells (not shown). We found that, in both cell lines, PFD increased p21 expression in a dose dependent manner (Fig. 3E,F). Interestingly, we have investigated ROS production following prolonged exposure to PFD and we found significant increase in ROS level after 24 h treatment, as detected using DCFH-DA assay (Fig. 2S). Next, we investigated whether furin inhibition, has a role in melanoma cell proliferation. We exposed melanoma cells to PFD in association with furin inhibitor BOS318. As expected, PFD increased p21 expression in a dose dependent manner. The treatment with furin inhibitor alone, did not increase p21 expression. Additionally, we found a synergism between the two agents, when PFD was used at the higher concentration (Fig. 3G,H). Next we evaluated the annexinV/PI distribution profile of the two BRAF mutated melanoma cells exposed to PFD, BOS318 or to their combination. Interestingly, we found that PFD increased the percentage of cell population entered in late apoptosis in a dose dependent manner, in both cell lines. In addition, the cotreatment with PFD 8.1 mM and BOS318 synergistically cooperate to the increased percentage of melanoma cells in late apoptosis (Fig. 3I).

Fig. 3
figure 3

Effect of PFD on ROS production in M21 cells exposed to different PFD treatments. (A) Representative histograms of ROS production using mitosox and corresponding relative fluorescence intensity associated to untreated M21 cell population (blue histogram), to M21 cells exposed to 2.7 mM (orange histogram) or to 8.1 mM (green histogram) concentration of PFD and compared to unstained cell (red histograms) (B) Mean fluorescent intensity of M21 melanoma cells. (C) Western blotting analysis of pNrf2 in melanoma cells exposed to 2.7 mM and 8.1 mM concentration of PFD. (D) Densitometric analysis is presented as the mean ± SD of the three independent experiments. (E) Western blotting of p21 in A375M6 (upper panel) and M21 cells (lower panel) exposed to PFD at 2.7 mM and 8.1 mM concentrations. (F) Densitometric analysis of western blotting presented as the mean ± SD of three independent experiments. (G) Western blotting analysis of p21 in M21 melanoma cells exposed to 2.7 mM and 8.1 mM concentration of PFD and/or furin inhibitor BOS318 (5 nM). (H) Densitometric analysis is presented as the mean ± SD of the three independent experiments. (I) Representative distribution of melanoma cells exposed to PFD and/or furin inhibitor across Q1/Q2/Q3/Q4 panels in annexinV/PI assay. Percentages are representative of one experiments out of three that gave similar results. Statistical analysis was performed using Graph Pad Prism 4 software and one way ANOVA test. Statistical significance was assumed for p values ≤ 0.05: *P < 0.05, **P < 0.01, and ***P < 0.001.

Full size image

Effect of PFD on furin convertase enzyme activities in melanoma cells

Finally, we proceeded to deeply explore the mechanistic insights involving furin convertase in melanoma cell aggressivenes. Furin is responsible for the transformation of precursor proteins into their mature forms. MT-MMP is one of the target of convertase activity of furin enzyme. In the Golgi apparatus furin removes the prodomain sequence of the MT1-MMP and then functionally active MT-MMP expressed on the cell surphace promote pro-gelatinase A (MMP-2) activation. Thus, we evaluated the biological activity of PFD on the inhibition of melanoma cell invasiveness through a reconstitued ECM. Melanoma cells were allowed to migrate toward a synthetic ECM in the presence or in the absence of PFD for 5 h avoiding any bias incoming from longer exposure which could lead to the reduction on cell proliferation. We found that PFD, in both cell lines, significantly reduced cancer cell invasiveness in a dose dependent manner (Fig. 4A).

Fig. 4
figure 4

Effect of PFD on furin dependent melanoma cell invasiveness and TGFβ production. (A) Inbibition of melanoma cell invasiveness in A375M6, and M21 human melanoma cells. Results are represented as a mean ± SD of tumor cells migrated for 5 h through GeltrexTM coated-filters. The results are presented as the mean ± SD of three independent experiments. (B) Western blotting of intracellular TGFβ (44 kDa) in A375M6, and M21 human melanoma exposed to PFD at 2.7 mM and 8.1 mM concentrations, in the presence or in the absence of BOS318. Each panel of the figure shows a typical experiment out of three that gave similar results. (C) Densitometric analysis is presented as the mean ± SD of the three independent experiments. Statistical analysis was performed using Graph Pad Prism 4 software and one way ANOVA test. Statistical significance was assumed for p values ≤ 0.05: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Full size image

Interestingly furin is also the key enzyme in the TGFβ1 intracellular maturation process53. In addition, the binding of TGFβ1 to its cognate receptor increases the expression of furin in a mothers against decapentaplegic homolog 2 (SMAD2/3-) and SMAD4-dependent manner54.

We next evaluated the intracellular levels of 44 kDa TGFβ1 pro-protein. Melanoma cells were treated with PFD for 24 h with or without furin inhibitor BOS318 at 5 nM concentration and we found that PFD reduced the expression of 44 kDa TGFβ1 pro-protein in A375M6 and M21 cells in a dose dependent manner (Fig. 4B,C) and we also found a reduction in cytoplasmic phospho-SMAD2/3 (data not shown). The furin inhibitor, as expected, reduced both TGFβ1 and phospho-SMAD2 expression.

Graphical representation of the antitumoral effect PFD-furin networks of interaction

figure a

Discussion

Over the past several decades, a significant increase in the global incidence of melanoma has been observed. This increase is both due to exogenous and endogenous factors that, alongside with an increase in life expectancy, differentially affect the molecular mechanisms involved in the onset and progression of the disease55. From the diagnostic and therapeutic point of view, omics strategies significantly contribute to the understanding of the molecular mechanisms underlying disease progression and became essential to the planning of more effective therapeutic strategies56.

The discovery of the single nucleotide mutation on BRAF proto-oncogene, which is present in almost the 50% of all melanomas, became a fundamental diagnostic biomarker in the development of personalized target therapy of melanoma57. Thus, in 2011 for the U.S and in 2012 for the European Union, vemurafenib, the first BRAFV600E inhibitor, has been approved for the treatment of BRAF-mutated metastatic melanoma58. However, clinical findings rapidly revealed that patients treated with BRAFV600E inhibitors, after an initial phase of tumor regression, undergo tumor relapse leading to the fatal outcome59. Another fundamental milestone in the treatment of metastatic melanoma is the introduction of ICI in the early 2010s60,61. Unfortunately, after initial enthusiastic results for the improvement of patient overall survival from 6 months to 6 years, ICI patients frequently experience adverse events and often resort to treatment discontinuation62. Moreover, there is still not a common consensus for ICI treatment, and specific immunotherapeutic biomarkers have failed to predict treatment response63,64,65,66.

In this work, we investigated whether it is worthwhile to apply the drug-repurposing strategy to treat melanoma using PFD. PFD has been approved firstly for non-cancer diseases such as IPF23. PFD is a well tolerated drug and, in IPF patients, the dose could be gradually increased up to 801–1200 mg/day37. This clinical dosage explains the use of concentrations in a millimolar range used in the present study and in others67. It is important to highlight that PFD treatment reduces Acute Exacerbation (AE) events in patients with IPF bearing lung cancer. This clinical evidence indicates that PFD reaches an effective concentration in the tumor fibrotic microenvironment and suggests that it may also be applicable in other fibrotic microenvironment, as those found in melanoma lesions68,69.

Recently, preclinical evicences and clinical trials involving patients with IPF and NSCLC, suggest that the use of PFD in cancer treatment could be beneficial35,36,70,71.

In the present study, we strayed from the common approach of other preclinical investigations in which the beneficial effect of PFD has been evaluated on mesenchymal cells focusing on the antifibrotic effect of PFD. Instead we assessed the effect of PFD directely on melanoma cells, aiming to gain insight on its mechanism of action.

Tumor cell proliferation and aggressiveness are controlled by different mechanisms. The emergence of resistance to targeted therapy in melanoma patients has revealed that these mechanisms can shift unpredictably, enabling melanoma cells, and in general cancer cells, to adapt and survive in challenging microenvironmental conditions, ultimately contributing to variable and inconsistent therapeutic responses. Tumor cells can exploit those different and often reciprocal independent mechanisms of survival to escape from microenvironmental stressing conditions and from the anticancer treatments. The complexity of signaling network of this intrinsic and extrinisc resistance justifies the use of different drugs delivered in association, with a meticolous administration timing61,62. We forsee that the small molecule of PFD may be particularly well-suited to counteracting the complex signaling network driving cancer cell aggressiveness, and this efficacy, observed in preclinical model of melanoma cells, appears to stem from PFD capacity to engage multiple targets, both well-characterized and yet to be fully elucidated.

Firstly, we evaluated PFD/furin interaction. In silico docking analysis revealed that PFD clearly occupies the catalytic domain of furin via its pyridone ring. This interference affects multiple downstream mechanisms regulated by furin. We next observed a significant inhibition of BRAF mutated melanoma cell proliferation, whether the proliferation of normal cell was not affected by the same treatments. We believe that, to fully understand these results, it is important to consider not only the mechanism of action of PFD but also its metabolic inactivation pathway. PFD is rapidly metabolized by liver enzyme systems such as cytochrome P45072. Moreover, it has been observed that, in rare cases, patients may develop idiosyncratic adverse reactions. It has been demonstrated that physiological inactivation by cytochrome enzymes leads to the formation of intermediates that remain biologically active. However, in some rare cases, these electrophilic intermediates species interact with cellular proteins, resulting in cell death and organ failure73.

We suggest that the inhibition of melanoma cell proliferation may depend on the ability of PFD to act as an electrophilic compound with potential toxicities. Moreover, the polymorphism of cytochrome enzymes, often observed in cancers74 could explain the different sensitivity between BRAF mutated melanoma cell lines and normal cells to PFD effects. This observation should be taken into consideration for a personalized approach of the treatment75.

We also observed that, PFD inhibition of melanoma cell proliferation and clonogenic ability did not exhibit a synergism when PFD was used in co-treatment with vemurafenib. We hypothesize that PFD inhibits furin activity and, consequently, affects the function of the BRAF oncogene. The lack of synergy between PFD and vemurafenib may depend on their competition for the same molecular target. Moreover, we observed a different trend of inhibition between the MTT and the Clonogenic assay. The MTT assay evaluates the metabolic activity of proliferating cells after an acute treatment with PFD. The colony formation assay, which involves approximately two weeks of observation, following the initial acute PFD treatment, evaluates the proliferative capacity of stem-like cells. The assay reflects the intrinsic capacity of the cell population to maintain a stem cell-like subset. This may explain the discrepancies observed between the two assays.

One of the most recognized mechanism of PFD action is the inhibition of reactive oxygen species (ROS) production. Cancer cells produce elevated levels of ROS compared to normal cells, mostly through NADPH oxidase mitochondrial pathway. Intrinsic anti-oxidant capacity of cancer cells guarantee the scavenging of excessive ROS and the mainteinance of the pro-tumorigenic role of ROS75,76,77,78. However, any perturbation of ROS balance in cancer cells reveals a lethal effect. Indeed, radiatherapy treatment exploits the induction of lethal amounts of ROS in the TME79,80. On the other hand, the subtraction of intrinsic anti-oxidant capacity drives cancer cells to death as well. We found that the reduction of ROS induced by PFD, associated with the reduction on phosphorylated nuclear factor erythroid 2-related factor 2 (pNRF2) expression, lead to a cell cycle arrest which was similar to that obtained with the use of vemurafenib.

Interestingly, prolonged exposure of melanoma cells to PFD did not result in a reduction of ROS, rather, it led to an overall increase in total ROS levels. This observation suggests that the elevated ROS production may be associated with the onset of cellular damage, as supported by the elevated percentage of late apoptotic cells.

Another well established mechanism of action of PFD is its inhibitory activity on TGFβ1 pathway, but little is known on the exact mechanism of action. TGFβ1 is secreted as an inactive precursor in the extracellular compartment bound to the latency-associated peptide (LAP). This complex (TGFβ1-LAP) is called small latent complex (SLC). SLC, at the extracellular level is covalently connected with specific binding proteins (latent TGFβ-binding proteins, LTBPs) and stored in the extracellular matrix. We hypothesize that the key point in the PFD-dependent inhibition of proliferation depends on the ability of PFD to inhibit furine converatse enxyme activity. This inhibition serves as point of convergence for the reduced amount of TGFβ available to keep the cells cycling, and available for promoting a downstream signalling for autocrine TGFβ release 46. At extracellular level, the beneficial effect of PFD relates with furin-dependent membrane-type matrix metallo-proteinase activation (MT-MMPs). MT-MMPs mediate the release of TGFβ1 from the extracellular storages81, and the activation of other extracellular proteinases involved in cancer cell invasiveness such as MMP2 and MMP982. Thus the inhibition of extracellular action of furin linked to PFD associates with a reduction of extracellular TGFβ1 release and signaling, and a reduction of invasiveness83.

The antiproliferative activity of PFD is associated with the enhancement of p21 expression, as a consequence of ROS release and cell damage response, but also with the inhibition of TGFβ1 pathway. Indeed, furin inhibitor treatment was not sufficient to induce p21 increase. Interestingly, we observed a synergistic effect of PFD in association with furin inhibitor probably dependent on the sustained reduction of TGFβ1 activation by these two agents.

Our observations suggest that PFD may be considered a potential treatment option for metastatic melanoma cases that are resistant to conventional therapeutic interventions through a drug repurposing approach84. In addition to the established use of PFD in IPF, PFD has demonstrated beneficial effects in other fibrotic conditions, including cardiac fibrosis85. However, recent reports have highlighted adverse effects, such as gastrointestinal intolerance and skin reactions86, indicating that off-target effects and idiosyncratic responses should not be underestimated. A key limitation of the present study is the absence of clinical data in melanoma patients, emphasizing the need for further preclinical and clinical investigation.