A (heat) shocking development: FBXW7 loss unleashes HSF1 to drive melanoma invasion and metastasis.

Abstract

Melanoma is the deadliest form of skin cancer because of its strong tendency for local invasion, distant metastasis, and resistance to traditional cytotoxic chemotherapy. A more complete understanding of the genetic landscape of melanoma will define and clarify the molecular pathways that drive tumor growth and will also reveal new therapeutic targets. Indeed, this is already being borne out: multiple studies have demonstrated that activating mutations in components of the RAS/RAF/MEK/ERK pathway occur in over half of melanoma tumor samples and targeted inhibitors of this pathway are currently used to treat patients with these cancers. Other common genetic events include mutations leading to the activation of the PI3K/AKT signaling pathway; deletion of the CDKN2A locus, which encodes the tumor suppressor proteins p16INK4a and p14ARF and likely impacts cell cycle control; and mutations in the tumor suppressor p53, which affects genomic stability and other processes (Kunz, 2014). Large-scale tumor genome projects have found that many putative cancer genes are mutated in relatively small numbers of tumors, and it can be challenging to determine whether these mutations represent ‘drivers’ or ‘passengers.’ However, it is increasingly clear that studying low frequency mutations often yields important insights into cancer biology and therapy. One recent example was published in the March issue of Nature Cell Biology (17, 2015:322-32). In this work, Kourtis and colleagues describe a critical interaction between two known cancer proteins, F-box and WD repeat domain containing 7 (FBXW7) and heat-shock factor 1 (HSF1). While the genes encoding these proteins are altered in only a small minority of melanoma cases, the authors show that this genetically defined patient subset has an increased incidence of tumor metastasis and decreased survival. Their work suggests therapeutic opportunities, as well. The FBXW7 protein functions as the substrate recognition subunit of an ubiquitin ligase complex that leads to degradation of a wide range of cancer promoting proteins, including cyclin E, c-Myc, Notch, KLF5, and TGIF. Thus, FBXW7 is an important negative regulator of a number of hallmarks of cancer, including cell growth, metabolism, differentiation, and invasion (Davis et al., 2014). It is frequently mutated or deleted in a wide range of human cancers, and low levels of FBXW7 have been associated with poorer cancer outcomes and with metastatic tumor growth. Collectively, these features make it a classical tumor suppressor gene. In human melanomas, FBXW7 mutations have been identified in up to 8% of patients, and epigenetic mechanisms may lower FBXW7 expression as well. Previous work has shown that FBXW7 knockdown can promote melanoma growth in model systems. This effect is mediated, at least in part, through stabilization of NOTCH1, which leads to upregulation of NOTCH1 target genes and promotion of tumor angiogenesis. Importantly, NOTCH1 pathway inhibitors can slow the growth of FBWX7-deficient tumor xenografts (Aydin et al., 2014). Kourtis et al. report their findings on a new role for FBXW7 in degrading HSF1, which is a transcription factor that is activated in response to diverse forms of cellular stress. In normal cells, the primary transcriptional targets of HSF1 are heat-shock proteins and related molecular chaperones that serve to restore protein homeostasis. Cancer cells are prone to many kinds of stress, including metabolic, oxidative, and proteotoxic stress. Recent work has demonstrated that the normal cellular functions of HSF1 can be commandeered by cancer cells to cope with these multiple stressors. A wide range of cancers show amplification of the HSF1 genetic locus as well as increased expression and nuclear localization of the HSF1 protein. These features are often associated with a poor prognosis. Importantly, the transcriptional program induced by HSF1 activity in cancer cells is fundamentally different from that seen in normal cells, with broad ranging effects on pathways that promote proliferation, survival, and invasion. In melanoma in particular, HSF1 was found to be a key mediator of tumor invasion and metastasis (Mendillo et al., 2012; Scott et al., 2011). HSF1 binds a much larger number of gene promoters in the FBXW7-deleted cells HSF1 overexpression is associated with increased invasion in melanoma models, and Kourtis et al. hypothesized that loss of FBXW7 would lead to similar phenotypes. As predicted, knockdown of FBXW7 in the melanoma cell line 451Lu led to significantly increased invasion in vitro and increased metastasis in vivo. Coordinate knockdown of HSF1 suppressed this increased invasion, supporting the idea that HSF1 is a critical FBXW7 substrate involved in this process. Interestingly, FBXW7 depletion did not affect primary tumor growth, which is in contrast to previously published reports using a different cell system (Aydin et al., 2014). These experimental results appear to be relevant to the human disease: the authors quantified HSF1 and FBXW7 by immunohistochemistry in human melanoma specimens and found an inverse relationship between HSF1 and FBXW7 protein expression. As predicted, metastatic samples had high levels of HSF1 and low levels of FBXW7, and high HSF1 protein expression was associated with poorer prognosis. Genetic analyses supported these results: HSF1 was found directly bound to the promoters of a number of genes that drive invasion and metastasis, and HSF1 transcriptional programs were highly enriched in metastatic disease. Other FBXW7 substrates have been implicated in melanoma growth and spread, including MYC and NOTCH1, and a criticism of the current paper is that these pathways were not adequately examined as alternative or cooperating drivers of phenotypes associated with FBXW7 knockdown. However, the data presented by Kourtis et al. clearly show that FBXW7 and HSF1 function can greatly impact melanoma biology and prognosis. How might this new information contribute to clinical practice? First, if broader studies confirm the prognostic value of HSF1 expression, this could be used to risk stratify patients in the future. Second, these genetic changes may impact response to targeted therapies. Indeed, Kourtis et al. show that a MEK inhibitor (U0125) markedly decreased the affinity of FBXW7 for HSF1 in HEK293T cells. They also reported that treatment with either the BRAF inhibitor vemurafenib or the MEK inhibitor trametinib can lead to nuclear HSF1 stabilization. Currently, the combination of BRAF and MEK inhibitors forms front-line therapy for patients with BRAF-mutant metastatic melanoma. Unfortunately, despite high initial response rates, all patients will eventually become resistant to these treatments. Perhaps the decreased affinity of FBXW7 for HSF1, and the resultant upregulation of HSF1 and its downstream targets, acts as a mechanism of tumor escape. To test this idea, it would be interesting to assess HSF1 levels in human tumor samples before and after treatment with BRAF and/or MEK inhibitors. The work by Kourtis et al. adds to the already abundant preclinical data showing that the modulation of FBXW7 or HSF1 can have clear effects on tumor invasion and spread, and agents that directly target these pathways are eagerly awaited. However, development of such drugs has proven challenging. In the case of FBXW7, effective treatments will need to restore the function of the mutated or absent protein. While small molecules or gene therapy approaches could be useful in this regard, it is generally much easier to counteract overexpressed proteins than to reactivate mutated or dormant ones. With this in mind, HSF1 inhibition may be a more viable anti-cancer treatment. Indeed, inhibitors of classical downstream targets of HSF1, including the heat-shock proteins HSP27 and HSP90, have shown encouraging results in early stage clinical trials. Because cancer-associated HSF1 overexpression has broad effects on many cellular processes in addition to the heat-shock response, there is significant concern that cross talk will lead to rapid resistance to these drugs. Therefore, the further development of direct inhibitors of HSF1 is paramount. Finally, it should be noted that the toxicities of these novel approaches are largely unknown and may be substantial.