Genetic Profiling of BRAF Inhibitor|[ndash]|Induced Keratoacanthomas Reveals No Induction of MAP Kinase Pathway Expression

TO THE EDITOR The new RAF inhibitor vemurafenib represents one of the first specifically targeted drugs for melanoma (Bollag et al., 2010; Flaherty et al., 2010; Joseph et al., 2010). However, some individuals receiving vemurafenib develop eruptive squamous neoplasms such as keratoacanthomas (KAs) (Chapman et al., 2011). KAs are usually considered low-grade squamous proliferations (Cribier et al., 1999; Clausen et al., 2006), and are often interpreted as a subtype of squamous cell carcinoma (SCC). Their clinical course involves rapid growth followed by spontaneous involution. The mechanism by which KAs develop is still unknown. Improved understanding of the molecular and genetic basis of KA formation would allow their improved management. To investigate genes involved in the development of KAs in patients receiving vemurafenib, we obtained biopsy specimens and studied gene expression profiles by a microarray approach; full details are available in the Supplementary Material online. These biopsies were originally obtained by the submitting clinician and preserved in formalin for standard paraffin embedding and sectioning and histopathological interpretation from the Pathology Department. We extracted RNA from the lesional cells in the paraffin block to create complementary DNA libraries, which were then screened using the Affymetrix U133plus 2.0 array (Santa Clara, CA). We compared profiles of KAs that arose after drug treatment to spontaneous KAs (arising in patients who did not receive vemurafenib), normal skin, and of classic cutaneous SCC after confirming the original histopathological diagnosis; five samples of each type were analyzed and all KAs were in the proliferative stage. Unsupervised clustering of the most varying gene expression probes demonstrated that KA tumors were separable from both cutaneous SCC and normal skin (Figure 1a), but KAs from patients with and without drug treatment were not distinguishable from each other (Figure 1b). Having observed these overall trends using unsupervised analysis, we used supervised analysis to confirm that although there were 260 significantly differentially expressed probes between drug-induced KA and SCC, and 41 significantly differentially expressed probes between drug-induced KA and normal skin, there were no probes that were significantly differentially expressed between drug-induced and spontaneous KAs (Bonferroni-corrected Student’s t-test P-value 0.01). We concluded that drug-induced KA and spontaneous KA tumors have the same expression profile. Recent reports have indicated that differences in mitogen-activated protein (MAP) kinase gene expression are potentially important for tumor progression (Packer et al., 2009; Joseph et al., 2010; Nazarian et al., 2010; Poulikakos et al., 2010). In our data sets, genes of the MAP kinase pathway were not overexpressed in either drug-induced or spontaneous KA (relative to each other or normal skin). To confirm this, we used quantitative reverse transcriptase–PCR to amplify several genes within the MAP kinase pathway and transforming growth factor-β receptor 1. The results in Table 1 indicated no significant difference in expression. We next examined downstream transcriptional targets of the MAP kinase pathway, including CCND1, DUSP6, SPRY2, ETV1, ETV5, FOSL1, and CXCL1, genes described as directly modulated by vemurafenib inhibition in BRAF V600E melanomas (Packer et al., 2009; Joseph et al., 2010). None was found to be differentially expressed. KAs and other squamous proliferations generally occur on sun-damaged skin. The precise mechanisms that underlie the initiation and growth of these neoplasms remain unclear. Vemurafenib can activate the MAP kinase pathway in wild-type BRAF melanoma in vitro, through either heterodimerization of CRAF with BRAF or activation of wild-type BRAF through NRAS (Heidorn et al., 2010; Joseph et al., 2010; Nazarian et al., 2010; Poulikakos et al., 2010). This effect is transient and may last only a few hours (Joseph et al., 2010). Side effects in melanoma cells may be due to unintended activation of the MAP kinase or other pathways in wild-type BRAF cells. We found that in vemurafenib-associated KAs mRNA levels downstream of BRAF are not increased relative to spontaneous KAs. These data suggest that vemurafenib may accelerate a neoplastic process in susceptible or sun-damaged keratinocytes, similar to the processes underlying spontaneous KAs, although any such activation is transient. Hence, by biopsy time, enhanced activation of the MAP kinase pathway is not seen and there is no increase in transcripts of members of the MAP kinase pathway or of downstream targets, despite continued therapy. Another possibility is that vemurafenib and similar molecules cause different “off-target” effects that cause activation through alternative mechanisms. The induced effects may be dose dependent and additional studies may define the threshold at which KAs occur in patients receiving vemurafenib or related agents. Recently, Oberholzer et al. (2012) and Su et al. (2012) have noted that HRAS mutations are overrepresented in vemurafenib-induced squamous neoplasms (in both KA and SCC) compared with spontaneous ones, and the drug may induce neoplastic proliferation in this setting. One limitation of this study is the small sample size analyzed; nonetheless, our findings indicate that KAs associated with vemurafenib treatment have similar genetic profiles to spontaneous KAs and similar expression of MAP kinase pathway genes. These data suggest that BRAF inhibitors may accelerate a neoplastic process in susceptible keratinocytes. However, any activation appears transient and drug-associated KAs behave similarly as spontaneous KAs. Our results strongly suggest that drug-induced KAs should be treated in a manner similar to spontaneous KAs. Although newer molecularly targeted drugs are presented as highly target specific, they may have “off-target” effects on other proteins. Further study and monitoring will be important to fully understand the underlying mechanisms. The authors state no conflict of interest. We thank Jamie Zhou for his assistance with the reverse transcriptase–PCR experiments and Alistair Cochran for a critical review of the manuscript. RPK acknowledges the Women's Dermatologic Society for support. DJL acknowledges financial support from the Joseph B. Gould Foundation. SWB has received departmental funds from the Department of Pathology. No other funding sources supported this research. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper