Dysregulated Alanine as a Potential Predictive Marker of Glioma-An Insight from Untargeted HRMAS-NMR and Machine Learning Data

Amino acid metabolism in glioma: in vivo MR-spectroscopic detection of alanine as a potential biomarker of poor survival in glioma patients

PURPOSE

Reprogramming of amino acid metabolism is relevant for initiating and fueling tumor formation and growth. Therefore, there has been growing interest in anticancer therapies targeting amino acid metabolism. While developing personalized therapeutic approaches to glioma, in vivo proton magnetic resonance spectroscopy (MRS) is a valuable tool for non-invasive monitoring of tumor metabolism. Here, we evaluated MRS-detected brain amino acids and myo-inositol as potential diagnostic and prognostic biomarkers in glioma.

METHOD

We measured alanine, glycine, glutamate, glutamine, and myo-inositol in 38 patients with MRI-suspected glioma using short and long echo-time single-voxel PRESS MRS sequences. The detectability of alanine, glycine, and myo-inositol and the (glutamate + glutamine)/total creatine ratio were evaluated against the patients' IDH mutation status, CNS WHO grade, and overall survival.

RESULTS

While the detection of alanine and non-detection of myo-inositol significantly correlated with IDH wildtype (p = 0.0008, p = 0.007, respectively) and WHO grade 4 (p = 0.01, p = 0.04, respectively), glycine detection was not significantly associated with either. The ratio of (glutamate + glutamine)/total creatine was significantly higher in WHO grade 4 than in 2 and 3. We found that the overall survival was significantly shorter in glioma patients with alanine detection (p = 0.00002).

CONCLUSION

Focusing on amino acids in MRS can improve its diagnostic and prognostic value in glioma. Alanine, which is visible at long TE even in the presence of lipids, could be a relevant indicator for overall survival.

Urinary D-asparagine level is decreased by the presence of glioblastoma

Gliomas, particularly glioblastomas (GBMs), pose significant challenges due to their aggressiveness and poor prognosis. Early detection through biomarkers is critical for improving outcomes. This study aimed to identify novel biomarkers for gliomas, particularly GBMs, using chiral amino acid profiling. We used chiral amino acid analysis to measure amino acid L- and D-isomer levels in resected tissues (tumor and non-tumor), blood, and urine from 33 patients with primary gliomas and 24 healthy volunteers. The levels of D-amino acid oxidase (DAO), a D-amino acid-degrading enzyme, were evaluated to investigate the D-amino acid metabolism in brain tissue. The GBM mouse model was created by transplanting GBM cells into the brain to confirm whether gliomas affect blood and urine chiral amino acid profiles. We also assessed whether D-amino acids produced by GBM cells are involved in cell proliferation. D-asparagine (D-Asn) levels were higher and DAO expression was lower in glioma than in non-glioma tissues. Blood and urinary D-Asn levels were lower in patients with GBM than in healthy volunteers (p < 0.001), increasing after GBM removal (p < 0.05). Urinary D-Asn levels differentiated between healthy volunteers and patients with GBM (area under the curve: 0.93, sensitivity: 0.88, specificity: 0.92). GBM mouse model validated the decrease of urinary D-Asn in GBM. GBM cells used D-Asn for cell proliferation. Gliomas induce alterations in chiral amino acid profiles, affecting blood and urine levels. Urinary D-Asn emerges as a promising diagnostic biomarker for gliomas, reflecting tumor presence and severity.

Synthesis and Biological Evaluation of Fluorine-18 and Deuterium Labeled l-Fluoroalanines as Positron Emission Tomography Imaging Agents for Cancer Detection

To fully explore the potential of 18F-labeled l-fluoroalanine for imaging cancer and other chronic diseases, a simple and mild radiosynthesis method has been established to produce optically pure l-3-[18F]fluoroalanine (l-[18F]FAla), using a serine-derivatized, five-membered-ring sulfamidate as the radiofluorination precursor. A deuterated analogue, l-3-[18F]fluoroalanine-d3 (l-[18F]FAla-d3), was also prepared to improve metabolic stability. Both l-[18F]FAla and l-[18F]FAla-d3 were rapidly taken up by 9L/lacZ, MIA PaCa-2, and U87MG cells and were shown to be substrates for the alanine-serine-cysteine (ASC) amino acid transporter. The ability of l-[18F]FAla, l-[18F]FAla-d3, and the d-enantiomer, d-[18F]FAla-d3, to image tumors was evaluated in U87MG tumor-bearing mice. Despite the significant bone uptake was observed for both l-[18F]FAla and l-[18F]FAla-d3, the latter had enhanced tumor uptake compared to l-[18F]FAla, and d-[18F]FAla-d3 was not specifically taken up by the tumors. The enhanced tumor uptake of l-[18F]FAla-d3 compared with its nondeuterated counterpart, l-[18F]FAla, warranted the further biological investigation of this radiotracer as a potential cancer imaging agent.

Integrative Magnetic Resonance Imaging and Metabolomic Characterization of a Glioblastoma Rat Model

Glioblastoma (GBM) stands as the most prevalent and lethal malignant brain tumor, characterized by its highly infiltrative nature. This study aimed to identify additional MRI and metabolomic biomarkers of GBM and its impact on healthy tissue using an advanced-stage C6 glioma rat model. Wistar rats underwent a stereotactic injection of C6 cells (GBM group, n = 10) or cell medium (sham group, n = 4). A multiparametric MRI, including anatomical T2W and T1W images, relaxometry maps (T2, T2*, and T1), the magnetization transfer ratio (MTR), and diffusion tensor imaging (DTI), was performed. Additionally, ex vivo magnetic resonance spectroscopy (MRS) HRMAS spectra were acquired. The MRI analysis revealed significant differences in the T2 maps, T1 maps, MTR, and mean diffusivity parameters between the GBM tumor and the rest of the studied regions, which were the contralateral areas of the GBM rats and both regions of the sham rats (the ipsilateral and contralateral). The ex vivo spectra revealed markers of neuronal loss, apoptosis, and higher glucose uptake by the tumor. Notably, the myo-inositol and phosphocholine levels were elevated in both the tumor and the contralateral regions of the GBM rats compared to the sham rats, suggesting the effects of the tumor on the healthy tissue. The MRI parameters related to inflammation, cellularity, and tissue integrity, along with MRS-detected metabolites, serve as potential biomarkers for the tumor evolution, treatment response, and impact on healthy tissue. These techniques can be potent tools for evaluating new drugs and treatment targets.

MALDI-MSI-LC-MS/MS Workflow for Single-Section Single Step Combined Proteomics and Quantitative Lipidomics

We introduce a novel approach for comprehensive molecular profiling in biological samples. Our single-section methodology combines quantitative mass spectrometry imaging (Q-MSI) and a single step extraction protocol enabling lipidomic and proteomic liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis on the same tissue area. The integration of spatially correlated lipidomic and proteomic data on a single tissue section allows for a comprehensive interpretation of the molecular landscape. Comparing Q-MSI and Q-LC-MS/MS quantification results sheds new light on the effect of MSI and related sample preparation. Performing MSI before Q-LC-MS on the same tissue section led to fewer protein identifications and a lower correlation between lipid quantification results. Also, the critical role and influence of internal standards in Q-MSI for accurate quantification is highlighted. Testing various slide types and the evaluation of different workflows for single-section spatial multiomics analysis emphasized the need for critical evaluation of Q-MSI data. These findings highlight the necessity for robust quantification methods comparable to current gold-standard LC-MS/MS techniques. The spatial information from MSI allowed region-specific insights within heterogeneous tissues, as demonstrated for glioblastoma multiforme. Additionally, our workflow demonstrated the efficiency of a single step extraction for lipidomic and proteomic analyses on the same tissue area, enabling the examination of significantly altered proteins and lipids within distinct regions of a single section. The integration of these insights into a lipid-protein interaction network expands the biological information attainable from a tissue section, highlighting the potential of this comprehensive approach for advancing spatial multiomics research.

PiDeeL: metabolic pathway-informed deep learning model for survival analysis and pathological classification of gliomas

MOTIVATION

Online assessment of tumor characteristics during surgery is important and has the potential to establish an intra-operative surgeon feedback mechanism. With the availability of such feedback, surgeons could decide to be more liberal or conservative regarding the resection of the tumor. While there are methods to perform metabolomics-based tumor pathology prediction, their model complexity predictive performance is limited by the small dataset sizes. Furthermore, the information conveyed by the feedback provided on the tumor tissue could be improved both in terms of content and accuracy.

RESULTS

In this study, we propose a metabolic pathway-informed deep learning model (PiDeeL) to perform survival analysis and pathology assessment based on metabolite concentrations. We show that incorporating pathway information into the model architecture substantially reduces parameter complexity and achieves better survival analysis and pathological classification performance. With these design decisions, we show that PiDeeL improves tumor pathology prediction performance of the state-of-the-art in terms of the Area Under the ROC Curve by 3.38% and the Area Under the Precision-Recall Curve by 4.06%. Similarly, with respect to the time-dependent concordance index (c-index), PiDeeL achieves better survival analysis performance (improvement of 4.3%) when compared to the state-of-the-art. Moreover, we show that importance analyses performed on input metabolite features as well as pathway-specific neurons of PiDeeL provide insights into tumor metabolism. We foresee that the use of this model in the surgery room will help surgeons adjust the surgery plan on the fly and will result in better prognosis estimates tailored to surgical procedures.

AVAILABILITY AND IMPLEMENTATION

The code is released at https://github.com/ciceklab/PiDeeL. The data used in this study are released at https://zenodo.org/record/7228791.

A comparison of different machine-learning techniques for the selection of a panel of metabolites allowing early detection of brain tumors

Metabolomics combined with machine learning methods (MLMs), is a powerful tool for searching novel diagnostic panels. This study was intended to use targeted plasma metabolomics and advanced MLMs to develop strategies for diagnosing brain tumors. Measurement of 188 metabolites was performed on plasma samples collected from 95 patients with gliomas (grade I-IV), 70 with meningioma, and 71 healthy individuals as a control group. Four predictive models to diagnose glioma were prepared using 10 MLMs and a conventional approach. Based on the cross-validation results of the created models, the F1-scores were calculated, then obtained values were compared. Subsequently, the best algorithm was applied to perform five comparisons involving gliomas, meningiomas, and controls. The best results were obtained using the newly developed hybrid evolutionary heterogeneous decision tree (EvoHDTree) algorithm, which was validated using Leave-One-Out Cross-Validation, resulting in an F1-score for all comparisons in the range of 0.476-0.948 and the area under the ROC curves ranging from 0.660 to 0.873. Brain tumor diagnostic panels were constructed with unique metabolites, which reduces the likelihood of misdiagnosis. This study proposes a novel interdisciplinary method for brain tumor diagnosis based on metabolomics and EvoHDTree, exhibiting significant predictive coefficients.

Integrating Multi-Omics Analysis for Enhanced Diagnosis and Treatment of Glioblastoma: A Comprehensive Data-Driven Approach

The most aggressive primary malignant brain tumor in adults is glioblastoma (GBM), which has poor overall survival (OS). There is a high relapse rate among patients with GBM despite maximally safe surgery, radiation therapy, temozolomide (TMZ), and aggressive treatment. Hence, there is an urgent and unmet clinical need for new approaches to managing GBM. The current study identified modules (MYC, EGFR, PIK3CA, SUZ12, and SPRK2) involved in GBM disease through the NeDRex plugin. Furthermore, hub genes were identified in a comprehensive interaction network containing 7560 proteins related to GBM disease and 3860 proteins associated with signaling pathways involved in GBM. By integrating the results of the analyses mentioned above and again performing centrality analysis, eleven key genes involved in GBM disease were identified. ProteomicsDB and Gliovis databases were used for determining the gene expression in normal and tumor brain tissue. The NetworkAnalyst and the mGWAS-Explorer tools identified miRNAs, SNPs, and metabolites associated with these 11 genes. Moreover, a literature review of recent studies revealed other lists of metabolites related to GBM disease. The enrichment analysis of identified genes, miRNAs, and metabolites associated with GBM disease was performed using ExpressAnalyst, miEAA, and MetaboAnalyst tools. Further investigation of metabolite roles in GBM was performed using pathway, joint pathway, and network analyses. The results of this study allowed us to identify 11 genes (UBC, HDAC1, CTNNB1, TRIM28, CSNK2A1, RBBP4, TP53, APP, DAB1, PINK1, and RELN), five miRNAs (hsa-mir-221-3p, hsa-mir-30a-5p, hsa-mir-15a-5p, hsa-mir-130a-3p, and hsa-let-7b-5p), six metabolites (HDL, N6-acetyl-L-lysine, cholesterol, formate, N, N-dimethylglycine/xylose, and X2. piperidinone) and 15 distinct signaling pathways that play an indispensable role in GBM disease development. The identified top genes, miRNAs, and metabolite signatures can be targeted to establish early diagnostic methods and plan personalized GBM treatment strategies.

Association between altered metabolism and genetic mutations in human glioma

BACKGROUND

Molecular markers for classification of gliomas include isocitrate dehydrogenase (IDH) mutations and codeletion of chromosomal arms 1p and 19q (1p/19q). While mutations in IDH enzymes result in the well-characterized production of oncometabolite 2-hydroxyglutarate, dysregulation of other metabolites in IDH tumors is less characterized. Similarly, the effects of 1p/19q codeletion on cellular metabolism are also unclear.

AIM

This study aimed to quantify changes in tumor metabolites in human glioma tissue as a function of both IDH mutation and 1p/19q codeletion.

METHODS AND RESULTS

Deidentified human glioma tissue and associated clinical data were obtained from the Emory University Winship Cancer Institute tissue biobank from 14 patients (WHO grades II, III, and IV; seven female and seven male). Proton (¹ H) high-resolution magic angle spinning (HR-MAS) nuclear magnetic resonance (NMR) spectroscopy data were acquired using a 600 MHz Bruker AVANCE III NMR spectrometer. Metabolite concentrations were calculated using LCModel. Differences in metabolite concentrations as a function of IDH mutation, 1p/19q codeletion, and survival status were determined using Mann-Whitney U tests. Concentrations of alanine, glutamine, and glutamate were significantly lower in glioma tissue with IDH mutations compared to tissue with IDH wildtype. Additionally, glutamate concentration was significantly lower in glioma tissue with 1p/19q codeletion compared to intact 1p/19q. Exploratory analysis revealed alanine concentration varied significantly as a function of survival status.

CONCLUSIONS

Given the emerging landscape of glioma treatments that target metabolic dysregulation, an improved understanding of altered metabolism in molecular sub-types of gliomas, including those with IDH mutation and 1p/19q codeletion, is an important consideration for treatment stratification and personalized medicine.

Assessing Metabolic Markers in Glioblastoma Using Machine Learning: A Systematic Review

Glioblastoma (GBM) is a common and deadly brain tumor with late diagnoses and poor prognoses. Machine learning (ML) is an emerging tool that can create highly accurate diagnostic and prognostic prediction models. This paper aimed to systematically search the literature on ML for GBM metabolism and assess recent advancements. A literature search was performed using predetermined search terms. Articles describing the use of an ML algorithm for GBM metabolism were included. Ten studies met the inclusion criteria for analysis: diagnostic (n = 3, 30%), prognostic (n = 6, 60%), or both (n = 1, 10%). Most studies analyzed data from multiple databases, while 50% (n = 5) included additional original samples. At least 2536 data samples were run through an ML algorithm. Twenty-seven ML algorithms were recorded with a mean of 2.8 algorithms per study. Algorithms were supervised (n = 24, 89%), unsupervised (n = 3, 11%), continuous (n = 19, 70%), or categorical (n = 8, 30%). The mean reported accuracy and AUC of ROC were 95.63% and 0.779, respectively. One hundred six metabolic markers were identified, but only EMP3 was reported in multiple studies. Many studies have identified potential biomarkers for GBM diagnosis and prognostication. These algorithms show promise; however, a consensus on even a handful of biomarkers has not yet been made.

Targeted metabolomics analyses for brain tumor margin assessment during surgery

MOTIVATION

Identification and removal of micro-scale residual tumor tissue during brain tumor surgery are key for survival in glioma patients. For this goal, High-Resolution Magic Angle Spinning Nuclear Magnetic Resonance (HRMAS NMR) spectroscopy-based assessment of tumor margins during surgery has been an effective method. However, the time required for metabolite quantification and the need for human experts such as a pathologist to be present during surgery are major bottlenecks of this technique. While machine learning techniques that analyze the NMR spectrum in an untargeted manner (i.e. using the full raw signal) have been shown to effectively automate this feedback mechanism, high dimensional and noisy structure of the NMR signal limits the attained performance.

RESULTS

In this study, we show that identifying informative regions in the HRMAS NMR spectrum and using them for tumor margin assessment improves the prediction power. We use the spectra normalized with the ERETIC (electronic reference to access in vivo concentrations) method which uses an external reference signal to calibrate the HRMAS NMR spectrum. We train models to predict quantities of metabolites from annotated regions of this spectrum. Using these predictions for tumor margin assessment provides performance improvements up to 4.6% the Area Under the ROC Curve (AUC-ROC) and 2.8% the Area Under the Precision-Recall Curve (AUC-PR). We validate the importance of various tumor biomarkers and identify a novel region between 7.97 ppm and 8.09 ppm as a new candidate for a glioma biomarker.

AVAILABILITY AND IMPLEMENTATION

The code is released at https://github.com/ciceklab/targeted_brain_tumor_margin_assessment. The data underlying this article are available in Zenodo, at https://doi.org/10.5281/zenodo.5781769.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.