Search tips
Search criteria 


Logo of oncotargetLink to Publisher's site
Oncotarget. 2017 January 24; 8(4): 6914–6928.
Published online 2016 December 28. doi:  10.18632/oncotarget.14321
PMCID: PMC5351679

Therapeutic potential of the metabolic modulator phenformin in targeting the stem cell compartment in melanoma


Melanoma is the most dangerous and treatment-resistant skin cancer. Tumor resistance and recurrence are due to the persistence in the patient of aggressive cells with stem cell features, the cancer stem cells (CSC). Recent evidences have shown that CSC display a distinct metabolic profile as compared to tumor bulk population: a promising anti-tumor strategy is therefore to target specific metabolic pathways driving CSC behavior. Biguanides (metformin and phenformin) are anti-diabetic drugs able to perturb cellular metabolism and displaying anti-cancer activity. However, their ability to target the CSC compartment in melanoma is not known. Here we show that phenformin, but not metformin, strongly reduces melanoma cell viability, growth and invasion in both 2D and 3D (spheroids) models. While phenformin decreases melanoma CSC markers expression and the levels of the pro-survival factor MITF, MITF overexpression fails to prevent phenformin effects. Phenformin significantly reduces cell viability in melanoma by targeting both CSC (ALDHhigh) and non-CSC cells and by significantly reducing the number of viable cells in ALDHhigh and ALDHlow-derived spheroids. Consistently, phenformin reduces melanoma cell viability and growth independently from SOX2 levels. Our results show that phenformin is able to affect both CSC and non-CSC melanoma cell viability and growth and suggests its potential use as anti-cancer therapy in melanoma.

Keywords: melanoma, cancer stem cells, phenformin, therapy, ALDH


Melanoma is the most aggressive, highly metastasizing and therapy-resistant skin tumor, with increasing incidence in the last decades. B-RAF inhibitors (B-RAFi) and the recently approved immune-checkpoint inhibitors ipilimumab and nivolumab/pembrolizumab have improved the progression-free survival of metastatic melanoma patients, however a large number of patients do not respond and many display disease relapse [13]. In the last few years, the contribution of cancer stem cells (CSC) to drug resistance has been clarified in many cancer types, including breast cancer and leukemia [46]. Conventional therapies kill mainly regular tumor cells while sparing CSC, leading to drug resistance and therapeutic failure. However, in melanoma, CSC existence and the selectivity/specificity of markers used for CSC isolation have been highly questioned in the last few years [7]. Evidence suggests that melanoma tumor growth is orchestrated by subpopulations of tumor-maintaining cells that can dynamically switch from a more differentiated state, and vice versa [8]. Nevertheless, many report also support the hierarchical CSC model in melanoma and several melanoma CSC markers have been identified so far, including the multi-drug resistance ABCB5 transporter and aldehyde dehydrogenase (ALDH) enzymes [9, 10]. We and others have recently shown the ability of ALDH activity to select for melanoma CSC [1012]. High ALDH levels protect melanoma cells from apoptosis while ALDH1 blockade prevents tumor relapse [1013].

It has become widely accepted that alterations of tumor cell metabolism are hallmarks of cancer [14, 15] and that targeting tumor cell metabolism is a promising anti-cancer strategy [16]. CSC seem to have specific metabolic profiles [17, 18], suggesting that targeting regulators of cancer cell metabolism might be a valuable CSC-eradicating approach. Biguanides, among which metformin and phenformin are the most widely known, are organic compounds with hypoglycaemic properties. Metformin is the first medication used to treat type II diabetes worldwide. Interestingly, biguanides display anti-cancer properties. Epidemiological evidences indicate a 30% lower risk for diabetic patients under metformin treatment to develop cancer [19, 20]. Our group has recently confirmed the anti-tumor effect of metformin and phenformin in breast cancer by targeting both tumor and microenvironment cells [2123]. Biguanides act mainly by inhibiting mitochondrial oxidative phosphorylation (OXPHOS) through the blockade of mitochondrial complex-1 [24]. In melanoma cells, OXPHOS plays an important role in ATP production [25] and recent preclinical studies targeting melanoma cell mitochondrial bioenergetic metabolism have shown to be effective [2629]. However, the effect of biguanides on melanoma growth and invasion is controversial. On one side, metformin seems to inhibit melanoma development [30, 31] and prevent invasion and metastasis [32]. On the other side, metformin accelerates B-RAF mutated melanomas growth in vivo by sustaining angiogenesis [33]. Different reports have shown the ability of metformin to selectively kill cancer stem cells [34, 35] also by reverting their quiescent state [36]. As a consequence, the combination of metformin with chemotherapy targeting the non-stem like compartment of the tumor is promising [37]. Recent findings suggest that other biguanides affect melanoma cell growth [38], possibly by reducing stem cell features [39]. Among these, phenformin strongly reduces melanoma growth and when combined with the B-RAFi PLX4720 gives a significant therapeutic advantage. Although phenformin seems to target specifically slow cycling melanoma cells [40], the direct effect on the CSC compartment of this tumor is unknown.

In the present work, we investigated the ability of phenformin to target the CSC compartment in melanoma by analyzing primary and metastatic melanoma cells both in monolayer cell cultures and 3D spheroids. We show that phenformin, but not metformin, abrogates melanoma cell growth and invasion in 2D and 3D models and affects both CSC and non-CSC cells in melanoma.


Phenformin decreases melanoma cell viability in both monolayer and spheroids cell cultures

First, we tested biguanides toxicity on melanoma cells. Besides SK-MEL-28 and A375 cells, we included the primary melanoma cell line BTC#2 in the analysis as a representative specimen of B-RAF-mutated melanoma cells established from a primary aggressive melanoma [41]. In accordance with previous findings [37], phenformin reduced melanoma cell viability by MTT (Figure (Figure1A,1A, upper panel) and cell proliferation by trypan blue cell counting starting from 24h after stimulus up to 72h (Figure (Figure1A,1A, lower panel). Interestingly, although biguanides interfere with cell metabolism, we observed similar results between MTT, a mitochondrial metabolism-sensitive viability assay, and trypan blue cell counting analyses. Since cell responses in 3D-cell cultures are similar to in vivo behavior [42], we also tested the effect of phenformin on melanoma spheroids by measuring cell viability by trypan blue cell counting 10 days after treatment. First of all, we observed a slight, but not significant, decrease in the number of viable cells/sphere over time in untreated SK-MEL-28 and BTC#2 spheroids (data not shown). This putatively reflects the different sensitivity of these cells to the microenvironmental conditions generated in the spheroid subcompartments, such as suboptimal nutrition and low oxygen supply [43]. When melanoma-derived spheroids were treated with phenformin, we observed a strong reduction in SK-MEL-28 and BTC#2 sphere size and morphology (Figure (Figure1B,1B, upper panel) as well as the number of viable cells in all cell lines upon treatment (Figure (Figure1B,1B, lower panel). Contrarily, the size and shape of A375-derived spheroids was only slightly affected by the treatment (Figure (Figure1B).1B). In line with the decrease in cell viability observed in monolayer cell cultures upon treatment with phenformin, we noticed a stronger effect of the drug on BTC#2-derived spheroids as compared to the other melanoma cell lines (Figure (Figure1B).1B). Interestingly, treatment of melanoma spheroids with a lower dose of phenformin (0.5mM) for 10 days was still able to reduce melanoma sphere-size (SK-MEL-28 and BTC#2) and the number of viable cells/sphere (Supplementary Figure 1A and 1B).

Figure 1
Phenformin reduces melanoma cell viability in both 2D and 3D models

As opposite to what observed for phenformin, no considerable effect of metformin on cell viability and proliferation was observed in monolayer cell cultures (Supplementary Figure 2A). In melanoma-derived spheroids, the number of viable cells/sphere at day10 was significantly reduced by metformin in BTC#2 and SK-MEL-28-derived spheroids, but not in A375 ones (Supplementary Figure 2B). Although the effect of phenformin and metformin in spheroids was similar in some cell lines (e.g. SK-MEL-28, Supplementary Figure 1B and Supplementary Figure 2B), all together our data suggest that phenformin is more effective than metformin in reducing melanoma cell viability.

To characterize phenformin effect on melanoma cell death, cells were treated with 0.5 and 1 mM phenformin for 72h, then harvested for Annexin V/7-AAD or PI staining. We observed a strong and significant induction of early-phase apoptosis by phenformin in all cell lines. This effect was dose-dependent for A375 and BTC#2 cells, while phenformin 0.5mM showed a stronger effect than 1mM in SK-MEL-28 cells (Figure (Figure1C).1C). These results were confirmed by PI-staining at the same timepoint (Figure (Figure1D)1D) and indicate that phenformin induces apoptosis in melanoma cells.

Phenformin, but not metformin, abrogates melanoma cell invasion in 3D spheroid models

In vivo, phenformin inhibits B-RAF-mutated melanoma cell growth [40], however its ability to reduce or inhibit melanoma metastatic process has not yet been assessed. To determine if phenformin affects melanoma cell metastasis, we performed a 3D-spheroid cell invasion assay on melanoma-spheroids in presence or absence of biguanides to evaluate the cell motility. Metformin failed to reduce, and even promoted, cell invasion both in SK-MEL-28 and BTC#2 derived spheroids (Figure (Figure2A).2A). By contrast, phenformin completely blocked melanoma cell invasion in both cell types (Figure (Figure2A).2A). Since phenformin affects melanoma cell viability and growth (Figure (Figure1A1A and and1B),1B), we analyzed invasion as well as cell viability in the same spheroid by enzymatic digestion of the physical bonds between the tumor cells and the extracellular matrix. This procedure allowed the quantification of cell viability in spheroids by trypan blue dye exclusion. As shown in Figure Figure2B,2B, the ability of phenformin to decrease melanoma spheroids invasion is independent from phenformin-induced decrease in cell viability. Moreover, lower doses of phenformin (0.5 mM) were still able to significantly decrease cell invasion both in SK-MEL-28 and BTC#2 cells (Figure (Figure2B).2B). Similar results were observed when spheroids were pre-treated with phenformin for 72h and allowed to invade in absence of the drug, indicating that phenformin fails to select for invasive cells in melanoma spheroids. In these conditions, metformin did not impair melanoma cell invasion (Supplementary Figure 3). All together these data demonstrate that phenformin decreases melanoma cell invasion in 3D models.

Figure 2
Phenformin, but not metformin, reduces melanoma spheroids invasion

MITF overexpression is not protective against phenformin inhibitory effects

In the attempt to define the molecular basis of phenformin inhibitory effect on melanoma cells, we tested whether it could affect the expression of MITF (Microphthalmia-associated transcription factor) that, besides being a melanocytic lineage-specific marker, it is also a pro-survival factor in melanoma [44, 45]. Indeed, phenformin induced a remarkable reduction of MITF protein and mRNA levels in all cell lines by western blotting and qRT-PCR (Figure (Figure3A3A and and3B).3B). To evaluate whether forced expression of MITF could protect from phenformin cytotoxic effects, we constitutively over-expressed MITF in melanoma cells; A375 cells, expressing a very low basal level of MITF [46], were transfected with a MITF-expressing vector or with an empty vector and single clones were generated by antibiotic selection. Two clones expressing different levels of MITF were selected: MITF#6 (high expression level) and MITF#9 (medium expression level) (Figure (Figure3C).3C). We used CTRL#3 and CTRL#4 clones as controls. First we verified that the MITF overexpressed protein was functional in our system. We analyzed the expression of well-established MITF target genes (Bcl-2, cMET and PGC1-α) in MITF clones by real time PCR. As shown in Figure Figure3C3C (lower panel) MITF overexpression increases the expression of the three genes in a MITF-dose dependent manner, thus confirming that MITF is functional in transfected cells. However, we failed to observe any difference in cell viability after treatment with phenformin between CTRL and MITF-overexpressing cells (Figure (Figure3D),3D), indicating that MITF is not sufficient to rescue phenformin cytotoxic activity in melanoma cells. No effect on MITF exogenous protein stability was observed upon treatment (Figure (Figure3E3E).

Figure 3
MITF is not protective against phenformin in melanoma cells

Phenformin decreases stem cell traits in melanoma

We have recently shown that ALDHhigh cells from melanoma specimens and cell lines retain CSC features and are highly resistant to conventional therapies [10]. Therefore, we asked whether phenformin is able to target CSC in melanoma. To this aim, we isolated ALDHhigh and ALDHlow melanoma cells by fluorescence-activated cell sorting as previously shown [10]. We chose to sort from SK-MEL-28 and A375 cells since only 0.8 % of BTC#2 cells are ALDHhigh (Supplementary Figure 4A). As expected, ALDHhigh melanoma cells express significantly higher levels of ALDH1A3, SOX2 and CD271 than ALDHlow cells by real time PCR (Figure (Figure4B),4B), western blotting (SOX2/ALDH1A1, Supplementary Figure 4B) and FACS analysis (CD271, Supplementary Figure 4C). ALDHlow cells slightly overexpressed MITF, consistent with its role in determining the differentiation towards the melanocyte lineage; as opposite, MITF main regulator, SOX10, was equally expressed in the two populations (Supplementary Figure 4D). Next, we evaluated the behavior of sorted ALDHhigh and ALDHlow SK-MEL-28 (Figure (Figure4)4) and A375 cells (Supplementary Figure 5) in 2D and 3D models. Sorted ALDHhigh and ALDHlow populations were seeded on ultra-low attachment plates to form spheroids, then monitored up to 14 days post-seeding. Both ALDHhigh, ALDHlow and bulk (unsorted) cells were able to form spheroids in these conditions, yet ALDHhigh cells generated slightly larger spheroids as compared to ALDHlow ones (Figure (Figure4C4C and Supplementary Figure 5A, left panels). At day 14, total cell output of ALDHhigh-derived spheroids was similar to that of ALDHlow ones (Figure (Figure4C4C and Supplementary Figure 5A, right panels). In accordance, the number of viable cells was not significantly different in ALDHhigh derived spheroids as compared to ALDHlow derived ones, as shown by trypan blue cell counting (Figure (Figure4C4C and Supplementary Figure 5A, right panels). When melanoma cells grown as monolayer were compared with spheroids for ALDH activity, we observed a significant increase in the number of ALDHhigh cells in 3D cell cultures as compared to monolayer cell cultures. This suggests that melanoma cells grown as spheroids may increase the CSC compartment by upregulating ALDH activity (Figure (Figure4D),4D), however further experiments are needed to clarify this point. We also observed a strong variation of ALDHhigh cell number in melanoma 2D cell cultures over time (Figure (Figure4D).4D). Although we don't have a full explanation on these results, it is possible that ALDH levels fluctuate in culture also in response to progressive metabolic changes that take place during the different phases of the cell cycle and as a function of cell density [47, 48].

Figure 4
ALDHhigh melanoma cells behavior in 2D vs 3D cell culture models

In order to assess phenformin effects on the CSC compartment in melanoma, we isolated ALDHhigh and ALDHlow melanoma cells by fluorescence activated cell sorting and analyzed them in 2D and 3D models in presence or absence of phenformin. In monolayer cell cultures, phenformin reduced ALDH isoforms expression in melanoma cells (Figure (Figure5A)5A) and ALDHhigh and ALDHlow melanoma cells viability at 48h, with no significant difference between the two populations (Figure (Figure5B5B and Supplementary Figure 5B). When seeded to form spheroids, both ALDHhigh and ALDHlow 96h-old spheroids were visibly similarly sensitive to phenformin at day 10 post-treatment (Figure 5B, 5C and Supplementary Figure 5B, 5C). This was confirmed by the evaluation of total cell output/spheroid (Figure (Figure5D5D and Supplementary Figure 5D) at the same timepoint, showing that phenformin leads to a comparable decrease in the overall number of cells in ALDHhigh and ALDHlow-derived spheroids. In our previous work, we have shown that ALDHhigh melanoma cells are resistant to paclitaxel, a drug currently employed in the treatment of advanced-stage melanoma patients [10]. Here we observed that while ALDHhigh cells were resistant to paclitaxel (Supplementary Figure 4E), phenformin was effective in targeting both ALDHhigh and ALDHlow melanoma cells. We then wanted to confirm these results by assessing the effect of phenformin in a cellular model of melanoma that overexpresses the widely known stem cell marker SOX2. SOX2 is a transcription factor essential for maintaining the tumorigenic ability of CSC in different tumor types [4951]. In melanoma, it promotes survival and self-renewal of CSC expressing high levels of ALDH [49]. qPCR (Figure (Figure6A)6A) and western blotting (Supplementary Figure 6A) analyses show that phenformin induces a significant down-regulation of SOX2 in both A375 and SK-MEL-28 cells. To evaluate whether forced expression of SOX2 could protect from phenformin cytotoxic effects, we transfected A375 cells with a SOX2-expressing vector or with an empty vector as control. Single clones were generated by antibiotic selection; SOX2 overexpression was confirmed by western blotting (Figure (Figure6B)6B) and immuno-fluorescence (Figure (Figure6C6C and Supplementary Figure 6B). One representative A375-SOX2 clone (in which 100% cells overexpressed SOX2, Figure Figure6C)6C) and one A375-CTRL clone were used. Phenformin strongly reduced the number of viable cells at 72h without any difference between SOX2-overexpressing and CTRL clone (Figure (Figure6D).6D). This was confirmed by viability assays, MTT and trypan blue cell counting, as shown in Figure Figure6E6E and and6F6F respectively. Interestingly, when we assessed the levels of ectopic SOX2 in SOX2-overexpressing cells treated with phenformin, we observed a strong decrease of SOX2 levels. This observation indicates that phenformin inhibits SOX2 expression through mechanisms that affect both its mRNA and protein.

Figure 5
Phenformin targets both ALDHhigh (CSC) and ALDHlow (non-CSC) melanoma cells
Figure 6
SOX2 overexpression in melanoma is not protective against phenformin


This study shows that modulation of melanoma cells energetic metabolism by the OXPHOS-inhibitor phenformin reduces melanoma cell viability, growth and invasion both in monolayer cell cultures and in multi-cellular spheroids. We show for the first time that phenformin reduces stem cell features in melanoma by downregulating ALDH and SOX2 expression levels. Phenformin suppresses both CSC and non-CSC cell viability and growth. By contrast metformin, a phenformin analogue, is less effective in reducing melanoma cell viability and fails to decrease cell invasion.

Using both 2D- and 3D-models we observed some discordant results in terms of drug responses [52, 53, 26], possibly due to different metabolic environments characterizing the two models. Contrarily to what observed in 2D-cell cultures, the metabolic activity of tumor cells is not homogeneous in spheroids and differs in the inner (glycolytic) vs outer (OXPHOS-dependent) layers of the spheroid structure. Since 3D cell culture approaches more accurately recapitulate solid tumor architecture, heterogeneity, differentiation and metabolism, the choice to use both 2D and 3D models in this study is of note and helps to overcome some methodological limitations of conventional 2D cell culture.

Phenformin, which is 50 fold more potent [54] and displays greater tissue bioavailability than metformin [55], markedly decreased melanoma cell viability by inducing apoptosis. Further, it reduced cell growth and invasion in both 2D and 3D models. This fits with what observed on tumor growth by Yuan and colleagues in vivo [40]. Interestingly, we also observed marked cytotoxic effects induced by phenformin on melanoma cells at doses lower than 1mM. Since high doses phenformin are toxic in vivo, our data as well as observations from other groups support its use in the clinic at lower, more tolerable doses, possibly in therapeutic combination regimens [27, 40, 51]. In addition, we observed for the first time a strong effect of phenformin on melanoma cell invasion thus indicating a potential phenformin anti-metastatic effect. Of note, phenformin activity on slow cycling BTC#2 cells is stronger as compared to fast growing cell lines. This is in line with the evidence that slow cycling, OXPHOS-dependent cells are more sensitive to drugs able to disrupt the mitochondrial respiration chain [26]. Consistently, phenformin seems to target slow cycling, JARID1B-high, melanoma cells [40].

In melanoma, MITF is both a pro-survival and a differentiating factor [57]. Interestingly, MITF drives metabolic reprogramming of melanoma cells towards mitochondrial metabolism by promoting the transcription of several genes involved in OXPHOS [58]. In line with this evidence, we observed that phenformin treatment strongly repressed MITF endogenous expression in A375 cells. This observation seems to indicate that phenformin has a transcriptional effect on MITF expression and suggests that MITF repression may be functional to phenformin activity in melanoma. However, when we tested the effect of phenformin in MITF-overexpressing melanoma cells, we did not observe any difference in drug response in MITF- and CTRL-cells. This seems to indicate that MITF is not sufficient to rescue phenformin cytotoxic effects in melanoma.

Melanoma is highly resistant to conventional treatments: even recently developed drugs show some degree of recurrence, this being mainly due to CSC persistence. Given melanoma heterogeneity and melanoma cell plasticity, the identification of a unique marker able to select cells with CSC features has been challenging [59]. We have previously shown that ALDHhigh cells in melanoma retain a CSC phenotype and recently, a role for SOX2 in sustaining the tumorigenic ability of ALDHhigh melanoma cells has been demonstrated [49]. Since SOX2 overexpression increases the self renewal ability of melanoma cells thus generating stem-like cells, we used this model as well as sorted ALDHhigh/low cells to study the effect of phenformin on the stem cell compartment in melanoma. We herein show for the first time that phenformin is among the few anti-cancer drugs able to target cells with the ALDHhigh CSC phenotype, although not specifically. Interestingly, the strong activity of phenformin on melanoma spheroids, which in turn display a higher ALDH activity as compared to monolayer cell cultures [60] (Figure (Figure4D),4D), seems to suggest that phenformin may have a preferential stronger cytotoxic effect on melanoma CSC. However, although ALDHhigh cells are resistant to paclitaxel, their sensitivity to phenformin is comparable with that of ALDHlow counterpart. This evidence is likely due to still unraveled molecular mechanisms activated by phenformin that might be shared by ALDHhigh and ALDHlow cells. Although we did not analyze glucose metabolism in melanoma CSC, it is reported that CSC often overexpress genes associated with glucose uptake, OXPHOS and fatty acid beta-oxidation [61], explaining, at least in part, the sensitivity of CSC to OXPHOS inhibitors.

Modulation of cell metabolism by inhibition of mitochondrial complex-1 has shown promising results [62]. In addition, resistance to single and combinatorial therapies is often associated with upregulation of OXPHOS [16]. Given the ability of the OXPHOS-inhibitor phenformin to target both CSC and non-CSC, our data provide evidence that phenformin might be a successful therapeutic option in melanoma, possibly in combination regimens.



Metformin (1,1-dimethylbiguanide hydrochloride), phenformin (N-(2-Phenylethyl)imidodicarbonimidic diamide monohydrochloride) and 3-(4,5-dimethylthiazol- 2-yl)-2,5-diphenyltetrazolium bromide were purchased from Sigma–Aldrich (St Louis, MO). Lipofectamine 2000 was purchased from Invitrogen (Eugene, OR). Paclitaxel was purchased from Accord Healthcare Italia.

Cell cultures

Human B-RAFV600E mutant melanoma cell line A375 was purchased from ATCC. SK-MEL-28 were a kind gift from Prof. Pincelli at the University of Modena and Reggio Emilia, Italy. Cell lines were authenticated by short tandem repeat DNA profiling after purchase. BTC#2 is a short-term melanoma cell line derived from a B-RAFVE600E mutated primary melanoma patient enrolled at Arcispedale S. Maria Nuova after signing an informed consent. Once isolated, BTC#2 cells were maintained in RPMI medium supplemented with 10% (vol/vol) heat inactivated fetal bovine serum (FBS). Expression of MART-1 by immuno-cytochemistry was used to confirm the identity of melanoma cells which were 98-100% positive. A375 and SK-MEL-28 cells were cultured in DMEM medium supplemented with 10% (vol/vol) FBS or BME medium supplemented with 10% (vol/vol) FBS containing penicillin/streptomycin, sodium pyruvate (1 mM), sodium bicarbonate (1.5 g/L) and NEAA (0.1mM), respectively. All cell lines were incubated at 37°C under 5% CO2. For SOX2 and MITF stable clones derivation see supplementary M&M.

Spheroid formation

We cultured melanoma cells either on flat bottom plates (2D-cell culture models) or round bottom ultra-low attach plates (3D-cell culture models/spheroids) in complete cell culture media. Culturing cells in ultra-low attachment plates permits growth of multi-cellular tumor spheroids that are organotypic models of solid tumor tissues. Melanoma cells were harvested from monolayer cultures, counted and resuspended into complete medium. 3000-4000 cells were seeded into each well of 96-well Corning Ultra-Low Attachment Plates (Thermo Fisher Scientific Inc., Waltham, MA) for 96h. Once formed, spheroids were incubated with vehicle or metformin/phenformin at the indicated doses up to 10 days. The stimuli were renewed every 3 days. At each timepoint, spheroids were harvested and single cell suspension was obtained by mechanical disaggregation procedure for further analyses.

Aldefluor assay and flow cytometry

The Aldefluor kit (Stem Cell Technologies, Vancouver, Canada) was used to isolate or quantify cells with high ALDH activity. For the isolation of ALDHhigh/low populations in melanoma, only SK-MEL-28 cells were used. For the quantification of ALDH activity in monolayer vs 3D cell cultures, A375, SK-MEL-28 and BTC#2-derived spheroids and monolayer cell cultures were analyzed at the indicated timepoints. Details on the sorting strategy and protocol are reported in the supplementary M&M section.

3D spheroid BME cell invasion assay

3D Spheroid BME Cell Invasion Assay was performed on SK-MEL-28 and BTC#2-derived spheroids, following manufacturer instructions (Trevigen 3-D spheroid Cell Invasion Assay, Trevigen, Gaithersburg, MD USA). We evaluated the ability of biguanides to inhibit spheroid invasion by drug-treating spheroids in invasion-permissive conditions (method I), or by pre-treating spheroids before embedding them into invasion-matrix (method II). The two methods are described in detail in the supplementary M&M section.


A375-SOX2 and CTRL clones were grown on chambers-slides for 48 hours, then washed in PBS and fixed in situ with buffered para-formaldehyde (4%) for 20 minutes at RT. Cells were permeabilized with 0.1% Triton X-100 for 5 minutes on ice, incubated with 2% bovine serum albumin/10% goat serum for 20 minutes, then for 60 minutes at 37°C with the mouse monoclonal anti-flag antibody (1:200, Thermo Scientific) or with FBS buffer alone as control. After three washes in PBS, cells were incubated for 60 minutes with the FITC-conjugated anti-mouse secondary antibody (Invitrogen) at 1:1000 dilution. Sections were then counterstained with Dapi (1:1000 diluted in PBS) for 5 minutes at room temperature and coverslipped with SlowFade® (Invitrogen) reagent.

Statistical analyses

Statistical analysis was performed by Student's t-test. One or two asterisks indicate a significant difference, * 0.01 < P < 0.05 and ** P < 0.01, respectively.



This work was supported by Fondazione Umberto Veronesi and Associazione Italiana per la Ricerca sul Cancro (AIRC).



The authors state no conflicts of interest.


1. Flaherty KT, Puzanov I, Kim KB, Ribas A, McArthur GA, Sosman JA, O'Dwyer PJ, Lee RJ, Grippo JF, Nolop K, Chapman PB. Inhibition of mutated, activated BRAF in metastatic melanoma. N Engl J Med. 2010;363:809–19. doi: 10.1056/NEJMoa1002011. [PMC free article] [PubMed] [Cross Ref]
2. Hodi FS, O'Day SJ, McDermott DF, Weber RW, Sosman JA, Haanen JB, Gonzalez R, Robert C, Schadendorf D, Hassel JC, Akerley W, van den Eertwegh AJ, Lutzky J, et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med. 2010;363:711–23. doi: 10.1056/NEJMoa1003466. [PMC free article] [PubMed] [Cross Ref]
3. Robert C, Ribas A, Wolchok JD, Hodi FS, Hamid O, Kefford R, Weber JS, Joshua AM, Hwu WJ, Gangadhar TC, Patnaik A, Dronca R, Zarour H, et al. Anti-programmed-death-receptor-1 treatment with pembrolizumab in ipilimumab-refractory advanced melanoma: a randomised dose-comparison cohort of a phase 1 trial. Lancet. 2014;384:1109–17. doi: 10.1016/S0140-6736(14)60958-2. [PubMed] [Cross Ref]
4. Bozorgi A, Khazaei M, Khazaei MR. New Findings on Breast Cancer Stem Cells: A Review. J Breast Cancer. 2015;18:303–312. doi: 10.4048/jbc.2015.18.4.303. [PMC free article] [PubMed] [Cross Ref]
5. Abdullah LN, Kai-Hua Chow E. Mechanisms of chemoresistance in cancer stem cells. Clin Transl Med. 2013;2:3. doi: 10.1186/2001-1326-2-3. [PMC free article] [PubMed] [Cross Ref]
6. Dick JE. Stem cell concepts renew cancer research. Blood. 2008;112:4793–807. doi: 10.1182/blood-2008-08-077941. [PubMed] [Cross Ref]
7. Quintana E, Shackleton M, Foster HR, Fullen DR, Sabel MS, Johnson TM, Morrison SJ. Phenotypic heterogeneity among tumorigenic melanoma cells from patients that is reversible and not hierarchically organized. Cancer Cell. 2010;18:510–23. doi: 10.1016/j.ccr.2010.10.012. [PMC free article] [PubMed] [Cross Ref]
8. Roesch A, Fukunaga-Kalabis M, Schmidt EC, Zabierowski SE, Brafford PA, Vultur A, Basu D, Gimotty P, Vogt T, Herlyn M. A temporarily distinct subpopulation of slow-cycling melanoma cells is required for continuous tumor growth. Cell. 2010;141:583–94. doi: 10.1016/j.cell.2010.04.020. [PMC free article] [PubMed] [Cross Ref]
9. Schatton T, Murphy GF, Frank NY, Yamaura K, Waaga-Gasser AM, Gasser M, Zhan Q, Jordan S, Duncan LM, Weishaupt C, Fuhlbrigge RC, Kupper TS, Sayegh MH, et al. Identification of cells initiating human melanomas. Nature. 2008;451:345–9. doi: 10.1038/nature06489. [PMC free article] [PubMed] [Cross Ref]
10. Luo Y, Dallaglio K, Chen Y, Robinson WA, Robinson SE, McCarter MD, Wang J, Gonzalez R, Thompson DC, Norris DA, Roop DR, Vasiliou V, Fujita M. ALDH1A isozymes are markers of human melanoma stem cells and potential therapeutic targets. Stem Cells. 2012;30:2100–13. doi: 10.1002/stem.1193. [PMC free article] [PubMed] [Cross Ref]
11. Luo Y, Nguyen N, Fujita M. Isolation of human melanoma stem cells using ALDH as a marker. Curr Protoc Stem Cell Biol. 2013;26 doi: 10.1002/9780470151808.sc0308s26. Unit 3.8. [PMC free article] [PubMed] [Cross Ref]
12. Boonyaratanakornkit JB, Yue L, Strachan LR, Scalapino KJ, LeBoit PE, Lu Y, Leong SP, Smith JE, Ghadially R. Selection of tumorigenic melanoma cells using ALDH. J Invest Dermatol. 2010;130:2799–808. doi: 10.1038/jid.2010.237. [PubMed] [Cross Ref]
13. Yue L, Huang ZM, Fong S, Leong S, Jakowatz JG, Charruyer-Reinwald A, Wei M, Ghadially R. Targeting ALDH1 to decrease tumorigenicity, growth and metastasis of human melanoma. Melanoma Res. 2015;25:138–48. doi: 10.1097/CMR.0000000000000144. [PubMed] [Cross Ref]
14. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646–74. doi: 10.1016/j.cell.2011.02.013. [PubMed] [Cross Ref]
15. Pavlova NN, Thompson CB. The Emerging Hallmarks of Cancer Metabolism. Cell Metab. 2016;23:27–47. doi: 10.1016/j.cmet.2015.12.006. [PMC free article] [PubMed] [Cross Ref]
16. Kroemer G, Pouyssegur J. Tumor cell metabolism: cancer's Achilles' heel. Cancer Cell. 2008;13:472–82. doi: 10.1016/j.ccr.2008.05.005. [PubMed] [Cross Ref]
17. Sancho P, Burgos-Ramos E, Tavera A, BouKheir T, Jagust P, Schoenhals M, Barneda D, Sellers K, Campos-Olivas R, Graña O, Viera CR, Yuneva M, Sainz B, Jr, et al. MYC/PGC-1α Balance Determines the Metabolic Phenotype and Plasticity of Pancreatic Cancer Stem Cells. Cell Metab. 2015;22:590–605. doi: 10.1016/j.cmet.2015.08.015. [PubMed] [Cross Ref]
18. De Luca A, Fiorillo M, Peiris-Pagès M, Ozsvari B, Smith DL, Sanchez-Alvarez R, Martinez-Outschoorn UE, Cappello AR, Pezzi V, Lisanti MP, Sotgia F. Mitochondrial biogenesis is required for the anchorage-independent survival and propagation of stem-like cancer cells. Oncotarget. 2015;6:14777–95. doi: 10.18632/oncotarget.4401. [PMC free article] [PubMed] [Cross Ref]
19. Evans JM1, Donnelly LA, Emslie-Smith AM, Alessi DR, Morris AD. Metformin and reduced risk of cancer in diabetic patients. BMJ. 2005;330:1304–5. doi: 10.1136/bmj.38415.708634.F7. [PMC free article] [PubMed] [Cross Ref]
20. Decensi A, Puntoni M, Goodwin P, Cazzaniga M, Gennari A, Bonanni B, Gandini S. Metformin and cancer risk in diabetic patients: a systematic review and meta-analysis. Cancer Prev Res. (Phila) 2010;3:1451–61. doi: 10.1158/1940-6207.CAPR-10-0157. Epub 2010 Oct 12. [PubMed] [Cross Ref]
21. Dallaglio K, Bruno A, Cantelmo AR, Esposito AI, Ruggiero L, Orecchioni S, Calleri A, Bertolini F, Pfeffer U, Noonan DM, Albini A. Paradoxic effects of metformin on endothelial cells and angiogenesis. Carcinogenesis. 2014;35:1055–66. doi: 10.1093/carcin/bgu001. [PMC free article] [PubMed] [Cross Ref]
22. Orecchioni S, Reggiani F, Talarico G, Mancuso P, Calleri A, Gregato G, Labanca V, Noonan DM, Dallaglio K, Albini A, Bertolini F. The biguanides metformin and phenformin inhibit angiogenesis, local and metastatic growth of breast cancer by targeting both neoplastic and microenvironment cells. Int J Cancer. 2015;136:E534–44. doi: 10.1002/ijc.29193. [PubMed] [Cross Ref]
23. Talarico G, Orecchioni S, Dallaglio K, Reggiani F, Mancuso P, Calleri A, Gregato G, Labanca V, Rossi T, Noonan DM, Albini A, Bertolini F. Aspirin and atenolol enhance metformin activity against breast cancer by targeting both neoplastic and microenvironment cells. Sci Rep. 2016;6:18673. doi: 10.1038/srep18673. [PMC free article] [PubMed] [Cross Ref]
24. Wheaton WW, Weinberg SE, Hamanaka RB, Soberanes S, Sullivan LB, Anso E, Glasauer A, Dufour E, Mutlu GM, Budigner GS, Chandel NS. Metformin inhibits mitochondrial complex I of cancer cells to reduce tumorigenesis. Elife. 2014;3:e02242. doi: 10.7554/eLife.02242. [PMC free article] [PubMed] [Cross Ref]
25. Ho J, de Moura MB, Lin Y, Vincent G, Thorne S, Duncan LM, Hui-Min L, Kirkwood JM, Becker D, Van Houten B, Moschos SJ. Importance of glycolysis and oxidative phosphorylation in advanced melanoma. Mol Cancer. 2012;11:76. doi: 10.1186/1476-4598-11-76. [PMC free article] [PubMed] [Cross Ref]
26. Cierlitza M, Chauvistré H, Bogeski I, Zhang X, Hauschild A, Herlyn M, Schadendorf D, Vogt T, Roesch A. Mitochondrial oxidative stress as a novel therapeutic target to overcome intrinsic drug resistance in melanoma cell subpopulations. ExpDermatol. 2015;24:155–7. doi: 10.1111/exd.12613. [PMC free article] [PubMed] [Cross Ref]
27. Livingstone E, Swann S, Lilla C, Schadendorf D, Roesch A. Combining BRAF(V) (600E) inhibition with modulators of the mitochondrial bioenergy metabolism to overcome drug resistance in metastatic melanoma. Exp Dermatol. 2015;24:709–10. doi: 10.1111/exd.12718. [PubMed] [Cross Ref]
28. Haq R, Fisher DE, Widlund HR. Molecular pathways: BRAF induces bioenergetic adaptation by attenuating oxidative phosphorylation. Clin Cancer Res. 2014;20:2257–63. doi: 10.1158/1078-0432.CCR-13-0898. [PMC free article] [PubMed] [Cross Ref]
29. Gopal YN, Rizos H, Chen G, Deng W, Frederick DT, Cooper ZA, Scolyer RA, Pupo G, Komurov K, Sehgal V, Zhang J, Patel L, Pereira CG, et al. Inhibition of mTORC1/2 overcomes resistance to MAPK pathway inhibitors mediated by PGC1α and oxidative phosphorylation in melanoma. Cancer Res. 2014;74:7037–47. doi: 10.1158/0008-5472.CAN-14-1392. [PMC free article] [PubMed] [Cross Ref]
30. Tomic T, Botton T, Cerezo M, Robert G, Luciano F, Puissant A, Gounon P, Allegra M, Bertolotto C, Bereder JM, Tartare-Deckert S, Bahadoran P, Auberger P, et al. Metformin inhibits melanoma development through autophagy and apoptosis mechanisms. Cell Death Dis. 2011;2:e199. doi: 10.1038/cddis.2011.86. [PMC free article] [PubMed] [Cross Ref]
31. Hirsch HA, Iliopoulos D, Struhl K. Metformin inhibits the inflammatory response associated with cellular transformation and cancer stem cell growth. Proc Natl AcadSci U S A. 2013;110:972–7. doi: 10.1073/pnas.1221055110. [PubMed] [Cross Ref]
32. Cerezo M, Tichet M, Abbe P, Ohanna M, Lehraiki A, Rouaud F, Allegra M, Giacchero D, Bahadoran P, Bertolotto C, Tartare-Deckert S, Ballotti R, Rocchi S. Metformin blocks melanoma invasion and metastasis development in AMPK/p53-dependent manner. Mol Cancer Ther. 2013;12:1605–15. doi: 10.1158/1535-7163.MCT-12-1226-T. [PubMed] [Cross Ref]
33. Martin MJ, Hayward R, Viros A, Marais R. Metformin accelerates the growth of BRAF V600E-driven melanoma by upregulating VEGF-A. Cancer Discov. 2012;2:344–55. doi: 10.1158/2159-8290.CD-11-0280. [PMC free article] [PubMed] [Cross Ref]
34. Gritti M, Würth R, Angelini M, Barbieri F, Peretti M, Pizzi E, Pattarozzi A, Carra E, Sirito R, Daga A, Curmi PMG, Mazzanti M, Florio T. Metformin repositioning as antitumoral agent: selective antiproliferative effects in human glioblastoma stem cells, via inhibition of CLIC1-mediated ion current. Oncotarget. 2014;5:11252–68. doi: 10.18632/oncotarget.2617. [PMC free article] [PubMed] [Cross Ref]
35. Janzer A, German NJ, Gonzalez-Herrera KN, Asara JM, Haigis MC, Struhl K. Metformin and phenformin deplete tricarboxylic acid cycle and glycolytic intermediates during cell transformation and NTPs in cancer stem cells. Proc Natl Acad Sci U S A. 2014;111:10574–9. doi: 10.1073/pnas.1409844111. [PubMed] [Cross Ref]
36. Kangwan N, Park JM, Kim EH, Hahm KB. Chemoquiescence for ideal cancer treatment and prevention: where are we now? J Cancer Prev. 2014;19:89–6. doi: 10.15430/JCP.2014.19.2.89. [PMC free article] [PubMed] [Cross Ref]
37. Hirsch HA, Iliopoulos D, Tsichlis PN, Struhl K. Metformin selectively targets cancer stem cells, and acts together with chemotherapy to block tumor growth and prolong remission. Cancer Res. 2009;69:8832. doi: 10.1158/0008-5472.CAN-09-2994. [PMC free article] [PubMed] [Cross Ref]
38. Petti C, Vegetti C, Molla A, Bersani I, Cleris L, Mustard KJ, Formelli F, Hardie GD, Sensi M, Anichini A. AMPK activators inhibit the proliferation of human melanomas bearing the activated MAPK pathway. Melanoma Res. 2012;22:341–50. doi: 10.1097/CMR.0b013e3283544929. [PubMed] [Cross Ref]
39. Kaushik G, Kwatra D, Subramaniam D, Jensen RA, Anant S, Mammen JM. Honokiol affects melanoma cell growth by targeting the AMP-activated protein kinase signaling pathway. Am J Surg. 2014;208:995–1002. doi: 10.1016/j.amjsurg.2014.09.014. [PMC free article] [PubMed] [Cross Ref]
40. Yuan P, Ito K, Perez-Lorenzo R, Del Guzzo C, Lee JH, Shen CH, Bosenberg MW, McMahon M, Cantley LC, Zheng B. Phenformin enhances the therapeutic benefit of BRAF (V600E) inhibition in melanoma. Proc Natl AcadSci U S A. 2013;110:18226–31. doi: 10.1073/pnas.1317577110. [PubMed] [Cross Ref]
41. Gandolfi G, Longo C, Moscarella E, Zalaudek I, Sancisi V, Raucci M, Manzotti G, Gugnoni M1, Piana S, Argenziano G, Ciarrocchi A. The extent of whole-genome copy number alterations predicts aggressive features in primary melanomas. Pigment Cell Melanoma Res. 2016;29:163–75. doi: 10.1111/pcmr.12436. [PubMed] [Cross Ref]
42. Smalley KSM, Lioni M, Noma K, Haass NK, Herlyn M. In vitro three-dimensional tumor microenvironment models for anticancer drug discovery. Expert Opinion Drug Discovery. 2008;3:1–10. doi: 10.1517/17460441.3.1.1. [PubMed] [Cross Ref]
43. Haass NK, Beaumont KA, Hill DS, Anfosso A, Mrass P, Munoz MA, Kinjyo I, Weninger W. Real-time cell cycle imaging during melanoma growth, invasion, and drug response. Pigment Cell Melanoma Res. 2014;27:764–76. doi: 10.1111/pcmr.12274. [PubMed] [Cross Ref]
44. Widlund HR, Fisher DE. Microphthalmia-associated transcription factor: a critical regulator of pigment cell development and survival. Oncogene. 2003;22:3035–3041. doi: 10.1038/sj.onc.1206443. [PubMed] [Cross Ref]
45. Hartman ML, Czyz M. Pro-survival role of MITF in melanoma. J Invest Dermatol. 2015;135:352–358. doi: 10.1038/jid.2014.319. [PubMed] [Cross Ref]
46. Kido K, Sumimoto H, Asada S, Okada SM, Yaguchi T, Kawamura N, Miyagishi M, Saida T, Kawakami Y. Simultaneous suppression of MITF and BRAF V600E enhanced inhibition of melanoma cell proliferation. Cancer Sci. 2009;100:1863–9. doi: 10.1111/j.1349-7006.2009.01266.x. [PubMed] [Cross Ref]
47. Lee MW, Kim DS, Yoo KH, Kim HR, Jang IK, Lee JH, Kim SY, Son MH, Lee SH, Jung HL, Sung KW, Koo HH. Human bone marrow-derived mesenchymal stem cell gene expression patterns vary with culture conditions. Blood Res. 2013;48:107–14. doi: 10.5045/br.2013.48.2.107. [PMC free article] [PubMed] [Cross Ref]
48. Marchitti SA, Brocker C, Stagos D, Vasiliou V. Non-P450 aldehyde oxidizing enzymes: the aldehyde dehydrogenase superfamily. Expert Opin Drug Metab Toxicol. 2008;4:697–720. doi: 10.1517/17425255.4.6.697. [PMC free article] [PubMed] [Cross Ref]
49. Santini R, Pietrobono S, Pandolfi S, Montagnani V, D'Amico M, Penachioni JY, Vinci MC, Borgognoni L, Stecca B. SOX2 regulates self-renewal and tumorigenicity of human melanoma-initiating cells. Oncogene. 2014;33:4697–708. doi: 10.1038/onc.2014.71. [PMC free article] [PubMed] [Cross Ref]
50. Basu-Roy U, Seo E, Ramanathapuram L, Rapp TB, Perry JA, Orkin SH, Mansukhani A, Basilico C. Sox2 maintains self renewal of tumor-initiating cells in osteosarcomas. Oncogene. 2012;31:2270–82. doi: 10.1038/onc.2011.405. [PMC free article] [PubMed] [Cross Ref]
51. Siegle JM, Basin A, Sastre-Perona A, Yonekubo Y, Brown J, Sennett R, Rendl M, Tsirigos A, Carucci JA, Schober M. SOX2 is a cancer-specific regulator of tumour initiating potential in cutaneous squamous cell carcinoma. Nat Commun. 2014;5:4511. doi: 10.1038/ncomms5511. [PMC free article] [PubMed] [Cross Ref]
52. Weiswald LB, Bellet D, Dangles-Marie V. Spherical cancer models in tumor biology. Neoplasia. 2015;17:1–15. doi: 10.1016/j.neo.2014.12.004. [PMC free article] [PubMed] [Cross Ref]
53. Kirk R. Targeted therapies: the toxic reality of new drugs. Nat Rev Clin Oncol. 2012;9:488. doi: 10.1038/nrclinonc.2012.134. [PubMed] [Cross Ref]
54. Owen MR, Doran E, Halestrap AP. Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochem J. 2000;348(Pt 3):607–14. [PubMed]
55. Hawley SA, Ross FA, Chevtzoff C, Green KA, Evans A, Fogarty S, Towler MC, Brown LJ, Ogunbayo OA, Evans AM, Hardie DG. Use of cells expressing gamma subunit variants to identify diverse mechanisms of AMPK activation. Cell Metab. 2010;11:554–65. doi: 10.1016/j.cmet.2010.04.001. [PMC free article] [PubMed] [Cross Ref]
56. Shackelford DB1.Unravelling the connection between metabolism and tumorigenesis through studies of the liver kinase B1 tumour suppressor. J Carcinog. 2013;12:16. doi: 10.4103/1477-3163.116323. [PMC free article] [PubMed] [Cross Ref]
57. Kordaß T, Weber CE, Oswald M, Ast V, Bernhardt M, Novak D, Utikal J, Eichmüller SB König R. SOX5 is involved in balanced MITF regulation in human melanoma cells. BMC Med Genomics. 2016;9:10. doi: 10.1186/s12920-016-0170-0. [PMC free article] [PubMed] [Cross Ref]
58. Haq R, Shoag J, Andreu-Perez P, Yokoyama S, Edelman H, Rowe GC, Frederick DT, Hurley AD, Nellore A, Kung AL, Wargo JA, Song JS, Fisher DE, et al. Oncogenic BRAF regulates oxidative metabolism via PGC1α and MITF. Cancer Cell. 2013;23:302–15. doi: 10.1016/j.ccr.2013.02.003. [PMC free article] [PubMed] [Cross Ref]
59. Shannan B, Perego M, Somasundaram R, Herlyn M. Heterogeneity in Melanoma. Cancer Treat Res. 2016;167:1–15. doi: 10.1007/978-3-319-22539-5_1. [PubMed] [Cross Ref]
60. Kunjithapatham R, Karthikeyan S, Geschwind JF, Kieserman E, Lin M, Fu DX, Ganapathy-Kanniappan S. Reversal of anchorage-independent multicellular spheroid into a monolayer mimics a metastatic model. Sci Rep. 2014;4:6816. doi: 10.1038/srep06816. [PMC free article] [PubMed] [Cross Ref]
61. Pastò A, Bellio C, Pilotto G, Ciminale V, Silic-Benussi M, Guzzo G, Rasola A, Frasson C, Nardo G, Zulato E, Nicoletto MO, Manicone M, Indraccolo S, et al. Cancer stem cells from epithelial ovarian cancer patients privilege oxidative phosphorylation, and resist glucose deprivation. Oncotarget. 2014;5:4305–19. doi: 10.18632/oncotarget.2010. [PMC free article] [PubMed] [Cross Ref]
62. Schöckel L, Glasauer A, Basit F, Bitschar K, Truong H, Erdmann G, Algire C, Hägebarth A, Willems PH, Kopitz C, Koopman WJ, Héroult M. Targeting mitochondrial complex I using BAY 87-2243 reduces melanoma tumor growth. Cancer Metab. 2015;3:11. doi: 10.1186/s40170-015-0138-0. [PMC free article] [PubMed] [Cross Ref]

Articles from Oncotarget are provided here courtesy of Impact Journals, LLC