Axl of Evil

Summary

Receptor tyrosine kinases (RTKs) play key roles in initiation and progression of human skin cancers, yet the importance of some members of the RTK family remains poorly understood. Findings from Sensi et al., published in this issue of Journal of Investigative Dermatology, provide exciting new insights into the expression and function of the RTK, Axl, in melanoma (Sensi et al., 2011). The authors show that Axl is frequently expressed in melanoma cell lines, particularly mutant NRAS harboring lines, and that Axl activation by its ligand, Gas6, likely occurs via both autocrine and paracrine mechanisms. Gene signatures from Axl-expressing cell lines are similar to published signatures from poorly differentiated tumors displaying high metastatic potential. Functionally, Axl was required for the invasive and migratory properties of an Axl-expressing melanoma cell line. These data emphasize that targeting the Gas6-Axl signaling axis should be investigated as a strategy to inhibit pro-metastatic properties in poorly differentiated melanomas.

Introduction

Axl was originally identified as a transforming gene in chronic myelogenous leukemia patients (Liu et al., 1988). Its name was derived from the Greek word “anexelekto” meaning uncontrolled (O’Bryan et al., 1991) and it was characterized as a receptor tyrosine kinase (RTK) in the TAM (Tyro-3, Axl, Mer) subfamily. Aberrant expression and activation of RTKs is implicated in melanoma progression (Tworkoski et al., 2011), as well as melanoma resistance to targeted therapies (Villanueva et al., 2010). Axl has been found to be upregulated in many cancer types; however, the potential importance of this RTK in melanoma has not been clearly elucidated. The manuscript from Sensi et al., published in this issue, seeks to fill this gap.

Sensi et al. detected high Axl expression in 22 of 58 of cell lines (38%) in a melanoma progression panel. The frequency of Axl expression increased with disease progression, with 30% of primary tumor lines and 40% of metastases displaying expression. Other TAM subfamily RTKs, Tyro3 and Mer, were expressed in a low percentage of melanomas and these lines lacked Axl expression. NRAS mutations are found in 15–20% of melanomas and are mutually exclusive from the approximately 50% of melanomas that harbor BRAF mutations. A high proportion of melanoma lines that highly expressed Axl were mutant NRAS genotype with an underrepresentation of mutant BRAF and wild-type BRAF/NRAS genotypes.

A role for Axl in melanoma migration and invasion

Gene signatures have been utilized to stratify melanomas into distinct subsets. An interesting aspect of the Sensi et al. study is their analysis of published gene sets to identify signatures that positively or negatively correlate with Axl expression. Of the genes positively correlated with Axl, there was high representation of genes associated with extracellular matrix interactions and remodeling, indicating a possible role for Axl in invasion. Furthermore, inhibiting Axl function in a high Axl-expressing melanoma cell line with short interfering RNAs or a pharmacological inhibitor led to reduced invasion and migration. Comparable observations have been reported recently by an independent group (Tworkoski et al., 2011). However, whereas Sensi and colleagues show that Axl depletion did not dramatically affect proliferation, the Tworkoski et al. study indicates that knockdown of Axl inhibited proliferation in some melanoma lines. Nevertheless, the consensus is that Axl mediates pro-invasive and metastatic behavior in melanoma cells.

Axl expression was also associated with poorly differentiated tumors. Additional biochemical and immunohistochemical analyses of tumors revealed a negative correlation between Axl and expression of microphthalmia-associated transcription factor (MITF). These data are consistent with a model of melanoma plasticity in which melanomas switch between proliferative and invasive states based on cues from the tumor microenvironment (Hoek et al., 2008). These distinct states can be distinguished by analysis of gene expression profiles. MITF, a transcription factor well known to regulate melanocyte differentiation and survival (Widlund and Fisher, 2003), represents a marker for proliferative, non-invasive cells, and Axl is one of the genes associated with the invasive phenotype (Hoek et al., 2008). While most of the tumors studied by Sensi et al. show high Axl and very low MITF staining, or vice versa, two tumors did show expression of both antigens. Thus, it would be interesting to perform staining in other melanoma collections to analyze for heterogeneous, but inversely correlated, expression of Axl and MITF within the same tumor, and whether Axl is preferentially upregulated at the tumor leading edge. A potential positive correlation between Brn2 and Axl should also be explored since Brn2 down-regulates MITF and promotes invasive properties (Goodall et al., 2008). However, it should be noted that decreased invasion and migration of Axl knockdown cells was independent of changes in MITF expression and that other factors, such as TCF4, have been implicated as a hub controlling invasive properties.

Targeting Axl signaling

As noted above, this study utilizes a pharmacological inhibitor of Axl, R428, which was originally developed by Rigel, Inc. R428 inhibits Axl at low nanomolar concentrations with less potent activity towards Tie-2, Ftl-1, Flt-3, Ret, and Abl, and is highly selective towards Axl compared to Tyro-3 and Mer (Holland et al., 2010). Based on the preclinical data in the Sensi et al. study, Axl inhibitors are unlikely to produce strong tumor shrinkage as a monotherapeutic approach in melanomas. Rather, they may prove useful to block the rapid dissemination that frequently occurs in melanoma and/or in combination with inhibitors that arrest proliferation and thus promote the switch to a highly invasive state. Notably, studies from the Hoek lab have shown that inhibition of RAF-MEK signaling induces an invasive-like morphology in melanoma cells cultured on a basement membrane-like matrix; however, whether Axl expression is modulated by ERK1/2 pathway inhibitors remains unclear. Another target is Gas6, the ligand for Axl. Gas6 is produced in a biologically active form in over half of the Axl-expressing melanomas and likely activates Axl signaling via both autocrine and paracrine mechanisms.

While Axl expression was detected with a significantly higher frequency in mutant NRAS melanomas, it may also be relevant to mutant BRAF cells. The RAF inhibitor, PLX4032 (Vemurafenib, Zelboraf®), has been FDA approved, but the clinical benefit is often short term in mutant BRAF V600 patients due to acquired resistance. RTKs are frequently up-regulated in melanoma cells that have acquired resistance to RAF inhibitors (Villanueva et al., 2010). Notably, Axl was a “hit” from a screen to identify kinases that when over-expressed promote resistance to PLX4720, the tool compound for PLX4032 (Johannessen et al., 2010). Sensi et al. show that Gas6 signaling via Axl promotes Akt signaling. Activation of Akt, albeit downstream of IGF-1R, has been implicated previously as one mechanism of acquired resistance to RAF inhibitors (Villanueva et al., 2010). The prevailing consensus is that multiple mechanisms will underlie resistance to RAF inhibitors; thus, it will be worthwhile examining changes in Axl expression in cell line-based models of resistance to RAF inhibitors, as well as in samples from relapsed patients. In summary, Gas6-Axl signaling may be a signaling “axis of evil” in melanoma and should be investigated further as a druggable pathway.

Clinical implications.

1. Axl expression may serve as a biomarker for poorly differentiated tumors with high metastatic potential.

2. Axl is druggable. Small molecule inhibitors to Axl may be an important component of combinatorial strategies to target a subset of melanomas that are poorly differentiated, as indicated by low MITF expression.

3. Upregulation of Axl and/or Axl-mediated signaling may mediate primary and acquired resistance to RAF inhibitors such as PLX4032 (Vemurafenib, Zelboraf®).