Significance This paper describes the mechanism by which copper mediates the interplay between the two energy-producing pathways, respiration and glycolysis. Many tumors produce increased levels of lactate, even when oxygen abounds, reflecting aerobic glycolysis (“Warburg effect”), whereas most normal tissues solely use respiration. We demonstrate that reducing systemic copper with a chelating drug impaired mitochondrial energy metabolism and decreased ATP levels despite induction of glycolysis. We propose that the metabolic phenotype of tumors is modulated in part by variable levels of copper in tumor microenvironment. Our work identifies copper as a tumor promoter by demonstrating that chronic exposure to elevated levels of copper in drinking water—to the maximum allowed in public water supplies—accelerates tumor growth in mice.

Abstract Copper is an essential trace element, the imbalances of which are associated with various pathological conditions, including cancer, albeit via largely undefined molecular and cellular mechanisms. Here we provide evidence that levels of bioavailable copper modulate tumor growth. Chronic exposure to elevated levels of copper in drinking water, corresponding to the maximum allowed in public water supplies, stimulated proliferation of cancer cells and de novo pancreatic tumor growth in mice. Conversely, reducing systemic copper levels with a chelating drug, clinically used to treat copper disorders, impaired both. Under such copper limitation, tumors displayed decreased activity of the copper-binding mitochondrial enzyme cytochrome c oxidase and reduced ATP levels, despite enhanced glycolysis, which was not accompanied by increased invasiveness of tumors. The antiproliferative effect of copper chelation was enhanced when combined with inhibitors of glycolysis. Interestingly, larger tumors contained less copper than smaller tumors and exhibited comparatively lower activity of cytochrome c oxidase and increased glucose uptake. These results establish copper as a tumor promoter and reveal that varying levels of copper serves to regulate oxidative phosphorylation in rapidly proliferating cancer cells inside solid tumors. Thus, activation of glycolysis in tumors may in part reflect insufficient copper bioavailability in the tumor microenvironment.

Copper is an essential trace element that is necessary for the activity of a number of metalloenzymes (1). Remarkably, the homeostatic balance of bioavailable copper is metastable and of evident importance, in that a number of tissue abnormalities and disease states in humans have been associated with either reduced or elevated levels of copper (2). Serum copper levels are elevated in cancer patients and correlate with the severity of the disease and response to therapies (3, 4). In animal models, copper-chelating drugs have been reported to have antiangiogenic activity (5⇓⇓–8). The molecular and cellular mechanisms by which copper levels modulate tumor angiogenesis and the generality of its effects on angiogenesis in different cancer types remain unclear, as are its potentially broader effects on tumor growth.

Using a genetically engineered mouse model of human cervical carcinoma, we previously observed and reported that cancer cells express higher levels of the copper transporter Ctr1 and that the tumors were differentially sensitive to reduction in systemic copper levels compared with normal tissues (9), leading us to hypothesize that cancer cells might have a greater demand for and dependence upon copper. To address this hypothesis, we turned to another genetically engineered mouse model of multistage tumorigenesis, the RIP1–Tag2 transgenic mouse line that expresses the simian virus 40 oncogenes under control of a rat insulin gene promoter (10). In this pancreatic neuroendocrine tumor model, islets undergo discrete steps of tumorigenesis, from hyperplasia, induction of angiogenesis, tumor growth, to invasive carcinoma, enabling us to investigate the effects of modulating bioavailable copper at distinct stages of tumor progression.

Here we provide evidence that differential copper intake levels modulate the activity of the copper-binding enzyme cytochrome c oxidase in cancer cells. Pharmacological suppression of systemic copper impairs oxidative phosphorylation and tumor growth, without attendant effects on ongoing tumor angiogenesis beyond a delay in initial angiogenic switching that may be consequent to impaired proliferation of cancer cells in incipient neoplasias. The results reveal that copper accessibility, much like that of oxygen and glucose, can modulate tumor phenotypes, suggesting that bioavailable copper may be amenable to therapeutic targeting of its heretofore-unappreciated role in tumor metabolism.

Discussion Our work provides direct evidence that varying levels of copper can modulate the proliferation of cancer cells and associated tumor growth, indicating that copper can be a rate-limiting nutrient for tumors, much like oxygen and glucose. We do not think, however, that copper is a carcinogen: exposure of wild-type mice to 20 μM copper in drinking water for up to 2 y did not result in increased cancer incidence. We demonstrate herein that the activity of cytochrome c oxidase, a key enzyme in oxidative phosphorylation, in tumors is affected by copper levels. Additional bioavailable copper evidently facilitates increased production of ATP, which is consumed to fuel rapid proliferation of cancer cells. Thus, copper may not initiate transformation, but may stimulate proliferation of transformed cells by providing energy needed for cell-cycle progression. In most eukaryotes, oxidative phosphorylation is the predominant source of ATP; glycolysis is used in an adaptive response to oxygen limitation. Tumors, however, produce more lactate than normal tissues, even in the presence of ample oxygen, indicative of increased ATP generation from glycolysis (30). A number of oncogenic pathways have been reported to contribute to such metabolic changes (31). Our findings suggest that an environmental factor may also play a critical role in shifting the balance between the two bioenergetic pathways: enhanced glycolysis of tumors may in part reflect insufficient copper bioavailability in the tumor microenvironment. Varying copper levels may have other effects on tumor growth, such as consequent changes in systemic iron distribution or altered function of other copper-dependent factors and pathways (1). Nevertheless, our demonstration that oxidative phosphorylation of tumors can be manipulated by copper levels has important implications for understanding tumor metabolism and for therapeutic strategies. Our studies indicate that the metabolic changes induced by copper limitation involve a mechanism distinct from the hypoxia-driven glycolytic shift. In hypoxia, HIF-1α is stabilized, resulting in transcriptional activation of a large set of genes, including those involved in oxygen delivery, glycolysis, invasion, and metastasis (28). In contrast, our data indicate that copper limitation does not activate HIF-1α or promote invasion. Notably, however, AMPK was activated by copper chelation. Activation of AMPK can also induce increased glycolysis, independent of the HIF-1α regulatory pathway (21), much as we observed in TM-treated tumors. Thus, copper chelation therapy may up-regulate glycolysis in an attempt to restore ATP in tumors but without necessarily involving activation of HIF-1α, which would likely have made cancers more invasive. Oxygen consumption is not quantitatively diminished in many tumors despite increased glycolysis, indicative of concurrent oxidative phosphorylation (32). The ATP contribution from oxidative phosphorylation accounts for 70–90% of total ATP produced in the majority of cancer cells, with glycolysis providing the remainder (33). Rapidly proliferating cancer cells exhibit increased sensitivity to respiratory inhibitors (34), and the tumor-forming capacity of transformed cells requires robust oxidative phosphorylation (35, 36). Our findings further support a continuing role for oxidative phosphorylation, despite the induction of aerobic glycolysis in tumor growth, and suggest that this process could be an important target in cancer therapy. Oxidative phosphorylation is essential for all cells in our body, and, as such, systemic inhibition of this pathway would be highly toxic unless tumors could be selectively targeted. Metformin, which is prescribed to type 2 diabetes patients, exerts its effects in part by inhibiting complex I of the electron transport chain (37, 38), and a number of retrospective clinical studies have found that metformin may offer a benefit in cancer prevention (39, 40). However, inhibition of complex I has been reported to induce ROS (41, 42), and cancer cells treated with metformin or its analog phenformin manifest elevated levels of ROS (43), which could potentially make them more invasive by activating HIF-1α. Copper chelators may offer several advantages over biguanides such as metformin and phenformin. First, copper chelators target complex IV, the rate-limiting enzyme of the mitochondrial respiratory chain and, as such, a key regulator of oxidative phosphorylation and ATP production (44). Second, in our study copper chelation did not increase ROS in tumors or invasiveness, consistent with the notion that complex IV is not a major site for ROS release (29). Third, cancers appear to have an increased need for copper compared with normal tissues (45, 46), and we have observed increased levels of the copper transporter Ctr1 protein in mouse cervical carcinoma (9) and increased Ctr1 mRNA levels in islets undergoing tumorigenesis (Fig. S4). Although TM exhibits cytostatic antitumor activity when used as a monotherapy both in the mouse model of pancreatic neuroendocrine tumor (this work) and in a mouse model of human papillomavirus 16-driven cervical carcinoma (9), we envision that this class of drug will have a greater therapeutic impact when used in rational combinations that warrant future investigation first in preclinical cancer models and then in focused, proof-of-concept clinical trials. One approach would involve combinations of TM or another copper chelator with platinum-based chemotherapy, wherein TM is predicted to have dual effects: (i) up-regulating the Ctr1 copper transporter in tumors in response to insufficient copper supply so as to increase the Ctr1-mediated import of the cytotoxic platinum drugs (9, 47), and (ii) directly impairing oxidative phosphorylation, ATP production, and consequent cancer cell proliferation and tumor growth (this work). An alternative therapeutic strategy, encouraged by the results of the cell-based combinatorial assays presented in this work, would involve the combination of a copper chelator with a glycolysis inhibitor, thereby simultaneously blocking copper-dependent respiration and glycolysis, the two major pathways for ATP production in highly ATP-demanding, hyperproliferative cancer cells.

Materials and Methods Mice, Tumor Analysis, and Cell Culture. RIP1–Tag2 mice have been described (10, 11). For TM treatment, 10 mg of ammonium tetrathiomolybdate (Sigma) was dissolved in 1 mL of H 2 O, and 100 μL of the solution was orally given to each mouse every 24 h with a gavage needle. For copper water treatment, 1 mL of 8 mM CuSO 4 was added to water bottles containing 400 mL of drinking water (final Cu concentration ∼20 μM; Cu concentration in drinking water without additional copper was 0.15–0.3 μM). Water was prepared fresh every week. βTC3 cells were derived from RIP1–Tag2 tumors (14). A2780, SKOV3, BxPC3, SUIT2, and MDAMB157 were from ATCC. Methods for tumor analysis and cell culture conditions are described in SI Materials and Methods. Cytochrome c Oxidase Activity, Metabolites, Mitochondrial Membrane Potential, and Lipid Peroxidation. Details of enzyme activity measurements and metabolites are described in SI Materials and Methods. Western Blotting and Quantitative PCR. Protocols and antibodies used for Western blotting and primers used for quantitative PCR are described in SI Materials and Methods. PET. See SI Materials and Methods for PET procedures. Tissue Copper Measurement. Tissues were weighed and digested in 10 μL of nitric acid per mg of tumor tissue for 2 h at 75 °C. Twenty-five microliters of digest and 25 μL of analytical grade H 2 O were mixed with 450 μL of 100 parts per billion cobalt, which served as an internal control for inductively coupled plasma mass spectrometry (ICP-MS) analysis. Fifty parts per billion copper was used as a standard.

Acknowledgments We thank R. Franks (University of California, Santa Cruz) for technical support with ICP-MS; F. Schaufele (University of California, San Francisco) for assistance with microscopy and image analysis; and F. McCormick (University of California, San Francisco) and members of our laboratories for comments and feedback. This study was supported by a Flight Attendant Medical Research Institute Clinical Innovator Award (D.H. and S.I.); grants from the Swiss Cancer League (D.H. and S.I.), National Cancer Institute (D.H.), the Swiss Federal Institute of Technology Lausanne (D.H. and J.A.), the European Research Commission Advanced Grant (J.A.), and the Swiss National Science Foundation (J.A.); and an award from the William K. Bowes, Jr. Charitable Foundation (D.H.). The work was supported in part by Biomedical Imaging Research Center of the Swiss Federal Institute of Technology Lausanne, University of Lausanne, University of Geneva, University Hospital of Geneva, Central University Hospital of Vaudois, and the Leenaards and Jeantet Foundations (C.P.-Y.). J.A. is the Nestlé Chair in Energy Metabolism at the Swiss Federal Institute of Technology Lausanne.

Footnotes Author contributions: S.I., J.A., and D.H. designed research; S.I. and C.P.-Y. performed research; P.A. and C.P.-Y. contributed new reagents/analytic tools; S.I., P.A., J.A., and D.H. analyzed data; S.I., J.A., and D.H. wrote the paper.

The authors declare no conflict of interest.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318431110/-/DCSupplemental.

Freely available online through the PNAS open access option.