Portal de Periódicos da CAPES

MITF: the power and the glory

2011; Wiley; Volume: 24; Issue: 2 Linguagem: Inglês

10.1111/j.1755-148x.2010.00809.x

1755-148X

Keith S. Hoek,

Retinal Development and Disorders

For regular readers of Pigment Cell and Melanoma Research, it goes without saying that the activities of the microphthalmia-associated transcription factor (MITF) are supremely important for pigmented cells both healthy and malignant. With commanding roles in controlling pigment cell development and function as well as in driving melanoma cell proliferation, MITF’s ‘master-regulator’ status is a clearly deserved moniker. For those wishing to catch up on the history of its discovery, how mouse mutants of MITF have revealed the complex subtlety of its function, how it is the ultimate translator of ultraviolet radiation into pigment and how it is thought to drive the progression of metastatic melanoma, a good place to begin is the December 2010 issue of this journal where a quartet of high priests (Arnheiter, Steingrimsson, Fisher, and Goding) have delivered the gospel on all things MITF – past, present and future. In this News and Views, two recent research papers exploring MITF’s function in melanoma are discussed, one detailing its role in protecting the integrity of melanoma cell DNA during mitosis and another styling MITF as the regulator of ‘stemness’ in melanoma cells. Both derive (to lesser or greater degrees) from an earlier publication, and they each present important results which further broaden our appreciation of MITF’s already manifold functions. In May of 2010, Giuliano et al. published a melanoma study in which reduction in MITF expression was shown to induce senescence and precipitate severe chromosome segregation defects (Giuliano et al., 2010). Those authors noted however that among the known target genes, none were factors involved in protecting against such damage. Not long afterwards, Irwin Davidson’s group, working with the 501Mel melanoma line, published a ground-breaking genome-wide study in which 1700 targets of the Brn2 transcription factor were identified using techniques of chromatin immunoprecipitation (ChIP) and array hybridization (Kobi et al., 2010). To identify additional targets of MITF, and perhaps along the way explain its role in preserving mitotic fidelity, the collaboration between these two teams was a natural strategy. Importantly, for target sequence identification, the authors substituted array hybridization with recently developed techniques of massively parallel sequencing. In this way, Strub et al. were able to precisely identify where along the 501Mel genome that MITF binds and thus annotate 5578 potential target genes. Mindful that mere occupancy of a DNA sequence does not necessarily entail a functional relationship, the authors went on to silence MITF expression and assess gene expression changes using RNA-seq. Integrating this with the ChIP-seq data revealed that MITF directly drives the transcription of 240 genes and suppresses that of a further 225. For additional validation, Strub et al. carefully reviewed existing literature and, gratifyingly, found significant agreement with most previous works. In doing so, they instil a firmly established confidence in a greatly expanded list of MITF target genes. The authors duly noted target genes that are likely to explain MITF’s involvement in preserving DNA replication and fidelity and employed functional assays to confirm their roles in facilitating DNA replication and mitosis. However, the data pool of new MITF target genes is so rich that Strub et al. are able to provide logical comment on a wide range of additional functions attributed to MITF, from suppression of senescence to phenotype switching. In fact, their results bring with them an increase in coherence, the sense that such distal processes are truly linked by their subservience to MITF regulation. Overall, their demonstration of exactly where MITF regulates gene expression across the genome is a laudable achievement. Perhaps the only thing missing is a comparison of the 12139 loci bound by MITF in 501Mel cells against Bill Pavan’s genome-wide list of over two million consensus sequence MITF-binding sites (Loftus et al., 2008). Such an analysis would provide an opportunity to revisit the rules for MITF–DNA interaction. The earlier-mentioned work by Giuliano et al. likely also inspired the second paper discussed here. In the Giuliano study, it was shown how MITF silencing pushed melanoma cells into G0–G1 growth arrest. This was explored further by Robert Ballotti’s team in a work communicated as an examination of the relationship between MITF expression and the stem cell-like features of melanoma cells. Their opening experiments showed that growing melanoma cells in stem cell medium give them (when injected into the flanks of nude mice) an in vivo growth advantage over cells grown in normal medium. Counterintuitively, the number of cells experiencing cell cycle arrest was shown to be greater in stem cell medium than in normal medium. In parallel, the authors demonstrated that MITF levels (typically associated with melanoma cell proliferation) are sharply reduced in stem cell medium. When they instead used siRNA to transiently abrogate MITF expression, they again saw an increase in both in vitro cell cycle arrest and in vivo tumourigenicity. As uncanny as those results were, the authors went on to show that an ostensibly MITF-positive melanoma population proliferating under normal culture conditions will harbour a small sub-population of cells which are MITF-negative and cell cycle-arrested. Further, upon subcutaneous injection, this non-proliferating sub-population transforms into an explosively proliferative (and MITF expressing) mass. By contrast, cultures depleted of this non-proliferating sub-population only infrequently formed tumours. So what is happening here? First consider we have learned that in a native culture of (MITF-positive) melanoma cells, there is a sub-population that is MITF-negative and cell cycle-arrested. How is such a sub-population derived and then maintained? That manipulation of MITF expression is enough to change the proportion of this sub-population suggests it is in the first place the result of an equilibrium dominated by extrinsic signalling. In other words, cells comprising the sub-population need not always be the same cells and that membership to the sub-population is a transient state to which any number of cells is subject. This is the most likely way such a combination of fast and slow cycling cells can be maintained in culture. Indeed, as the siRNA experiments show, it must be the transient nature of this state which is critical for generating what the authors term ‘melanoma-initiating cells’ that drive rapid tumourigenesis. The importance of this transience is highlighted by comparing the authors’ results against a very different outcome obtained from melanoma cell cultures for which MITF expression is generally low. In such a case, tumourigenesis is significantly retarded by several weeks (Hoek et al., 2008), which is a strong contrast to the accelerated in vivo expansion of cells for which MITF expression is briefly abrogated. The sense is that MITF-expressing melanoma cells must pass through the eye of a needle, in which they temporarily shut down MITF expression and enter G0–G1 arrest, before undergoing a virulent burst of proliferation. It is a scenario that prompts many questions concerning what in melanoma cells is an evidently dynamic relationship between the cell cycle and MITF expression. That Cheli et al. refer to a shifting sub-population of transiently MITF-negative melanoma cells in terms reminiscent of cancer stem cells is a reflection of how stretched the concept has become. Following hard on the heels of our abandoning the notion that melanoma stem cells represent a diminishingly small population, Ballotti’s group now furnishes us with additional evidence repudiating the idea that they are any sort of fixed population at all. Considering the data emerging from Meenhard Herlyn’s laboratory, wherein expression of a melanoma tumourigenesis marker (JARID1B) is shown to flit puckishly from cell to cell, the parallels are unmistakable (Roesch et al., 2010). We should seriously consider ascribing the hierarchies we are evidently so fond of to phenotypic states rather than to particular subsets of cells. To map out in fine detail MITF’s basic function as a transcription factor across the entire genome, Davidson’s team have successfully brought to bear high-throughput technologies of unprecedented power. Meanwhile, the discovery by Ballotti’s group that MITF regulation (and melanoma cell proliferation) is far more dynamic than previously supposed reminds us that we still have a long way to go before we may understand MITF in all its glory.