N-Acetylaspartate and 2-Hydroxyglutarate Assessed in Human Brain Tissue by Mass Spectrometry as Neuronal Markers of Oncogenesis

To the Editor: Mass spectrometry (MS)1 assays are becoming crucial inputs into clinical decisions in personalized medicine (1). Several assays are finding application for the diagnosis of cancers because they can rapidly reveal the molecular content of minute biopsy tissue samples and provide diagnostic information that complements histopathological examination. For gliomas, which represent 75% of all malignant brain tumors, N-acetylaspartate (NAA) and 2-hydroxyglutarate (2HG) are markers of oncogenesis, and their MS measurement in brain tissue has the potential to improve glioma diagnosis. Here, we describe new results of a direct MS method applied to NAA and 2HG and discuss diagnostic and prognostic implications. We performed electrospray ionization (ESI) of brain tissue extracts with a triple quadrupole mass spectrometer in multiple reaction monitoring (MRM) mode. Quantitative measurements were made on 28 human banked neurological specimens obtained from the Biorepository of the Methodist Hospital in Indianapolis (IRB # 1410015344). Adjacent tissue was examined by histopathology to identify the presence of the tumor, estimate degree of infiltration, which was measured as tumor cell percentage, and assess isocitrate dehydrogenase (IDH) mutation status. The tissue was weighed wet and extracted using methanol:water (3:2 v/v, 2 g/L) spiked with NAA-d3 and 2HG-d3. Extracts were homogenized and centrifuged. The supernatant was infused into the instrument and MRM transitions were monitored for a total of 72 s in negative ionization mode. Quantification was achieved using internal standard calibration and validated for recovery (<20% error), precision (CV < 10%), lower limits of detection (5 and 20 ng/mL, respectively, for NAA and 2HG), and for carry-over and matrix effects. Concentrations of NAA (Fig. 1) were found to decrease significantly with the increase in tumor cell percentage (low, medium, and high; median concentrations 824, 313, and 43 ng/mg, respectively). Concentrations up to 1500 ng/mg were detected in low-glioma–infiltrated tissue, matching independent results (2); concentrations of <80 ng/mg were measured in highly infiltrated tissue. Concentrations of 2HG (Fig. 1) were consistently below the lower limit of detection in wild-type gliomas, whereas IDH-mutant gliomas showed approximately a 10-fold increase (median 254 ng/mg). We attribute the wide distribution of concentrations in IDH-mutant gliomas to variations in tissue cellularity and heterogeneity of tumor density. Receiver operating characteristic curve analysis provided 100% clinical sensitivity and specificity for 2HG MS in identifying IDH mutation, showing better performance than magnetic resonance spectroscopy (MRS), largely due to the superior detection capability and molecular specificity of MS. We suggest a cutoff of 45 ng/mg based on preliminary MS results, a value that is lower than that based on MRS results (2 mmol/L, approx. 300 ng/mg) (3). The box represents the interquartile range with a median line and whiskers at ±1.5 SD. Quantitative measurements were made using the MRM transitions m/z 174.0→88.2 for NAA, 177.1→91.3 for NAA-d3, 147.1→129.1 for 2HG, and 150.1→132.1 for 2HG-d3. Population medians for NAA and 2HG concentrations were statistically different (P values equal to 0.004 and 0.0001, respectively, using Kruskal–Wallis nonparametric test). Random numbers between 0 and the limit of detection (20 ng/mL) were assigned to wild-type glioma specimens for which no 2HG was detected. NAA is an abundant neurometabolite in healthy tissue but glioma cells do not express the biosynthetic enzyme L-aspartate N-acetyltransferase (2), accounting for the decrease of NAA concentration with increasingly glioma-infiltrated tissue. Our findings support the hypothesis that tumor invasion can be assessed through neuronal cell damage in its path by measuring NAA depletion (4). Assessment of glioma infiltration is difficult with current diagnostic tools (e.g., histopathology) and is not usually performed during surgical removal. MRS has been investigated as noninvasive pre- and postoperative strategy to track global NAA depletion (4). MS can be used intraoperatively to maximize glioma resection, which is positively prognostic. Complete resection is unattainable because gliomas infiltrate diffusely. Unresected tumor at the surgical margins causes recurrence and malignant progression and is one reason for glioma treatment failure. Assessment of tumor infiltration in biopsied tissue at neurosurgeon-defined points along the surgical margins can corroborate neurosurgeons' acumen and guide resection maneuvers, which currently rely on preoperative MRI images roughly showing tumor location. Intraoperative NAA MS technology is readily available. Desorption electrospray ionization (DESI) is a modified ESI strategy that has been used to rapidly analyze biopsied tissue smears in the operating room without interfering with surgical procedures (5). Initial results from 10 surgical cases show that high-glioma infiltration assessed via NAA DESI-MS could be seen even at margins that appeared clear by postoperative MRI (5). IDH mutations result in neomorphic activity of the enzyme to produce 2HG in glioma cells, which inhibits normal cellular processes. IDH mutation is a powerful prognostic factor in malignant gliomas. We propose its intraoperative assessment via quantitative 2HG MS as an adjunct to standard immunohistochemistry assays that are laborious and can be performed only postoperatively. Our initial results suggest that intraoperative quantitative 2HG MS will improve the accuracy of diagnosis because IDH mutation is part of the glioma classification system. Also, it can influence clinical decisions regarding the aggressiveness of glioma removal to better balance surgical benefits and costs: more aggressive resection of IDH-mutant gliomas is associated with increased survival rate, whereas for wild-type gliomas it is not (3), thereby moving toward customized glioma patient treatment and management. mass spectrometry N-acetylaspartate 2-hydroxyglutarate electrospray ionization isocitrate dehydrogenase magnetic resonance spectroscopy Desorption electrospray ionization. The authors thank Dr. Eyas M. Hattab (Neuropathologist, University of Louisville, KY) and Dr. Aaron A. Cohen-Gadol (Neurosurgeon, IU School of Medicine, IN) for ongoing collaboration.