The T-box transcription factor, TBX3, is sufficient to promote melanoma formation and invasion

Malignant melanoma is a highly aggressive and intractable disease which has a rising incidence surpassing all other cancers[1]. It is therefore imperative to identify new targets for the development of improved therapies to treat melanoma. To this end, an understanding of the molecular mechanisms underpinning the stepwise progression of melanocytes to metastatic melanoma is important. Indeed, in the last decade, several transcription factors have been identified as potential therapeutic targets including MITF, PAX3, SNAIL and the developmentally important T-box transcription factor TBX3[2‐4].

TBX3 is overexpressed in a number of cancers including a subset of melanomas but its precise role in the progression of this disease still needs to be clarified. Earlier studies pointed to TBX3 contributing to the oncogenic process by functioning as an anti-senescence factor by directly repressing the negative cell cycle regulator p19/p14ARF[5, 6]. More recent studies however suggest that TBX3 may function in advanced stages of melanoma. Hoek and colleagues demonstrated that TBX3 was upregulated in approximately 55% of cell lines obtained from advanced melanoma lesions[7]. Furthermore, we have shown that knocking down TBX3 in vertical growth phase (VGP) and metastatic melanoma cells inhibits cell migration and tumour formation[8] and TBX3 was demonstrated to repress the E-cadherin gene which is characteristically downregulated in VGP melanoma[9]. Furthermore, Boyd and co-workers demonstrated that oncogenic B-RAF is able to upregulate TBX3 expression which in turn represses E-cadherin levels[10].

To determine if TBX3 is not only essential but sufficient to promote melanoma formation, we stably transfected the non-tumourigenic human WM1650 radial growth phase (RGP) melanoma cell line, which expresses undetectable levels of TBX3, with the pEGFP-N3 vector in which the cytomegalovirus promoter drives the expression of a green fluorescent protein tagged human TBX3.

Western blot and qRT-PCR analyses confirmed that TBX3 was successfully overexpressed in the WM1650-TBX3 cell line established (Figure 1a, b).

We previously reported that VGP and metastatic melanoma cells in which TBX3 was depleted maintained a higher proliferative ability than their control cells[8]. It was therefore hypothesised that the overexpression of TBX3 in RGP cells would lead to a decrease in their proliferative ability.

Indeed growth curves and pulsing the cells with 5-bromo-2-deoxyuridine (BrdU), a synthetic thymidine analogue confirmed this to be the case (Figure 1c, d). To further explore the reason(s) for the decrease in cell proliferation, the cell cycle profile of WM1650-TBX3 and WM1650-ctrl cells were compared by flow cytometry. The results showed that there were 10% more WM1650-TBX3 cells in the G1 phase, but 10.5% fewer in the S-phase (Figure 1e). Furthermore, while the WM1650-TBX3 cells had reduced levels of p21, p16 and TBX2, they had increased levels of the tumour suppressor proteins, p53 and p14 (Figure 1f). This is significant because TBX2, a homologue of TBX3, was shown to be a powerful pro-proliferative factor in melanoma and breast cancer cell lines which express both TBX2 and TBX3[8, 11] and there has been an indication that TBX2 and TBX3 can repress each other[9]. Together these results show that TBX3 impacts negatively on melanoma cell proliferation.

To further explore the possibility that TBX3 may drive RGP cells to a more aggressive melanoma phenotype the anchorage independent growth and colony forming ability of the WM1650-ctrl and WM1650-TBX3 cells were analysed using soft agar and methyl cellulose cell viability methylthiazol tetrazolium (MTT) assays. The results show that TBX3 promoted anchorage independent growth since the WM1650-TBX3 cells formed more and larger colonies (Figure 2a, b). Similarly, the two-dimensional scratch (Figure 2c) and transwell motility (Figure 2d) assays showed that the WM1650-TBX3 cells migrated significantly faster than the WM1650-ctrl cells. To confirm that the mechanism by which TBX3 affects melanoma cell adhesion and migration is through repressing E-cadherin, we investigated the levels of this cell adhesion molecule. Our results show that, compared to WM1650-ctrl cells, WM1650-TBX3 cells had lower levels of E-cadherin mRNA and protein (Figure 2e). To confirm the above findings in vivo, the tumour forming ability of WM1650-TBX3 and WM1650-ctrl cells was compared in nude mice. While 100% of mice injected with WM1650-TBX3 cells formed tumours ranging from 8 to 17.84 mm in diameter, 0% of mice injected with WM1650-ctrl cells formed tumours (Figure 2f). The ability of TBX3 to drive non-tumourigenic cells to form tumours is shown to be highly sensitive and specific (100% in both cases). H & E staining of tumour sections confirmed evidence of tumour cell infiltration into the skeletal muscle of the hind limb of mice injected with WM1650-TBX3 cells (Figure 2g black arrows).