Metabolomics of Human Cerebrospinal Fluid Identifies Signatures of Malignant Glioma

Abstract

Cerebrospinal fluid is routinely collected for the diagnosis and monitoring of patients with neurological malignancies. However, little is known as to how its constituents may change in a patient when presented with a malignant glioma. Here, we used a targeted mass-spectrometry based metabolomics platform using selected reaction monitoring with positive/negative switching and profiled the relative levels of over 124 polar metabolites present in patient cerebrospinal fluid. We analyzed the metabolic profiles from 10 patients presenting malignant gliomas and seven control patients that did not present malignancy to test whether a small sample size could provide statistically significant signatures. We carried out multiple unbiased forms of classification using a series of unsupervised techniques and identified metabolic signatures that distinguish malignant glioma patients from the control patients. One subtype identified contained metabolites enriched in citric acid cycle components. Newly diagnosed patients segregated into a different subtype and exhibited low levels of metabolites involved in tryptophan metabolism, which may indicate the absence of an inflammatory signature. Together our results provide the first global assessment of the polar metabolic composition in cerebrospinal fluid that accompanies malignancy, and demonstrate that data obtained from high throughput mass spectrometry technology may have suitable predictive capabilities for the identification of biomarkers and classification of neurological diseases. Cerebrospinal fluid is routinely collected for the diagnosis and monitoring of patients with neurological malignancies. However, little is known as to how its constituents may change in a patient when presented with a malignant glioma. Here, we used a targeted mass-spectrometry based metabolomics platform using selected reaction monitoring with positive/negative switching and profiled the relative levels of over 124 polar metabolites present in patient cerebrospinal fluid. We analyzed the metabolic profiles from 10 patients presenting malignant gliomas and seven control patients that did not present malignancy to test whether a small sample size could provide statistically significant signatures. We carried out multiple unbiased forms of classification using a series of unsupervised techniques and identified metabolic signatures that distinguish malignant glioma patients from the control patients. One subtype identified contained metabolites enriched in citric acid cycle components. Newly diagnosed patients segregated into a different subtype and exhibited low levels of metabolites involved in tryptophan metabolism, which may indicate the absence of an inflammatory signature. Together our results provide the first global assessment of the polar metabolic composition in cerebrospinal fluid that accompanies malignancy, and demonstrate that data obtained from high throughput mass spectrometry technology may have suitable predictive capabilities for the identification of biomarkers and classification of neurological diseases. Patients with malignant gliomas (MG) 1The abbreviations used are:MGmalignant gliomasCNScentral nervous systemsSRMselected reaction monitoringCSFcerebrospinal fluidHILIChydrophilic interaction chromatographyTCAcitric acid cycleLC-MS/MStandem mass spectrometryMRImagnetic resonance imaging. 1The abbreviations used are:MGmalignant gliomasCNScentral nervous systemsSRMselected reaction monitoringCSFcerebrospinal fluidHILIChydrophilic interaction chromatographyTCAcitric acid cycleLC-MS/MStandem mass spectrometryMRImagnetic resonance imaging. have poor prognosis. Those with glioblastomas have a median survival of 14.6 months and 5-year survival is only 9.8% despite aggressive treatment with temozolomide chemo-irradiation (1Stupp R. Hegi M.E. Mason W.P. van den Bent M.J. Taphoorn M.J.B. Janzer R.C. Ludwin S.K. Allgeier A. Fisher B. Belanger K. Hau P. Brandes A.A. Gijtenbeek J. Marosi C. Vecht C.J. Mokhtari K. Wesseling P. Villa S. Eisenhauer E. Gorlia T. Weller M. Lacombe D. Cairncross J.G. Mirimanoff R.O. European Org R. Treatment; Canc Brain Tumour Grp; Radiat Oncol G., and Natl Canc Inst Canada Clin T Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial.Lancet Oncol. 2009; 10: 459-466Abstract Full Text Full Text PDF PubMed Scopus (5550) Google Scholar). Patients with anaplastic gliomas have a slightly better prognosis with a median survival of 3 to 10 years depending on the molecular genetics of the tumor (2Cairncross G. Berkey B. Shaw E. Jenkins R. Scheithauer B. Brachman D. Buckner J. Fink K. Souhami L. Laperierre N. Mehta M. Curran W. Phase III trial of chemotherapy plus radiotherapy compared with radiotherapy alone for pure and mixed anaplastic oligodendroglioma: Intergroup Radiation Therapy Oncology Group Trial 9402.J. Clin. Oncol. 2006; 24: 2707-2714Crossref PubMed Scopus (628) Google Scholar, 3Louis D.N. Holland E.C. Cairncross J.G. Glioma classification - A molecular reappraisal.Am. J. Pathol. 2001; 159: 779-786Abstract Full Text Full Text PDF PubMed Scopus (245) Google Scholar). Therefore, having a means of inquiring the biological state of the malignant glioma would guide the application of proper treatment. Despite the high resolution of magnetic resonance imaging (MRI), only macroscopic structural information is obtained and this measurement does not reveal the underlying biology of the malignant glioma. Although tumor tissue analysis performed serially over time is possible, brain biopsies have inherent sampling errors because of the heterogeneous nature of the tumor and resections may lead to neurological deficits (4Chandrasoma P.T. Smith M.M. Apuzzo M.L.J. Stereotactic biopsy in the diagnosis of brain masses - Comparison of results of biopsy and resected surgical specimen.Neurosurgery. 1989; 24: 160-165Crossref PubMed Scopus (153) Google Scholar, 5Glantz M.J. Burger P.C. Herndon 2nd, J.E. Friedman A.H. Cairncross J.G. Vick N.A. Schold Jr., S.C. Influence of the type of surgery on the histologic diagnosis in patients with anaplastic gliomas.Neurology. 1991; 41: 1741-1744Crossref PubMed Google Scholar). malignant gliomas central nervous systems selected reaction monitoring cerebrospinal fluid hydrophilic interaction chromatography citric acid cycle tandem mass spectrometry magnetic resonance imaging. malignant gliomas central nervous systems selected reaction monitoring cerebrospinal fluid hydrophilic interaction chromatography citric acid cycle tandem mass spectrometry magnetic resonance imaging. These limitations led us to search for an improved and/or complementary method of evaluating malignant gliomas using biomarkers in the cerebrospinal fluid (CSF). Because the CSF bathes the tissues of the central nervous system (CNS), it provides an attractive source of material for clinical diagnostics. CSF is readily accessible either by lumbar puncture or reservoir sampling, which is less invasive and can be performed serially and may potentially yield a more integrated view of the tumor's activity (6Swanson K. Lok E. Song S. Dessain J. Wong E.T. Cerebrospinal fluid (CSF) Biomarkers for breast cancer (BC) brain metastases.Ann. Neurol. 2009; 66: S58Google Scholar, 7Lehtinen M.K. Zappaterra M.W. Chen X. Yang Y.J. Hill A.D. Lun M.L. Maynard T. Gonzalez D. Kim S. Ye P. D'Ercole A.J. Wong E.T. LaMantia A.S. Walsh C.A. The cerebrospinal fluid provides a proliferative niche for neural progenitor cells.Neuron. 2010; 69: 893-905Abstract Full Text Full Text PDF Scopus (439) Google Scholar). As a result, CSF-derived biomarkers may provide an earlier diagnosis than MRI, obviate invasive procedures, offer information on the tumor's biological state, prognosticate patient survival, and/or predict treatment responses. CSF derived from patients presenting malignant gliomas may contain signatures of altered metabolism known to occur in tumor cells and may further harbor signatures of secondary, noncell-autonomous effects that arise from the tumor microenvironment (8Locasale J.W. Cantley L.C. Vander Heiden M.G. Cancer's insatiable appetite.Nat. Biotechnol. 2009; 27: 916-917Crossref PubMed Scopus (40) Google Scholar, 9Luo J. Solimini N.L. Elledge S.J. Principles of Cancer Therapy: Oncogene and Non-oncogene Addiction.Cell. 2009; 136: 823-837Abstract Full Text Full Text PDF PubMed Scopus (1359) Google Scholar, 10Vander Heiden M.G. Cantley L.C. Thompson C.B. Understanding the Warburg effect: the metabolic requirements of cell proliferation.Science. 2009; 324: 1029-1033Crossref PubMed Scopus (10213) Google Scholar, 11Michelakis ED S.G. Dromparis P Webster L Haromy A Niven E Maguire C Gammer TL Mackey JR Fulton D Abdulkarim B McMurtry MS Petruk KC. Metabolic modulation of glioblastoma with dichloroacetate.Sci. Translation Med. 2010; 2: 31-34Crossref Scopus (546) Google Scholar). Such physiological processes include inflammation, endocrine pathophysiology, and cellular debris resulting from necrotic cells that accumulates as a byproduct of disrupted tissue architecture (12Tlsty T.D. Coussens L.M. Tumor stroma and regulation of cancer development.Annu. Rev. Pathol.-Mech. Dis. 2006; 1: 119-150Crossref PubMed Scopus (804) Google Scholar). Metabolic profiling has also been used as a diagnostic tool in the setting of human cancer. Studies of metabolites from patient urine that were discovered using metabolomics technology have recently come into focus as biomarkers for cancer and cell A. L.M. J. B. R. M.K. A. S. B. J. D. S. A. S. C. profiles for in cancer 2009; PubMed Scopus Google Scholar, L. W. based analysis for cell and 10: PubMed Scopus Google Scholar). used mass spectrometry to the metabolic composition of urine in patients with cancer different of the The identified a of metabolites that with cancer and A. L.M. J. B. R. M.K. A. S. B. J. D. S. A. S. C. profiles for in cancer 2009; PubMed Scopus Google Scholar). is for the of metabolism in the biological fluid of cancer of altered cellular metabolism is that of the that to the of gliomas in cellular metabolism S. X. P. P. A. R. S. T. J. Smith R. N. G. B. integrated analysis of human glioblastoma PubMed Scopus Google Scholar). in glioma as or and have multiple metabolic G. Thompson C.B. The biology of metabolic cell and Full Text Full Text PDF PubMed Scopus Google Scholar). the metabolic and in of gliomas a that the activity of the that results in the of L. S. S. L.M. Cantley L.C. Thompson C.B. Heiden 2009; PubMed Scopus Google Scholar). is that the metabolic in malignant glioma cells may to global in the composition of metabolic profiling of CSF has a composition of metabolites present in high in patient CSF A. T. in the cerebrospinal fluid in by Scopus Google Scholar, T. C. L. S. J. N. R. T. van A.J. R. The of on of the and selected metabolites in cerebrospinal fluid 2009; PubMed Scopus Google Scholar, M.J. K. D. R. B. D. S. R. L. The human cerebrospinal fluid B. PubMed Scopus Google Scholar, B. E. J. G. analysis of human and fluid of in mass metabolomics 2009; PubMed Scopus Google Scholar, L. T. S. L. K. C. A. B. R. T. van A.J. and metabolomics analysis of human cerebrospinal fluid 2010; Full Text Full Text PDF PubMed Scopus Google Scholar, T. A. B. L. S. A. C. L.M. R. T. van A.J. R. The of on the and of human cerebrospinal Scopus Google Scholar). clinical have in the composition of CSF One identified in metabolic profiles in patients diagnosed with E. D. S. D. S. Metabolic profiling of that may on and in Med. 2006; Scopus Google Scholar). that with in the composition of CSF G. G. analysis of the cerebrospinal fluid in in the of Clin. Google Scholar). that metabolomics technology may for the of in the CSF of patients with malignant gliomas that could as biomarkers and/or with a mass spectrometry based platform selected reaction monitoring with and switching using a mass M. Yang X. A for and Scopus Google Scholar, A.D. M. J. D. from Tumor in Scopus Google Scholar, J.W. T. C.A. A.J. G. T. D. E. M. R. G. A.H. M. L. G. Cantley L.C. Vander Heiden M.G. and to Scopus Google Scholar, D. G. M. G. J.W. C.J. Vander Heiden M.G. Cantley L.C. of by to cellular Scopus Google Scholar, J. S.G. Vander Heiden M.G. Yang R. J.W. B. R. Cantley L.C. G. J. Metabolic regulation of by cell Full Text Full Text PDF Scopus Google we whether we could molecular in the CSF of patients with malignant gliomas using a sample The platform hydrophilic interaction chromatography J. C. and of cellular metabolites by hydrophilic interaction mass A. 2006; PubMed Scopus Google Scholar, W. for targeted B. PubMed Scopus Google Scholar). polar metabolites from were targeted from a The platform to include as polar metabolites that we could that metabolic the acid metabolism, metabolism, to cancer cell We the relative levels of 124 metabolites using of patient CSF in a of patients with malignant gliomas and a control of metabolites with in CSF metabolomics multiple to the data in an unbiased we identified significant in the composition patients with the and as as with diagnosed and malignant our demonstrate that the composition of the CSF may provide biomarkers and may provide into the underlying the of malignant would provide a for clinical that can the and as as the predictive of the biomarkers identified in the of of CSF were obtained from and patients and of state in the Brain Tumor by and CSF were collected the time of neurological when an for lumbar puncture or sampling from a were collected from lumbar puncture 3 in CSF from a were obtained for CSF and CSF analysis were the time of the clinical CSF as cell and were also versus survival tumor size MRI, and survival of and were also of patients with malignant gliomas and CSF anaplastic and anaplastic and Newly from diagnosis to sample tumor size from tumor size from or - - in a and of CSF were to in by for 10 were collected and and in of and by for were to a using a Fisher and for were using for mass were and analyzed using a mass to a system of a of using and switching to metabolites were targeted in and in and in with a source of The time and the cycle time for metabolites resulting in data were to the using a hydrophilic interaction chromatography were from B to B from B to B from B from B to B from B for to the A of in from the for were integrated using the using a of data with a time were carried out using the The resulting from the a for metabolites in sample were from the We this more because we it to be more than for which can lead to the of metabolites to analysis on the carried out using the method in the As is data were to the and data to the we used a We a of with of and to that of the We the of and a for were obtained by the for in the data to the obtained from were to be and for the components. analysis using the carried out using J. data and using Scholar). The of the for obtained and from to The were used for the The of the carried out using the method in R. 3 were obtained with using to performed using method in the method that the in to an Scopus Google Scholar). were obtained from were obtained using test for the analysis for which a The performed that the by the for human requirements for the have been CSF collected from patients as of clinical an the 10 patients with malignant diagnosed and were anaplastic and anaplastic CSF from seven control malignancy were also were collected from lumbar puncture 3 in CSF from a the system that the first a and it into the for the a selected for and switching because of the time and time in the The cycle for were performed in to data were collected with our metabolomics we patient sample having a over 124 polar metabolites from malignant gliomas and seven in from patient CSF analysis of the of of the a of the of the a of of the of in the the of or for analysis of the that from analysis of the of the malignant glioma control in the metabolites were on in the malignant glioma and metabolites were on in the control These can be better when the of in control and compared with the it is that in the of malignant gliomas and obtained that metabolites levels on the of in control versus These in could further be by the in the of as by of and patients. the of the exhibited a that the patients more in that of composition be obtained in than in control These also indicate that a of metabolites can be identified in a small than or to of a mass-spectrometry based metabolomics analysis of CSF could potentially distinguish malignant glioma patients from malignancy as as the analysis of the and that in the small composition of the CSF from led us to global in the small One used method to global in data using J. analysis of Rev. 2001; 2: PubMed Scopus Google Scholar, T. R. T. S. T. van M. T. D. of breast distinguish tumor with clinical S. A. 2001; PubMed Scopus Google Scholar, T. R. J. T. A. S. R. S. J. D. of breast tumor in data S. A. PubMed Scopus Google Scholar). Such have been to the of profiles in different in of A. P. M. J. J. Kim L. A.H. P.T. D. J.W. of and to Med. PubMed Scopus Google Scholar). These in the for diagnosis and prognosis D. M. N. K. C. N. M. S. L. and predictive for adjuvant chemotherapy in resected Clin. Oncol. 2010; PubMed Scopus Google Scholar). unsupervised using method to an Scopus Google that the carried of the resulting classification of metabolites and in malignant glioma CSF relative to the control The to of cancer patients and to the control the CSF from patients that have diagnosed malignant gliomas and a that is in from the signatures that to the from control Therefore, with a this classification not only segregated malignant glioma from also patients have from with diagnosed malignant gliomas be from because and into our to malignant glioma patients based on the metabolic composition of unsupervised were Although different classification different may reveal different the of the we also carried out a analysis in the levels and the of is into that for the sample M. is Biotechnol. PubMed Scopus Google Scholar). analysis of provides an unbiased means of the by for the relative of to that the of can a biological to the by of metabolites that We first the of a analysis on the A used to a of obtained from data of and to that of the A for obtained from this that the and were statistically significant and and and of the A which of the first is in of it is that is in the to further the we used the to the by the the metabolic an analysis of the involved in it is that metabolites from metabolism to from the citric acid cycle and metabolism in the first from the cycle and metabolism in the The metabolites from the cycle and that biological the glioma and for its from control A of the of sample the first is in the it is that of malignant glioma patients were identified and from the control a sample by with a different a is the and method identified different the We the metabolites that from to glioma patients. Although the classification were to in the data which to the of the of the of and not We an an analysis could have the clinical application for diagnostic metabolites in the CSF of the malignant gliomas relative to the control using a more test These metabolites from metabolic as and central One recently identified in glioma patients is the of is also in patient 3 which could indicate the of an in the tumor L. S. S. L.M. Cantley L.C. Thompson C.B. Heiden 2009; PubMed Scopus Google in malignant glioma control change glioma in a The different of metabolic profiles in the CSF of malignant glioma patients a by diagnosis and and a of patients 3 and with that a have profiles from the CSF of the of metabolites that patients 3 and 10 a in metabolites derived from the citric acid cycle in cycle metabolites not in patients that this may a of malignant glioma patients of metabolites to patients seven metabolites that change in this relative to the patients with of metabolites involved in tryptophan and These metabolites include acid and in this is used as a for activity M. S.J. B. C. of by tryptophan PubMed Scopus Google Scholar). activity of this is also in neurological R. to and when the system the Rev. PubMed Scopus Google Scholar). is the in the of metabolites of activity in patients. clinical of the metabolomics analysis that the be compared with known used in clinical Because clinical were for the patients in this we whether the relative levels of metabolites with of tumor that used to tumor size in glioma patients of of and the using fluid M. D. S.J. Holland E. K. C. D. R. A analysis of patients with glioblastoma of and 2001; PubMed Scopus Google Scholar). Although as for to results would sample we We of with metabolites and that metabolites with and with the levels of of with The levels of metabolites and with tumor These metabolites acid and of clinical is patient We metabolites that with patient metabolites were to with patient survival of metabolites and that the of can be in the The metabolites in and acid metabolism We present a global polar analysis of human CSF from a small of patients with malignant this we and that statistically significant signatures of glioma and patients can be obtained from a small sample The an and switching platform for the and of unsupervised classification to of metabolites patient and the inherent of for The based platform of a MS and for with and with and is because metabolites may and in to inherent in biological we used multiple classification and which provides of signatures that distinguish malignant gliomas from control as as patients with and diagnosed an for the of our be that metabolites resulting from metabolic can in the CSF in an with to the The of biomarkers from the metabolomics platform is different from that in and the of metabolites is and with to the of Therefore, despite the small sample size in our the in levels that is a for in CSF when in signatures may also have the to be used as diagnostic Because for small than for cells and the of CSF lumbar versus is less than the levels of biological CSF biomarkers for and that from either lumbar or source as to or levels J. R. and CSF and in PubMed Scopus Google Scholar). of our patient from the lumbar for 3 CSF collected from a Therefore, our method is to our analysis of the in the However, a of whether or not the of a would be different lumbar and would the and analysis of CSF from and a is in clinical the that 3 and 10 different from the of the malignant have and However, it to be whether the in the from the in the versus a of tumor or treatment on the A to more the and sampling of CSF the of composition with the of tumor or treatment a of to the metabolites in the the and inflammatory cells that have into the tumor malignant gliomas have altered metabolism by as or and The metabolic that in the of metabolites in the CSF the tumor is less than of the of the brain and cells can into the cell to a L. S. S. L.M. Cantley L.C. Thompson C.B. Heiden 2009; PubMed Scopus Google Scholar, R. A. L. M. S. C. M. A. P. of glioma cell in and in Neurol. 69: PubMed Scopus Google Scholar). we were to high levels of in patient 3 that this may have an malignant glioma. We that the CSF levels to tumor size as by and to the is a to be in glioblastomas and anaplastic than in gliomas and brain tissue as by magnetic resonance K. G. K. A. A. J. P. M. glioblastoma and levels in the 2009; PubMed Scopus Google Scholar, M. Smith L. of levels and of J. Google Scholar). with the that with of the glioma our data also that the tumor size in our with levels in the is an of the cell and it is into the fluid as the C.A. in human it Full Text Full Text PDF PubMed Scopus Google Scholar). The CSF of our that with the tumor of in human cells in and for control in Google and this that this may arise from the of tryptophan metabolism is for the and of M. S.J. B. C. of by tryptophan PubMed Scopus Google Scholar, R. to and when the system the Rev. PubMed Scopus Google Scholar, L. A. C. C. T. C. P. tryptophan in PubMed Scopus Google Scholar, M. and 2010; Full Text Full Text PDF PubMed Scopus Google Scholar). our with malignant we and which metabolites involved in the tryptophan metabolism, as compared with diagnosed patients. an acid and a for the also to be in our patients with These metabolites may be of the The of CSF metabolites from the cycle may indicate Patients 3 and 10 malignant with CSF and months from and tumor were as by and on is to the tumor to the nature of the malignant glioma because tumor size is based on the of from tumor Because glioma cells of the of tumor as on the of the tumor size because not only tumor also and in the brain E.T. Tumor and in malignant 2006; PubMed Scopus Google Scholar). Together our provides the first of profiling in the CSF of malignant Glioma patients. that mass metabolomics offer a technology for the of biomarkers of malignant Glioma from sampling biological fluid. biomarkers from metabolites as as from signatures were is our that this the of metabolomics technology and the metabolic analysis of CSF for further in as a clinical trial with a patient We Vander Heiden and for with