Loss of AMPK|[alpha]|1 Expression Is Associated with Poor Survival in Melanoma Patients

TO THE EDITOR Metastatic melanoma is notorious for its rapid progression, resistance to conventional chemotherapy, and poor prognosis, as only 14% of patients with metastatic melanoma survive for 5 years (Miller and Mihm, 2006). There has been limited success in the chemotherapy of melanoma, with only a handful of Food and Drug Administration-approved drugs (Finn et al., 2012). Although BRAFV600E-specific inhibitor, vemurafinib showed promise in treatment of melanoma, its effectiveness has been limited to patients harboring the V600E mutation in BRAF gene; moreover, recent information suggest the development of lethal resistance to vemurafinib (Finn et al., 2012). It was shown earlier that BRAF inhibits the activation of AMP-activated protein kinase (AMPK; Zheng et al., 2009). AMPK is a key regulator of cell metabolism and its activation has been reported to have antineoplastic effects (Kim and He, 2013). However, the role of AMPK in cancer has been controversial with studies showing the ‘prosurvival’ effects of AMPK (Liang and Mills, 2013), and only a few studies reported on the relationship between AMPK expression and cancer progression. To our knowledge, AMPK expression has never been analyzed in melanoma patients. AMPK consists of catalytic α-subunit and regulatory β- and γ-subunits, and in mammals, each subunit occurs as multiple isoforms (Kim and He, 2013). AMPKα1 is reportedly associated with tumor suppressor functions of AMPK (Lee et al., 2010; Liang and Mills, 2013; Zheng et al., 2013) and the present study was therefore undertaken to analyze the correlation between AMPKα1 expression and patient survival using the tissue samples collected from melanoma patients. We performed tissue microarray analysis of AMPKα1 expression in 128 melanoma patients (Supplementary Table S1 online), and the staining pattern is illustrated in Supplementary Figure S1 online. On the basis of the area under the receiver operating characteristic curve values, the best cutoff immunoreactive score (IRS) was determined as IRS-4 (Supplementary Figure S2 online; Remmele and Stegner, 1987). The Kruskal–Wallis and χ2-test analysis of expression using IRS-4 as the cutoff showed that the expression of AMPKα1 was significantly increased from dysplastic nevi to primary melanoma, and then decreased significantly from primary to metastatic melanoma (Figure 1). Furthermore, high staining of AMPKα1 tended to correlate with presence of ulcerated tumors, whereas low AMPKα1 staining was more frequently seen in tumors at sun-exposed sites (Supplementary Table S2 online). The increase in expression of AMPKα1 in primary melanoma patients is in agreement with the previous study on chromosomal alterations in desmoplastic melanoma, which reported amplification in AMPK gene (Pryor et al., 2011). Interestingly, evidence from literature presents a dual role of AMPK in melanoma. On one hand, activation of AMPK by metformin has been shown to inhibit melanoma invasion and metastasis both in vitro and in vivo (Cerezo et al., 2013), on the other hand, AMPK was shown to regulate the expression of micropthalmia-induced transcription factor, and reported to positively influence the growth of BRAFV600E-driven melanoma by upregulating vascular endothelial growth factor (Martin et al., 2012; Borgdorff et al., 2013). The Kaplan–Meier analysis of survival in the patients (primary and metastatic melanoma combined) revealed that the patients with low AMPKα1 expression had significantly worse overall and disease-specific 5-year survival (Supplementary Figures S3A and B online), but when only primary melanoma patients were analyzed the association between AMPKα1 expression and patient survival was not statistically significant (Supplementary Figures S3C and D online). Intriguingly, multivariate but not univariate Cox proportional hazards regression analysis showed that AMPKα1 was an independent prognostic marker when age, gender, tumor subtype, thickness, and ulceration were used as variables (Table 1 and Supplementary Table S3 online). We think that the reason could be due to interaction of AMPKα1 with the variables such as subtype and ulceration, and possibly also because of variability of the correlation between AMPKα1 expression and patient survival when analyzed alone, and when analyzed together with other variables. Further studies are therefore warranted to clarify the association between AMPKα1 expression and patient survival. Although we could not find any reports on AMPKα1 expression and patient survival, previous studies in hepatocellular carcinoma and lung cancer patients showed that low expression of activated AMPK (phospho-AMPK) correlated with poor prognosis, which support our present findings (William et al., 2012; Zheng et al., 2013). Recently, AMPK was reported to phosphorylate BRAF at Ser729 position, leading to association of BRAF with 14-3-3 proteins and to attenuation of the MEK-ERK signaling (Shen et al., 2013). Our group has previously reported the association between BRAF expression and melanoma progression and patient survival, and the present study had 116 cases that were also used for the analysis of BRAF expression (see Supplementary Information online for reference). Analysis of correlation between BRAF and AMPKα1 expression using the common cases revealed that there was a tendency for inverse correlation between the two markers, but the association was not statistically significant (Supplementary Figure S3 online). The Kaplan–Meier analysis of the survival data from these common cases showed that irrespective of BRAF expression, patients with high AMPKα1 expression had comparatively better overall and disease-specific 5-year survival indicating the importance of AMPK in patient prognosis (Supplementary Figure S4 online). The cross talk between mitogen-activated protein kinase pathway and energy sensing pathways is gaining importance, and our findings shed further light on the relationship between BRAF and AMPK. Our results support the recent findings on therapeutic benefits of phenformin in melanoma (Yuan et al., 2013) and encourage for further research on the possible effects of combining BRAF inhibitors and AMPK activators in melanoma treatment. Owing to limitations in sample availability, we could not use a larger database to analyze the AMPKα1 expression, and we think that the correlation between AMPKα1 expression and patient survival needs to be reproduced in a larger cohort. More studies with more patient information would conclusively determine the usefulness of AMPKα1 expression in melanoma prognosis and treatment. Nevertheless, our study identifies the role of AMPKα1 in melanoma prognosis and underscores the importance for further research. The use of human skin tissues and the waiver of patient consent in this study were approved by the Clinical Research Ethics Board of the University of British Columbia and were done in accordance with the Declaration of Helsinki guidelines. The authors state no conflict of interest. We thank Dr Ladan Fazli and Scott Kwong for the help in imaging, Dr Gholamreza Safaee Ardekani for providing the BRAF database, and Canadian Institutes of Health Research for financial support. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper