Although early primary melanoma is curable through surgery, treatment of advanced disease remains a challenge and strategies employed in the past 30 years have not significantly improved cure rates, which are less than 3%. All major chemotherapy drugs, immunotherapies and radiotherapies have failed in large-scale clinical trials. Unlike other cancers, drug resistance in melanoma is not acquired selectively following drug therapy, suggesting that chemo-resistance in melanoma is intrinsic. After years of research we now understand a great deal about the underlying biology of melanoma. Through the work of many groups including our own many of the critical signaling pathways responsible for melanoma growth, survival and chemo-resistance have now been identified. Our laboratory has for the last 25 years extensively studied the biology of melanoma development and progression. In recent years we have placed major emphasis on the development of orthotopic models for melanoma growth, invasion and metastasis using a novel human/mouse chimeric model. This model also provided valuable information for transformation of melanocytes within the tissue context of normal human skin. The in vitro/in vivo skin reconstruction model demonstrated how melanoma cells escape from the control of epidermal keratinocytes to develop intimate interactions with stromal fibroblasts and endothelial cells in the tumor microenvironment. The accumulated knowledge in melanoma biology can now be leveraged for exploring the best pathways to target for therapy. Targeted therapy is a new field in melanoma. In the past, therapeutic approaches in melanoma were almost exclusively directed towards immunotherapy or biological response modifier therapies. Those approaches have provided important information, but only now is the field beginning to appreciate that melanoma cells can be killed through a rational approach which inactivates the most critical signaling pathways. This project intends to provide a solid understanding of our options for targeted therapy of melanomas and we are focusing on two major pathways, the MAPK and PI3K pathways, which are important for cell growth and survival.
BRAF in melanoma. We previously demonstrated that virtually all melanomas have high constitutive activity in the MAPK pathway. The reasons for this increased signaling activity are manyfold and include activating mutations in BRAF in over 65% of melanomas (8), activating N-Ras mutations in 15% of melanomas (9-11), autocrine stimulation of growth factor receptors, such as c-met and FGFR1 (12), activation through ?v?3 integrin binding (7), and Notch1 signaling (13). Constitutive activity in the Ras/Raf/MEK/ERK MAPK pathway contributes to the oncogenic phenotype of melanoma through its effects on cell proliferation, invasion and survival (14). In experimental systems, the role of BRAF in melanoma looks convincing. In vitro studies have shown that V600E BRAF is an oncogene in immortalized mouse melanocytes (15) and that selective downregulation of V600E BRAF using RNAi causes cell death and reversal of the melanoma phenotype (16). Increased BRAF activity also suppresses the activity of the melanocyte-specific transcription factor micropthalmia (MITF), diverting the melanoma cells from a differentiation pathway into a highly proliferative state (17). Activation of the MAPK pathway through the growth factors SCF/c-kit ligand and bFGF, in combination with activation of the G-protein coupled receptor EDNRB, drives melanocytes in human skin grafted to SCID mice to uncontrolled proliferation resembling malignant transformation (3). Thus, the BRAF/MEK pathway is an ideal candidate for targeted melanoma therapy (18).
PI3 kinase pathway. Activity in the PI3K/Akt pathway arises either through paracrine growth factor loops, for example through IGF-1 stimulation (19), PI3K gene mutations or loss of PTEN expression. Melanomas generally lack PI3K (20) and Akt (21, 22) mutations, but PTEN is lost in 30% of cell lines (23-25) and approximately 10% of clinical melanoma specimens (26, 27). Transduction of melanocytes with the p85 gene of PI3K leads, in combination with other oncogenes, to malignant transformation (28). Akt is the central player in the PI3K/Akt-mediated oncogenic behavior of melanoma and it phosphorylates a number of important downstream targets in the regulation of apoptosis. Of the three Akt family members, Akt1-3, (12), about 50% of melanoma cells have constitutive activity in Akt3 (29). Inhibition of Akt activity in melanoma, using either PI3K inhibitors or selective RNAi to Akt3, both reduce growth and induce some degree of apoptosis (29, 30).
3D spheroid model development and responses to BRAF/MEK and PI3K inhibition There is evidence that the tumor microenvironment modulates the response to therapy. We drew upon our long experience in developing organotypic models to investigate whether modulating the microenvironment and growing the melanoma cells in 3D culture would impact their response to MEK/PI3K inhibitors. For assays, melanoma cells were first grown under non-adherent conditions on top of agar until they formed 3D spheroids. After 48 to 72 hours, the spheroids were harvested, mixed with collagen, plated out, overlayed with media and left to grow. Cells emerged from the spheroids to invade into the surrounding collagen in a stage-specific manner. Spheroids grown from cells derived from earlier stage lesions (RGP and VGP; WM35 and WM793) were less invasive into the collagen than those from metastases (1205Lu) (Figure 4A), and all remained viable (Figure 4B). Spheroids were treated with either U0126 or LY294002 and left to grow for 72 hours. Under these conditions it was noted that the most aggressive melanoma cells were completely resistant to U0126, no longer underwent apoptosis and continued to invade into the collagen (Figure 5). The apparent resistance to the MEK inhibitors of metastatic melanoma cells is indeed supported by early clinical studies on MAPK inhibitors which demonstrate little benefit to melanoma patients.
Similar experiments were performed with the 3D spheroid model using the PI3K inhibitor LY294002. Again it was noted that only the melanoma cell lines from the early-stage RGP and VGP lesions had reduced cell viability (data not shown). As with U0126, the cell lines derived from metastatic lines were resistant to the effects of LY294002 in 3D culture. The same results were also seen following treatment of melanoma cells with inhibitors of Src (SU6656) and Rho kinase (ROCK). Implantation of the spheroids into collagen was critical to the drug resistance, as spheroids left on-top of the agar were not drug resistant.
Smalley, K.S., Haass, N.K., Brafford, P.A., Lioni, M., Flaherty, K.T., Herlyn, M.: Multiple signaling pathways must be targeted to overcome drug resistance in cell lines derived from melanoma metastases. Molec Cancer Therapy 5: 1136-1144, 2006. PMID 16731745
Smalley, K.S., and Herlyn, M.: Towards the targeted therapy of melanoma. Mini Rev. Med. Chem. 6: 387-393, 2006. PMID16613575
Herlyn, M.: Molecular targets in melanoma: strategies and challenges for diagnosis and therapy. Int J Cancer 118:523-526, 2006. PMID16258898
Smalley, K.S., Herlyn, M.: Targeting intracellular signaling pathways as a novel strategy in melanoma therapeutics. Ann N Y Acad Sci 1059:16-25, 2005. PMID16382039
Inhibition of melanoma cells in spheroids implanted in collagen using a MEK inhibitor. Melanoma cells were grown under non-adherent conditions for 72 hours until spheroids had formed. Spheroids were then harvested and implanted into a collagen gels before being treated with UO126 (1 and 10?M). After 72 hours cells were treated with a cell viability kit, where living cells stain green and dead cells red. In all cases, data shown are representative of three experiments. Magnification (WM35, WM793) X 10. Magnification (1205Lu, C8161) X4. U0126 inhibited only the growth of early-stage melanoma cell lines (WM35 and WM793) but not of metastatic melanoma cells (1205Lu and C8161). Note: Three of the lines (WM35, WM793, 1205Lu) were V600E BRAF mutated and one, (C8161) was wild-type for both N-Ras and BRAF. |