SOX2 as a novel contributor of oxidative metabolism in melanoma cells
Deregulated metabolism is a hallmark of cancer and recent evidence underlines that targeting tumor energetics may improve therapy response and patient outcome. Despite the general attitude of cancer cells to exploit the glycolytic pathway even in the presence of oxygen (aerobic glycolysis or “Warburg effect”), tumor metabolism is extremely plastic, and such ability to switch from glycolysis to oxidative phosphorylation (OxPhos) allows cancer cells to survive under hostile microenvironments. Recently, OxPhos has been related with malignant progression, chemo-resistance and metastasis. OxPhos is induced under extracellular acidosis, a well-known characteristic of most solid tumors, included melanoma.
To evaluate whether SOX2 modulation is correlated with metabolic changes under standard or acidic conditions, SOX2 was silenced and overexpressed in several melanoma cell lines. To demonstrate that SOX2 directly represses HIF1A expression we used chromatin immunoprecipitation (ChIP) and luciferase assay.
In A375-M6 melanoma cells, extracellular acidosis increases SOX2 expression, that sustains the oxidative cancer metabolism exploited under acidic conditions. By studying non-acidic SSM2c and 501-Mel melanoma cells (high- and very low-SOX2 expressing cells, respectively), we confirmed the metabolic role of SOX2, attributing SOX2-driven OxPhos reprogramming to HIF1α pathway disruption.
SOX2 contributes to the acquisition of an aggressive oxidative tumor phenotype, endowed with enhanced drug resistance and metastatic ability.
KeywordsMelanoma Tumor extracellular acidosis SOX2 HIF1α Oxidative metabolism
ATP Synthase F1 Subunit Alpha
cytochrome c oxidase subunit 4 isoform 1
cytochrome c oxidase subunit 5B
hypoxia-inducible factor α
hexokinase isoform 2
lactate dehydrogenase A
pyruvate dehydrogenase kinase 1
pyruvate dehydrogenase phosphatase 2
peroxisome proliferator-activated receptor gamma coactivator 1-α
sex-determining region Y (SRY)-Box2
tricarboxylic acid cycle
transcription starting site
In the last decades, tumor metabolism has drawn increasing attention in the scientific world and deregulating cellular energetics has recently become a hallmark of cancer . Instead of using an oxidative metabolism like most of normal cells, cancer cells convert glucose into lactate even in the presence of high oxygen tension, exploiting the so-called aerobic glycolysis or “Warburg effect”. Despite the energetic gain in terms of ATP production is lower than during the oxidative phosphorylation (OxPhos), the Warburg metabolism is about 100-fold faster than OxPhos and ensures biomass formation and DNA duplication, that is crucial for cancer cell proliferation . Indeed, fermentation to lactic acid and the glycolytic breakdown of glucose generate a number of substrates which turn into “anabolic” precursors for the synthesis of different compounds, such as glucose-6-phosphate for glycogen and ribose 5-phosphate, dihydroxyacetone phosphate for triacylglyceride and phospholipids, and pyruvate for alanine and malate. Metabolite accumulation upstream pyruvate production is further increased by the up-regulation of the low activity M2 isoform of pyruvate kinase (PKM2), that slows down the last step of glycolysis. In this respect, intermediate components of the glycolytic pathway appear to be more significant than its final product pyruvate. Given the limited pyruvate supply, to replenish the tricarboxylic acid cycle (TCA) cancer cells increase glutamine consumption, a key nutrient that provides carbon for acetyl-CoA, citrate production and lipogenesis, nitrogen for purine, pyrimidine and DNA synthesis, and reducing power in the form of NADPH to support cell proliferation . The particular attitude of proliferating cancer cells to use aerobic glycolysis favors a microenvironment enriched in lactate and protons, with a subsequent pH reduction. Moreover, the large amount of lactate released by tumor cells can be taken up by normal stromal cells to regenerate pyruvate, which in turn can be extruded to refuel cancer cells . The reduction in oxygen tension that characterizes proliferating tumor tissues, stimulates the hypoxia-inducible factor α (HIF1α), which drives the anaerobic glycolysis. This leads to lactate dehydrogenase A (LDH-A)-dependent lactic acid production, and the upregulation of monocarboxylated transporter (MCT)-4 and of sodium-proton exporters to avoid intracellular acidosis. As a direct consequence, both aerobic and anaerobic glycolysis adopted by cancer cells contribute to the acidification of tumor microenvironment. Dysregulated pH is emerging as a hallmark of cancer, since cancer cells show a ‘reversed’ pH gradient with a constitutively increased intracellular pH (pHi) that is higher than the extracellular pH (pHe). Indeed, while normal differentiated adult cells show pHi of ∼7.2 and pHe of ∼7.4, cancer cells have a higher pHi (> 7.4) and a lower pHe (6.7–7.1). This ‘reversed’ pH gradient creates a perfect storm for metastatic progression  by promoting malignant phenotype endowed with apoptosis resistance, radio- and chemotherapy resistance, immune surveillance escape programs, increased migration and ability of secondary organs colonization . As an additional aspect, we have recently reported that acidic cancer cells undergo a metabolic change characterized by the acquisition of a more OxPhos phenotype through the inhibition of HIF1α expression, associated with a reduced proliferation compared to standard pH condition .
Tumor cells are extremely plastic even in terms of cellular energetics and may shift their metabolic phenotypes to adapt to microenvironmental changes, giving a selective advantage to cancer cells under unfavourable environments . Most of solid tumors, including melanoma, undergo such plastic changes in metabolism. Cutaneous melanoma, despite representing less than 5% of all skin cancers, is responsible for the majority of skin cancer-related deaths . The incidence of malignant melanoma in most developed countries has risen faster than any other cancer type since the mid-1950s. It is estimated that the annual increase in the incidence rate of melanoma has been approximately 3–7% per year worldwide for Caucasians. Detection and surgical treatment of early-stage disease seems to prevent progression in most cases. However, patients with deep primary tumors or tumors that metastasize to regional lymph nodes frequently develop distant metastases. Median survival after the onset of distant metastases is only 6–9 months, and the 5-year survival rate is less than 5% .
Recent studies have pointed out the crucial role of the transcription factor SOX2 (sex-determining region Y (SRY)-Box2) in melanoma and cancer in general. SOX2 has been correlated with growth, tumorigenicity, drug resistance, and metastasis in at least 25 different tumors, including cancers of the ovary, lung, skin, brain, breast, prostate, and pancreas . In the majority of these cancers, SOX2 has been found to have increased expression or gene amplification in tumor tissues. Moreover, SOX2 has been associated with stemness and tumor initiating cells (TICs), proposed to explain origin and heterogeneity of many tumors , including cervical, lung, ovarian, head and neck squamous cell carcinoma, medulloblastoma, skin squamous-cell carcinoma, and melanoma . Indeed, SOX2 has been reported to regulate self-renewal and tumorigenicity of human melanoma-initiating cells [13, 14]. Previous reports indicate that SOX2 is expressed in 50% of melanomas and a minority of nevi [15, 16, 17], and is associated with dermal invasion and primary tumor thickness . However, the role of SOX2 in melanoma growth and progression is more controversial. While an early paper reported that SOX2 silencing reduces in vivo growth of A2058 melanoma cells , recent studies suggest that SOX2 is dispensable for melanomagenesis and metastasis formation [19, 20].
Here we show for the first time that SOX2 is highly expressed in melanoma cells exposed to extracellular acidosis, where it modulates cell metabolism in order to favor an oxidative phenotype, possibly interfering with HIF1α expression. This additional attitude of SOX2 might add new information on its crucial importance in malignant progression.
A375-M6 , commercial 501-Mel, SK-Mel-2, SK-Mel-5, SK-Mel-28 and patient-derived SSM2c  melanoma cell lines were maintained in DMEM 4.5 g/l glucose, 2 mM L-glutamine, and 10% FBS (Euroclone, Milan Italy). 24-h medium acidification was obtained by adding HCl 1 N in complete culture medium to reach pH 6.7 ± 0.1. pH value was monitored by using Orion pH meter 520A-1. pH was monitored for the first hour after medium acidification to check the maintenance of a pH value at 6.7, and then at the end point of each experiment. Cells were treated with 50 mM 2-deoxyglucose (Calbiochem, San Diego, CA, USA) or 10 mM Metformin (Sigma-Aldrich, Milan, Italy) for 24 h.
SOX2 silencing and overexpression
SOX2-silenced A375-M6 cells were obtained by siRNA transfection with Sox-2 siRNA (sc-38408, Santa Cruz Biotechnology, Dallas, Texas, USA) or control siRNA-A (sc-37007, Santa Cruz Biotechnology), according to manufacturer’s instructions. SOX2 silencing in SSM2c cells was obtained by lentiviral transduction. Lentiviruses were produced in HEK-293 T cells. Lentiviral vectors used were pLKO.1-puro (LV-c) (Open Biosystems, Lafayette, CO, USA) and pLKO.1-puro-shSOX2–1 (LV-shSOX2–1) targeting the 3′ untranslated region of SOX2 (targeting sequence 5’-CTGCCGAGAATCCATGTATAT-3′) as previously reported . SOX2 overexpression in 501-Mel cells was obtained by retroviral transduction. Retroviruses were produced in HEK-293 T cells. Retroviral vectors used were generated by co-transfection of 1 μg pBABE (Addgene, Cambridge, MA, USA, #1764) or pBABE-SOX2 (cloned into the BamHI/SalI restriction sites of pBABE vector using the following primers: SOX2-F 5’-ATGTACAACATGATGGAGACGG-3′ and SOX2-R 5’-TCACATGTGTGAGAGGGGC-3′), 0.9 μg pUMVC packaging plasmid (Addgene, #8449) and 0.1 μg pCMV-VSV-G envelope (Addgene, #8454).
Western blot analysis
Cells were lysed in RIPA buffer (Merck Millipore) containing PMSF (Sigma-Aldrich), sodium orthovanadate (Sigma-Aldrich), and protease inhibitor cocktail (Calbiochem), sonicated and centrifuged 15 min at 14,000 rpm at 4 °C. Equal amounts of protein were separated on Bolt® Bis-Tris Plus gels, 4–12% precast polyacrylamide gels (Life Technologies, Milan, Italy). Fractionated proteins were transferred to a PVDF membrane using the iBlot 2 System (Life Technologies). Following 1-h blocking with Odyssey blocking buffer (Dasit Science, Milan, Italy), membrane was probed overnight at 4 °C with the following primary antibodies: anti-SOX2 mouse monoclonal antibody (R&D System, Minneapolis, MN, USA), anti-HIF-1α rabbit polyclonal antibody (Novusbio, Milan, Italy), anti- GLUT-1, GLUT-3, MCT-1, MCT-4 and PGC1α rabbit polyclonal antibodies (Santa Cruz Biotechnology). After that, membrane was incubated 1 h at room temperature with goat anti-mouse IgG Alexa Fluor 680 antibody (Invitrogen) or goat anti-rabbit IgG Alexa Flour 750 antibody (Invitrogen- Life Technologies, Milan, Italy). Membrane was visualized by the Odyssey Infrared Imaging System (LI-COR® Bioscience, Lincoln, Nebraska USA). Anti-HSP90 (Santa Cruz Biotechnology), β-actin (Sigma-Aldrich) and HDAC2 (Santa Cruz Biotechnology) antibodies were used to assess equal amount of protein loaded in each lane.
Cells were harvested by using Accutase (Euroclone), collected in flow cytometer tubes (2 × 105 cells/tube), permeabilized for 15 min with 0.25% Tryton X-100 PBS, and incubated 1 h at 4 °C with anti-SOX2 antibody (Santa Cruz Biotechnology). Cells were washed in PBS and incubated 1 h in the dark at 4 °C with anti-goat antibody conjugated with FITC (Merk Millipore, Milan, Italy). Samples were washed in PBS and the analyzed at BD FACSCanto (BD Biosciences, Milan, Italy). The flow cytometer was calibrated using cells incubated with secondary antibody only. For each sample, 1 × 104 events were analysed.
Lactate production by cancer cells was evaluated in 24-h conditioned medium by using D-Lactate Colorimetric Assay Kit (Biovision, CA, USA) according to manufacturer’s instructions. The analysis was performed at the microplate reader (Bio-Rad, Milan, Italy) and data normalized for the cell number of each sample, to get a final result of lactate production (nM) by 1 × 105 cells.
Glucose uptake detection
Glucose uptake by melanoma cells was evaluated by using Glucose Uptake Cell-Based Assay Kit (Cayman Chemical, Michigan, USA) according to manufacturer’s instructions. Briefly, melanoma cells were glucose-starved for 1 h by using RPMI medium without glucose (Euroclone), then incubated for 15 min in the dark with 2-NBDG, a FITC-labeled deoxyglucose analog, harvested and analyzed at BD FACSCanto (BD Biosciences). The flow cytometer was calibrated using untreated cells. For each sample, 1 × 104 events were analyzed.
Quantitative real time PCR (qPCR)
List of forward and reverse primers used for qPCR analysis
Annexin V/PI flow cytometer analysis
Cell death was determined by flow cytometer analysis using Annexin V FITC-conjugated (Immunotools GmbH, Friesoythe, Germany) and PI (Sigma-Aldrich) according to the manufacturer’s protocol. Briefly, cells were harvested with Accutase (Eurolone), collected in flow cytometer tubes (1 × 105 cells/tube), washed in PBS, and incubated 15 min at 4 °C in the dark with 100 μl Annexin binding buffer (100 mM HEPES, 140 mM NaCl, 25 mM CaCl2, pH 7.4), 1 μl of 100 μg/ml PI working solution, and 5 μl Annexin V FITC-conjugated. Each sample was added with Annexin binding buffer to reach 500 μl volume/tube. Samples were analyzed at BD FACSCanto (BD Biosciences). Cellular distribution depending on Annexin V and/or PI positivity allowed the measure of the percentage of viable cells (Annexin V and PI negative cells), early apoptosis (Annexin V-positive and PI negative cells), late apoptosis (Annexin V and PI-positive cells), and necrosis (Annexin V-negative and PI-positive cells).
Melanoma cells were fixed with 1% formaldehyde for 10 min and lysed in cell Lysis Buffer (5 mM PIPES pH 8, 85 mM KCl, 0.5% NP-40) added with protease inhibitors. Nuclei were collected by centrifugation at 4500 rpm for 10 min and lysed in nuclear lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl pH 8) added with protease inhibitors. Chromatin was sonicated to an average size of 200–600 bp, diluted with ChIP Dilution Buffer (10 mM Tris-HCl pH 8, 1% Triton X-100, 2 mM EDTA, 140 mM NaCl) and incubated overnight with 20 μl of protein G magnetic dynabeads pre-conjugated with mouse anti-SOX2 (MAB2018; R&D System) or normal mouse IgG (sc-2025; Santa Cruz Biotechnology) antibodies. DNA was purified and qPCR was carried out at 60 °C using FastStart SYBR Green Master (Roche Diagnostic, Monza, Italy) in a Rotorgene-Q (Qiagen, Milan, Italy). Primer sequences are listed in Table 1.
Luciferase reporter assays
Luciferase reporters were used in combination with Renilla luciferase pRL-TK reporter vector (Promega, Madison, WI) to normalize luciferase activities. pCS2 + MT vector (Promega) was used to equal DNA amounts; pCS2 + SOX2 was cloned into pCS2 + MT using the following primers: Fwd 5’-ATGTACAACATGATGGAGACGG-3′, Rev. 5’-CACATGTGTGAGAGGGGC-3′ after digestion with XhoI/SnaBI restriction enzymes; pGL4.20-HIF1αprom was purchase from Addgene (Plasmid #40173). Luminescence was measured using the Dual-Glo Luciferase Assay System (Promega) and the GloMax® 20/20 Luminometer (Promega).
The experiments were performed at least three times for a reliable application of statistics. Statistical analysis was performed with GraphPad Prism software. Values are presented as mean ± SD. ANOVA or Student’s T test were used to evaluate the statistical significance.
Extracellular acidosis promotes SOX2 expression in melanoma cells contributing to OxPhos metabolism
Metabolic drugs sustain SOX2 contribution to OxPhos metabolism in acidic A375-M6 melanoma cells
Modulation of SOX2 expression and metabolic adaptation of SSM2c and 501-Mel melanoma cells under standard conditions
SOX2-driven metabolic adaptation to OxPhos is due to HIF1α pathway disruption
Cancer cells are characterized by a deregulated metabolism since, unlike normal cells, they largely depend on glycolysis even in the presence of oxygen, a phenomenon referred to as “Warburg effect” or aerobic glycolysis . Nevertheless, cancer metabolism does not exclusively depend on aerobic glycolysis. Indeed, tumor cells can rather shift between different metabolic phenotypes or be in a hybrid state utilizing both glycolytic and oxidative metabolism . This plasticity is also referred to cancer bioenergetics and contributes to positively select cancer cells in order to survive besides any environmental changes and hostile conditions . We and others have already reported that the acidic microenvironment, that characterizes most of solid tumors and is associated with aggressive tumor phenotypes , favors OxPhos at the expense of glycolysis [7, 28, 29, 30, 31, 32]. Here we correlate for the first time SOX2 expression in acidic melanoma cells with a more oxidative metabolism, that is in turn associated with tumor progression and poor prognosis. In this regard, OxPhos metabolism has recently regained its role in cancer progression, given its association with occurrence of chemo-resistance and development of metastasis . Moreover, despite the existence of controversial opinions , recent studies suggest that cancer stem cells are more reliant upon an oxidative metabolism than the non-stem bulk in different tumor types, including leukemia, ovarian, pancreatic, and breast cancer. OxPhos metabolism has been also shown to be privileged by circulating tumor cells compared to primary tumor cells of melanoma and breast cancers , and to be correlated with chemo-resistance in glioma , lung , pancreatic , prostate , and ovarian cancers . Furthermore, several cases of metabolic shift to OxPhos following targeted therapies have been reported . This is the case of melanomas carrying activating BRAF mutations, where BRAF inhibitors induce PGC1α, a master regulator of mitochondrial biogenesis, which in turn promotes oxidative metabolism .
By exploiting an in vitro model of extracellular acidosis, we demonstrated that SOX2 is induced by an acidic microenvironment and, importantly, that SOX2 depletion in acidic melanoma cells reprograms their metabolism to a more glycolytic phenotype, also reducing OxPhos-related genes that characterize acidosis-exposed melanoma cells. The reprogramming toward a more glycolytic profile is also evident in SOX2-silenced cells grown in standard condition. A tightly correlation between SOX2 and OxPhos emerges when SOX2-silenced melanoma cells, either grown in acidic or standard pH medium, are treated with 2-DG and Metformin. 2-DG targets glucose metabolism inducing a decrease of ATP generation, whereas Metformin blocks complex I of the respiratory chain. Interestingly, epidemiological and retrospective studies have revealed a lower incidence of cancer and better outcomes in diabetic patients taking Metformin compared to non-diabetics or diabetics using alternative drugs . We found that 2-DG promotes cell death in SOX2-silenced cells grown in standard pH conditions and, most importantly, also in acidic melanoma cells depleted of SOX2. On the other hand, Metformin was effective only in acidosis-exposed cancer cells, since its efficacy is significantly reduced in SOX2-silenced cells, even though a further cell death reduction could be expected. This could be probably due to the high levels of SOX2 in acidic melanoma cells associated with an only partial SOX2 silencing efficacy.
To better understand SOX2 contribution in OxPhos metabolism, we determined metabolic markers in SSM2c cells, characterized by high SOX2 levels, and in 501-Mel, which show low/no SOX2 expression, upon knock-down or ectopic expression of SOX2, respectively. We confirmed the ability of SOX2 to contribute to an oxidative metabolism. Indeed, SOX2 knock-down leads to the suppression of the master regulator of mitochondrial metabolism PGC1α, a phenomenon associated with the promotion of critical steps of the glycolytic pathway, i.e. GLUTs, HK2, PDP2/PDK1 axis and LDH-A. Furthermore, SOX2 silencing induces a switch of MCT genes from type 1 to type 4, indicating a preferred lactate efflux characteristic of a glycolytic metabolism. Consistently, in SOX2 overexpressing 501-Mel cells the reduction of most glycolytic markers came together with a promotion of MCT-1, a promoter of lactate influx. The clinical importance of MCT expression levels derived from their tightly correlation with shorter overall survival of advanced melanoma patients .
SOX2 has been already associated with tumor initiation, growth, drug resistance, and metastasis. Chemo-resistant cancer cells that appear to preferably exploit oxidative metabolism, have been also associated with enhanced SOX2 expression in gastric, lung, prostate, colorectal , and breast  cancers. These findings prompted us to verify whether SOX2 influence in cell metabolism might be related to HIF1α activity, considering that HIF1α strongly induces a glycolytic phenotype. Our results indicate that HIF1α and SOX2 are inversely correlated in normoxic condition, and this effect might be functionally sufficient to reprogram melanoma cells toward OxPhos. This is likely true mainly in conditions when SOX2 exceeds HIF1α in terms of protein expression, as in the case of acidosis-exposed melanoma cells. Instead, under hypoxia, HIF1α stabilization, despite the presence of SOX2 , likely represents the leading factor that causes cancer cell metabolic switch to anaerobic glycolysis. Among the so-called non-canonical HIF1α regulation , it is quite interesting to recall that an increased lactate production is able to promote HIF1α stabilization, although the mechanism has not been yet clarified . Thus, the lactate increase observed in SOX2-silenced acidic and non-acidic melanoma cells could be able to contribute to HIF1α expression and glycolytic re-conversion. Furthermore, quite recently it was demonstrated that HIF1α represses PGC1α expression in renal cell carcinoma, suggesting a regulatory loop among these transcriptional factors, involving oxygen sensing to mitochondrial biogenesis . This is in line with our findings, i.e. PGC1α reduction and HIF1α promotion upon SOX2 silencing.
In conclusion, with this study we would propose a thigh correlation between SOX2 expression and OxPhos metabolism in melanoma cells, under a condition of reduced HIF1α expression. Oxidative metabolism might be of a crucial importance for melanoma progression. Indeed, cancer cells may take advantage of this metabolic reprogramming toward OxPhos contributing to the development of an aggressive tumor phenotype endowed with an enhanced drug resistance and metastatic ability .
this study was financially supported by Istituto Toscano Tumori (ITT) (Decreto Dirigenziale Regione Toscana n. 5254 of 04/12/2013 to LC), Ente Cassa di Risparmio di Firenze, Associazione Italiana per la Ricerca sul Cancro (AIRC) (grant IG-14184 to BS; fellowship to EA*, SP* and AB) and Università degli Studi di Firenze.
Availability of data and materials
All data generated or analysed during this study are included in this published article and its supplementary information files.
LC, EA*, SP* and BS designed the experiments and write the manuscript. EA* and SP* performed the experiments helped by SP, AB, JR and FB. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 38.Dar S, Chhina J, Mert I, Chitale D, Buekers T, Kaur H, et al. Bioenergetic adaptations in Chemoresistant ovarian Cancer cells. Sci Rep. 2017;7(1):8760.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.