AurkA inhibitors enhance the effects of B-RAF and MEK inhibitors in melanoma treatment

AurkA inhibitor (MLN8054) was purchased from Selleck Chemicals (Munich, Germany), while GSK2118436 (B-RAF inhibitor) and GSK1120212 (MEK inhibitor) were kindly provided by Glaxo SmithKline (London, UK). The B-RAF and MEK inhibitors were tested at a concentration of 30 nM in all experiments while the AurkA inhibitor was tested at a concentration of 1 μM concentration.

A human melanoma cell line with the B-RAF (V600E) mutation (A375mel) was kindly provided by Dr. M. Bettinotti (NIH, Bethesda, USA). The cells were cultured in RPMI 1640 medium (Lonza, Milan, Italy), supplemented with 3 mM L-glutamine (Invitrogen-Gibco®/Life Technologies, Monza, Italy), 2% penicillin/streptomycin and 10% fetal bovine serum (FBS). All cultures were incubated at 37°C in a humidified 5% CO₂ atmosphere.

Cell proliferation was assessed using E-16-well plates and the xCELLigence technology (Acea Bioscience, San Diego, CA, USA, distributed by Roche Diagnostics) [35]. Briefly, cells (1 × 10³ A375 cells/well) were seeded in E-16-well plates in complete medium and grown for 48 hours. Inhibitor drugs were then added as single agents or in different combinations (B-RAF inhibitor [GSK2118436] plus MEK inhibitor [GSK1120212], BRAF inhibitor plus AurkA inhibitor [MLN8054], MEK inhibitor plus AurkA inhibitor [MLN8054] or triple-combination of all three drugs) and the cell growth was monitored for an additional 72 hours. Microelectrodes, placed on the bottom of plates, were used to detect impedance changes proportional to the number of adherent cells and expressed as the cell index. The impedance value of each well was automatically monitored by the xCELLigence system and expressed as a cell index value. Doubling times for each cell line were calculated from the cell growth curve during the exponential phase. The experiments were conducted in triplicate and repeated twice.

Western blot was performed according to standard procedures. Mouse monoclonal antibody against p53 (DO-1; diluted 1:1000) (Santa Cruz Biotechnology, Inc. Dallas, TX, USA), rabbit antibody monoclonal to c-Myc (1:5000; Abcam, Cambridge, UK), and rabbit polyclonal antibody against β-actin (1:1000, Cell Signaling Technology, Danvers, MA, USA) were used. Detection was achieved by HRP-conjugated anti-mouse (Biorad, Hercules, CA, USA, 1:2,000) or HRP-conjugated anti-rabbit (Cell Signaling Technology, 1:1000) antibodies. Immune complexes were visualized by an enhanced chemiluminescence system (ECL Advance™, Amersham Pharmacia Biotech, Piscataway, NJ, USA). Actin was used as a loading control. The image analysis was performed by ImageJ software (http://​rsbweb.​nih.​gov/​ij/​). Results represent the means (±SEM) of three independent experiments performed in triplicate. P-values were determined by using t-tests and a value ≤0.005 is reported (in figures with the symbol ***).

A 3D culture system of a differentiated and full-thickness skin reconstruction model of A375 melanoma cells was purchased from MatTek (Ashland, MA, USA). The 3D tissues were fed through the basolateral (bottom) surface, and incubated in duplicate with serum-free medium containing DMSO 0.2% as control, B-RAF inhibitor plus MEK inhibitor, MEK inhibitor plus AurkA inhibitor or a triple-combination of all three drugs.

The medium was replenished every other day, and the 3D tissues were collected on days 0, 5, 9 and 12 from the beginning of treatment and fixed with 4% paraformaldehyde. Tissues were paraffin-embedded, serial-sectioned, deparaffinized in xylene and rehydrated through graded decreasing concentrations of alcohol. Finally, sections were stained by hematoxylin and eosin (H&E) for cell morphology analysis. For immunohistochemistry, antigen retrieval was carried out in citrate buffer (pH6) by heating in a pressure cooker. The sections were then stained with antibodies specific for the detection of S-100 (rabbit polyclonal; Dako, Milan, Italy) and Ki-67 (Mib-1clone; Dako). Appropriate positive and negative controls were included for each antibody test. Serial sections, stained with H&E, S100 and Ki-67 were observed under a 10x/0.30 NA and 40x/0.75 NA objective lens, using a DM6000B upright microscope (Leica Microsystems, Milan, Italy). Images were captured using a high-resolution Leica DFC480 camera and digitally transferred by the Leica Application Suite 4.0 software. AdobePhotoshop CS5 software was used for image analysis. Ki-67 and S-100 positively stained cells were counted in six different fields on each slide (observed under a 40x/0.75 NA objective lens using the Leica DM6000B microscope) and their quantification was performed on digital images with the ImageJ program and AdobePhotoshop CS5. Results were expressed as percentages with respect to the controls.

A375 (BRAFV600E) cell proliferation rates were assessed upon exposure to the different inhibitor drugs alone and in combination. As shown in Figure 1A, A375 cell proliferation rates were, as expected, reduced with B-RAF inhibitor alone and the combined B-RAF inhibitor plus MEK inhibitor. AurkA inhibitor in combination with B-RAF inhibitor and as a triple combination with B-RAF plus MEK inhibitors also reduced cell proliferation rate, although it was not very effective on cell proliferation rate, as single agent (Figure 1A). This suggests combined B-RAF/AurkA inhibition might be an alternative to B-RAF/MEK inhibition and that triple B-RAF/MEK/AurkA inhibitor therapy could be considered as a therapeutic option. Interestingly, the MEK plus AurkA inhibitor combination demonstrated higher anti-tumor efficacy than the B-RAF plus MEK inhibitor combination. This finding seems to suggest that a MEK/AurkA combination could represent an alternative treatment strategy for melanoma patients without B-RAF mutations.

To confirm that B-RAF and MEK inhibitors were inhibiting their respective proteins, we analyzed the expression of c-Myc, a downstream protein of B-RAF/MEK activated by the MAPK pathway [36]. A reduction in c-Myc protein levels in A375 melanoma cells was observed after 72 hours of exposure to B-RAF inhibitor plus MEK inhibitor, MEK inhibitor plus AurkA inhibitor and the triple drug combination (Figure 1B).

The tumor protein p53 has been reported to be phosphorylated by AurkA, leading to its increased degradation and downregulation of checkpoint-response pathways [37]. Thus, to confirm that the AurkA inhibitor (MLN8054) was inhibiting AurkA protein, we analyzed p53 protein level in the A375 cell line. After 72 hours of drug exposure, p53 protein level increased (Figure 1C). Levels of p53 were lower in cell lines exposed to double and triple drug combinations compared with single agent AurkA inhibitor, suggesting that the down-regulation of c-Myc, following B-RAF/MEK inhibition, affected p53 protein levels.

At baseline (day 0), H&E staining of cultures revealed keratinocytes in the upper epidermal layer, organized in the basal, spinous, granular, and corneum stratum; a second distinct layer of cells was represented by A375 melanoma cells (Figure 2A). At this early time point, this layer was only a few cells thick, with these cells distinguished by their dark nuclear staining. A third distinct layer is represented by dermal stratum consisting of fibroblast-contracted collagen. Analysis of S-100 protein expression, a marker of the melanocytic cell lineage, confirmed the presence of melanoma cells in the tissues (Figure 2A).

In the control tissue, metastatic melanocytes proliferated over time and developed nodules at the dermal/epidermal junction by day 5 (data not shown), which were further enlarged at days 9 (Figure 2B) and 12 (Figure 2C), as confirmed by their S-100 and Ki67 immunostaining. In contrast, in tissues treated with the triple drug combination there were fewer tumor nodules close to the epidermis with less invasion of the dermal structures at day 9 (Figure 2B). Exposure to B-RAF inhibitor plus MEK inhibitor or MEK inhibitor plus AurkA inhibitor provided results intermediate between control and the three drugs combined, as showed in Figure 2B; however, less tumor nodules were observed in tissues treated with the MEK plus AurkA inhibitors compared with B-RAF plus MEK inhibitors.

At day 12, the effect of the triple drug combination on melanoma growth was more pronounced (Figure 2C). Clusters of melanoma cells were detected on the other side of the dermal stratum in non-treated tissue, as shown in detail in Figure 2D, suggesting vertical growth tumor development.

Interestingly, in all treated and non-treated tissues, individual proliferating melanoma cells were observed infiltrating the dermis, as suggested by their Ki-67 (data not shown) and S-100 antigen expression (Figure 2E). These cells showed a different morphology to those observed at the dermal/epidermal junction. The small nest cells were polygonal in shape and remained confined at the junction site (Figure 2D, section a). In comparison, the individual melanoma cells in the dermal stratum appeared as spindle-shaped cells connected to one another and forming a 3D-cellular network (Figure 2E). These data suggest that melanoma dermal cells might be more able to escape the drug anti-tumor effect than small nest melanoma cells localized at the epidermal/dermal junction. The observation that the drugs penetrated to the other side of the tissues from the dermal stratum and killed the small nest cells suggests that the ability of individual melanoma dermal cells to escape the drug effect was dependent on their intrinsic properties and not on the dermal stratum being a barrier to drug exposure.

It has been previously reported that the epithelial-to-mesenchymal transition (EMT) mechanism plays an important role in promoting chemoresistance, invasion, and stem cell-like properties [38],[39]. The observed different shape of the cells suggested a potential EMT in these tissues, and it may explain the more resistant phenotype of these cells to the drugs used. This provides support to the idea that specific targeting of EMT could potentially serve to decrease metastasis and overcome drug resistance. However, further studies are needed to explain the potential role of these cells.

Observations at both the epidermal/dermal junction and the dermal stratum were also quantified with the densitometric results shown in Figure 3A and B, respectively. To determine whether the decrease in S100-positive cells might account for the inhibition of melanocyte proliferation, tissue sections were stained with the proliferation marker Ki-67. This clearly revealed a decrease in the number of proliferative cells at the epidermal/dermal junction site after triple drug exposure compared with control. Intermediate results were obtained with both double drug combinations. Quantitative analysis of the Ki-67 immunohistochemical results are presented in Figure 3A. No significant differences were found in Ki-67 immunostained cells after triple or either double drug treatments at the dermal stratum compared with control (Figure 3B).