MSACL 2018 US Abstract

Topic: Metabolomics

Metabolomic Profiling of a Large Cohort of Glioblastoma Stem Cell Lines Identifies Metabolites Associated with Clinical Outcome

Amy Caudy (Presenter)
University of Toronto

Bio: Dr. Caudy is an Associate Professor at the University of Toronto in the Donnelly Centre for Molecular and Cellular Biology and the Department of Molecular Genetics. She is an organizer of the Cold Spring Harbor Summer Course in metabolomics, and uses full scan metabolomics to discover the function of proteins and pathways.

Authorship: Amy A. Caudy (1), Olga Zaslaver (1), Adam P. Rosebrock (2)
(1) University of Toronto, Toronto, Canada. (2) Stony Brook School of Medicine, Stony Brook, USA

Short Abstract

Glioblastoma multiforme tumors are driven by a rare stem cell population. In an effort to identify metabolic pathways active in these tumors as new therapeutic targets for a disease with few available drug therapies, we have profiled the metabolome of patient-derived glioblastoma stem cell lines cultured in vitro. To enable this large-scale project with samples arriving over a period of years, we have developed stable, reliable LC-MS methods and a biologically derived stable isotope reference material to enable comparison across different sample sets. We observe differences among patients in levels of 2-hydroxyglutarate and alpha-aminoadipic-acid, two metabolites whose levels have been linked to the progression of glioblastoma, and we observe many other known and unidentified metabolites whose levels vary across patients.

Long Abstract


Glioblastoma multiforme (GBM) is a deadly cancer with extremely limited options for treatment. GBM tumors contain a population of stem cells that are drug resistant and are responsible for tumor regrowth following treatment. These stem cells can be cultured under controlled conditions, and will initiate tumors that recapitulate the phenotype of the original patient tumor when transplanted into the brains of rodents as orthotopic xenografts. We hope to identify new drug targets for cancer by determining which metabolic pathways are active in tumors as compared with normal, nontumorigenic neural stem cells. As part of the Stand Up to Cancer Canada Brain Tumor Stem Cell Dream Team, we have used LC-MS full scan metabolomics to quantify metabolite levels in stem cell lines derived from over 40 patients’ tumors, comprising more than 800 samples including replicates.


Analysis of this large sample set in which samples have been supplied over a period of years required development of robust, reproducible methods for metabolomics analysis. Each sample is analyzed with two liquid chromatographic methods: an acidic reverse phase method and an ion-paired method. The acidic reverse phased method fully resolves many isomeric compounds in the TCA cycle, including citrate and isocitrate. The ion-paired reverse phase method that permits robust, reproducible analysis of polar metabolites with little interference from salt. In order to quantify relative metabolite levels across samples, we mix each sample with a 13C and 15N fully labeled biological reference material. We find that this stable isotope reference allows robust comparison across sample sets over large periods of time.


Analysis of this sample set reveals that tumor stem cell lines fall into three different metabolic patterns that result from changes in levels of many metabolites. Among the notable changes we observe is variation in the levels of the lysine metabolite alpha-amino-adipic acid, a metabolite whose levels have been linked to glioblastoma patient survival in other studies.

Production of D-2-hydroxyglutarate (D-2-HG) from 2-oxoglutarate by mutant IDH is a common feature in glioblastoma. This oncometabolite inhibits DNA demethylation and histone demethylation. One patient-derived line in our cohort overproduces D-2-HG as the result of an IDHR100Q mutation, but in a recurrent tumor the metabolite is no longer produced even though the mutation remains and is expressed. This down-regulation of D-2-HG suggests that D-2-HG may not be advantageous to tumors following the remodeling of the epigenome. Similarly, we observe another patient in which stem cells from the primary tumor show unexplained accumulation of L-2-HG; this metabolic change is not present in recurrent tumors. We differentiate and quantify the entantiomers of D- and L-2-HG using a tartaric acid derivatization method run on the same column as the acidic reverse phase method.

Neural stem cells can be cultured in suspension as neurospheres or adherent to plates. There are large changes in steady state metabolite levels between these conditions, even though metabolite levels of biological replicates under a single condition are well-correlated. This observation underscores the importance of maintaining defined culture conditions.

Conclusions & Discussion

Our study demonstrates the potential for our full-scan metabolomics approach to discover clinically relevant metabolites. We anticipate that the other metabolites that systematically vary across different groups of patients will provide new avenues for targeting pathways active in these tumors.

References & Acknowledgements:

The Stand Up to Cancer Canada Brain Tumor Stem Cell Dream Team is led by Peter Dirks and Sam Weiss. Michelle Kushida and Artee Luchman assisted with cell culture.

Financial Disclosure

GrantsyesAgilent (husband)
Board MemberyesMSACL (husband)

IP Royalty: yes

IP Desc:RNA interference (unrelated to talk)

Planning to mention or discuss specific products or technology of the company(ies) listed above: