Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Letter
  • Published:

Metabolic determinants of cancer cell sensitivity to glucose limitation and biguanides

Abstract

As the concentrations of highly consumed nutrients, particularly glucose, are generally lower in tumours than in normal tissues1,2, cancer cells must adapt their metabolism to the tumour microenvironment. A better understanding of these adaptations might reveal cancer cell liabilities that can be exploited for therapeutic benefit. Here we developed a continuous-flow culture apparatus (Nutrostat) for maintaining proliferating cells in low-nutrient media for long periods of time, and used it to undertake competitive proliferation assays on a pooled collection of barcoded cancer cell lines cultured in low-glucose conditions. Sensitivity to low glucose varies amongst cell lines, and an RNA interference (RNAi) screen pinpointed mitochondrial oxidative phosphorylation (OXPHOS) as the major pathway required for optimal proliferation in low glucose. We found that cell lines most sensitive to low glucose are defective in the OXPHOS upregulation that is normally caused by glucose limitation as a result of either mitochondrial DNA (mtDNA) mutations in complex I genes or impaired glucose utilization. These defects predict sensitivity to biguanides, antidiabetic drugs that inhibit OXPHOS3,4, when cancer cells are grown in low glucose or as tumour xenografts. Notably, the biguanide sensitivity of cancer cells with mtDNA mutations was reversed by ectopic expression of yeast NDI1, a ubiquinone oxidoreductase that allows bypass of complex I function5. Thus, we conclude that mtDNA mutations and impaired glucose utilization are potential biomarkers for identifying tumours with increased sensitivity to OXPHOS inhibitors.

This is a preview of subscription content, access via your institution

Access options

Rent or buy this article

Prices vary by article type

from$1.95

to$39.95

Prices may be subject to local taxes which are calculated during checkout

Figure 1: Nutrostat design and metabolic characterization of cancer cells under chronic glucose limitation.
Figure 2: Barcode-based cell competition assay and RNAi screen in Nutrostats.
Figure 3: Deficiencies in glucose utilization or complex I underlie low-glucose sensitivity of cancer cells.
Figure 4: Cancer cells with deficiencies in glucose utilization or complex I are sensitive to phenformin.

Similar content being viewed by others

References

  1. Hirayama, A. et al. Quantitative metabolome profiling of colon and stomach cancer microenvironment by capillary electrophoresis time-of-flight mass spectrometry. Cancer Res. 69, 4918–4925 (2009)

    Article  CAS  PubMed  Google Scholar 

  2. Gullino, P. M., Grantham, F. H. & Courtney, A. H. Glucose consumption by transplanted tumors in vivo. Cancer Res. 27, 1031–1040 (1967)

    CAS  PubMed  Google Scholar 

  3. Owen, M. R., Doran, E. & Halestrap, A. P. Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochem. J. 348, 607–614 (2000)

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. El-Mir, M. Y. et al. Dimethylbiguanide inhibits cell respiration via an indirect effect targeted on the respiratory chain complex I. J. Biol. Chem. 275, 223–228 (2000)

    Article  CAS  PubMed  Google Scholar 

  5. Seo, B. B., Matsuno-Yagi, A. & Yagi, T. Modulation of oxidative phosphorylation of human kidney 293 cells by transfection with the internal rotenone-insensitive NADH-quinone oxidoreductase (NDI1) gene of Saccharomyces cerevisiae. Biochim. Biophys. Acta 1412, 56–65 (1999)

    Article  CAS  PubMed  Google Scholar 

  6. Cairns, R. A., Harris, I. S. & Mak, T. W. Regulation of cancer cell metabolism. Nature Rev. Cancer 11, 85–95 (2011)

    Article  CAS  Google Scholar 

  7. Urasaki, Y., Heath, L. & Xu, C. W. Coupling of glucose deprivation with impaired histone H2B monoubiquitination in tumors. PLoS ONE 7, e36775 (2012)

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  8. Luo, B. et al. Highly parallel identification of essential genes in cancer cells. Proc. Natl Acad. Sci. USA 105, 20380–20385 (2008)

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  9. Tasseva, G. et al. Phosphatidylethanolamine deficiency in Mammalian mitochondria impairs oxidative phosphorylation and alters mitochondrial morphology. J. Biol. Chem. 288, 4158–4173 (2013)

    Article  CAS  PubMed  Google Scholar 

  10. Haack, T. B. et al. Exome sequencing identifies ACAD9 mutations as a cause of complex I deficiency. Nature Genet. 42, 1131–1134 (2010)

    Article  CAS  PubMed  Google Scholar 

  11. Crabtree, H. G. Observations on the carbohydrate metabolism of tumours. Biochem. J. 23, 536–545 (1929)

    CAS  PubMed Central  PubMed  Google Scholar 

  12. Barretina, J. et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 483, 603–607 (2012)

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  13. Jones, J. B. et al. Detection of mitochondrial DNA mutations in pancreatic cancer offers a “mass”-ive advantage over detection of nuclear DNA mutations. Cancer Res. 61, 1299–1304 (2001)

    CAS  PubMed  Google Scholar 

  14. Santidrian, A. F. et al. Mitochondrial complex I activity and NAD+/NADH balance regulate breast cancer progression. J. Clin. Invest. 123, 1068–1081 (2013)

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Shackelford, D. B. et al. LKB1 inactivation dictates therapeutic response of non-small cell lung cancer to the metabolism drug phenformin. Cancer Cell 23, 143–158 (2013)

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Larman, T. C. et al. Spectrum of somatic mitochondrial mutations in five cancers. Proc. Natl Acad. Sci. USA 109, 14087–14091 (2012)

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  17. Evans, J. M., Donnelly, L. A., Emslie-Smith, A. M., Alessi, D. R. & Morris, A. D. Metformin and reduced risk of cancer in diabetic patients. Br. Med. J. 330, 1304–1305 (2005)

    Article  Google Scholar 

  18. DeCensi, A. et al. Metformin and cancer risk in diabetic patients: a systematic review and meta-analysis. Cancer Prev. Res. (Phila.) 3, 1451–1461 (2010)

    Article  CAS  Google Scholar 

  19. Birsoy, K., Sabatini, D. M. & Possemato, R. Untuning the tumor metabolic machine: targeting cancer metabolism: a bedside lesson. Nature Med. 18, 1022–1023 (2012)

    Article  CAS  PubMed  Google Scholar 

  20. Pollak, M. Metformin and pancreatic cancer: a clue requiring investigation. Clin. Cancer Res. 18, 2723–2725 (2012)

    Article  CAS  PubMed  Google Scholar 

  21. Hall, A. et al. Dysfunctional oxidative phosphorylation makes malignant melanoma cells addicted to glycolysis driven by the (V600E)BRAF oncogene. Oncotarget 4, 584–599 (2013)

    Article  PubMed  PubMed Central  Google Scholar 

  22. Ben Sahra, I., Tanti, J. F. & Bost, F. The combination of metformin and 2 deoxyglucose inhibits autophagy and induces AMPK-dependent apoptosis in prostate cancer cells. Autophagy 6, 670–671 (2010)

    Article  PubMed  Google Scholar 

  23. Javeshghani, S. et al. Carbon source and myc expression influence the antiproliferative actions of metformin. Cancer Res. 72, 6257–6267 (2012)

    Article  CAS  PubMed  Google Scholar 

  24. Jain, M. et al. Metabolite profiling identifies a key role for glycine in rapid cancer cell proliferation. Science 336, 1040–1044 (2012)

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  25. Possemato, R. et al. Functional genomics reveal that the serine synthesis pathway is essential in breast cancer. Nature 476, 346–350 (2011)

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

  26. Birsoy, K. et al. MCT1-mediated transport of a toxic molecule is an effective strategy for targeting glycolytic tumors. Nature Genet. 45, 104–108 (2013)

    Article  CAS  PubMed  Google Scholar 

  27. Clerc, P. & Polster, B. M. Investigation of mitochondrial dysfunction by sequential microplate-based respiration measurements from intact and permeabilized neurons. PLoS ONE 7, e34465 (2012)

    Article  ADS  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank G. Stephanopoulos for assistance with Nutrostat design, M. Holland for mtDNA sequencing consultation, L. Garraway for assistance identifying mtDNA mutant cell lines, T. Yagi for the NDI1 antibody, T. DiCesare for diagrams, and members of the Sabatini Laboratory for assistance (particularly A. Saucedo, C. Koch, O. Yilmaz, Y. Gultekin and A. Hutchins for technical assistance, and D. Lamming and W. Comb for critical reading of the manuscript). This research is supported by fellowships from The Leukemia and Lymphoma Society and The Jane Coffin Childs Fund to K.B., the Council of Higher Education Turkey and Karadeniz T. University Scholarships to B.Y. and grants from the David H. Koch Institute for Integrative Cancer Research at MIT, The Alexander and Margaret Stewart Trust Fund, the NIH (K99 CA168940 to R.P. and CA103866, CA129105, and AI07389 to D.M.S.) and the Starr Cancer Consortium. D.M.S. is an investigator of the Howard Hughes Medical Institute.

Author information

Authors and Affiliations

Authors

Contributions

K.B., R.P. and D.M.S. conceived the project and designed the experiments. K.B. and R.P. designed and engineered the Nutrostat and performed the screening, knockdown, cell proliferation, extracellular flux, glucose consumption, and tumour formation experiments and processed and analysed sequencing and expression data. F.K.L., E.C.B., B.Y., and W.W.C. assisted with experiments. C.B.C. performed the metabolite profiling experiments. T.W. provided bioinformatic support for shRNA abundance deconvolution, P.T. assisted in identifying mtDNA mutations. K.B., R.L.P and D.M.S. wrote and all authors edited the manuscript.

Corresponding author

Correspondence to David M. Sabatini.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Extended data figures and tables

Extended Data Figure 1 Model of the metabolic determinants of sensitivity to low glucose and biguanides.

This diagram outlines the interplay between reserve oxidative phosphorylation (OXPHOS) capacity, sensitivity to biguanides, and sensitivity to culture in low glucose. Most cancer cell lines and normal cells tested exhibited an ability to respond to glucose limitation by upregulating OXPHOS, rendering them less sensitive to biguanides and low-glucose conditions. In contrast, cell lines harbouring mutations in mtDNA-encoded complex I subunits or exhibiting impaired glucose utilization have a limited reserve OXPHOS capacity and are therefore unable to properly respond to biguanides and low glucose, rendering them sensitive to these perturbations. At the extreme, cells artificially engineered to have no OXPHOS (Rho cells) exhibit extreme low glucose sensitivity, but resistance to further inhibition of OXPHOS. Thus, mtDNA mutant cancer cells exist at an intermediate state of OXPHOS functionality that renders them sensitive to treatment with biguanides in vitro and in vivo. Similarly, cell lines with impaired glucose utilization exhibit biguanide sensitivity specifically under the low glucose conditions seen in the tumour microenvironment.

Extended Data Figure 2 Proliferation and media glucose levels in standard culture conditions.

a, Jurkat cell proliferation under 10 mM (black) versus 1 mM (blue) glucose in standard culture conditions. b, Media glucose concentrations over time from cultures in a. Error bars are s.e.m., n = 3. Replicates are biological, means reported. *P < 0.05 by two-sided student’s t-test.

Extended Data Figure 3 Additional data supporting the RNAi screen.

a, Genes scoring as preferentially required for growth in 10 mM glucose compared to 0.75 mM glucose (top). shRNAs scores (given in per cent) and pathway classifications are indicated. Immunoblot analyses depict suppression of PKM by shRNAs (PKM1, PKM2) compared to control (RFP). Proliferation of cells in 0.75 mM (blue) relative to 10 mM glucose (black) harbouring shRNAs targeting PKM or control is also shown (PKM1). *P < 0.05 relative to RFP 0.75 mM glucose. b, Nuclear-encoded core complex I genes are shown in the grey box, with those that scored as differentially required under 0.75 mM glucose in red. The dot plot reports the degree of differential requirement for growth in 0.75 mM glucose (as the log2 difference in shRNA abundance in 0.75 mM glucose versus 10 mM glucose) of individual shRNAs targeting non-core complex I genes, core complex I genes, or non-targeting controls. Red bar, population median. c, Top, mRNA levels of the non-scoring OXPHOS genes (black) and the scoring OXPHOS gene (blue) indicated upon suppression, with the shRNAs indicated as measured by qPCR, relative to a non-targeting shRNA (RFP). Bottom, cell number from 7-day proliferation assay of cells in 0.75 mM glucose relative to 10 mM glucose (not shown) harbouring the indicated shRNAs. shRFP control normalized to 1. Error bars are s.e.m., n = 3. Replicates are biological, means reported. *P < 0.05 by two-sided student’s t-test.

Extended Data Figure 4 Additional data characterizing mitochondrial dysfunction and impaired glucose utilization in cancer cell lines.

a, Ratio of oxygen consumption rate (OCR) to extracellular acidification rate (ECAR) (left) or OCR normalized to protein content (right) for glucose-limitation-resistant (black) or glucose-limitation-sensitive (blue) cell lines. b, Left, mtDNA content for indicated cell lines by qPCR using primers targeting ND1 (black) or ND2 (grey) normalized to gDNA repetitive element (Alu) relative to KMS-12BM. Right, mitochondrial mass measured by fluorescence intensity of Mitotracker Green dye for indicated cell lines. c, Per cent change from baseline (second measurement) of ECAR or OCR in Jurkat cells in conditions in which glucose concentration was maintained at 0.75 mM (blue) or increased to indicated concentrations (black). d, Uptake of 3H-labelled 2-DG (counts per min per ng protein) in 0.75 mM glucose at indicated time points in GLUT3-high (grey) or GLUT3-low (blue) cell lines. e, Heatmap of gene expression values for genes indicated at top and cell lines (left). Genes organized by P value with lowest expressed genes in NCI-H929 and KMS-26 to the left, those with significantly lower expression are in red. Expression values reported as log2-transformed fold difference from the median (scale colour bar shown to the right). f, Immunoblot analyses for GLUT3 and NDI1 expression in indicated cell lines (β-actin loading control). g, i, Proliferation of cell number in cells overexpressing GLUT3 or NDI1 relative to control vector (4 days). h, OCR of permeabilized cells indicated upon addition of indicated metabolic toxins and substrates. j, Fold change in OCR in indicated cells expressing NDI1 relative to control vector. k, l, Proliferation for 4 days of control (Vector) or NDI1 expressing cell lines indicated (NDI1) under 10 mM (black) and 0.75 mM glucose (blue). Error bars are s.e.m. n = 4 for ac, h, j; n = 3 for d, g, i, k, l. Replicates are biological, means reported. *P < 0.05 by two-sided student’s t-test.

Extended Data Figure 5 Gene expression signature for identifying cell lines with impaired glucose utilization.

Heatmap of gene expression values for the genes indicated on the right for the cell lines in the CCLE set. Gene expression values are reported as the difference from the median across the entire sample set according to the scale colour bar to the top right. Genes 1–8 comprised the gene expression signature used to identify samples with impaired glucose utilization. Samples are sorted based upon this signature, with those predicted to exhibit impaired glucose utilization at the top. The order of samples and all values are reported in Supplementary Table 4.

Extended Data Figure 6 GLUT3 overexpression increases tumour xenograft growth and cell proliferation in low-glucose media.

a, KMS-26 cell lines infected with GLUT3-overexpressing vector or infected with control vector were mixed in equal proportions and cultured under different glucose concentrations. In addition, these mixed cell lines were injected into NOD/SCID mice subcutaneously. After 2.5 weeks, genomic DNA was isolated from tumours as well as cells grown in vitro under the indicated glucose concentrations. Using qPCR, relative abundance of control vector and GLUT3 vector were determined and plotted relative to 10 mM glucose in culture (n = 9). b, Average volume of unmixed tumour xenografts from KMS-26 cell lines infected with GLUT3-overexpressing vector relative to control vector (2.5 weeks) (n = 6). Replicates are biological, means reported. *P < 0.05 by two-sided student’s t-test.

Extended Data Figure 7 Sanger sequencing traces validating mtDNA mutations.

The table summarizing mtDNA mutations in complex I subunits from Fig. 3j is reproduced (bottom right). Traces for each cell line (left) are shown in the order indicated by the table. In cases in which the sequence shown is in the reverse orientation to the revised Cambridge Reference Sequence, these are indicated by ‘reverse str’. For each trace, the gene sequenced is at the bottom left, the DNA sequence is at the top, and the nucleotide alteration is in red text.

Extended Data Figure 8 Additional data supporting the hypersensitivity of cell lines with the identified biomarkers to biguanides.

a, b, Viability (a, 10 mM glucose) or relative change in cell number (b, 4 days, glucose concentration indicated in key) of indicated cell lines at phenformin concentrations indicated. Viability measured by ATP levels on day 3 at phenformin concentrations indicated by the black–blue scale, compared to ATP levels on day 0. Value of 1 indicates fully viable cells (untreated). Value of 0 indicates no change in ATP level compared to day 0 (cytostatic). Negative values indicate decrease in ATP levels (−1 indicates no ATP). c, Viability as in a of indicated cell lines under 0.75 mM and 10 mM glucose at indicated phenformin concentrations. d, Left, relative change in cell number in 0.75 mM glucose, 2 mM metformin relative to untreated in glucose limitation resistant (black) and sensitive (blue) cell lines. Right, relative size of tumour xenografts derived from the indicated cell lines in mice injected with PBS or metformin (intraperitoneal, 300 mg kg−1 day−1). e, Viability as in a of NCI-H929 cells at the indicated concentrations of phenformin and glucose. f, Relative size of indicated cell line xenografts in mice treated with PBS or phenformin (1.7 mg ml−1 in drinking water). g, Per cent change in OCR of control (Vector) or NDI1-expressing lines (NDI1) relative to the second basal measurement and at indicated phenformin concentrations. h, Proliferation of 143B wild-type or 143B rho (no mtDNA) cell lines under 0.75 mM or 10 mM glucose with or without phenformin treatment. Error bars are s.e.m. n = 4 for a, c, e, g; n = 3 for b, d, h (left); n = 5 for d (right), f. Replicates are biological, means reported. *P < 0.05 by two-sided student’s t-test.

Extended Data Figure 9 Long-term treatment of mtDNA mutant cells with phenformin.

a, Sanger sequencing traces of mtDNA-encoded ND1 and ND4 genes from Cal-62 cells expressing NDI1 or control vector cultured under 5 to 20 µM phenformin or no phenformin for 1.5 months. Regions containing mutant sequence are indicated by the red box. b, Heteroplasmy levels for mutation in ND1 or ND4 were assessed by measuring the relative areas under the curve from Sanger-sequencing and plotted. c, Cal-62 cell lines cultured with or without phenformin for 1.5 months assessed for their ability to proliferate in 0.75 mM glucose (blue) relative to 10 mM glucose (black). The proliferation assay was for 4 days in the absence of phenformin. d, Heteroplasmy levels of ND1 and ND4 as in b of Cal-62 tumour xenografts in mice treated with or without phenformin for 28 days. Error bars are s.e.m., n = 3. Replicates are biological (c) or technical (b, d), means reported. *P < 0.05 by two-sided student’s t-test.

Extended Data Figure 10 Schematic of Nutrostat setup.

Part numbers, sizes and dimensions for the Nutrostat assembly are indicated (see Methods for additional details).

Supplementary information

Supplementary Table 1

Primary Cell Competition Data. Primary Data for Figure 2b, Proliferation of barcoded cell lines in high and low glucose. (XLS 14 kb)

Supplementary Table 2

Primary Screening Data. Primary Data for Figure 2d, RNAi Screen under 10 mM or 0.75 mM glucose conditions. (XLS 2858 kb)

Supplementary Table 3

Alternative Hit Lists from Pooled shRNA Screen. Data from Supplementary Table 2 analyzed using the GENE-E software to generate lists of scoring genes alternative to those in Figure 2e and Extended Data Figure 2a. (XLS 370 kb)

Supplementary Table 4

Impaired Glucose Import Gene Signature Data. Primary Data for Extended Data Figure 5. (XLS 746 kb)

PowerPoint slides

Rights and permissions

Reprints and permissions

About this article

Cite this article

Birsoy, K., Possemato, R., Lorbeer, F. et al. Metabolic determinants of cancer cell sensitivity to glucose limitation and biguanides. Nature 508, 108–112 (2014). https://doi.org/10.1038/nature13110

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/nature13110

This article is cited by

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links

Nature Briefing: Cancer

Sign up for the Nature Briefing: Cancer newsletter — what matters in cancer research, free to your inbox weekly.

Get what matters in cancer research, free to your inbox weekly. Sign up for Nature Briefing: Cancer