Gliomas phenocopy an inborn error of metabolism to drive neuronal activity and tumor growth

93 Cancer metabolism, cancer neuroscience, metabolite signaling, guanidinoacetate, creatine, 94 glioma, GABA, GAMT deficiency, inborn error of metabolism 95 96

97 Gliomas are the most common type of central nervous system malignancy diagnosed in adults. 98 Its most aggressive forms, high grade gliomas (HGG, Grade 4), carry a median life expectancy 99 of only 1-2 years. Therefore, there is an urgent need for insights into molecular mechanisms 100 underlying aggressive glioma behavior and new therapeutic strategies that disrupt these 101 processes. Adult HGGs encompass two diagnostic categories that are defined by isocitrate 102 dehydrogenase (IDH) mutational status1: 1) Glioblastoma (GBM), IDH-wildtype, Grade 4, and 103 2) Astrocytoma, IDH-mutant, Grade 4. IDH mutant enzymes display neomorphic synthesis of the 104 oncometabolite (R)-2-hydroxyglutarate [(R)-2HG], which reprograms the epigenome2–9 to 105 promote glioma development and maintenance10–12. This metabolic alteration underpins the 106 antitumor activity of the mutant IDH1/2 inhibitor vorasidenib, which was recently approved by 107 the United States Food and Drug Administration for the treatment of IDH-mutant low-grade 108 glioma13. These advances provide proof-of-principle that altered metabolism can serve as both an 109 oncogenic driver and a therapeutic target in glioma. However, in contrast to our understanding of 110 metabolic reprogramming by IDH oncogenes, which represent an uncommon subset of HGG, 111 knowledge of metabolic mechanisms that sustain HGGs across genetic and transcriptomic 112 subtypes is limited. 113 114 Prior research has provided vital insights into the biochemical landscape of adult-type gliomas14–115 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 6 17. HGGs display reduced creatine content15,18,19 and harbor elevated levels of nucleotides and 116 their precursors14,20,21, tryptophan15, lysine catabolites22, lipid species14,17, and redox 117 metabolites14 including glutathione15. Nevertheless, comprehensive metabolomic analyses of 118 HGG, LGG, and non-malignant brain specimens have not kept pace with comparable genomics 119 and transcriptomics analyses of these tissues23–25. This limited momentum has impeded insights 120 into metabolic alterations that are causally linked to aggressive glioma behavior. 121 122 Growth-promoting interactions of gliomas with the neural circuits they invade have emerged as a 123 major axis of pathophysiology. The activity of neurons, especially glutamatergic neurons, 124 robustly drives the growth26–31 and invasion32,33 of multiple forms of gliomas through paracrine 125 signaling factors26 and bona fide neuron-to-cancer synapses28,31. In turn, gliomas increase the 126 excitability of neurons34–37, contributing to glioma-associated seizures and further augmenting 127 glioma growth-promoting, activity-regulated neuronal interactions. The role that metabolic 128 mechanisms may play in this pathophysiology, however, is largely unexplored. 129 130 To complement previous studies and enhance the depth of metabolomic characterization of 131 HGGs, we collected a set of 91 primary human brain tissue samples and used them to conduct 132 multi/i1 omics analyses including quantification of 563 polar metabolites and lipids. To our 133 knowledge, this dataset represents the most detailed metabolic portrait of HGG constructed to 134 date. These specimens spanned non/i1 malignant cortex, brain metastases, lower grade gliomas 135 (LGG, Grades 2–3), and HGGs. We integrated polar metabolomics, lipidomics, and 136 transcriptomics data with clinical and molecular tissue features. Through this effort, we sought to 137 discover biochemical patterns unique to the HGG metabolome and define their contributions to 138 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 7 HGG maintenance and progression. We report the discovery of a metabolic phenotype present in 139 HGG that spans genetic and transcriptomic disease subtypes and recapitulates the biology of a 140 rare inborn error of creatine metabolism, GAMT deficiency. We further demonstrate that this 141 metabolic phenotype drives hyperexcitability of the HGG microenvironment and tumor growth, 142 thereby revealing a tractable target for therapeutic intervention in patients with HGG. 143 144

145 Metabolomic and lipidomic profiling of human brain and brain tumor tissues 146 To generate new insights into brain tumor metabolism, we prospectively collected 91 adult 147 human brain tissue samples from tumor or seizure foci resection surgeries. These samples 148 included LGGs, HGGs, brain metastases, and non-malignant brain specimens (Figure S1A). 149 Each tissue was divided and processed for metabolomic, lipidomic, and transcriptomic analyses. 150 We used defined gene expression programs25 to classify glioma samples as Proneural, Classical, 151 or Mesenchymal (Figure S1B). Next, we integrated data for 563 polar metabolites and lipid 152 species with clinical and molecular information associated with each specimen and performed 153 unsupervised clustering to detect shared biochemical profiles in an unbiased manner (Figure 1A 154 and Table S1). This approach revealed seven dominant clusters, including subgroups enriched in 155 Mesenchymal GBMs, recurrent LGGs, index LGGs, brain metastases, and non-malignant brain 156 samples, as well as two mixed populations that did not display clearly defined correlates with 157 molecular or clinical features. Importantly, supervised dimensionality reduction analysis 158 demonstrated that HGGs, LGGs, brain metastases, and non-malignant brain samples were readily 159 distinguishable by biochemical profiles (Figure 1B). Moreover, glioma samples of different 160 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 8 histological subtypes, grades, and transcriptional subtypes could be discriminated based on 161 metabolite content (Figures S1C-E). 162 163 Before evaluating novel biochemical hallmarks of brain tumor classes, we first assessed expected 164 metabolic differences among glioma and non-malignant brain samples. 2HG levels were 165 markedly elevated in IDH-mutant glioma specimens relative to IDH-wildtype gliomas and non-166 malignant brain controls (Figure S1F). L-DOPA, a precursor to dopamine, was enriched in non-167 malignant brain samples compared with HGGs (Figure S1G), consistent with prior work38. 168 Additionally, intermediates in lysine degradation22 (aminoadipate) and nucleotide 169 metabolism20,21 (orotate and deoxyguanosine) pathways were elevated while the ratio of N-170 acetylaspartate to creatine (an imaging biomarker of HGG18,19) decreased in HGG versus non-171 malignant brain tissues (Figures S1H-K). Observing these anticipated biochemical patterns 172 affirmed the validity of our dataset. 173 174 Next, we conducted pathway analyses to identify biochemical processes that are altered in 175 specific brain tumor subtypes. Changes in metabolism of amino acids, including arginine, 176 proline, glutamate, histidine, alanine, and aspartate, were observed independent of tumor subtype 177 in all cancer versus non-malignant brain tissue comparisons (Figures 1C-1E). HGGs showed 178 specific changes in in the pentose-glucuronate interconversion pathway (Figure 1C), which were 179 driven by enrichment of levels of hexose monophosphates and glucuronic acid and depletion of 180 pentitols relative to non-malignant brain. Comparing non-malignant brain tissues with LGGs, 181 which are defined by the presence of IDH1 or IDH2 mutations1, glutathione metabolism ranked 182 among the top differentially regulated pathways (Figure 1D). This finding aligns with our prior 183 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 9 study showing that 2HG inhibits branched chain amino acid transaminase-dependent synthesis of 184 glutamate39, a precursor to glutathione, as well as others’ work describing an inverse correlation 185 between 2HG and glutathione abundance in human gliomas40,41. Finally, butanoate metabolism 186 differed markedly between brain metastases and non-malignant brain samples (Figure 1E), 187 principally reflecting lower levels of the neurotransmitter GABA in the former. This alteration 188 may be attributable to lower neuronal content of metastases relative to primary brain tumors 189 given that metastases tend not to display the same degree of diffuse infiltration and integration 190 into the brain that is characteristic of gliomas. 191 192 Guanidinoacetate accumulates in HGG and is associated with a bottleneck in the creatine 193 synthesis pathway 194 Accumulation of the oncometabolite 2HG is a unique and nearly universal metabolic hallmark of 195 IDH-mutant LGGs, as well as the HGGs that develop from them42. This discovery has opened 196 new avenues for diagnosing43,44 and monitoring45 IDH-mutant gliomas and prompted approval of 197 the mutant IDH1/2 inhibitor vorasidenib to treat these tumors by the U.S. Food and Drug 198 Administration. Considering these transformative effects on the neuro-oncology field, we 199 questioned whether similar patterns of recurrent metabolite accumulation may also be present in 200 predominantly IDH-wildtype HGGs. We rank-ordered metabolites according to the degree of 201 enrichment or depletion they display in HGG specimens compared to non-malignant brain 202 tissues (Figure 2A). We found that an intermediate in the creatine synthesis pathway, 203 guanidinoacetate (GAA), markedly accumulated in HGGs (Figures 2B and 2C). Upon extending 204 our analysis to all metabolites in this pathway, incorporating data from LGG samples, and 205 focusing on tumor samples from index surgical cases, we found that preferential upregulation of 206 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 10 GAA in HGG was not associated with a global increase in creatine or its products. GAA levels 207 were elevated ~100-fold in HGG versus non-malignant brain (Figure 2C). In contrast, creatine 208 and creatinine were decreased while phosphocreatine showed a similar, albeit statistically 209 insignificant, trend (Figures 2D-F). Interestingly, the effect size of GAA accumulation in HGG 210 was similar to 2HG enrichment in IDH-mutant gliomas when compared with relevant controls 211 (Figures 2C and S1F). 212 213 We next evaluated patterns of GAA accumulation across glioma subgroups. GAA, but not other 214 creatine synthesis pathway intermediates, was numerically but not significantly increased in 215 recurrent relative to index LGG samples (Figures S2A-D). These data imply that GAA 216 enrichment is associated with aggressive glioma disease and may manifest during progression of 217 some LGGs in addition to occurring broadly in HGGs. Stratifying HGGs by IDH mutational 218 status and gene expression programs, we found that GAA robustly and selectively accumulates 219 across all subgroups of Grade 4 IDH-mutant astrocytoma and GBM (Figures S2E-H). These 220 effects were unique to glioma because we observed <3-fold enrichment of GAA in lung 221 squamous tumors relative to benign adjacent lung tissues46, an effect that was associated with a 222 modest increase in tumoral creatine (Figures S2I-L). 223 224 To more directly address the prevalence of GAA upregulation in HGGs versus LGGs and non-225 malignant brain, we constructed receiver operating characteristic (ROC) curves describing the 226 relationship between GAA content and glioma presence and type in surgical specimens from 227 index and recurrent resections or those from index cases alone. GAA content served as a robust 228 tissue biomarker, effectively discriminating HGG tissue from LGG or non-malignant brain 229 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 11 comparators (Figures 2G, 2H, S2M, and S2N). In contrast, GAA levels weakly distinguished 230 LGG and non-malignant brain samples (Figures S2O and S2P). Our data align conceptually with 231 a prior study showing that GAA was elevated in the interstitial fluid of 8 adult-type HGGs47. The 232 statistical power provided by our collection of primary tissue samples establishes that GAA 233 accumulation is a highly recurrent metabolic hallmark of GBM tumors and likely of Grade 4 234 IDH-mutant astrocytomas as well. 235 236 Observing simultaneous enrichment of GAA and depletion of creatine in HGG suggested that the 237 creatine synthesis pathway may be reprogrammed in these tumors. To obtain a comprehensive 238 view of substrate-product interactions involving this pathway, we leveraged a platform for 239 nitrogen metabolism profiling recently developed by our group48. This platform supports parallel 240 stable isotope tracing of 31 distinct 15N-labeled nutrients in Human Plasma-like Medium49 241 (HPLM) in two cell cultures. We used this platform to probe global nitrogen metabolism 242 programs in NHA non-transformed, immortalized astrocytes and TS516 GBM glioma stem-like 243 cells (GSCs) (Figures 2I and 2J, Table S2). In the creatine synthesis pathway, arginine and 244 glycine serve as substrates for the enzyme arginine:glycine amidinotransferase (AGAT), which 245 produces both GAA and ornithine (Figure 2B). GAA is then methylated by the enzyme 246 guanidinoacetate N-methyltransferase (GAMT) to produce creatine in an S-adenosylmethionine 247 (SAM)-dependent manner. Ranking tracer-metabolite interactions based on differential labeling 248 in NHA and TS516 lines showed that 15N4-arginine-dependent labeling of GAA was strongly 249 enriched in GBM cells but 15N4-arginine-dependent labeling of creatine was not (Figure 2K). 250 These data indicate that HGGs harbor a bottleneck in the creatine synthesis pathway, displaying 251 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 12 robust AGA T-dependent GAA formation that is uncoupled from GAMT-dependent creatine 252 generation. 253 254 To further investigate this model and validate nitrogen metabolism profiling results, we 255 performed low-throughput tracing experiments with 15N4-arginine, 15N-glycine, 15N-serine (a 256 precursor to glycine), and 15N3-creatine in NHA and TS516 cultures. Only TS516 cells displayed 257 strong GAA labeling by 15N4-arginine, 15N-glycine, and 15N-serine (but not 15N3-creatine) tracers 258 (Figure 2L-2O). In contrast, neither NHA nor TS516 cells used 15N4-arginine, 15N-glycine, or 259 15N-serine tracers to produce creatine (Figures 2P-2R). Rather, both cell types predominantly 260 consumed extracellular 15N3-creatine to sustain their intracellular creatine pools (Figure 2S). 261 These data revealed an unexpected pattern of metabolism: TS516 GBM cells robustly synthesize 262 GAA but do not use it to produce creatine. Because GAMT is the only enzyme in human cells 263 known to use GAA as a substrate, we hypothesized that GBM cells may constitutively export 264 newly synthesized GAA. Indeed, we found that TS516 cells readily accumulate and secrete GAA 265 while NHA cells do not (Figures 2T and 2U). 266 267 Creatine synthesis pathway reprogramming and GAA secretion are recurring features of HGG 268 We next sought to evaluate the generalizability of GAA accumulation and secretion in HGG. We 269 assessed metabolite profiling data from two large independent collections of human brain tissue 270 specimens. The first included HGG, LGG, and non-malignant brain specimens (Figures 3A-3D) 271 while the second comprised only HGG and LGG samples (Figures 3E-3G). Using these 272 validation cohorts, we confirmed preferential upregulation of GAA, but not other creatine 273 synthesis pathway intermediates, in HGG versus LGG or non-malignant brain tissues. Therefore, 274 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 13 pervasive GAA accumulation in HGGs cannot be explained by our patient sample collection or 275 metabolite quantification approach. 276 277 Differential amino acid metabolism activities in TS516 and NHA cells (Figures 2I-2U) suggested 278 that GAA enrichment and creatine depletion in HGGs may be driven by cell-autonomous 279 metabolic alterations in tumor cells. If true, evidence of creatine synthesis pathway 280 reprogramming should be readily observable in diverse patient-derived GSC lines. To address 281 this possibility, we quantified intracellular GAA content, rates of GAA secretion, and 15N4-282 arginine-dependent labeling of GAA and creatine in a panel of IDH-mutant and IDH-wildtype 283 GSC lines, NHA non-transformed, immortalized astrocyte lines derived from two donors, and 284 non-transformed neural stem cell (NSC) lines (ENSA and NSC11). IDH-mutant GSC lines were 285 derived from Grade 4 IDH-mutant astrocytomas (MGG152, HK211, HK213 and HK252 cells) or 286 Grade 3 IDH-mutant and 1p/19q-codeleted oligodendrogliomas (BT054 and TS603 cells). All 287 IDH-wildtype GSC lines were derived from GBMs except for BT260 cells, which were 288 generated from a Grade 3 IDH-wildtype oligodendroglioma prior to the World Health 289 Organization’s update to CNS tumor classification criteria in 20211. All non-malignant NHA and 290 NSC lines displayed low GAA cellular content, negligible GAA secretion, and minimal de novo 291 creatine synthesis pathway activity (Figures 3H-3K). In contrast, the vast majority of GSC lines 292 demonstrated clear evidence of creatine synthesis pathway reprogramming, including GAA 293 accumulation and secretion as well as robust labeling of GAA, but not creatine, by 15N4-arginine. 294 The two exceptions were MGG152 and BT260 cells. Notably, HK308 GBM cells displayed 295 robust GAA synthesis and secretion but not intracellular accumulation. These data suggest that 296 GAA is rapidly exported by these cells upon synthesis. Importantly, differences in 15N4-arginine 297 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 14 metabolism between GSC lines and non-transformed neural cells were not confounded by 298 reduced tracer uptake in the latter (Figure S3). 299 300 These findings imply that HGGs display high rates of GAA secretion, but non-malignant brain 301 tissues do not. To test this prediction, we leveraged recent advances in organoid explant 302 modeling of non-malignant brain and brain tumor tissues50–53 to evaluate GAA secretion from 303 primary human brain tissues specimens. We created a series of human brain tissue explants from 304 non-malignant brain and GBM samples (Table S3). Next, we cultured these explants in a 305 customized formulation of HPLM50 immediately upon surgical resection. Then we collected 306 conditioned media and measured GAA released by each explant. As anticipated, GBM but not 307 non-malignant brain explants displayed robust GAA secretion (Figure 3L). To determine whether 308 GAA release is broadly observed in brain tissue from patients with GBM or if this is a specific 309 feature of the GBM microenvironment, we used stereotactic neurosurgical navigation to produce 310 spatially defined brain tissue explants from a patient who underwent surgical tumor resection of 311 a GBM. We established explants from normal-appearing non-malignant tissue, tumor-adjacent 312 edematous tissue marked by Fluid-Attenuated Inversion Recovery (FLAIR) hyperintensity, and 313 tissues from the tumor margin and tumor core (Figures 3M-Q). Explants with high tumor cell 314 content derived from the GBM margin and core displayed elevated GAA secretion relative to 315 FLAIR-positive and normal-appearing tissues with low tumor cell content. Considering data 316 from human tissue explants together with those from GSC lines, our findings indicate that HGG 317 cells synthesize GAA and release it into tumor interstitial fluid, leading to 10- to 100-fold 318 increases in GAA content in the HGG microenvironment. 319 320 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 15 Guanidinoacetate secretion is associated with imbalanced AGAT and GAMT expression 321 To begin to define the molecular mechanism underlying aberrant GAA secretion by HGG cells, 322 we evaluated expression of creatine synthesis pathway enzymes in human HGGs and non-323 malignant brain specimens by querying TCGA and GTEx RNA sequencing datasets (Figure 4A). 324 We found that AGA T expression was broadly upregulated while expression of GAMT and distal 325 enzymes in the pathway were downregulated in HGG tissues relative to non-malignant brain. 326 These differences were observed across tumor transcriptional subtypes and IDH genotypes. We 327 hypothesized that an increase in AGAT expression without a commensurate change in GAMT 328 may saturate GAMT catalytic activity and result in GAA accumulation and release from HGG 329 cells. This conceptual model would imply that the ratio of AGAT:GAMT expression may serve 330 as a biomarker of creatine synthesis pathway reprogramming and GAA enrichment in HGG. 331 Indeed, the AGAT:GAMT mRNA expression ratio was markedly elevated in HGGs versus non-332 malignant human brain tissues (Figure 4B). 333 334 These data provided a global portrait of creatine synthesis pathway gene expression in brain 335 tissues but did not address how AGAT and GAMT expression are regulated within discrete cell 336 populations. Therefore, we assessed AGAT and GAMT transcript abundance in a published 337 single-cell RNA sequencing dataset of human GBM. Tumor cells and oligodendrocytes displayed 338 the highest levels of AGAT expression and AGAT:GAMT ratios (Figures S4A-C). Myeloid cells 339 in the GBM microenvironment synthesize creatine and export it to neighboring tumor cells, 340 thereby supporting tumoral bioenergetics54. While AGAT and GAMT transcripts were expressed 341 by macrophage and microglia populations, we did not observe GAA secretion from cultured 342 microglia (Figure S4D). Taken together, these data and findings from others suggest that myeloid 343 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 16 cells in the HGG TME display canonical creatine synthesis pathway activity whereas tumor cells 344 repurpose this pathway to drive GAA secretion. Reinforcing this idea, evaluation of transcript- 345 and protein-level AGAT and GAMT expression in GSC lines and NHA cells revealed tumor cell-346 autonomous increases in AGAT abundance and the AGAT:GAMT ratio in nearly all GSC 347 cultures compared to astrocytes (Figures 4C-4F). Exceptions included MGG152 and BT260 cell 348 lines that do not synthesize or secrete GAA (Figures 3H-3K). 349 350 We next asked if the pervasive increase in AGAT transcription in HGG cells was associated with 351 changes in chromatin accessibility or transcription factor activity at the GATM genomic locus 352 (note that the GATM gene encodes the AGAT enzyme). We analyzed the expression distribution 353 of AGAT splice isoforms in AGAThigh (BT054, HK211, HK213, HK252, HK157, and HK308) 354 and AGATlow (MGG152 and BT260) GSC lines and in NHA cells (Figures S4E and S4F). 355 AGAThigh cells nearly exclusively expressed one transcript (ENST00000396659), which is 356 associated with promoter and enhancer elements surrounding exon 1 and an intragenic enhancer 357 located between exons 2 and 3 of the GATM gene. We performed ATAC-sequencing of AGAThigh 358 GSCs, AGATlow GSCs, and NHA cells (Figure 4G). AGAThigh GSCs displayed more open 359 chromatin at these promoter and enhancer sequences. Moreover, footprinting analysis 360 demonstrated increased transcription factor occupancy at these loci in AGAThigh GSCs. 361 Therefore, overexpression of the AGAT enzyme in HGGs is linked to chromatin remodeling 362 within the GATM gene and enhanced transcription factor recruitment to constituent cis-363 regulatory elements. 364 365 Imbalanced AGAT and GAMT activities are necessary and sufficient for GAA accumulation 366 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 17 Having shown that GAA accumulation and secretion correlate with an increase in the ratio of 367 AGAT to GAMT expression, we next asked if this relationship extends to primary tumor 368 samples. Using matched proteomics and metabolomics data from 50 human gliomas, we 369 conducted a correlation analysis, comparing protein-level AGAT:GAMT expression with tumoral 370 GAA content (Figure 5A). We observed a positive association between these properties that 371 tracked with tumor grade. 372 373 To test whether increased AGAT activity relative to GAMT causes GAA accumulation, we 374 transduced natively AGATlow NHA cells with wildtype AGAT, a catalytically dead AGA T point 375 mutant (C407A55), or an empty vector control (Figure 5B). To further modulate the ratio of 376 AGAT to GAMT activities, we expressed Cas9 and one of two GAMT-targeted sgRNAs or a 377 control sgRNA targeting the AAVS1 safe harbor locus. Overexpressing wildtype AGA T was 378 sufficient to increase GAA levels (Figure 5C) and triggered robust labeling of GAA, but not 379 creatine, by 15N4-arginine (Figures 5D, 5E, and S5A). These effects were dependent on the 380 catalytic activity of AGA T because they were abolished by the C407A point mutation. Increasing 381 the ratio of AGAT to GAMT activity further by simultaneously overexpressing wildtype AGA T 382 and knocking out GAMT evoked higher levels of GAA accumulation relative to AGAT 383 overexpression alone. Therefore, imbalanced expression and activity of AGAT and GAMT 384 enzymes is sufficient to explain GAA enrichment in human HGGs. 385 386 To ask if imbalanced AGAT and GAMT activities are necessary to sustain high cellular GAA 387 levels, we further engineered NHA lines stably expressing AGAT wildtype or C407A mutant 388 enzymes or an empty vector control. We transduced these lines to overexpress wildtype GAMT, 389 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 18 a catalytically dead GAMT point mutant (E45S56), or an empty vector control (Figure 5F). 390 Increasing GAMT activity in AGAT-overexpressing cells depleted intracellular GAA (Figure 391 5G). This effect was associated with substantial 15N4-arginine labeling of both GAA and creatine 392 pools (Figures 5H, 5I, and S5B). In all experiments, the catalytically dead mutants of AGAT and 393 GAMT failed to recapitulate effects of the respective wildtype enzyme. These studies show that 394 low GAMT activity is required for GAA to accumulate upon AGAT upregulation. Moreover, our 395 findings indicate that GAMT occupies a critical node in the creatine synthesis pathway, 396 ultimately dictating whether AGA T activity leads to creatine production or GAA accumulation 397 and secretion. 398 399 GAA-dependent GABAA agonism and chloride dysregulation drive neuronal activity in the 400 HGG microenvironment 401 Recurrent enrichment of GAA in HGGs but not index LGGs or non-malignant brain tissues 402 (Figures 2C, 3A, and 3E) suggested that this metabolite may play a functional role in promoting 403 aggressive brain tumor behavior. Because HGGs display a bottleneck in the creatine synthesis 404 pathway, we posited that GAA’s role in tumor promotion may be distinct from its canonical role 405 as a creatine precursor. GAA is known to accumulate broadly in tissues of patients with GAMT 406 deficiency, an inborn error of creatine metabolism caused by loss-of-function mutations in the 407 GAMT gene (Figure 6A). These patients suffer from intractable seizures, a symptom that is not 408 commonly observed in patients with other inborn errors of creatine metabolism that do not result 409 in GAA accumulation57. Seizure promotion suggests that GAA elicits excitatory actions in the 410 brain. Indeed, preclinical studies have shown that GAA regulates neuronal electrochemical 411 signaling by stimulating GABAA receptors58,59 as well as inhibiting synaptic glutamate uptake60 412 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 19 and Na+/K+ ATPase activity61. These properties were intriguing because patients with HGG, like 413 those with GAMT deficiency, also experience seizures, cognitive dysfunction, and neurological 414 deficits. Glioma-infiltrated tissues are more electrically active than non-malignant brain tissues28, 415 and microenvironmental neuronal activity in turn drives glioma growth through neuron-to-416 glioma synapses and activity-regulated signaling26–28,33,36,62,63. Although glutamate release by 417 tumor cells represents one process by which glioma cells augment the activity of local 418 neurons28,35,64, we do not have a comprehensive understanding of the molecular mechanisms 419 driving electrical activity in the HGG microenvironment. Thus, we hypothesized that GAA 420 secretion by HGG cells may represent a critical yet unappreciated mechanism of glioma-neuron 421 crosstalk. 422 423 To better understand the pathophysiological relevance of GAA accumulation in the contexts of 424 GAMT deficiency and HGG, we quantified concentrations of GAA in tissue specimens from 425 HGG and non-malignant brain samples, comparing these values with reported GAA 426 concentrations in brain tissues of Gamt-wildtype and Gamt-knockout mice (Figures 6B and 6C). 427 Although control samples showed species-specific numerical differences in GAA content, we 428 observed ~100-fold increases in GAA levels in brain tissues from both GAMT-knockout mice 429 and HGG patients relative to these controls. Therefore, it is plausible that GAA accumulation in 430 the HGG microenvironment surpasses a threshold that elicits neuronal excitation in the setting of 431 GAMT deficiency. To further elucidate local GAA concentrations in human GBM, we again 432 leveraged stereotactic neurosurgical navigation to collect normal-appearing, FLAIR-positive, and 433 tumor margin and core tissue samples from a patient undergoing GBM resection (Figure 6D-6H). 434 We found that GAA levels sharply rose with increasing proximity to the tumor core, an effect 435 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 20 that was associated with an increase in the ratio of AGAT to GAMT protein expression (as 436 measured by validated immunohistochemistry assays, Figures S6A and S6B). Critically, GAA 437 content of the tumor margin, where interactions between tumor cells and neurons are heightened, 438 was orders of magnitude higher than normal-appearing brain tissue. 439 440 We investigated mechanisms through which GAA may contribute to increased electrochemical 441 signaling in the glioma microenvironment. GAA bears striking structural similarity to the 442 neurotransmitter GABA (Figure 6I), a neurotransmitter than can be hyperpolarizing and 443 inhibitory or depolarizing and excitatory, depending on the intracellular concentration of 444 chloride. Whole-cell voltage-clamp recordings (Vhold = 0 mV) from CA1 pyramidal neurons in 445 acute mouse hippocampal slices revealed that exogenous GAA application (but not creatine) 446 induced a large outward current that was completely blocked by the GABAA receptor inhibitor 447 gabazine (GBZ) (Figures 6J and 6K). Similarly, current-clamp recordings from cortical 448 pyramidal neurons in acute slices from adolescent murine brain revealed that GAA reduced 449 neuron input resistance, consistent with ion channel opening (Figures 6L and 6M). Co-treatment 450 with blockers of GABAergic, but not glutamatergic, receptors fully prevented the decrease in 451 input resistance. GABAA receptor blockade explained this effect, because treating slice cultures 452 with the GABAA-selective inhibitor picrotoxin alone was sufficient to completely reverse the 453 activity of GAA on input resistance/channel opening (Figures 6L and 6M). Together, these 454

support the function of GAA as a GABAA receptor agonist, opening the GABAA receptor 455 channel. 456 457 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 21 Our findings presented an apparent paradox: GAA engaged GABA receptors and elicited 458 hyperpolarization of mature murine neurons but also accumulates dramatically in the 459 hyperexcitable HGG microenvironment. Intracellular chloride ion concentrations ([Cl-]i) in 460 neurons dictate whether GABA receptor agonism is depolarizing or hyperpolarizing. NKCC1 461 and KCC2 are chloride transporters that play important roles in importing and exporting Cl- from 462 neurons, respectively. Healthy adult neurons express high levels of KCC2 and low levels of 463 NKCC1, thereby reducing [Cl-]i and resulting in chloride influx and neuronal hyperpolarization 464 upon GABA receptor opening (Figure 6N). Interestingly, prior research has shown that GBM-465 infiltrated brain displays hyperexcitability in part due to dysregulation of [Cl-]i homeostasis65,66. 466 Neurons in the GBM-infiltrated brain display reduced expression of KCC2, increased chloride 467 levels, and depolarizing (excitatory) GABA signaling (Figure 6O). Therefore, we asked if GAA 468 accumulation in HGGs cooperates with tumor-specific [Cl-]i dysregulation to drive excitatory 469 GABAergic neurotransmission in glioma-associated neurons. We established orthotopic TS516 470 xenografts in immunodeficient mice displaying tumor infiltration throughout one hemisphere but 471 not the other. In acute cortical slices from these mice, high-density multi-electrode array (MEA) 472 analysis revealed that GAA wash-on led to an increase in neuronal firing in tumor-infiltrated 473 brain tissue but not in the contralateral hemisphere (Figures 6P, 6Q, and S6C). Treating tumor-474 infiltrated brain slices with the KCC2 activator CLP257 to decrease [Cl-]i abolished the 475 excitatory effect of GAA in the tumor microenvironment (Figure 6R). In agreement with prior 476 work65,66, neuronal cells exhibited markedly lower expression of KCC2 in the brain hemisphere 477 infiltrated by TS516 GBM cells (marked by human antigen Ku80) compared to the contralateral 478 side (Figures 6S and 6T). Together, our findings establish that interstitial GAA accumulation 479 promotes neuronal activity in gliomas in a GABAA receptor- and [Cl-]i -dependent manner. 480 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 22 481 AGAT inhibition represses electrochemical signaling and glioma growth 482 Neuronal activity plays an important role in driving glioma cell proliferation26. Therefore, we 483 sought to determine whether inhibiting AGAT-dependent GAA synthesis and consequent 484 neuronal activation attenuates GBM growth. To genetically repress AGA T, we sequentially 485 transduced TS516 GBM GSCs with adenoviral expression vectors to transiently express Cas9 486 and two sgRNAs targeting the GATM gene (which encodes the AGAT enzyme) or two sgRNAs 487 targeting the AAVS1 safe harbor locus. We established isogenic AGAT-expressing (AGA T WT) 488 and AGAT-knockout (AGAT KO) bulk TS516 cell cultures (Figure 7A). This approach allowed 489 us to achieve nearly complete silencing of AGA T in the latter without resorting to single cell 490 cloning, which can be challenging in GSC lines that display long doubling times. As expected, 491 AGAT knockout reversed intracellular GAA accumulation and GAA secretion (Figures 7B and 492 7C). Although AGAT inhibition did not affect the growth of TS516 monocultures in vitro (Figure 493 7D), it did prolong survival of mice bearing TS516 orthotopic GBM xenografts (Figure 7E), 494 underscoring the central role of the neural tumor microenvironment to this mechanism. To assess 495 the effects of AGA T knockout on tumor metabolite content, we developed assays to assess tissue 496 and GAA and creatine content (Figure S7A, S7B). The survival benefit seen in AGA T KO 497 tumors was associated with depletion of tumoral GAA but not creatine (Figures 7F and 7G), 498 consistent with minimal de novo creatine synthesis in HGG cells (Figure 3K). These data 499 indicate that GAA accumulation drives HGG growth in a microenvironment-dependent manner. 500 Importantly, this oncogenic function of AGAT activity in glioma cells cannot be explained by 501 canonical effects on creatine synthesis. 502 503 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 23 We next evaluated whether AGA T knockout blocks neuronal hyperactivity in the HGG 504 microenvironment. We quantified nuclei expressing cFOS+, an immediate early gene indicative 505 of recent neuronal activity, in cortical sections of tumor-infiltrated brain tissues harvested from 506 mice bearing AGAT WT or AGA T KO xenografts. Electrochemically active cells marked by 507 cFOS expression were enriched in GAA-replete, AGAT WT glioma xenograft tissues and 508 reduced in GAA-depleted, AGAT KO comparators (Figures 7H-7J and S7C). The majority of 509 cFOS+ cells in the microenvironment of AGA T WT xenografts also expressed NeuN (Figures 510 S7C and S7D), affirming their neuronal identity. Consistently, we also observed increased levels 511 of vGLUT1, a marker of excitatory synapses, in tumor-infiltrated brain regions of mice with 512 AGAT WT versus AGAT KO xenografts (Figures 7K-7M, S7E, and S7F), suggesting that local 513 hyperexcitability induced by GAA promotes interactions between glutamatergic neurons and 514 tumor cells in the HGG microenvironment. 515 516

from studies involving genetic manipulation of AGAT support therapeutically targeting 517 AGAT for HGG treatment but do not outline a tractable means to do so. Interestingly, a dietary 518 therapy that attenuates AGA T activity and GAA accumulation is effective in reducing neuronal 519 hyperactivity and seizure incidence in patients with GAMT deficiency67,68. This diet features 520 reduced arginine content and high levels of ornithine, which reduce AGAT substrate availability 521 and promote end-product inhibition of AGAT, respectively. Given the efficacy of this diet in 522 GAMT deficiency, we tested its therapeutic activity against HGGs. We first modeled the impact 523 of ornithine supplementation on AGAT function by measuring 15N4-arginine labeling of GAA 524 and found that high levels of ornithine reduced AGAT activity in cultured TS516 cells (Figure 525 7N). Next, we tested tolerability and antitumor efficacy of an arginine-depleted and ornithine-526 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 24 supplemented diet (hereafter the “GAMT deficiency diet”). Mice administered the GAMT 527 deficiency diet did not display weight loss nor reduced chow consumption relative to mice fed a 528 standard diet (Figures S7G and S7H). Critically, placing mice bearing TS516 orthotopic 529 xenografts on the GAMT deficiency diet prolonged their survival relative to mice given a 530 standard diet (Figure 7O). The survival effect size of the GAMT deficiency diet was similar to 531 the effect size associated with genetic inhibition of AGAT (Figure 7E). Therefore, this dietary 532 therapy may constitute an effective, clinically validated approach to block aberrant AGA T 533 activity and oncogenic GAA signaling in patients with HGG. 534 535

536 We generated a deep metabolite profiling dataset that spanned primary brain tumors, brain 537 metastases, and non-malignant brain specimens to discover GAA accumulation as a recurrent 538 metabolic feature of HGG that manifests independently of canonical genetic and transcriptomic 539 disease subtypes. We went on to validate this finding in two independent cohorts. Unlike GAA, 540 creatine and creatine derivatives were downregulated or unchanged in HGGs compared to non-541 malignant brain tissues. Stable isotope tracing studies revealed that these metabolic changes were 542 tied to a bottleneck in the de novo creatine synthesis pathway in GSC lines that functionally 543 disconnects GAA production from creatine synthesis. Our findings reveal conceptual parallels 544 between accumulation of GAA in HGGs and enrichment of the oncometabolite (R)-2HG in IDH-545 mutant gliomas. Both metabolites reach levels in glioma tissues that are at least 1-2 orders of 546 magnitude higher than those in non-malignant brain and display enrichment profiles that map 547 onto clinically relevant glioma features. Specifically, increased GAA content serves as a 548 surrogate marker of tumor grade while (R)-2HG accumulation indicates the presence of an IDH1 549 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 25 or IDH2 mutation. These properties suggest that GAA, like (R)-2HG, may have utility as a 550 glioma tissue biomarker in diagnostic and therapeutic applications, including spectroscopic 551 imaging and intraoperative mass-spectrometry based margin detection43,44,69–71. From a 552 biological standpoint, GAA and (R)-2HG share another feature: both metabolites function as 553 signaling molecules rather than serving as substrates for anabolic or bioenergetic processes in 554 tumor cells. These insights reinforce the importance of metabolite signaling in the molecular 555 pathogenesis of glioma. 556 557 Findings from our work and prior research indicate that the creatine synthesis pathway is 558 regulated in a highly cell type-specific manner in HGGs. Tumor-associated macrophages and 559 microglia provide a local supply of creatine to GBM cells that promotes tumor tolerance to 560 microenvironmental stressors, including hypoxia and acidification54. This interaction may allow 561 HGG cells to repurpose the creatine synthesis pathway to produce and secrete GAA, a finding 562 from our work that aligns with observation of GAA accumulation in HGG interstitial fluid in a 563 previous study47. Our data demonstrate that GAA secretion by HGG cells is explained by AGAT 564 upregulation in the absence of a commensurate increase in GAMT. RNA/i1 seq comparisons of 565 human HGG and reference brain showed broad AGAT upregulation with reduced expression of 566 GAMT and distal enzymes across transcriptional subtypes and IDH genotypes. In 567 patient/i1 derived GSCs, AGAT levels and the ratio of AGAT:GAMT expression rose relative to 568 non-malignant neural cells, which was associated with increased chromatin accessibility and 569 transcription factor occupancy at key cis-regulatory elements in the GATM gene. In primary 570 tumors, the ratio of AGA T:GAMT protein expression positively correlated with tissue GAA 571 content and tumor grade. Gain/i1 and loss/i1 of/i1 function studies support this model: AGAT 572 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 26 overexpression raised cellular GAA and 15N/i1 arginine flux into the GAA pool in a catalytic 573 activity/i1 dependent fashion, while concurrent GAMT knockout or overexpression amplified or 574 reversed these effects, respectively. Therefore, we propose that imbalanced AGAT and GAMT 575 activities saturate the catalytic capacity of GAMT to convert GAA to creatine, leading to GAA 576 buildup and secretion from HGG cells. 577 578 Human genetics and pathophysiology underscore the clinical relevance of GAA accumulation. In 579 GAMT deficiency, an inborn error of creatine metabolism marked by systemic GAA 580 accumulation, patients present with seizures, cognitive impairment, and psychomotor deficits72 581 that mirror the clinical phenotype commonly seen in glioma. GAA has been shown to regulate 582 neuronal activity by serving as an agonist for GABAA receptors58,59 and repressing Na+/K+ 583 A TPase activity60 and synaptic glutamate uptake61. The standard therapy for GAMT deficiency 584 involves restricting arginine (an AGAT substrate) and supplementing creatine and ornithine (an 585 end-product AGAT inhibitor). This treatment causes GAA depletion, electroencephalogram 586 normalization, and decreased seizure frequency in GAMT deficiency patients67,68. These results 587 are not achieved by creatine supplementation alone and intractable seizures are uncommon in 588 other inborn errors of creatine metabolism caused by loss of function mutations in genes 589 encoding the SLC6A8 creatine transporter or AGAT50, thereby reinforcing the link between GAA 590 accumulation and neuronal hyperactivity. We found that inhibiting GAA synthesis by genetic or 591 dietary approaches constrained HGG growth in vivo: CRISPR knockout of AGAT in orthotopic 592 GBMs decreased tumoral GAA and markers of neuronal activity and neuron-glioma interactions 593 and extended host survival without altering the growth of in vitro GBM cell monocultures. 594 Inspired by clinical management of GAMT deficiency, we modeled an arginine/i1 restricted, 595 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 27 ornithine/i1 supplemented diet. After confirming that ornithine accumulation inhibited GAA 596 synthesis in vitro, we tested this diet in mice and found that it was well tolerated and 597 recapitulated effects of genetic AGAT inhibition on survival of glioma-bearing mice. These 598 observations underscore the neuroactivity of GAA and suggest that dietary interventions used in 599 the clinical management of GAMT deficiency may display therapeutic activity in patients with 600 HGG. 601 602 Advances in cancer neuroscience demonstrate that gliomas elevate neuronal activity of glioma-603 infiltrated circuits to further augment neuron-glioma interactions and accelerate tumor growth26–604 28,31,33,62,63. Concordantly, tumor-infiltrated brain tissues are more electrically active than healthy 605 tissue31. However, the mechanisms through which glioma cells enhance neuronal firing are only 606 beginning to be elucidated. Release of glutamate28,34 and synaptogenic factors36,37,62 are 607 implicated in this process, but the full complement of molecules involved in glioma-to-neuron 608 communication is still emerging. Our data implicate GAA accumulation as a novel mechanism of 609 neuronal activation in HGG. GAA is structurally similar to GABA, a classically inhibitory 610 neurotransmitter. Our results in acute cortical and hippocampal brain slices confirm previous 611 studies that GAA is a GABAA receptor agonist58,59. Recent work has shown that GABAergic 612 signaling promotes progression of diffuse midline gliomas73 and that [Cl-]i dysregulation 613 converts GABAergic signaling from inhibitory to excitatory within the GBM 614 microenvironment65,66. Mechanistically, these changes in GBM are tied to KCC2 downregulation 615 in tumor-resident neurons. We therefore hypothesized that chloride dysregulation and GAA 616 accumulation cooperate to drive excitatory GABAergic signaling in HGG-infiltrated brain. In 617 glioma/i1 bearing brain slices, GAA preferentially increased firing within tumor/i1 infiltrated 618 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 28 regions. This effect of GAA was abrogated with pretreatment of the KCC2 activator CLP-257. 619 Finally, we observed diminished KCC2 expression and a GAA-dependent increase in 620 electrochemical signaling markers in xenograft HGG tissues. These findings support a model in 621 which extracellular GAA produced by HGG cells activates GABAA currents that are excitatory 622 under an aberrant local chloride set/i1 point. 623 624 Together, our results define how GAA becomes enriched in HGGs, establish its origin in tumor 625 cells, connect its accumulation to neuronal hyperexcitability, and outline a tractable treatment 626 strategy to disrupt metabolite signaling that drives neuronal activity-dependent glioma growth. 627 628

629 This study was supported by National Institutes of Health (NIH) grants R01CA258586 and 630 R01CA289260 to S.K.M. and K.G.A., R01NS142141 to S.K.M., P50CA165962 and 631 U19CA264504 to S.K.M., K99CA277576 to Y .X., and R35CA220449 to R.J.D. This work was 632 also supported by awards from Oligo Nation to S.K.M. and K.G.A., Cancer Prevention and 633 Research Institute of Texas (CPRIT) grants RR190034 and RP230344 to S.K.M., CPRIT grant 634 RP2400489 to S.K.M. and B.L., a Distinguished Scientist Award from the Sontag Foundation to 635 S.K.M., and by a Human Frontier Science Program (HFSP) postdoctoral fellowship award 636 (LT0018/2022-L) to Y .X. D.D.S. was supported by NIH/NCI K12CA0903354, a Burroughs 637 Wellcome Career Award for Medical Scientists, and a Lubin Family Foundation Scholar Award. 638 L.G.Z., T.P.M., and the Children’s Research Institute Metabolomics facility are supported by 639 CPRIT grant RP240494. M.R.S. was supported by NIH/NCI grant F30CA271634. K.G.A 640 received funding from CPRIT RP210140 as well as NIH grant P30CA047904 awarded to UPMC 641 Hillman Cancer Center. S.A.W. is supported by CPRIT Training Grant RP210041. This project 642 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 29 used the service of the University of Pittsburgh Small Molecule Biomarker Core facility, which 643 was supported, in part, by the University of Pittsburgh Office of the Senior Vice Chancellor, 644 Health Sciences, and NIH grants S10RR023461 and S10OD028540. This project used Zeiss 645 Axioscan 7 in the Whole Brain Microscopy Facility (WBMF) at the UT Southwestern Medical 646 Center, which is supported by the NIH grant 1S10OD032267-01 (to Denise Ramirez). Some data 647 used were generated by the TCGA Research Network (https://www.cancer.gov/tcga) or by the 648 Genotype-Tissue Expression (GTEx) Project, which was supported by the Common Fund of the 649 Office of the Director of the National Institutes of Health, and by NCI, NHGRI, NHLBI, NIDA, 650 NIMH, and NINDS. Some figures were constructed using BioRender. The authors wish to 651 express their gratitude to all patients who contributed to this study. 652 653 AUTHOR CONTRIBUTIONS 654 Conceptualization: K.G.A. and S.K.M. 655 Methodology: K.G.A., K.M., C.K.E., S.A.W., Y .X., M.R.S., M.T.G., Y .-T.H., J.I.T., D.D.S., 656 M.T., N.M., V .T.P., R.E.W.III, A.K., T.D.N., B.J.K., N.M.C., C.W., D.R.M., M.A.G., T.P.M., 657 J.R.G., K.M.H. and S.K.M. 658 Investigation: K.M., C.K.E., S.A.W., Y .X., M.R.S., M.T.G., S.E.K., Y .-T.H., J.I.T., K.G., 659 W.H.H., M.T., M.M.L., M.E.-S., B.C.S., V .T.P., P.K., T.S., C.L., L.G.Z., F.C. and S.A. 660 Formal analysis: K.M., C.K.E., S.A.W., Y .X., M.R.S., M.T.G., Y .-T.H., J.I.T., L.G., N.M., K.J.H., 661 T.E.R. and S.K.M. 662 Data curation: S.A.W., Y .X., M.R.S., B.J.K., N.M.C. and C.W. 663 Software: M.R.S. and L.G. 664 Validation: K.M., C.K.E., S.A.W. and Y .X. 665 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 30 Resources: K.G.A., T.R.P ., B.C.L., P.O.Z., D.R.M., M.A.G., M.A.C., L.X. and O.H.K. 666 Supervision: K.G.A., P.R., S.M., T.D.N., D.R.M., M.A.G., L.X., J.-A.L., K.M.H., S.C., R.J.D. 667 and S.K.M. 668 Project administration: K.G.A., S.O., P.K. and S.K.M. 669 Writing-Original draft: K.G.A., K.M., C.K.E., S.A.W., X.Y . and S.K.M. 670 Writing-Review & Editing: K.G.A., C.K.E., S.A.W., Y .X., M.R.S., D.D.S., T.D.N., M.A.C., 671 B.L., M.M., K.M.H., S.C., R.J.D. and S.K.M. 672 Funding acquisition: B.L., K.G.A. and S.K.M. 673 674 DECLARATION OF INTERESTS 675 S.K.M. receives research funding from Servier Pharmaceuticals. S.K.M. and K.G.A. have 676 intellectual property interests related to brain tumor metabolism and are co-founders of Gliomet. 677 T.E.R. has received consulting fees from Servier Pharmaceuticals, which is unrelated to the 678 current work. R.J.D. is a founder and advisor at Atavistik Bioscience, and an advisor for Vida 679 Ventures and Faeth Therapeutics. 680 681 FIGURE LEGENDS 682 Figure 1. Metabolomic stratification of human brain tissues. (A) Profiling of polar 683 metabolites and lipids (n=563) in tissue specimens from adult HGG (n=39), LGG (n=25), brain 684 metastases (n=18), and non-malignant brain (“Brain”; n=9). Metabolite peak intensities were 685 log2-transformed and median normalized. Data are stratified by k-means clustering with k=7. (B) 686 Partial least-squares discriminant analysis (PLS-DA) of metabolite content of patient tissue 687 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 31 samples grouped by sample class. (C-E) Metabolite set enrichment analyses in (C) HGG, (D) 688 LGG, or (E) metastases relative to non-malignant brain tissue. See also Figure S1. 689 690 Figure 2. Identification of guanidinoacetate accumulation as a metabolic hallmark of high-691 grade glioma. (A) Waterfall plot of relative metabolite abundance in HGG vs non-malignant 692 brain samples. (B) Schema depicting the creatine synthesis pathway. (C-F) Peak intensities of 693 (C) guanidinoacetate (GAA), (D) creatine, (E) phosphocreatine, and (F) creatinine in index 694 samples of HGG (n=26), LGG (n=15), and non-malignant (“Brain”, n=9). (G-H) ROC analysis 695 of GAA content as a discriminant marker of index HGG tissue (n=26) compared to (G) non-696 malignant brain tissue (n=9) or (H) index LGG tissue (n=15). (I-J) Sankey diagrams depicting 697 labeling of 15N-labeled tracers (top bars) in nitrogen metabolism pathway intermediates in 698 immortalized astrocytes (NHA Donor #1) (I) or TS516 GSCs (J). (K) Waterfall plot of all tracer 699 to metabolite labeling interations ranked by differential labeling score. (L-S) Fractional 700 enrichment of label from indicated tracer in (L-O) GAA or (P-S) creatine following incubation of 701 NHA Donor #1 or TS516 cells with 15N-labeled tracer for 18 hours (n=3 per cell line). (T-U) 702 Intracellular (T) and secreted (U) GAA content of NHA Donor #1 and TS516 cultures at 48 703 hours in HPLM (n=3 per cell line). Data are means ± SEM. n.s.= not significant, n.d. = not 704 detected, *p<0.05, **p<0.01, ***p<0.001 (Welch’s t-test). See also Figure S2. 705 706 Figure 3. HGG cells display a bottleneck in the creatine synthesis pathway and GAA 707 secretion. (A-D) Peak intensities of (A) GAA, (B) creatine, (C) phosphocreatine, and (D) 708 creatinine in Validation Cohort #1 of HGG (n=28), LGG (n=9), and non-malignant brain (n=6) 709 tissue samples. (E-G) Peak intensities of (E) GAA, (F) creatine, and (G) creatinine in Validation 710 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 32 Cohort #2 of HGG (n=9) and LGG (n=28) human surgical specimens. (H-I) (H) Intracellular 711 GAA content or (I) GAA secretion in non-malignant cells (n=4), IDH-mutant GSCs (n=6), and 712 IDH WT GSCs (n=7) following 48-hour culture in HPLM (n=3 per cell line). (J-K) Fractional 713 enrichment of label from 15N4-arginine in (J) GAA and (K) creatine in the same cell line panel 714 following 18-hour incubation in tracer-containing HPLM (n=3 per cell line). (L) GAA secretion 715 following 18-hour culture of explanted human largely non-malignant neocortex (brain, n=7) or 716 GBM (n=5) tissues in HPLM. (M-Q) Spatially resolved GAA secretion in human GBM. (M) 717 Explant GAA secretion in samples generated from four distinct points within a single GBM 718 corresponding to (N) largely non-malignant brain (“Normal”), (O) FLAIR-enhancing tumor-719 adjacent neocortex (“FLAIR”), (P) tumor margin (“Margin”), and (Q) tumor core (“Core”) as 720 assessed on intraoperative magnetic resonance imaging (MRI). Scale bar = 20 µm. Data are 721 means (A-D and H-L) or medians (E-G) ± SEM. n.d. = not detected. n.s.= not significant, 722 *p<0.05, **p<0.01, ***p<0.001 (Welch’s t-test in A-D, Mann-Whitney test in E-G, 723 Kolmogorov-Smirnoff test in L). See also Figure S3. 724 725 Figure 4. Imbalanced expression of AGAT and GAMT is associated with GAA 726 accumulation in glioma. (A) RNAseq analysis of transcript levels of AGAT, GAMT, and 727 creatine kinase isoforms (CKMT2, CKB, CKMT1A/B) in HGG surgical specimens (n=151) and 728 non-malignant brain specimens (n=255). Data are z-scores of log2(transcripts per million +1) 729 values. Data for HGG samples were reanalyzed from the Cancer Genome Atlas (TCGA)74. Data 730 for non-malignant brain specimens were reanalyzed from the Genome Tissue Expression Project 731 (GTEX)75,76. (B) Relative expression of AGAT and GAMT, quantified as ratio of transcript 732 abundances for a given sample, in HGG and non-malignant brain RNAseq from TCGA and 733 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 33 GTEX. (C-D) RNAseq analysis of AGAT and GAMT transcript abundance in a panel of IDH 734 mutant GSCs (n=6), IDH WT GSCs (n=7), and non-malignant (n=1) cells (n=3 per cell line). 735 Data are fragments per kilobase of transcript per million mapped reads (FPKM). (E) 736 Representative immunoblots of AGAT and GAMT expression in a panel of IDH mutant GSCs 737 (n=6), IDH WT GSCs (n=7), and non-malignant (n=4) cells. (F) Densitometry analysis of the 738 ratio of AGAT and GAMT protein expression from (E). Data are presented as means. (G) 739 Aggregated chromatin accessibility and transcription factor footprint tracks from AGAThigh 740 GSCs (n=6), AGATlow GSCs (n=2), and NHA Donor #1 cells. ***p<0.001 (Welch’s t-test). See 741 also Figure S4. 742 743 Figure 5. Imbalanced AGAT and GAMT expression is necessary and sufficient for GAA 744 accumulation. (A) Correlation of tumor GAA content with AGAT and GAMT expression 745 profiles from Validation Cohort #2. AGAT : GAMT Scores and GAA abundance values are 746 derived from proteomics and metabolomics analyses of primary tissues, respectively (n=50). (B) 747 Immunoblot analyses of NHA Donor #1 cells engineered to express either empty vector (EV), 748 wild-type AGAT cDNA (AGAT WT), or catalytically-dead AGAT cDNA (AGAT C407A) and 749 Cas9 with an sgRNA targeting either the AAVS1 safe-harbor locus control or one of two sgRNAs 750 targeting GAMT (n=3), (C) Intracellular GAA content of engineered NHA Donor #1 cells 751 following 48-hour culture in HPLM (n=9 for all lines except AGAT WT, sgGAMT #1 cells, for 752 which n=8). (D-E) Fractional enrichment of cellular (D) GAA and (E) creatine by 15N4-arginine 753 tracer in engineered NHA Donor #1 cells following 18-hour incubation in HPLM (n=3 per line). 754 (F) Immunoblot analyses of NHA cells engineered to express either empty vector (EV), wild-755 type AGAT (AGAT WT), or catalytically-dead AGAT cDNA (AGAT C407A) and either EV, 756 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 34 wild-type GAMT cDNA (GAMT WT), or catalytically-dead GAMT cDNA (GAMT E45S) 757 (n=3). (G) Intracellular GAA content of engineered NHA Donor #1 lines following 48-hour 758 culture in HPLM (n=3 for all lines). (H-I) Fractional enrichment of cellular (H) GAA and (I) 759 creatine by 15N4-arginine tracer in engineered NHA Donor #1 cell lines following 18-hour 760 incubation in HPLM (n=3 per line). Data are means ± SEM. n.s.= not significant, *p<0.05, 761 **p<0.01, ***p<0.001 (Welch’s t-test). See also Figure S5. 762 763 Figure 6. GAA accumulation in the glioma microenvironment activates GABAergic 764 signaling. (A) Schema depicting metabolic alterations and phenotypes in GAMT deficiency and 765 HGG. (B-C) Tissue GAA levels in (B) a murine model of GAMT deficiency (n=10) and 766 littermate controls (n=6) or (C) human non-malignant brain (n=7) and HGG (n=9) tissues. In 767 (B), data are from Schmidt et al77. (D-H) Spatially resolved GAA content in human GBM. (D) 768 GAA content in samples taken from four distinct points within a single GBM corresponding to 769 (E) largely non-malignant brain (“Normal”), (F) FLAIR-enhancing tumor-adjacent neocortex, 770 (G) tumor margin, and (H) tumor core as assessed on intraoperative magnetic resonance imaging 771 (MRI). (E-H) MRI, hematoxylin and eosin (H&E), and immunohistochemistry (IHC) for AGAT 772 and GAMT of sampled tissues in (D). Scale bar = 50 µm. (I) Chemical structures of GAA and γ-773 aminobutyric acid (GABA). (J) Representative traces of patch-clamp recordings of neurons in 774 murine brain slices treated with 100 µM GAA or creatine in 0 µM or 10 µM of the GABAA 775 receptor inhibitor gabazine. (K) Integrated area under the curve (AUC) of patch-clamp traces for 776 slices treated with 100 µM GAA (n=5), 100 µM GAA + 10 µM gabazine (n=6), or 100 µM 777 creatine (n=5). (L-M) Normalized input resistance of neurons treated with GAA alone (n=3), 10 778 µM of the AMPA receptor inhibitor 3-(2-carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP) 779 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 35 and 20 µM of the NMDA receptor inhibitor 6,7-dinitroquinoxaline-2,3-dione (DNQX; E 780 blockers, n=5), 100 µM of the GABAA receptor inhibitor picrotoxin and 2 µM of the GABAB 781 receptor inhibitor CGP52432 (I blockers, n=5), all inhibitors (E+I blockers, n=8), or picrotoxin 782 alone (n=8). In (L), data depict changes in normalized input resistance over time following GAA 783 wash-on. In (M), data depict average post-wash-on normalized input resistance. Data are 784 normalized to average input resistance of each neuron prior to GAA wash-on. (N-O) Schemas 785 depicting maintenance of [Cl-]i by NKCC1 and KCC2 in (N) the brain under normal physiologic 786 conditions and (O) HGGs that display pathologic KCC2 downregulation. (P-R) Multielectrode 787 array (MEA) assay of neuronal activity in response to GAA treatment in tumor-bearing brain 788 slices. Spike rates normalized to mean pre-treatment spiking rates, prior to and following 789 treatment with 20 µM GAA for (P) tumor-infiltrated right hemisphere (n=3 slices), (Q) tumor-790 free contralateral cortex (n=3 slices), and (R) tumor-infiltrated neocortex treated with 30 µM 791 CLP-257 (n=3 slices). (S-T) Immunofluorescence analysis of cell markers and KCC2 expression 792 in the tumor microenvironment. (S) Representative images of markers Ku80 (human GBM 793 cells), cFOS (neuronal activity), and DAPI (nuclei) as well as KCC2 expression in tumor-794 infiltrated and contralateral neocortex. (T) Quantification of KCC2 expression in tumor-795 infiltrated and contralateral neocortex (n=3 mice). Data are means ± SEM. n.s. = not significant, 796 *p<0.05, **p<0.01, ***p<0.001 (Welch’s t-test in B-C, unpaired t-test in M and T, paired t-test 797 in P-R). See also Figure S6. 798 799 Figure 7. GAA accumulation drives neuronal hyperexcitability and glioma aggressiveness. 800 (A) Representative immunoblots of AGAT, GAMT, and GAPDH in AGAT wild-type (WT) and 801 AGAT knockout (AGAT KO) TS516 GSCs. (B) Intracellular GAA in AGAT WT and AGAT 802 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 36 KO TS516 GSCs after 48-hour culture in HPLM (n=3 per cell line). (C) GAA secretion by 803 AGAT WT and AGAT KO TS516 GSCs during 48-hour culture in HPLM (n=3 per cell line). 804 (D) In vitro growth assay of AGAT WT and AGAT KO TS516 GSCs (n=3 per time point per 805 cell line). Growth curves are generated by nonlinear exponential regression. (E) Overall survival 806 of mice bearing AGAT WT or AGAT KO TS516 xenografts (n=10 per arm). (F-G) Absolute (F) 807 GAA and (G) creatine content for tumors from (E) (n=7 for AGAT WT, n=8 for AGAT KO). 808 (H-I) Representative H&E and cFOS immunofluorescence for (H) AGAT WT and (I) AGAT 809 KO TS516 tumors. Scale bars = 1 mm, 500 µm, or 50 µm. (J) Quantification of cFOS+ nuclei in 810 peritumoral cortex in AGAT WT and AGAT KO TS516 tumors (n=6 per tumor type). (K-L) 811 Immunofluorescent imaging of AGAT WT and AGAT KO TS516 tumors. Scale bars = 20 µm. 812 (M) Quantification of vGLUT1 positive foci in predetermined regions of interest within tumor 813 beds of AGAT WT and AGAT KO TS516 tumors (n=7 per genotype). (N) Intracellular content 814 of the M+2 isotopologue of GAA following 15N4-arginine tracing in the presence or absence of 815 150 µM ornithine supplementation (n=3 per condition). (O) Survival of TS516 xenograft bearing 816 mice receiving a standard (n=10) or an ornithine-supplemented, arginine-restricted (GAMT 817 deficiency, n=10) diet. Data are means ± SEM. n.s.= not significant, *p<0.05, **p<0.01 818 (Welch’s t-test in B, C, and J; log-rank test in E and O; unpaired t-test in F, G, and M). See also 819 Figure S7. 820 821 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 37 STAR METHODS 822 823 RESOURCE A V AILABILITY 824 Lead Contact 825 • Further information and requests for resources and reagents should be directed to and will 826 be fulfilled by the Lead Contact, Samuel K. McBrayer 827 ([email protected]). 828 829

Availability 830  Plasmids generated in this study will be deposited to Addgene prior to publication. 831 832 Data and Code Availability 833  Metabolomics data will be deposited to the National Metabolomics Data Repository 834 (NMDR) and will be publicly available as of the date of publication. Accession numbers 835 will be listed in the key resources table. All other data reported in this paper will be 836 shared by the Lead Contact upon request. 837  This paper does not report original code. 838  Any additional information required to reanalyze the data reported in this paper is 839 available from the Lead Contact upon request. 840 841 Key Resources Table 842 REAGENT or RESOURCE SOURCE IDENTIFIER Antibodies Rabbit monoclonal anti-GAPDH (D16H11) Cell Signaling Technology Cat# 5174S, RRID: AB_10622025 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 38 Mouse monoclonal anti-VINCULIN Sigma Cat# V9131, RRID: AB_477629 Mouse monoclonal anti-AGAT Abcam Cat# 119269, RRID: AB_10902241 Rabbit polyclonal anti-AGAT ATLAS Cat# HPA026077, RRID: AB_1849528 Rabbit polyclonal anti-GAMT Proteintech Cat# 10880-1-AP, RRID: AB_2109304 Rabbit polyclonal anti-GAMT Thermo Fisher Scientific Cat# PA5-119778, RRID: AB_2913350 Rabbit monoclonal anti-cFOS Cell Signaling Technology Cat# 2250, RRID: AB_2247211 Chicken monoclonal anti-cFOS Synaptic Systems Cat# 226 009, RRID: AB_2943525 Rabbit monoclonal anti-KU80 Cell Signaling Technology Cat# 2180, RRID: AB_2218736 Guinea pig polyclonal anti-VGLUT1 Millipore Cat# AB5905, RRID: AB_2301751 Mouse monoclonal anti-KCC2 Millipore Cat# MABN88, RRID: AB_11213323 Goat polyclonal anti-Mouse IgG (H+L), HRP Thermo Fisher Scientific Cat# 31430, RRID: AB_228307 Goat polyclonal anti-Rabbit IgG (H+L), HRP Thermo Fisher Scientific Cat# 31460, RRID: AB_228341 Alexa Fluor 488 AffiniPure donkey anti-rabbit IgG (H+L) Jackson ImmunoResearch Cat# 711-546-152, RRID:AB_2340619 Alexa Fluor 488 donkey anti-chicken IgΥ Invitrogen Cat# A78948 Alexa Fluor 594 goat anti-rabbit IgG (H+L) Invitrogen Cat# A11012, RRID:AB_2534079 Alexa Fluor 647 AffiniPure donkey anti-guinea pig IgG (H+L) Jackson ImmunoResearch Cat# 706-606-148, RRID:AB_2340477 Alexa Fluor 647 donkey anti-rabbit IgG (H+L) Invitrogen Cat# A32795 Bacterial and virus strains HB101 Competent Cells Promega Cat# L2015 XL10-Gold Ultracompetent Cells Agilent Cat# 200315 Biological samples 8656 nonmalignant human brain tissue sample This study N/A 6377 nonmalignant human brain tissue sample This study N/A 3218 nonmalignant human brain tissue sample This study N/A 3242 nonmalignant human brain tissue sample This study N/A 7346 nonmalignant human brain tissue sample This study N/A 6518 nonmalignant human brain tissue sample This study N/A 7768N nonmalignant human brain tissue sample This study N/A 2739 human glioblastoma tissue sample This study N/A 1207 human glioblastoma tissue sample This study N/A 4564 human glioblastoma tissue sample This study N/A 8025 human glioblastoma tissue sample This study N/A 3280 human glioblastoma tissue sample This study N/A 4617 human glioblastoma tissue sample This study N/A 7757 human glioblastoma tissue sample This study N/A 8547 human glioblastoma tissue sample This study N/A 7768T human glioblastoma tissue sample This study N/A (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 39 5852 spatially-resolved human glioblastoma tissue sample This study N/A 1068 nonmalignant brain explant Savani et al48 N/A 3649 nonmalignant brain explant Savani et al48 N/A 4875 nonmalignant brain explant This study N/A 6777 nonmalignant brain explant Savani et al48 N/A 8995 nonmalignant brain explant This study N/A 9230 nonmalignant brain explant Savani et al48 N/A 9413 nonmalignant brain explant Savani et al48 N/A 1232 glioblastoma brain explant Savani et al48 N/A 3313 glioblastoma brain explant Savani et al48 N/A 5565 glioblastoma brain explant Savani et al48 N/A 7602 glioblastoma brain explant Savani et al48 N/A 9041 glioblastoma brain explant Savani et al48 N/A Chemicals, peptides, and recombinant proteins Recombinant Human Epidermal Growth Factor GoldBio Cat# 1150-04 Recombinant Human Fibroblast Growth Factor 2 GoldBio Cat# 1140-02 Heparin STEMCELL Technologies Cat# 07980 Penicillin/Streptomycin Thermo Fisher Cat# 15140148 Amphotericin B GeminiBio Cat# 400104 Plasmocin InvivoGen Cat# ant-mpp L-Glutamine Thermo Fisher Cat# 25030081 B27 Supplement Thermo Fisher Cat# 17504044 N2 supplement Thermo Fisher Cat# 17502048 BenchMark Fetal Bovine Serum GeminiBio Cat# 100-106 Accutase STEMCELL Technologies Cat# 07922 Gabazine EMD Millipore Cat# SR95531 Guanidinoacetic acid Sigma-Aldrich Cat# G11608 Guanidinoacetic acid TCI Cat# G0167 Creatine monohydrate Thermo Scientific Cat# B25009.22 CLP-257 Bio-Techne Cat# 5242/10 3-(2-Carboxypiperazin-4-yl)propyl-1-phosphonic acid (CPP) Hello Bio Cat# HB0036 CGP 52432 Tocris Cat# 1246 6,7-dinitroquinoxaline-2,3-dione (DNQX) Tocris Cat# 2312 Picrotoxin Hello Bio Cat# HB0506 Esp3I FastDigest Thermo Fisher Cat# FD0454 BsmBI-v2 New England Biolabs Cat# R0739 TransIT Transfection Reagent Mirus Bio Cat# MIR2 B-304 LentiX Concentrator Takara Cat# 631232 PacI New England Biolabs Cat# R0547S Polybrene MedChemExpress Cat# HY-112735 G418 Sulfate GoldBio Cat# G-418 Hygromycin GoldBio Cat# H-270-EZ50 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 40 Glucose Thermo Fisher Cat# 15023021 (R)-(–)-3-hydroxybutyric acid sodium salt Millipore Sigma Cat# 298360 2-hydroxybutyric acid sodium salt Millipore Sigma Cat# 220116 Acetone Millipore Sigma Cat# AX0120 Ammonium chloride Millipore Sigma Cat# A9434 Betaine Millipore Sigma Cat# 61962 Calcium chloride Millipore Sigma Cat# C5670 Calcium nitrate tetrahydrate Millipore Sigma Cat# C1396 Citric acid Millipore Sigma Cat# 251275 Creatine Millipore Sigma Cat# C0780 Creatinine Millipore Sigma Cat# C4255 D-(–)-fructose Millipore Sigma Cat# F3510 D-(+)-galactose Millipore Sigma Cat# G5388 Formic acid Millipore Sigma Cat# 1002640100 Glycerol Millipore Sigma Cat# G2025 Glycine Millipore Sigma Cat# G7126 Hypoxanthine Millipore Sigma Cat# H9377 L-(–)-malic acid Millipore Sigma Cat# M7397 L-2-aminobutyric acid Millipore Sigma Cat# A2536 L-alanine Millipore Sigma Cat# A7627 L-arginine monohydrochloride Millipore Sigma Cat# A5131 L-asparagine Millipore Sigma Cat# A0884 L-aspartic acid Millipore Sigma Cat# A9256 L-carnitine hydrochloride Millipore Sigma Cat# C0283 L-citrulline Millipore Sigma Cat# C7629 L-cysteine hydrochloride Millipore Sigma Cat# C1276 L-cystine Millipore Sigma Cat# C8755 L-glutamic acid Millipore Sigma Cat# G1251 L-glutamine Millipore Sigma Cat# G3126 L-glutathione reduced Millipore Sigma Cat# G6013 L-histidine monohydrochloride monohydrate Millipore Sigma Cat# H5659 L-isoleucine Millipore Sigma Cat# I2752 L-leucine Millipore Sigma Cat# L8000 L-lysine monohydrochloride Millipore Sigma Cat# L5626 L-methionine Millipore Sigma Cat# M9625 L-ornithine monohydrochloride Millipore Sigma Cat# O2375 L-phenylalanine Millipore Sigma Cat# P2126 L-proline Millipore Sigma Cat# P0380 L-serine Millipore Sigma Cat# S4500 L-threonine Millipore Sigma Cat# T8625 L-tryptophan Millipore Sigma Cat# T0254 L-tyrosine Millipore Sigma Cat# 93829 L-valine Millipore Sigma Cat# V0500 Magnesium chloride Millipore Sigma Cat# M8266 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 41 Magnesium sulfate Millipore Sigma Cat# M2643 Malonic acid Millipore Sigma Cat# M1296 N-acetylglycine Millipore Sigma Cat# A16300 O-acetyl-L-carnitine hydrochloride Millipore Sigma Cat# A6706 Phenol red sodium salt Millipore Sigma Cat# P5530 Potassium chloride Millipore Sigma Cat# P5405 RPMI 1640 vitamins solution Millipore Sigma Cat# R7256 Sodium acetate Millipore Sigma Cat# S5636 Sodium bicarbonate Millipore Sigma Cat# S5761 Sodium chloride Millipore Sigma Cat# S7653 Sodium L-lactate Millipore Sigma Cat# L7022 Sodium phosphate dibasic heptahydrate Millipore Sigma Cat# S9390 Sodium pyruvate Millipore Sigma Cat# P2256 Succinic acid Millipore Sigma Cat# S3674 Taurine Millipore Sigma Cat# T0625 trans-4-hydroxy-L-proline Millipore Sigma Cat# H5534 Urea Millipore Sigma Cat# U5378 Uric acid Millipore Sigma Cat# U2625 α -ketoglutaric acid Millipore Sigma Cat# K1128 Ammonium chloride (15N) Cambridge Isotope Laboratories Cat# NLM-467 Hypoxanthine (15N4) Cambridge Isotope Laboratories Cat# NLM-8500 L-alanine (15N) Cambridge Isotope Laboratories Cat# NLM-454 L-arginine hydrochloride (15N4) Cambridge Isotope Laboratories Cat# NLM-396 L-asparagine hydrochloride (15N2) Cambridge Isotope Laboratories Cat# NLM-3286 L-aspartic acid (15N) Cambridge Isotope Laboratories Cat# NLM-718 L-citrulline (ureido-15N1) Cambridge Isotope Laboratories Cat# NLM-6850 L-creatine (15N3) Cambridge Isotope Labs Cat# NLM-9218 L-cysteine (15N) Cambridge Isotope Laboratories Cat# NLM-2295 L-cystine (15N2) Cambridge Isotope Laboratories Cat# NLM-3818 L-glutamic acid (15N) Cambridge Isotope Laboratories Cat# NLM-135 L-glutamine (15N2) Cambridge Isotope Laboratories Cat# NLM-1328 L-glycine (15N) Cambridge Isotope Laboratories Cat# NLM-202 L-histidine hydrochloride monohydrate (15N3) Cambridge Isotope Laboratories Cat# NLM-1513 L-isoleucine (15N) Cambridge Isotope Laboratories Cat# NLM-292 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 42 L-leucine (15N) Cambridge Isotope Laboratories Cat# NLM-142 L-lysine dihydrochloride (15N2) Cambridge Isotope Laboratories Cat# NLM-1554 L-methionine (15N) Cambridge Isotope Laboratories Cat# NLM-752 L-ornithine hydrochloride (15N2) Cambridge Isotope Laboratories Cat# NLM-3610 L-phenylalanine (15N) Cambridge Isotope Laboratories Cat# NLM-108 L-proline (15N) Cambridge Isotope Laboratories Cat# NLM-835 L-serine (15N) Cambridge Isotope Laboratories Cat# NLM-2036 L-threonine (15N) Cambridge Isotope Laboratories Cat# NLM-742 L-tryptophan (15N2) Cambridge Isotope Laboratories Cat# NLM-800 L-tyrosine (15N) Cambridge Isotope Laboratories Cat# NLM-590 Sodium nitrate (15N) Cambridge Isotope Laboratories Cat# NLM-157 Taurine (15N) Millipore Sigma Cat# 491330 Urea (15N2) Cambridge Isotope Laboratories Cat# NLM-233 Uric acid (1,3-15N2) Cambridge Isotope Laboratories Cat# NLM-1697 Uridine (15N2) Cambridge Isotope Laboratories Cat# NLM-812 L-valine (15N) Cambridge Isotope Laboratories Cat# NLM-316 L-glutamine (amide-15N) Cambridge Isotope Laboratories Cat# NLM-557 Dialyzed Fetal Bovine Serum GeminiBio Cat# 100-108 tert-Butyl methyl ether Millipore Sigma Cat# 650560 L-glutamine (13C5) Cambridge Isotope Laboratories Cat# CLM-1822 Chloramphenicol Millipore Sigma Cat# C0378 Isopropanol Thermo Fisher Cat# A461 SPLASH™ LIPIDOMIX™ Sph/Cer mix standard Avanti Research Cat# 330707 L-ornithine monohydrochloride Millipore Sigma Cat# O2375 Acetonitrile (Optima LC/MS Grade) Thermo Fisher Cat# A9554 1,2-¹³C₂ , 3-¹/i3 N-guanidinoacetic acid Cambridge Isotope Labs Cat# CNLM-8300 2,2-D2-guanidinoacetic acid Cambridge Isotope Labs Cat# DLM-9998 Guanidineacetic acid Sigma-Aldrich Cat# G11608 Methanol Thermo Fisher Cat# A456 Creatine-d5 H2O (N-methyl-d3; glycine-2,2-d2) LGC Standards Cat# TRC-C781496- 50MG (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 43 Guanidineacetic-2,2-d2 Acid CDN Isotopes Cat# D-6320 Creatine LGC Standards Cat# TRC-C781483 Guanidinoacetic Acid LGC Standards Cat# TRC-G821250- 25G Hibernate A BrainBits Cat# HA RBC Lysis Buffer Thermo Fisher Cat# 00433357 cOmplete Mini Protease Inhibitor Cocktail Millipore Sigma Cat# 11836153001 DAPI Cayman Chemical Cat# 14285 Sucrose Sigma Aldrich Cat# S9378 Sodium bicarbonate Thermo Fisher Cat# BP328 Potassium chloride Millipore Sigma Cat# P9333 D-glucose Millipore Sigma Cat# G8270 Sodium phosphate monobasic Sigma Aldrich Cat# S8282 Magnesium chloride Millipore Sigma Cat# M9272 Calcium chloride Millipore Sigma Cat# 223506 Sodium chloride Millipore Sigma Cat# S9888 Choline chloride Millipore Sigma Cat# C1879 Sodium bicarbonate Millipore Sigma Cat# S6014 D-glucose Millipore Sigma Cat# G8270 Ascorbic acid Millipore Sigma Cat# A92902 Potassium chloride Millipore Sigma Cat# P5405 Sodium phosphate dibasic Millipore Sigma Cat# S9638 Sodium pyruvate Millipore Sigma Cat# P5280 Magnesium chloride Millipore Sigma Cat# M8266 Calcium chloride Millipore Sigma Cat# C7902 Kynurenic acid Millipore Sigma Cat# K3375 Sodium chloride Millipore Sigma Cat# S7653 Sucrose Thermo Fisher Cat# S5 Potassium chloride Thermo Fisher Cat# P330 Sodium phosphate monobasic Thermo Fisher Cat# S397 Sodium bicarbonate Thermo Fisher Cat# P233 D-glucose Gibco Cat# 15023-021 Calcium chloride Thermo Fisher Cat# 012316.A1 Magnesium sulfate Thermo Fisher Cat# 033337.36 Cesium methanesulfonate Millipore Sigma Cat# C1426 HEPES Millipore Sigma Cat# H3375 Adenosine 5′ -triphosphate magnesium salt Millipore Sigma Cat# A9187 Guanosine 5′ -triphosphate sodium salt hydrate Millipore Sigma Cat# G8877 Phosphocreatine di(tris) Millipore Sigma Cat# P7936 EGTA Millipore Sigma Cat# E3889 Tetrodotoxin Hello Bio Cat# HB1034 Potassium gluconate Millipore Sigma Cat# P1847 Phosphocreatine di(tris) salt Millipore Sigma Cat# P1937 HEPES Millipore Sigma Cat# H4034 Guanosine 5′ -triphosphate sodium salt solution Millipore Sigma Cat# G3776 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 44 Sodium chloride Thermo Fisher Cat# P271 GlutaMAX Thermo Fisher Cat# 35050061 B27 Supplement minus Vitamin A Thermo Fisher Cat# 12587010 2-mercaptoethanol Thermo Fisher Cat# BP176100 Insulin Millipore Sigma Cat# I9278 Critical commercial assays e-Myco PCR Detection Kit Bulldog Bio Cat# 2523348 e-Myco PLUS PCR Detection Kit Bulldog Bio Cat# 25233 MycoAlert Mycoplasma Detection Kit Lonza Cat# LT07-318 Q5 Hot Start High-Fidelity 2X Master Mix New England Biolabs Cat# M0494 RNeasy Mini Kit QIAGEN Cat# 74004 QIAquick Gel Extraction Kit QIAGEN Cat# 28706 Monarch PCR & DNA Cleanup Kit New England Biolabs Cat# T1030 Gateway BP Clonase II Thermo Fisher Cat# 11789100 Gateway LR Clonase II Thermo Fisher Cat# 11791020 In-Fusion HD Cloning Kit Takara Bio Cat# 102518 Virabind Adenovirus Miniprep Kit Cell Biolabs Cat# VPK-099 Adeno-X GoStix Takara Cat# 632270 Lenti-X GoStix Plus Takara Cat# 631280 Zymo Clean and Concentrator-5 Kit Zymo Research Cat# D4013 NEBNext 2x MasterMix New England Biolabs Cat# M0541S RNeasy Plus Universal Kit Qiagen Cat# 73404 Kapa RNA HyperPrep Kit with RiboErase Roche Cat# KK8561 Leica Bond Polymer Refine Detection Kit Leica Cat# DS9800,RRID:AB_2 891238 Deposited data Metabolomics analysis of human brain tumor tissue specimens This study To be deposited prior to publication Metabolomics analysis of TS516 and NHA Donor #1 cells in nitrogen metabolism profiling platform This study To be deposited prior to publication TCGA-GBM RNA sequencing analysis of human GBM tissue TCGA https://www.cancer.g ov/tcga, https://portal.gdc.can cer.gov/projects/TC GA-GBM GTEx RNA sequencing analysis of human brain tissue GTEX https://www.ncbi.nlm .nih.gov/projects/gap /cgi- bin/study.cgi?study_i d=phs000424.v10.p 2 RNA sequencing of glioma stem-like cells and NHA Donor #1 cells Wu et al7, Savani et al50 The Multi-Layered Transcriptional Architecture of Glioblastoma Ecosystems; single-cell RNA sequencing of human glioblastoma Nomura et al78 https://www.ncbi.nlm .nih.gov/geo/query/a cc.cgi?acc=GSE274 546 Experimental models: Cell lines (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 45 NHA Donor #1 (immortalized with HPV E6 and E7 and hTERT) Sonoda et al.79; Provided by R. Pieper N/A NHA Donor #2 (immortalized with HPV E6 and E7 and hTERT) Shi et al11. N/A HEK293T ATCC Cat# CRL-3216, RRID: CVCL_0063 293AD CellBioLabs Cat# AD-100, RRID: CVCL_KA63 BT054 Kelly et al.80; Provided by S. Weiss N/A BT260 Koivunen et al6; provided by K. Ligon N/A HK157 Laks et al.81; Provided by H. Kornblum N/A HK211 Laks et al.81; Provided by H. Kornblum N/A HK213 Laks et al.81; Provided by H. Kornblum N/A HK252 Laks et al.81; Provided by H. Kornblum N/A HK308 Laks et al.81; Provided by H. Kornblum N/A TS516 Rohle et al.82; Provided by I. Mellinghoff N/A TS603 Rohle et al.82; Provided by I. Mellinghoff N/A UTSW5 This study N/A UTSW63 Nguyen et al83 N/A UTSW71 Nguyen et al83 N/A ENSA Wang et al84 N/A NSC11 Wang et al84 N/A MGG152 Wakimoto et al85 N/A HMC3 ATCC Cat# CRL-3304, RRID: RRID:CVCL_II76 Experimental models: Organisms/strains C57BL/6J mice; 5-7 weeks old Jackson Labs Cat# 000664, RRID:IMSR_JAX:00 0664 C57BL/6J mice; P18-23 Jackson Labs Cat# 000664, RRID:IMSR_JAX:00 0664 ICR-SCID mice; 11 weeks old, female Taconic Cat# ICRSC-F ICR-SCID mice; 7 weeks old, female Taconic Cat# ICRSC-F ICR-SCID mice; 8 weeks old, female Taconic Cat# ICRSC-F Oligonucleotides sgAAVS1 #1: GTCACCAATCCTGTCCCTAG Chen et al86 N/A sgAGAT #1: TGTGGGCAATGAGATTATCG This study N/A sgGAMT #1: GGCCAGCGCGTGCATATAGG This study N/A (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 46 sgGAMT #2: GTACGACACGTACCCACTCT This study N/A sgAAVS1 #2: TGTTAGGCAGATTCCTTATC Chen et al86 N/A sgAGAT #2: ACTTCAATGACCAGTCAATG This study N/A attB1_U6_fwd: GGGGACAAGTTTGTACAAAAAAGCAGGCTCTGAGG GCCTATTTCCCATG This study N/A bGHpolyA_attB2_Rev_v2: GGGGACCACTTTGTACAAGAAAGCTGGGTTCCATA GAGCCCACCGCAT This study N/A GATM_C407A_InvPCR_F: TTCCATGCCTGGACCTGCGATGTC This study N/A GATM_C407A_InvPCR_R: GGTCCAGGCATGGAAGCCTCCT This study N/A attB_AGAT_F: GGGGACAAGTTTGTACAAAAAAGCAGGCTCTGCCA CCATGCTGCGGGT This study N/A attB_AGAT_R: GGGGACCACTTTGTACAAGAAAGCTGGGTTTCAGTC CAAGTAGGACTGTAAGGTGCC This study N/A attB_GAMT_F: GGGGACAAGTTTGTACAAAAAAGCAGGCTCTGCCA CCATGAGCGCCCC This study N/A attB_GAMT_R: GGGGACCACTTTGTACAAGAAAGCTGGGTTTCAGC CTTTGGTCACCAGGGG This study N/A Recombinant DNA lentiCRISPRv2-GFP Addgene 82416 lentiCRISPRv2-mCherry Addgene 99154 pDONR223 Thermo Fisher N/A pAd/PL-DEST Thermo Fisher V49420 pENTR221-AGAT WT from Ultimate ORF Human Clone Library Life Technologies Cat# HORF01; RRID:SCR_005371 pLenti-EF1α -DEST-IRES-Neo Savani et al48 Addgene, Cat# pending pLenti-EF1α -EV-IRES-Neo Savani et al48 Addgene, Cat# pending pLenti-EF1α -AGAT_WT-IRES-Neo This study Addgene, To be deposited prior to publication pLenti-EF1α -AGAT_C407A-IRES-Neo This study Addgene, To be deposited prior to publication pLenti-EF1α -EV-PGK-Hygro This study Addgene, To be deposited prior to publication pLenti-EF1α -GAMT_WT-PGK-Hygro This study Addgene, To be deposited prior to publication pLenti-EF1α -GAMT_E45S-PGK-Hygro This study Addgene, To be deposited prior to publication psPAX2 Addgene Cat# 12260 pMD2.G Addgene Cat# 12259 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 47 Software and algorithms El-MAVEN (0.12.0) Thermo Fisher https://elucidata.io/el -maven/, RRID: SCR_022159 TraceFinder (5.1 SP2) Thermo Fisher Cat# OPTON-31001, RRID: SCR_023045 Multiquant v2.1.1 Sciex https://sciex.com/pro ducts/software/multi quant-software R (4.4.2) The R Project for Statistical Computing https://www.r- project.org/, RRID: SCR_001905 RStudio (2024.09) Posit Software https://posit.co/produ cts/open- source/rstudio/, RRID: SCR_000432 AccuCor (0.3.0) Su et al.87 https://github.com/lp arsons/accucor, RRID: SCR_023046 Metabolomics and Stable Isotope Tracing Postprocessing Script Savani et al48 Deposition pending GraphPad Prism (10.4.1) GraphPad https://www.graphpa d.com/scientific- software/prism/, RRID: SCR_002798 Astrocyte 2.2.0 UT Southwestern Bioinformatics Core n/a ImageJ Schneider et al88 RRID:SCR_003070 Trimgalore v 0.6.4 Krueger et al89 https://github.com/Fe lixKrueger/TrimGalor e, RRID:SCR_011847 BWA Li et al90 https://bio- bwa.sourceforge.net/ , RRID:SCR_010910 Sambamba Tarasov et al91 http://www.open- bio.org/wiki/Sambam ba.Sambamba v0.5.0, RRID:SCR_024328 SamTools Danecek et al92 https://www.htslib.or g/, RRID:SCR_002105 bedtools Quinlan et al93 https://github.com/ar q5x/bedtools2, RRID:SCR_006646 PhantomPeakQualTools Kharchenko et al94, Landt et al https://code.google.c om/p/phantompeakq ualtools/, RRID:SCR_005331 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 48 MACS2 Zhang et al95 https://github.com/m acs3-project/MACS, RRID:SCR_013291 Seurat Satija et al96 https://satijalab.org/s eurat/, RRID:SCR_016341 STAR v2.7.3a Dobin et al97 https://github.com/al exdobin/STAR, RRID:SCR_004463 Subread v1.6.3 Liao et al98 https://subread.sourc eforge.net/, RRID:SCR_009803 GSVA H/i3 nzelmann et al99 https://www.biocond uctor.org/packages/r elease/bioc/html/GS VA.html, RRID:SCR_021058 Clampex v11.3 Molecular Devices https://www.molecul ardevices.com/produ cts/axon-patch- clamp- system/acquisition- and-analysis- software/pclamp- software-suite, RRID: SCR_011323 Clampfit v11.4 Molecular Devices https://www.molecul ardevices.com/produ cts/axon-patch- clamp- system/acquisition- and-analysis- software/pclamp- software-suite, RRID: SCR_011323 Igor Pro v9.05 WaveMetrics https://www.waveme trics.com/, RRID:SCR_000325 Labview 8.6 National Instruments https://www.ni.com/e n/support/downloads /software- products/download.l abview.html#570679 , RRID: SCR_014325 BrainWave 5 3Brain https://www.3brain.c om/products/softwar e/brainwave5 Tidyverse Wickham et al100 https://www.tidyvers e.org/, RRID:SCR_019186 Other NeuroCult NS-A Basal Medium (Human) with Proliferation Supplement STEMCELL Technologies Cat# 05751 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 49 Neurobasal Medium Thermo Fisher Cat# 21103049 DMEM Thermo Fisher Cat# 11995-065 EMEM ATCC Cat# 30-2003 Clear Flat Bottom Ultra-Low Attachment 6-well Plates Corning Cat# 3471 Clear Flat Bottom Ultra-Low Attachment 24-well Plates Corning Cat# 3473 100 mm Ultra-Low Attachment Culture Dishes Corning Cat# 4615 FACS Aria II SORP 4-Laser BD Biosciences Cat# N/A SpeedVac Thermo Fisher Cat# SPD2030 TissueLyser II QIAGEN RRID: SCR_018623 MultiTherm Shaker with heating and cooling Benchmark Scientific Cat# H5000-HC ViCELL XR Cell Viability Analyzer Beckman Coulter RRID: SCR_019664 Q-Exactive HF-X Thermo Fisher RRID: SCR_020425 Orbitrap Exploris 480 Thermo Fisher RRID: SCR_027000 Vanquish Flex UHPLC Thermo Fisher N/A AB QTRAP 5500 Applied Biosystems Sciex RRID: SCR_020517 AB QTRAP 6500+ Applied Biosystems Sciex RRID:SCR_021831 Nexera X2 LC-30AD HPLC Shimadzu N/A NextSeq 500 System Illumina RRID: SCR_014983 Leica Bond RX Automated Stainer Leica RRID:SCR_025548 BOND Epitope Retrieval Solution 1 Leica Cat# AR9961 BOND Epitope Retrieval Solution 2 Leica Cat #AR9640 Tissue-Tek Film® Automated Coverslipper Sakura Finetek Cat# 4740, 4743 Leica Aperio AT2 Scanner Leica Microsystems RRID:SCR_021256 NanoZoomerS60 Digital Slide Scanner Hamamatsu RRID:SCR_023762 Zeiss LSM880 with Airyscan Confocal Laser Scanning Microscope Zeiss RRID:SCR_020925 VT1200S Vibratome Leica Cat# N/A; RRID:SCR_018453 VT1000P Vibratome Leica N/A Borosilicate glass capillary Sutter Instrument Cat# BF150-86- 10HP P-1000 micropipette puller Sutter Instrument RRID:SCR_021042 Multiclamp 700B Microelectrode Amplifier Molecular Devices RRID:SCR_018455 Digidata 1550B Molecular Devices N/A Multiclamp 700A Microelectrode Amplifier Molecular Devices RRID:SCR_021040 BioCam DupleX with CorePlate™ 3D 3Brain N/A ISO Pro Rodent 3000 Diet Purina Prolab Cat# 5P76 Teklad Irradiated Global 16% Protein Rodent Diet Inotiv Cat# 2916 Teklad Amino Acid Diet Inotiv Cat# TD.01084 Teklad 0.665% Ornithine, Arginine-restricted diet Inotiv Cat# TD.240691 24-Well Ultra-Low Adherence Plate Corning Cat# 3473 843 Experimental model and study participant details 844 Human Subjects 845 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 50 The study was conducted according to the principles of the Declaration of Helsinki. Patient 846 tissue and blood were collected following ethical and technical guidelines on the use of human 847 samples for biomedical research after informed patient consent under an Institutional Review 848 Board (IRB)-approved protocol. All patient samples were de-identified before processing. All 849 patient samples and organoid explants were diagnosed and graded according to the 2021 WHO 850 Classification of Tumors of the Central Nervous System (CNS), 5th edition1. Patient samples for 851 the Discovery Cohort were collected under a protocol approved by the IRB of the University of 852 Texas Southwestern Medical Center (STU 022011-070). Patient samples for Validation Cohort 853 #1 were collected under a protocol approved by the IRB of Northwestern University (17-048-854 CDH). Patient samples for the Validation Cohort #2 are part of a collection of specimens 855 associated with a study supported by the NCI CPTAC group and are described in a manuscript in 856 preparation. Patient samples used to generate organoid explants were collected under a protocol 857 approved by the IRB of the University of Pittsburgh (STUDY19080321). Information related to 858 primary samples can be found in Tables S1 and S3. 859 860 Cell lines 861 NHA Donor #1 cells (non-malignant human astrocytes, immortalized with human papillomavirus 862 E6, E7, and hTERT; sex unknown)79 were obtained from R. Pieper at the University of 863 California in San Francisco. NHA Donor #2 cells were generated from commercially available 864 astrocytes (Lonza CC-3187) as described previously11. HEK293T cells (female, ATCC CRL-865 3216) and 293AD cells (female, CellBioLabs AD-100) were obtained commercially. BT05480 866 (female) were provided by S. Weiss at the University of Calgary. BT2606 cells (sex unknown) 867 were provided by K. Ligon at Dana Farber Cancer Institute. HK157 (female), HK211 (female), 868 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 51 HK213 (male), HK252 (male), and HK308 (female) cells were provided by H. Kornblum at the 869 University of California in Los Angeles81. TS516 (sex unknown, RRID: CVCL_A5HY) and 870 TS603 (sex unknown, RRID: CVCL_A5HW) were provided by I. Melinghoff at Memorial 871 Sloan-Kettering Cancer Center82. ENSA (sex unknown) and NSC11 (sex unknown) cells were 872 provided by J. Rich at University of North Carolina-Chapel Hill84. MGG152 (male) cells were 873 provided by D. Cahill at MGH85. UTSW63 (male) and UTSW71 (male) were generated as 874 previously described83. HMC3 cells (Sex unknown, A TCC CRL-3304, RRID: CVCL_II76) were 875 acquired commercially. 876 877 UTSW5 GBM cells (female) were generated as described previously83. Briefly, fresh tumor 878 tissue from the operating room was suspended in ice cold Hibernate A (BrainBits HA) with 100 879 U/mL and 100 μ g/mL, respectively, of penicillin/streptomycin (Thermo Fisher 15-140-148) for 880 transfer on ice to the laboratory, then transferred to RBC lysis buffer (ThermoFisher 00433357) 881 for 10 minute incubation at room temperature before washing in Hibernate A supplemented with 882 2mM Glutamax (ThermoFisher 35050061), penicillin/streptomycin (100 U/mL and 100 μ g/mL, 883 respectively; Thermo Fisher 15-140-148), and Amphotericin B (0.25 μ g/mL, Gemini Bio-884 Products 400104). Following washing, tissue was dissociated using the Brain Tumor 885 Dissociation Kit (Miltenyi Biotec) per manufacturer instructions, and single-cell suspensions 886 were cultured in NeuroCult NS-A Basal Medium (Human) with 1x Proliferation Supplement 887 (STEMCELL Technologies 05751), supplemented with 20 ng/mL EGF (GoldBio 1150-04); 20 888 ng/mL bFGF (GoldBio 1140-02); 2 µg/mL heparin (STEMCELL Technologies 07980); 100 889 U/mL and 100 μ g/mL, respectively, of penicillin/streptomycin (Thermo Fisher 15-140-148); 250 890 ng/mL amphotericin B (GeminiBio 400104); and 0.25 µg/mL Plasmocin (InvivoGen ant-mpp) 891 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 52 on ultra-low adherence plates (6-well plates: Corning 3471, 10 cm dishes: Corning 4615) in 5% 892 CO2 and at ambient oxygen at 37°C. 893 894 All cell lines were routinely evaluated for mycoplasma contamination with the e-Myco 895 Mycoplasma PCR Detection Kit (Bulldog Bio 2523348), e-Myco PLUS Mycoplasma PCR 896 Detection Kit (Bulldog Bio 25233), or MycoAlert Mycoplasma Detection Kit (Lonza L T07-318), 897 per manufacturer’s instructions, and routinely tested negative throughout the course of this study. 898 As reference short term tandem repeat profiles have not been established for these lines, no cell 899 line authentication was performed. Sex and source of each line is stated above or listed as 900 unknown if unreported in the original publication describing its derivation. 901 902 Cell Culture 903 BT054, BT260, HK157, HK211, HK213, HK252, HK308, TS516, TS603, UTSW5, UTSW63, 904 and UTSW71 human GSCs were cultured in NeuroCult NS-A Basal Medium (Human) with 1x 905 Proliferation Supplement (STEMCELL Technologies 05751), supplemented with 20 ng/mL EGF 906 (GoldBio 1150-04); 20 ng/mL bFGF (GoldBio 1140-02); 2 µg/mL heparin (STEMCELL 907 Technologies 07980); 100 U/mL and 100 μ g/mL, respectively, of penicillin/streptomycin 908 (Thermo Fisher 15-140-148); 250 ng/mL amphotericin B (GeminiBio 400104); and 0.25 µg/mL 909 Plasmocin (InvivoGen ant-mpp) on ultra-low adherence plates (6-well plates: Corning 3471, 10 910 cm dishes: Corning 4615) in 5% CO2 and at ambient oxygen at 37°C. 911 912 ENSA, NSC11, and MGG152 cells were cultured in Neurobasal Medium (Thermo Fisher 913 21103049) supplemented with 3mM glutamine (Thermo Fisher 25030081); 1× B27 supplement 914 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 53 (Thermo Fisher 17504044); 0.25× N2 supplement (Thermo Fisher 17502048); 20 ng/mL EGF 915 (GoldBio 1150-04); 20 ng/mL bFGF (GoldBio 1140-02); 2 µg/mL heparin (STEMCELL 916 Technologies 07980); 50 U/mL and 50 μ g/mL, respectively, of penicillin/streptomycin (Thermo 917 Fisher 15-140-148); 125 ng/mL amphotericin B (GeminiBio 400104); and 0.25 µg/mL 918 Plasmocin (InvivoGen ant-mpp) on ultra-low adherence plates (6-well plates: Corning 3471, 10 919 cm dishes: Corning 4615) in 5% CO2 and at ambient oxygen at 37°C. 920 921 NHA Donor #1, NHA Donor #2, HMC3, HEK293T, and 293AD were cultured in DMEM 922 (Thermo Fisher 11995-065) supplemented with 10% fetal bovine serum (FBS; GeminiBio 100-923 106) and 100 U/mL and 100 μ g/mL, respectively, of penicillin/streptomycin (Thermo Fisher 15-924 140-148) on tissue culture-treated plates in 5% CO2 and at ambient oxygen at 37°C. 925 HMC3 microglia were cultured in EMEM (ATCC 30-2003) supplemented with 10% fetal bovine 926 serum (FBS; GeminiBio 100-106) and 100 U/mL and 100 μ g/mL, respectively, of 927 penicillin/streptomycin (Thermo Fisher 15-140-148) on tissue culture-treated plates in 5% CO2 928 and at ambient oxygen at 37°C. 929 930 Animals 931 All care and treatment of experimental animals were carried out in strict accordance with Good 932 Animal Practice as defined by the US Office of Laboratory Animal Welfare and approved by the 933 University of Pittsburgh Medical Center (protocols 25066832 and 25087128) and UT 934 Southwestern Medical Center (protocols 2019-102795 and 2015-101252) Institutional Animal 935 Care and Use Committee. Animal welfare assessments were carried out daily during treatment 936 periods. Animals were housed in a pathogen-free environment between 20-26°C and at 30-70% 937 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 54 humidity, with a 12 hour:12 hour light:dark cycle. ICR-SCID (Taconic) mice were obtained from 938 Taconic at 6-7 weeks of age. C57BL/6J mice were obtained from Jackson Laboratories. Mice 939 were housed together (2-5 mice of the same sex per cage) and provided free access to water and 940 chow diet (Purina ISO Pro Rodent 3000 “5P76” or Teklad 2916). 941 942 Generation of patient-derived xenograft (PDX) mouse models 943 TS516 orthotopic xenografts for experiments assessing neuronal network activity in tumor-944 bearing and non-tumor-bearing brain hemispheres were established by intracranial injection of 945 5,000 cells into 11-week-old female ICR SCID mice (IcrTac:ICR-Prkdcscid, Taconic ICRSC). 946 TS516 orthotopic xenografts for xenograft studies of AGAT expression were established by 947 intracranial injection of 1,000 TS516 cells into 8-week-old female ICR SCID mice (Taconic 948 ICRSC). TS516 orthotopic xenografts for xenograft studies of dietary ornithine supplementation 949 and arginine restriction were established by intracranial injection of 1,000 TS516 cells into 7-950 week-old female ICR SCID mice (Taconic ICRSC). Mice were anesthetized with isoflurane and 951 immobilized using a stereotactic frame. An incision was made to expose the skull surface, and a 952 hole was drilled into the skull. Cells suspended in 5 µL cell culture medium were injected into 953 the brain through the hole using a 5 μ L syringe (Hamilton) at 0.5-1 mm anterior and 2 mm lateral 954 to the bregma and a depth of 3 mm from the brain surface. The skin was closed with surgical 955 clips and analgesia was administered. Mice receiving experimental diets received either Teklad 956 amino acid diet (Inotiv TD.01084) or diet supplemented with 0.665% ornithine and arginine-free 957 (Inotiv TD.240691). Survival analyses were performed by researchers who were not blinded to 958 the treatment arms or genotypes of the mice. Mice were euthanized when they displayed 959 neurological symptoms or became moribund. 960 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 55 961

Details 962 Chemicals 963 Where indicated, cells or brain slices were treated with gabazine (EMD Millipore SR95531), 964 GAA (Millipore Sigma G11608, or TCI G0167), creatine monohydrate (Thermo Fisher 965 B25009.22), CLP-257 (Bio-Techne 5242/10), 10 µM 3-(2-carboxypiperazin-4-yl)propyl-1-966 phosphonic acid (CPP) (Hello Bio HB0036), 20 µM (6,7-dinitroquinoxaline-2,3-dione (DNQX) 967 (Tocris 2312), 100 µM picrotoxin (Hello Bio HB0506), and/or 2 µM CGP 52432 (Tocris 1246). 968 969 V ectors 970 lentiCRISPRv2-GFP (Addgene 82416) and lentiCRISPRv2 mCherry (Addgene 99154) were 971 digested with FastDigest Esp3I (Thermo Fisher FD0454) or BsmBIv2 (New England Biolabs 972 R0739). sgRNAs targeting the AA VS1 safe harbor locus86 (sg #1: 973 GTCACCAATCCTGTCCCTAG), AGA T (sg #1: TGTGGGCAATGAGATTATCG), or GAMT 974 (sg #1: GGCCAGCGCGTGCA TA TAGG; sg #2: GTACGACACGTACCCACTCT) were ligated 975 into lentiCRISPRv2-GFP to generate a construct expressing Cas9 and the desired sgRNA. 976 sgRNAs targeting the AA VS1 safe harbor locus (sg #2: TGTTAGGCAGATTCCTTATC) or 977 AGAT (sg #2: ACTTCAATGACCAGTCAATG) were ligated into lentiCRISPRv2-mCherry to 978 generate a construct expressing Cas9 and the desired sgRNA. Reaction mixtures were 979 transformed into XL10-Gold Ultracompetent Cells (Agilent 200315) and validated by whole 980 plasmid sequencing (Plasmidsaurus). 981 982 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 56 CRISPR/Cas9 and fluorophore expression cassettes for lentiCRISPRv2-GFP-sgAA VS1 #1, 983 lentiCRISPRv2-GFP-sgAGA T #1, lentiCRISPRv2-mCherry-sgAA VS1 #2, and lentiCRISPRv2-984 mCherry-sgAGAT #2 were amplified from these constructs and appended with 5’ attB1 and 3’ 985 attB2 sites using the following primers: 986 attB1_U6_Fwd_v2: 987 GGGGACAAGTTTGTACAAAAAAGCAGGCTCTGAGGGCCTATTTCCCA TG 988 bGHpolyA_attB2_Rev_v2: 989 GGGGACCACTTTGTACAAGAAAGCTGGGTTCCA TAGAGCCCACCGCAT 990 PCR product was gel-purified with the QIAquick Gel Extraction Kit (Qiagen 28706), purified 991 with the Monarch PCR & DNA Cleanup Kit (New England Biolabs T1030) and cloned into the 992 Gateway vector pDONR223 by BP reaction (Thermo Fisher 11789100) overnight. The resultant 993 products were transformed into HB101 Competent Cells (Promega L2015), and identity of the 994 final products were validated by whole-plasmid sequencing (Plasmidsaurus). 995 996 pENTR223 plasmids containing CRISPR/Cas9 expression cassettes for GFP-sgAA VS1 #1, GFP-997 sgAGAT #1, mCherry-sgAA VS1 #2, and mCherry-sgAGA T #2 were cloned into the Gateway 998 vector pAd/PL-DEST (Thermo Fisher V49420) by LR reaction (Thermo Fisher 11791020) 999 overnight. Resultant pAdCRISPR-GFP-sgAA VS1 #1, pAdCRISPR-GFP-sgAGAT #1, 1000 pAdCRISPR-mCherry-sgAA VS1 #2 and pAdCRISPR-mCherry-sgAGAT #2 vectors for 1001 adenoviral transduction of mammalian cells were transformed into HB101 Competent Cells 1002 (Promega L2015) and validated by whole plasmid sequencing (Plasmidsaurus). 1003 1004 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 57 AGAT (RefSeq NM_001482.2) WT cDNA was generated from the Ultimate ORF Human Clone 1005 Library (Life Technologies HORF01) in a pENTR221 Gateway vector. The C407A point 1006 mutation was generated by inverse PCR of pENTR221-AGAT using CloneAmp HiFi PCR 1007 Master Mix (Takara 639298) and the following primers: 1008 GATM_C407A_InvPCR_F: TTCCA TGCCTGGACCTGCGATGTC 1009 GATM_C407A_InvPCR_R: GGTCCAGGCA TGGAAGCCTCCT 1010 The amplified inverse PCR product was then gel-purified with the QIAquick Gel Extraction Kit 1011 (Qiagen 28706), PCR purified with the Monarch PCR & DNA Cleanup Kit (New England 1012 Biolabs T1030) and assembled using the In-Fusion HD Cloning Kit (Takara Bio 102518). The 1013 assembled product was then transformed into XL10-Gold Ultracompetent Cells (Agilent 1014 200315). Final pENTR221-AGA T_C407A product was validated by whole plasmid sequencing 1015 (Plasmidsaurus). 1016 1017 5’ attB1 and 3’ attB2 sites were appended to AGA T WT and AGAT C407A cDNA using Q5 Hot 1018 Start High-Fidelity 2X Master Mix polymerase (New England Biolabs M0494) and the following 1019 primers: 1020 attB_AGAT_F: 1021 GGGGACAAGTTTGTACAAAAAAGCAGGCTCTGCCACCA TGCTGCGGGT 1022 attB_AGAT_R: 1023 GGGGACCACTTTGTACAAGAAAGCTGGGTTTCAGTCCAAGTAGGACTGTAAG1024 GTGCC 1025 PCR product was gel-purified with the QIAquick Gel Extraction Kit (Qiagen 28706), purified 1026 with the Monarch PCR & DNA Cleanup Kit (New England Biolabs T1030) and cloned into the 1027 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 58 Gateway vector pDONR223 by BP reaction (Thermo Fisher 11789100) overnight. The resultant 1028 products were transformed into HB101 Competent Cells (Promega L2015). pENTR223-1029 AGAT_WT and pENTR223-AGA T_C407A were validated by whole plasmid sequencing 1030 (Plasmidsaurus). 1031 1032 GAMT WT and GAMT E45S cDNA were synthesized as gBlocks by Twist Biosciences. attB 1033 sites were added and sequences amplified using Q5 Hot Start High-Fidelity 2X Master Mix 1034 polymerase (New England Biolabs M0494) and the following primers: 1035 attB_GAMT_F: 1036 GGGGACAAGTTTGTACAAAAAAGCAGGCTCTGCCACCA TGAGCGCCCC 1037 attB_GAMT_R: 1038 GGGGACCACTTTGTACAAGAAAGCTGGGTTTCAGCCTTTGGTCACCAGGGG 1039 PCR product was gel-purified with the QIAquick Gel Extraction Kit (Qiagen 28706), PCR 1040 purified with the Monarch PCR & DNA Cleanup Kit (New England Biolabs T1030) and cloned 1041 into the Gateway vector pDONR223 by BP reaction (Thermo Fisher 11789100) overnight to 1042 generate pENTR223-GAMT_WT and pENTR223-GAMT_E45S plasmids. The resultant 1043 products were transformed into HB101 Competent Cells (Promega L2015). pENTR223-1044 GAMT_WT and pENTR223-GAMT _E45S were validated by whole plasmid sequencing 1045 (Plasmidsaurus). 1046 1047 pENTR223-EV , pENTR223-AGAT_WT, and pENTR223-AGAT_C407A were cloned into 1048 pLenti-EF1α -DEST-IRES-Neo48 (Addgene deposition pending) by Gateway LR reaction 1049 (Thermo Fisher 11791020) overnight. The lentiviral vectors pLenti-EF1α -EV-IRES-Neo, pLenti-1050 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 59 EF1α -AGAT_WT-IRES-Neo, and pLenti-EF1α -AGAT_C407A-IRES-Neo were transformed into 1051 HB101 Competent Cells (Promega L2015) and validated by whole plasmid sequencing 1052 (Plasmidsaurus). 1053 1054 pENTR223-EV , pENTR223-GAMT_WT, and pENTR223-GAMT_E45S were cloned into 1055 pLenti-EF1α -DEST-PGK-Hygro by Gateway LR reaction (Thermo Fisher 11791020) overnight. 1056 The lentiviral vectors pLenti-EF1α -EV-PGK-Hygro, pLenti-EF1α -GAMT_WT-PGK-Hygro, and 1057 pLenti-EF1α -GAMT_E45S-PGK-Hygro were transformed into HB101 Competent Cells 1058 (Promega L2015) and validated by whole plasmid sequencing (Plasmidsaurus). 1059 1060 Transfection and viral transduction 1061 Lentiviral particles were produced by transfection of HEK293T cells with expression vectors and 1062 packaging plasmids psPAX2 (Addgene 12260, gift of Didier Trono) and pMD2.G (Addgene 1063 12259, gift of Didier Trono) in a ratio of 4:3:1 using TransIT-LT1 transfection reagent (Mirus Bio 1064 MIR2 B-304). Media was discarded and replaced at 24 hours post-transfection, then virus-1065 containing media was collected 48 and 72 hours post-transfection and passed through a 0.45 μ m 1066 filter (Corning 431220). 1 volume LentiX Concentrator (Takara 631232) was added per 3 1067 volumes clarified supernatant. The mixture of viral supernatant and LentiX Concentrator was 1068 incubated at 4°C overnight, then centrifuged at 1500 × g for 45 minutes. After centrifugation, 1069 supernatant was discarded, and the remaining pellet was resuspended in /i1 original viral 1070 supernatant volume of DMEM supplemented with 10% fetal bovine serum (FBS; GeminiBio 1071 100-106) and 100 U/mL and 100 μ g/mL, respectively, of penicillin/streptomycin (Thermo Fisher 1072 15-140-148). 1073 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 60 1074 To generate adenovirus, pAdCRISPR constructs were linearized by PacI (New England Biolabs 1075 R0547S) digestion at 37°C overnight. 293Ad cells were subsequently transfected with 2.5 µg 1076 linear pAdCRISPR constructs using TransIT-LT1 transfection reagent (Mirus Bio MIR2 B-304). 1077 Media was changed at 24 hours after transfection, and cells were expanded to a 10 cm dish and 1078 incubated for 10-14 days with regular media changes. 10-14 days following transfection, 293Ad 1079 cells and media were harvested and centrifuged at 3739 × g for 5 minutes. Supernatant was 1080 discarded and the pellet was resuspended in 1 mL media. Resuspended pellet was subjected to 3 1081 freeze-thaw cycles with 30 minutes at -80°C followed by 15 minutes at 37°C. After final thaw, 1082 suspensions were centrifuged at 2093 × g for 15 minutes to generate crude adenovirus. Presence 1083 of adenovirus capsid in crude adenovirus prep was confirmed using Adeno-X GoStix (Takara 1084 632270). Crude adenovirus was then amplified 2-3 times by inoculating 293Ad cells with crude 1085 virus and harvesting as above after 2-4 days. Amplified virus was purified using Virabind 1086 Adenovirus Miniprep Kit (Cell Biolabs VPK-099) according to manufacturer’s instructions. 1087 1088 NHA Donor #1 cells were plated at a density of 1.5 × 105 cells per well in 2 mL DMEM in a six 1089 well plate and allowed to adhere overnight. The next day, 3 uL polybrene (8 µg/mL, 1090 MedChemExpress HY-112735) was added to each well along with 1 mL viral pLenti-EF1a-EV-1091 IRES-Neo, pLenti-EF1a-AGAT_WT-IRES-Neo, or pLenti-EF1a-AGAT_C407A-IRES-Neo 1092 supernatant. Plates were centrifuged at 4,000 × g for 30 minutes at room temperature, then 1093 incubated overnight. The following day, cells were expanded and replated in a 10 cm dish. After 1094 initial selection at 1500 µg/mL G418 (GoldBio G-418), stable cell lines were maintained in 600 1095 µg/mL G418. 1096 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 61 1097 NHA Donor #1 EV , AGA T WT, and AGA T C407A cells were plated at a density of 1.5 × 105 1098 cells per well in 2 mL DMEM without G418 in a six well plate and allowed to adhere overnight. 1099 The next day, 3 uL polybrene (8 µg/mL, MedChemExpress HY-112735) was added to each well 1100 along with 1 mL virus for pLenti-EF1a-EV-PGK-hygro, pLenti-EF1a-GAMT_WT-PGK-Hygro, 1101 or pLenti-EF1a-GAMT_E45S-PGK-Hygro; or 1 mL lentiCRISPRv2-GFP-sgAA VS1 #1, 1102 lentiCRISPRv2-GFP-sgGAMT #1, or lentiCRISPRv2-GFP-sgGAMT #2 lentivirus. Plates were 1103 centrifuged at 4,000 × g for 30 minutes at room temperature, then incubated overnight. The 1104 following day, cells were expanded and replated in a 10 cm dish. For cells transduced with 1105 pLenti-EF1a-EV-PGK-hygro, pLenti-EF1a-GAMT_WT-PGK-Hygro, or pLenti-EF1a-1106 GAMT_E45S-PGK-Hygro, following initial selection at 250 µg/mL hygromycin (GoldBio H-1107 270-EZ50), stable cell lines were maintained in 100 µg/mL hygromycin. Cells transduced with 1108 lentiCRISPRv2-GFP constructs were sorted for GFP positivity 14 days post-transduction using a 1109 FACS Aria II SORP Four-Laser instrument (BD Biosciences) in the Children’s Research Institute 1110 Flow Cytometry Core. 1111 1112 For adenoviral transduction, TS516 cells were plated at 1 × 107 cells per dish in 10 mL 1113 NeuroCult NS-A Basal Medium (Human) and treated with AdCRISPR GFP sgAA VS1 #1 or 1114 sgAGAT #1 adenovirus at a concentration empirically determined to yield 30-50% transduction 1115 efficiency for each viral preparation. At 16-18 hours, media was changed to fresh NeuroCult NS-1116 A Basal Medium (Human) and cells were split out to 2 dishes each. At 7 days post-transduction, 1117 cells were sorted for GFP positivity using a FACS Aria II SORP Four-Laser instrument (BD 1118 Biosciences) in the Children’s Research Institute Flow Cytometry Core. Sorted GFP+ cells were 1119 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 62 transduced with AdCRISPR mCherry sgAA VS1 #2 or sgAGA T #2 adenovirus 14 days following 1120 initial sort as previously done with GFP-containing virus, then sorted 7 days following secondary 1121 transduction using a FACS Aria II SORP Four-Laser instrument (BD Biosciences) in the 1122 Children’s Research Institute Flow Cytometry Core. After two weeks in culture, sorted mCherry+ 1123 cells were sorted for GFP and mCherry dual-negative populations to establish stably-transduced 1124 AGAT WT and AGAT KO lines. 1125 1126 Preparation of HPLM and tracer HPLM 1127 HPLM was prepared as previously described49. Tracer-containing HPLM library was as 1128 previously described48, with the addition of a medium containing 15N3-creatine (Cambridge 1129 Isotope Laboratories NLM-9218). Tracer-containing HPLM used in nitrogen metabolism 1130 profiling, profiling validation, and 15N4-arginine tracing experiments were generated identically 1131 to the unlabeled version, with labeled metabolites replacing their unlabeled counterparts at 1132 equimolar concentrations. HPLM for 15N4-arginine tracing in TS516 under ornithine-1133 supplemented conditions was generated without arginine or ornithine, then 400 µM 15N4-arginine 1134 (Cambridge Isotope Laboratories NLM-396) and either 0 µM or 150 µM ornithine (Millipore 1135 Sigma O2375) were added. Tracer HPLM was assembled by mixing thawed frozen or freshly-1136 prepared stocks of unlabeled HPLM pools, replacing one pool per tracing condition with a 1137 version containing an individual 15N tracer. Media was adjusted to pH 7.4 using NaOH or HCl, 1138 then diluted to the final volume with ultrapure water, sterile filtered with a 0.22 µm PES filter 1139 (Millipore Sigma SCGP00525, Corning 431153, Corning 431097, or Corning 431098), and 1140 supplemented as described below. To prevent glutamate toxicity in culture, HPLM was prepared 1141 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 63 without glutamate for experiments in GSCs, NSCs, and immortalized astrocyte cells, except for 1142 conditions in which glutamate was traced. 1143 1144 For experiments in GSCs and NSCs, HPLM was supplemented with 1 × B27 supplement 1145 (Thermo Fisher 17504044); 0.25 × N2 supplement (Thermo Fisher 17502048); 20 ng/mL EGF 1146 (GoldBio 1150-04); 20 ng/mL bFGF (GoldBio 1140-02); 2 µg/mL heparin (STEMCELL 1147 Technologies 07980); 50 U/mL and 50 μ g/mL, respectively, penicillin/streptomycin (Thermo 1148 Fisher 15-140-148); 125 ng/mL amphotericin B (GeminiBio 400104); and 0.25 µg/mL 1149 Plasmocin (InvivoGen ant-mpp). For experiments in differentiated cells, HPLM was 1150 supplemented with 10% dialyzed FBS (GeminiBio 100-108) and 100 U/mL and 100 μ g/mL, 1151 respectively, of penicillin/streptomycin (Thermo Fisher 15140148). 1152 1153 Nitrogen metabolism profiling platform 1154 Nitrogen metabolism profiling was conducted as described previously48. Briefly, NHA Donor #1 1155 cells were plated in 6-well plates (2.5 × 104 cells per well) in 2 mL DMEM prepared as described 1156 above. TS516 cells were plated in 6-well ultra-low adherence plates (2 × 105 cells per well) in 2 1157 mL NeuroCult NS-A Basal Medium (Human) prepared as described above. After 24 hours, 2 mL 1158 unlabeled HPLM prepared as described above was added to produce a mixture of 50% native 1159 medium and 50% HPLM. 24 hours later, media was changed to 100% unlabeled HPLM (10 mL 1160 for NHA Donor #1 cells and 4.5 mL for TS516 cells). 24 hours later, media was changed to fresh 1161 100% HPLM with one nitrogen-containing metabolite per sample exchanged for an equimolar 1162 amount of its 15N-labeled counterpart (10 mL for NHA Donor #1 cells, 4.5 mL for TS516 cells). 1163 After 18 hours, samples were harvested and prepared for LC-MS analysis as described below. 1164 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 64 1165 Creatine synthesis pathway tracing in cells 1166 For tracing of creatine synthesis pathway substrates, differentiated cells (NHA Donor #1 and #2) 1167 were plated at 2.5 × 104 cells per well for NHA Donor #1 and 1 × 105 cells per well for NHA 1168 Donor #2 in 2 mL DMEM. GSCs with exception of MGG152 were plated at 2 × 105 cells per 1169 well in 2 mL NeuroCult NS-A Basal Medium (Human) prepared as described above in 6-well 1170 ultra-low adherence plates. NSCs and MGG152 were plated at 2 × 105 cells per well in 2 mL 1171 NeuroBasal prepared as described above in 6-well ultra-low adherence plates. 24 hours after 1172 plating, 2 mL unlabeled HPLM prepared as described above was added to produce a 1:1 mixture 1173 of HPLM and native medium. 24 hours later, media was changed to 10 mL 100% HPLM for 1174 NHA Donor #1 and #2 or 4.5 mL HPLM for GSCs or NSCs. 24 hours later, media was changed 1175 to 10 or 4.5 mL fresh HPLM in which the unlabeled metabolite was exchanged for an equimolar 1176 concentration of 15N4-arginine (Cambridge Isotope Laboratories NLM-396), 15N3-creatine 1177 (Cambridge Isotope Laboratories NLM-9218), 15N-glycine (Cambridge Isotope Laboratories 1178 NLM-202), or 15N-serine (Cambridge Isotope Laboratories NLM-2036). Cells were incubated 1179 with tracer for 18 hours, then harvested and prepared for LC-MS as described below. For each 1180 cell line, counts and diameter were quantified from parallel cell cultures using a Beckman 1181 Coulter Vi-CELL XR cell viability analyzer (RRID: SCR_019664) at the time of sample harvest. 1182 1183 Intracellular GAA content in cultured cells 1184 Differentiated cells (NHA Donor #1 and #2) were plated at 5 × 104 cells per well for NHA Donor 1185 #1 and 2 x 105 cells per well for NHA Donor #2 in 2 mL DMEM. GSCs with exception of 1186 MGG152 were plated at 4 × 105 cells per well in 2 mL NeuroCult NS-A Basal Medium (Human) 1187 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 65 prepared as described above in 6-well ultra-low adherence plates. NSCs and MGG152 were 1188 plated at 4 × 105 cells per well in 2 mL NeuroBasal medium prepared as described above in 6-1189 well ultra-low adherence plates. 24 hours after plating, 2 mL unlabeled HPLM prepared as 1190 described above was added to produce a 1:1 mixture of HPLM and native medium. 24 hours 1191 later, media was changed to 10 mL 100% HPLM for NHA Donor #1 and #2 or 4.5 mL HPLM for 1192 GSCs or NSCs. 24 hours later, media was changed to 2 mL fresh HPLM. 24 hours later, media 1193 was changed to 2 mL fresh HPLM per well. 24 hours later, cells were collected, washed in 1194 Optima saline, and snap frozen. For each cell line, cell counts and cell diameter were quantified 1195 from parallel cell cultures using a Beckman Coulter Vi-CELL XR cell viability analyzer (RRID: 1196 SCR_019664) at the time of sample harvest. 1197 1198 Secretion of GAA in cell culture 1199 Differentiated cells (NHA Donor #1 and #2) were plated at 5 × 104 cells per well for NHA Donor 1200 #1 and 2 × 105 cells per well for NHA Donor #2 in 2 mL DMEM in 6-well plates. HMC3 1201 microglia were plated at 5 × 104 cells per well in 2 mL EMEM in 6-well plates. GSCs with 1202 exception of MGG152 were plated at 4 × 105 cells per well in 2 mL NeuroCult NS-A Basal 1203 Medium (Human) prepared as described above in 6-well ultra-low adherence plates. NSCs and 1204 MGG152 were plated at 4 × 105 cells per well in 2 mL NeuroBasal prepared as described above 1205 in 6-well ultra-low adherence plates. 24 hours after plating, 2 mL unlabeled HPLM prepared as 1206 described above was added to produce a 1:1 mixture of HPLM and native medium. 24 hours 1207 later, media was changed to 10 mL 100% HPLM for NHA Donor #1 and #2 or 4.5 mL HPLM for 1208 GSCs or NSCs. 24 hours later, media for all cells was changed to 2 mL fresh HPLM. 48 hours 1209 later, media samples were collected from wells, spun at 4,000 × g for 3 minutes, 100 µL 1210 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 66 transferred to a fresh microcentrifuge tube, and snap frozen in liquid nitrogen. For each cell line, 1211 cell counts and cell diameter were quantified from parallel cell cultures using a Beckman Coulter 1212 Vi-CELL XR cell viability analyzer (RRID: SCR_019664) at the time of sample harvest. 1213 1214 Preparation of cultured cells for LC-MS 1215 Prior to LC-MS analysis, cell samples were suspended in 1 µL ice-cold 80% acetonitrile 1216 (Thermo Fisher A9554) per 1,000 cells. Adherent cells (NHA Donor #1 and NHA Donor #2) 1217 were lifted from plates by scraping to facilitate resuspension. Cell samples resuspended in 1218 acetonitrile were platform-vortexed for 20 minutes at 4°C and centrifuged for 10 minutes at 1219 17,000-21,100 × g at 4°C. The supernatant was transferred to a fresh tube and again centrifuged 1220 for 10 minutes at 17,000-21,100 × g at 4°C. The resultant supernatant was then analyzed by LC-1221 MS. For experiments in which absolute quantification of GAA in cell samples was performed, 1222 internal standards 1,2-¹³C/i1 , 3-¹/i1 N-guanidinoacetic acid (Cambridge Isotope Laboratories 1223 CNLM-8300) or 2,2-D2-guanidinoacetic acid (Cambridge Isotope Laboratories DLM-9998) at 1224 concentrations of 20, 100, or 200 nM were added. Standard curves for GAA (Millipore Sigma 1225 G11608) were created fresh for each submission in neat 80% acetonitrile over a range of 1-1226 10,000 nM. All reagents used were Optima-grade. 1227 1228 Preparation of media for LC-MS 1229 Prior to LC-MS analysis, media was diluted 1:10 in ice-cold 80% acetonitrile (Thermo Fisher 1230 A9554) and vortexed for 20 minutes at 4°C, then centrifuged for 10 minutes at 17,000-21,100 × 1231 g at 4°C. The supernatant was transferred to a fresh tube and again centrifuged for 10 minutes at 1232 17,000-21,100 × g at 4°C. The resultant supernatant was then analyzed by LC-MS. For 1233 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 67 experiments in which absolute quantification of GAA was performed in media, an internal 1234 standard of 1,2-¹³C/i1 , 3-¹/i1 N-guanidinoacetic acid (Cambridge Isotope Laboratories CNLM-1235 8300) was added to all samples and standards at concentrations of 20, 100, or 200 nM. Standard 1236 curves for GAA (Millipore Sigma G11608) were created fresh for each submission in neat 80% 1237 acetonitrile over a range of 1-10,000 nM. All reagents used were Optima-grade. 1238 Preparation of tissue for LC-MS 1239 Tissues collected for metabolomic profiling of human brain tumors were obtained in the 1240 operating room, suspended in ice-cold Hibernate A (BrainBits HA), and snap frozen. Samples 1241 were prepared as previously described101, with some modifications. Briefly, samples were 1242 thawed, washed in 1 mL ice-cold normal saline, and resuspended in 50 µL ice-cold 80% 1243 methanol (Thermo Fisher A456) per mg of tissue. Tissue samples were homogenized in a 1244 TissueLyser II (Qiagen 85300, RRID: SCR_018623) at 4°C at an oscillation rate of 25 Hz. 520 1245 µL tert-Butyl methyl ether (Millipore Sigma 650560) and 380 µL water were added to 500 µL of 1246 tissue homogenate in a glass vial, then vortexed at room temperature for 1 hour at 1,000 rpm in a 1247 thermal mixer (Benchmark Scientific H5000-HC). Samples were centrifuged for 10 minutes at 1248 1,000 × g at 4°C. 1249 1250 For polar metabolomics analysis, 800 µL of the lower aqueous phase was combined with 300 µL 1251 of the upper organic phase, avoiding interlayer debris. A SpeedVac (Thermo Fisher SPD2030) 1252 was used to dry down polar metabolite extracts, and dried metabolites were then resuspended in 1253 100 µL ice-cold 80% acetonitrile (Thermo Fisher A9554). A mixture of internal standards, 1254 composed of 3 µM 13C5-glutamine (Cambridge Isotope Laboratories CLM-1822), 3 µM 15N-1255 valine (Cambridge Isotope Laboratories NLM-316), 3 µM 15N-methionine (Cambridge Isotope 1256 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 68 Laboratories NLM-752), and 3 µM chloramphenicol (Millipore Sigma C0378) was added. 1257 Samples were sonicated in a room temperature water bath for 10 minutes, then centrifuged for 10 1258 minutes at 21,100 × g at 4°C. The resultant supernatant was then analyzed by LC-MS. 1259 For lipidomics analysis, 100 µL of upper organic phase was transferred to a new glass vial. The 1260 lipid extract was dried using nitrogen to avoid oxidation, and dried lipids were resuspended in 1261 500 µL isopropanol and methanol mixture (65:35, V/V) with 15 µL of Avanti Sph/Cer mix 1262 standard (Avanti Research, 330707). Samples were then vortexed at 4°C for 10 minutes, then 1263 centrifuged for 10 minutes at 1000 × g at 4°C. The resultant supernatant was then analyzed by 1264 LC-MS. 1265 1266 For experiments in which absolute quantification of GAA in human tissue samples was 1267 performed, tissues were homogenized in 50 µL ice-cold 80% Optima methanol (Thermo Fisher 1268 A456) per mg of tissue using a TissueLyser II (Qiagen 85300, RRID: SCR_018623) with two 1269 carbide beads per tube (Qiagen 69997) at an oscillation rate of 25 Hz. Samples were vortexed for 1270 20 minutes at 4°C, then centrifuged for 10 minutes at 17,000-21,100 × g at 4°C. The supernatant 1271 was transferred to a fresh tube then centrifuged for 10 minutes at 17,000-21,100 × g at 4°C. 200 1272 or 250 µL of supernatant was transferred to a fresh tube, then dried using a SpeedVac (Thermo 1273 Fisher SPD2030). Dried metabolites were resuspended in 7.5 µL ice-cold 80% acetonitrile 1274 (Thermo Fisher A9554) per mg tissue, vortexed for 20 minutes at 4°C, and centrifuged for 10 1275 minutes at 17,000-21,100 × g at 4°C. The resultant supernatant was then diluted 1:10 and 1276 analyzed by LC-MS. For experiments in which absolute quantitation of GAA was performed in 1277 tissues, 1,2-¹³C/i1 , 3-¹/i1 N-guanidinoacetic acid (Cambridge Isotope Laboratories CNLM-8300) 1278 was added as an internal standard to all samples and standards. Standard curves for GAA 1279 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 69 (Millipore Sigma G11608) were created fresh for each submission in neat 80% acetonitrile over 1280 a range of 1-10,000 nM. All reagents used were Optima-grade. 1281 1282 Extraction of TS516 AGAT WT and AGAT KO xenograft tissues collected for quantification of 1283 GAA and creatine was performed using pre-chilled 80% methanol (Thermo Fisher A456). A 1284 volume of 80% MeOH equivalent to 50 µL per mg of tissue was added to each sample tube. Two 1285 tungsten carbide beads (Qiagen 69997) were added to each tube for homogenization. Tissue 1286 homogenization was carried out using a TissueLyser set to 25 Hz for 1-minute cycles. Following 1287 homogenization, samples were vortexed for 20 minutes at 4/i1 °C. Samples were centrifuged 1288 twice at maximum speed (~13,000–15,000 × g) for 10 minutes at 4/i1 °C. The resulting 1289 supernatant was transferred and analyzed by LC-MS/MS and internal standards for both GAA 1290 (Guanidineacetic-2,2-d2 Acid, CDN Isotopes D-6320) and creatine [Creatine-d5 H2O (N-1291 methyl-d3; glycine-2,2-d2), LGC Standards TRC-C781496] were added to final supernatant. A 1292 standard curve of creatine (LGC Standards TRC-C781483) and GAA (TRC-G821250) was 1293 created for each run with duplicate QCs (low, medium, and high) to ensure the results for each 1294 day were accurate. The linear range for GAA and creatine in brain was 1-25,000 ng/mL. 1295 1296 Metabolite Quantification by LC-MS 1297 10-20 µL of metabolite extract in acetonitrile was injected and analyzed with a Q-Exactive HF-X 1298 (Thermo, RRID: SCR_020425), Orbitrap LUMOS (Thermo, RRID: SCR_020562), or Orbitrap 1299 Exploris 480 (Thermo, RRID: SCR_027000) hybrid quadrupole-orbitrap mass spectrometer 1300 coupled to a Vanquish Flex UHPLC system (Thermo Fisher), as described previously102,103; or 1301 with an AB QTRAP 5500 (Applied Biosystems Sciex) or AB QTRAP 6500+ (Applied 1302 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 70 Biosystems Sciex) with ESI source coupled to a Nexera X2 LC-30AD HPLC (Shimadzu 1303 Corporation). Chromatographic separation of samples acquired on the Q-Exactive, LUMOS, or 1304 Exploris Orbitrap instruments was accomplished using a Millipore-Sigma ZIC-pHILIC column 1305 using a binary gradient of 10 mM ammonium formate pH 9.8 and acetonitrile. Spectra acquired 1306 on Orbitrap instruments were acquired with resolving power of 60,000, 120,000, or 240,000 full 1307 width at half maximum (FWHM), and a scan range set to 60-1,050 or 80-1,200 m/z, collecting 1308 data in both positive and negative polarities. To improve signal resolution of GAA and its 1309 isotopologues, a targeted selected ion monitoring (tSIM) scan event using a 7 Da window around 1310 GAA was used in tandem with full scan data acquisition to capture all relevant stable isotope 1311 labeling events with an AGC targeted of 1 × 105 ions. 1312 1313 Chromatographic separation of samples acquired on the AB QTRAP 5500 or AB QTRAP 6500+ 1314 was accomplished with a SeQuant® ZIC®-pHILIC HPLC column (150/i1 ×/i1 2.1/i1 mm, 5/i1 µm, 1315 polymeric) or a BEH Z-HILIC VanGuardTM Fit HPLC column (150/i1 ×/i1 2.1/i1 mm, 5/i1 µm), 1316 using an electrospray ionization (ESI) source in multiple reaction monitoring (MRM) mode. 1317 Samples acquired on the AB QTRAP 5500 were resolved using a binary gradient of 10 mM 1318 ammonium acetate in water (pH 9.8 adjusted with ammonium water; mobile phase A) and 100% 1319 ACN (mobile phase B), with gradient elution as follows: 0-15/i1 minutes, linear gradient 90-30% 1320 B, followed by 3 minute wash with 30% B before reconditioning for 6/i1 minutes using 90% B. 1321 Samples acquired on the AB QTRAP 6500 were resolved using a binary gradient of 10 mM 1322 ammonium acetate in water (pH 9.8 adjusted with ammonium water; mobile phase A) and 90% 1323 ACN:10% mobile phase A (mobile phase B), with gradient elution as follows: 0-15/i1 minutes, 1324 linear gradient 100-33% B, followed by 3 minute wash with 33% B before reconditioning for 1325 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 71 6/i1 minutes using 90% B. A flow rate of 0.25 mL/minute was used with injection volume 10-20 1326 µL. MRM data was acquired with Analyst 1.6.3 software (SCIEX). 1327 1328 For samples acquired on the AB QTRAP 5500 in which isotopologue species of GAA and 1329 creatine pathway-associated metabolites were quantified, the MRMs used for each metabolite 1330 were as follows: For GAA, M+0 (Q1/Q3: 118/76, CE: 16), M+1 (Q1/Q3: 119/76 and 119/77, 1331 CE: 16), M+2 (Q1/Q3: 120/76, 120/77, 120/78; CE: 16), M+3 (Q1/Q3: 121/76, 121/77, 121/78, 1332 CE: 16), internal standard M+3 (Q1/Q3: 121/79, CE: 16); for arginine, M+0 (Q1/Q3: 175/70, 1333 CE: 25), M+1 (Q1/Q3: 176/70, 176/71; CE: 25), M+2 (Q1/Q3: 177/70, 177/71, 177/72; CE: 25), 1334 M+3 (Q1/Q3: 178/70, 178/71, 178/72, 178/73; CE: 25), M+4 (Q1/Q3: 179/70, 179/71, 179/72, 1335 179/73, 179/74; CE: 25); for ornithine, M+0 (Q1/Q3: 133/70, CE: 25), M+1 (Q1/Q3: 134/70, 1336 134/71; CE: 25), M+2 (Q1/Q3: 135/70, 135/71, 135/72; CE: 25); for creatine, M+0 (Q1/Q3: 1337 132/90, CE: 34), M+1 (Q1/Q3: 133/90, 133/91; CE: 34), M+2 (Q1/Q3: 134/90, 134/91, 134/92; 1338 CE: 34), M+3 (Q1/Q3: 135/90, 135/91, 135/92, 135/93; CE: 34); for creatinine, M+0 (Q1/Q3: 1339 114/86, CE: 15), M+1 (Q1/Q3: 115/86, 115/87; CE: 15), M+2 (Q1/Q3: 116/86, 116/87, 116/88; 1340 CE: 15), M+3 (Q1/Q3: 117/86, 117/87, 117/88, 117/89; CE: 15); and for phosphocreatine, M+0 1341 (Q1/Q3: 210/79, 210/97; CE: -26), M+1 (Q1/Q3: 211/79, 211/97; CE: -26), M+2 (Q1/Q3: 1342 212/79, 212/97; CE: -26), M+3 (Q1/Q3: 213/79, 213/97; CE: -26). 1343 1344 Peaks for full scan data were integrated using El-Maven 0.12.0 software (Elucidata) using a 1345 targeted in-house spectral library detecting metabolites using precursor and product ion 1346 spectra104, and peaks for GAA tSIM data were integrated using a targeted method in TraceFinder 1347 5.1 SP2 (Thermo Fisher OPTON-31001). Peaks from data obtained on the AB QTRAP 5500 and 1348 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 72 6500+ were integrated using MultiQuant v2.1.1 (Sciex) software. For tracing experiments 1349 acquired on the AB QTRAP 5500, peak areas for each MRM corresponding to a given 1350 isotopologue species, as outlined above, were summed to yield the total abundance of each 1351 isotopologue. Total ion counts for full scan data were quantified using TraceFinder 5.1 SP2 1352 software (Thermo Fisher OPTON-31001, RRID: SCR_023045). Peaks were normalized using 1353 probabilistic quotient normalization105 and total ion counts using R48 (RRID: SCR_001905) or by 1354 normalization to an internal standard. For stable isotope tracing studies, correction for natural 1355 abundance of metabolite labeling was performed using AccuCor 0.3.087 (RRID: SCR_023046) in 1356 R (RRID: SCR_001905) or by manual correction. 1357 1358 Tissue Collection, Processing, and Extraction – Glioma Metabolomics V alidation Cohort 1 1359 De-identified tumor and matched normal brain tissue specimens were collected intraoperatively, 1360 immediately snap-frozen in liquid nitrogen, and stored at -80°C until shipment. Approximately 1361 10 mg of frozen tissue per sample was submitted to Metabolon, Inc. (Morrisville, NC) for global 1362 metabolomic profiling. Recovery standards were added to each specimen prior to extraction to 1363 enable quality control monitoring of instrument performance and data normalization. Protein was 1364 precipitated by vigorous shaking with methanol, followed by centrifugation to recover 1365 chemically diverse metabolites. The resulting supernatant was split into multiple fractions 1366 optimized for four parallel UPLC-MS/MS methods: (1) reverse-phase positive ion electrospray 1367 ionization (ESI) optimized for hydrophilic species, (2) reverse-phase positive ion ESI optimized 1368 for hydrophobic species, (3) reverse-phase negative ion ESI, and (4) hydrophilic interaction 1369 chromatography (HILIC) negative ion ESI. A fifth aliquot was reserved as backup. Following 1370 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 73 drying under nitrogen, extracts were concentrated with a TurboVap® and maintained under inert 1371 gas overnight until analysis by MS. 1372 1373 Each fraction was reconstituted in a solvent system compatible with the intended UPLC-MS/MS 1374 method, each containing internal standards at fixed concentrations to verify injection and 1375 chromatographic consistency. Analyses were performed on a Waters ACQUITY UPLC system 1376 coupled to a Thermo Fisher Q-Exactive high-resolution mass spectrometer (Orbitrap, 35,000 1377 FWHM resolution) equipped with a heated electrospray ionization (HESI-II) source. For reverse-1378 phase methods, extracts were eluted from a Waters BEH C18 column using gradients of water, 1379 methanol, and acetonitrile containing 0.05% perfluoropentanoic acid and 0.1% formic acid. 1380 HILIC separations employed a BEH Amide column with water/acetonitrile and 10 mM 1381 ammonium formate at pH 10.8. MS analysis alternated between full MS and data-dependent MSn 1382 scans across 70-1000 m/z, with dynamic exclusion applied to enhance coverage. 1383 Metabolites were identified by comparison with Metabolon’s spectral library, using retention 1384 time, accurate mass, and MS/MS data. Relative quantitation was performed by peak area 1385 integration, with values normalized to tissue mass. Missing data were imputed with the minimum 1386 observed value. Quality control included pooled technical replicates, blanks, and internal 1387 standards in every run. Instrument performance and reproducibility were monitored by 1388 calculating relative standard deviations, with experimental samples randomized and QC 1389 injections interspersed. 1390 1391 Absolute quantification of GAA in cells, conditioned media, and tissues 1392 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 74 Metabolite preparation for absolute quantification of GAA in cells, media, and tissue was 1393 performed as described above. Standard curves for absolute quantification were generated using 1394 MultiQuant v2.1.1 or GraphPad Prism 10.4.1 by straight-line, non-linear regression with 1/x2 1395 weighting. Parallel cultures for each cell line evaluated were used to obtain cell counts and 1396 average cellular diameter values using a Vi-CELL XR cell viability analyzer (Beckman Coulter). 1397 Total cellular volume was calculated using the formula /g1848/g3404 /g2872 /g2871 /g1499/g2024/g1499/g1866/g1499/g4672 /g3031 /g2870 /g4673 /g2871 , where V represents 1398 cell volume per µL acetonitrile extract, d = average cell diameter, and n = number of cells per µL 1399 extract. Media GAA content was calculated as 1400 /g1833/g1827/g1827 /g3046/g3032/g3030/g3045/g3032/g3047/g3032/g3031 /g3404 /g4670 /g3008/g3002/g3002 /g4671 /g3252/g3262 /g3254/g3299/g3295/g3293/g3276/g3278/g3295 /g1499/g3005/g1499/g3023/g2879 /g4670 /g3008/g3002/g3002 /g4671 /g3270/g3252/g3262 /g3254/g3299/g3295/g3293/g3276/g3278/g3295 /g1499/g3005/g1499/g3023 /g3041 , where [GAA] is the calculated 1401 concentration of GAA in conditioned media (CM) and unconditioned media (UCM) extract, D is 1402 the dilution factor of media in acetonitrile, V is the volume of media in which cells were 1403 incubated for the assay, and n represents the cell count in thousands. For quantification of media 1404 GAA secretion, a signal:noise threshold of conditioned > 2.5 × unconditioned media was applied 1405 to all samples, values below which were considered undetectable. For tissue samples, volumes 1406 were calculated using tissue weights and a previously published value of brain tissue density106. 1407 1408 LC-MS/MS assays for absolute quantification of GAA and creatine in tumor xenograft tissue 1409 LC-MS/MS assays for the quantification of GAA and creatine in AGAT KO tumor studies were 1410 developed and validated according to the FDA guidelines for bioanalytical method validation107. 1411 Briefly, a LC-MS/MS system consisting of a Thermo Fisher Vanquish UPLC and Thermo Fisher 1412 TSQ Quantis Plus that was equipped with a heated ESI (HESI) source was used. The SRM 1413 transitions used for quantitation are as follows: for GAA, m+0 (Q1/Q3: 118.1/76.0), m+2 ISTD 1414 (Q1/Q3: 120.1/78.0); for creatine, m+0 (Q1/Q3: 132.1/90.0), m+5 ISTD (Q1/Q3: 137.1/95.2). 50 1415 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 75 μ L of plasma or brain tissue was protein precipitated using 450 μ L of methanol containing the 1416 internal standards. The sample was vortexed and then centrifuged at 10,000 × g for 10 minutes 1417 before the supernatant was transferred and analyzed by LC-MS/MS. Chromatographic separation 1418 of the samples were accomplished with a Waters Acquity BEH c18 (2.1 x 100 mm, 1.7 μ m) 1419 column with an isocratic elution using 80:20 (A:B, v:v) of water with 0.1% formic acid (A) and 1420 methanol (B) at a flow rate of 0.25 mL/minute. The total runtime was 3 minutes. 1421 1422 A standard curve was created for each run with duplicate QCs (low, medium, and high) to ensure 1423 the results for each day were accurate. The linear range of the standard curves for GAA and 1424 creatine in plasma was 50-25,000 ng/mL and 1-25,000 ng/mL in brain tissue. For both matrices, 1425 each analyte showed excellent linearity (r2 ≥ 0.9955). All FDA guideline criteria were met for 1426 each analyte in brain tissue. High analyte recovery was demonstrated (94% to 112%), and the 1427 intra- and inter- day accuracy (±11.2%) and precision (±8.7%) for quality controls were 1428 excellent. Other criteria, including dilution (±9.6%), stability after three 24 hour freeze thaw 1429 cycles (±4.7%), autosampler stability for 72 hours (±10.2%), bench top stability at RT for 4 1430 hours (±5.3%), and matrix effect (±11.0%) for GAA and creatine also passed. 1431 1432 Immunoblot analysis of protein expression 1433 Cells were lysed in EBC lysis buffer containing a protease inhibitor cocktail (Millipore Sigma 1434 11836153001). Lysates were resolved by SDS-PAGE and transferred to nitrocellulose 1435 membranes (Bio-Rad 1620112 or Bio-Rad 1704158). Primary antibodies used included: anti-1436 GAPDH (D16H11, Cell Signaling Technology 5174S, 1:5000, rabbit monoclonal, RRID: 1437 AB_10622025), anti-vinculin (Sigma V9131, 1:100,000, mouse monoclonal, 1438 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 76 RRID:AB_477629), anti-AGAT (Abcam 119269, 1:1000, mouse monoclonal, RRID: 1439 AB_10902241), anti-GAMT (PTG 10880-1-AP, 1:1000, rabbit polyclonal, RRID: AB_2109304). 1440 HRP-conjugated secondary antibodies used included: anti-Mouse IgG (Thermo Fisher 31430, 1441 1:2,000, goat polyclonal, RRID: AB_228307) and anti-Rabbit IgG (Thermo Fisher 31460, 1442 1:2,000, goat polyclonal, RRID: AB_228341). Densitometry analysis was performed in ImageJ. 1443 1444 ATAC-sequencing and Transcription Factor Footprinting analysis of GATM 1445 Omni-A TAC-seq was performed as described by Corces et al108. 50,000 cells were collected and 1446 DNase treated for 30 minutes at 37°C prior to nuclei extraction and transposition. After 1447 transposition, cleanup was performed using Zymo Clean and Concentrator-5 Kit (Zymo 1448 Research D4013) and DNA was eluted in 21 uL of elution buffer. To add adaptor sequences, 5 1449 cycles of amplification were performed using NEBNext 2× Master Mix and qPCR (New 1450 Engalnd Biolabs M0541S) was used to determine the addition of 0-3 cycles based on sample 1451 amplification profiles. Samples were purified using Zymo Clean and Concentrator-5 Kit. 1452 Libraries were sequenced on the NextSeq 500 (Illumina, RRID:SCR_014983) using paired-end 1453 sequencing of 75 base pair reads by the Molecular Biology Core Facilities at Dana Farber Cancer 1454 Institute. 1455 1456 Fastq files associated with NHA, AGA T-low, and AGA T-high GSCs were analyzed using the 1457 Nextflow-based BICF Astrocyte 2.2.0 A TAC-seq analysis workflow from the UT Southwestern 1458 Bioinformatics Core. Briefly, adaptors were trimmed with TrimGalore89, and reads mapped with 1459 BW A90, filtered with SamTools92, and sorted with Sambamba91. Duplicates were then marked 1460 with Sabamba, reads filtered with SamTools, and percentage of reads in mitochondria calculated 1461 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 77 and library complexity calculated with SamTools and bedtools93. Cross-correlation was 1462 calculated using PhantomPeakQualTools94,110, peaks were called with MACS295 from overlaps 1463 of pooled replicates, and consensus peaks were called and annotated from ChipSeeker output 1464 files. 1465 1466 Footprinting analysis was performed using TOBIAS through UT Southwestern BioHPC. Peaks 1467 were filtered for only Chr peaks and to remove blacklisted peaks. .narrowPeak output from the 1468 above A TACseq workflow was then merged from all samples and replicate bam files were 1469 merged with SamTools. Bam files were then merged according to AGAT expression and 1470 malignancy status. 1471 1472 Splice isoform analysis of GATM 1473 Isoform-specific GATM gene expression data were generated by reanalysis of RNAseq fastq files 1474 using the hg38 build through Astrocyte 2.2.0 to derive relative expression of transcript variants. 1475 Differential gene expression analysis was then performed through Astrocyte 2.2.0. 1476 1477 Single cell RNA sequencing analysis of human GBM samples 1478 Human GBM single cell RNA sequencing analysis was conducted using a previously published 1479 dataset (GSE274546)78. Single cell UMI count files from all primary tumors were downloaded 1480 from the NCBI Sequence Read Archive (SRA) and processed using the R package Seurat. Cells 1481 with less than 1000 or more than 7000 genes, or with greater than 3% mitochondrial fraction 1482 were excluded from further analysis. Subsequent steps including normalization, finding highly 1483 variable genes, scaling, cell clustering and UMAP construction were performed. Marker genes 1484 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 78 for each Seurat cluster were identified by using the Seurat “FindAllMarkers” function, cell 1485 clusters were annotated by marker gene expression as in the original publication78. Gene 1486 expression was visualized by using the Seurat “VlnPlot” function. 1487 1488 RNA sequencing and transcriptional subtype classification of human primary glioma samples 1489 Total RNA from primary human glioma samples was extracted using RNeasy Plus Universal kit 1490 (Qiagen 73404) according to the manufacturer’s instructions. RNA libraries were generated 1491 using Kapa RNA HyperPrep Kits with RiboErase (HMR) (Roche KK8561) 1492 following manufacturer instructions. Briefly, all RNA samples underwent rRNA depletion, 1493 RNAse H and DNase treatment and fragmentation. The RNA fragments were used as template 1494 for cDNA synthesis. The cDNA fragments went through the ligation process with unique 1495 molecular identifier adapters synthesized by IDT. The products were purified and enriched with 1496 PCR amplification to create the final cDNA library. The library quality was verified on Agilent 1497 2100 Bioanalyzer and sequenced on Illumina NovaSeq 6000 sequencer platform using 150 bp 1498 paired-end protocol at the UT Southwestern Genomics & Microarray core. 1499 1500 Sequencing adapters from raw reads were trimmed using Trim Galore (v0.6.4). Trimmed reads 1501 were aligned to the human genome (hg38) using STAR (v2.7.3a). Mapped reads were quantified 1502 using featureCounts from the Subread package (v1.6.3). Transcriptional subtype classification 1503 was done by scoring each glioma samples with gene signatures25,111 of each GBM subtype by the 1504 ssGSEA method from the R package GSV A. 1505 1506 Immunohistochemistry of human brain samples 1507 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 79 H&E and immunohistochemistry procedures for the spatially-resolved human glioma case 1508 (patient ID #5852) were performed at HistoWiz, Inc, using the Leica Bond RX automated stainer 1509 (Leica Microsystems), a Standard Operating Procedure, and a fully automated workflow. 1510 Samples were fixed in 10% formalin for 24 hours, washed in 70% ethanol, and embedded in 1511 paraffin. Samples were shipped as paraffin embedded blocks and sectioned at 4 μ m. One section 1512 per sample was stained in hematoxylin and eosin (H&E). Immunohistochemistry (IHC) was 1513 conducted using a Leica Bond RX automated stainer (Leica Microsystems). HistoWiz in-house 1514 murine skeletal muscle and kidney tissue were used as negative and positive controls, 1515 respectively, for optimization of IHC staining for both AGA T and GAMT. Tissue slides were 1516 dewaxed with xylene and alcohol-based solutions. Heat-induced epitope retrieval (HIER) was 1517 performed for 20 minutes using BOND Epitope Retrieval Solution 1 (H1-20; citrate-based 1518 buffer, pH 6.0; Leica Biosystems AR9961) for the GATM antibody, and Epitope Retrieval 1519 Solution 2 (H2-20; EDTA-based buffer, pH 9.0 AR9640) for the GAMT antibody. Slides were 1520 then incubated with anti-GATM (HPA026077, Atlas Antibodies, RRID: AB_1849528) at a 1:100 1521 dilution and anti-GAMT (PA5-119778, Thermo Fisher, RRID:AB_2913350) at a 1:5,000 dilution 1522 for 30 minutes at room temperature. Signal detection was carried out using the Leica Bond 1523 Polymer Refine Detection Kit (Leica Biosystems DS9800, RRID:AB_2891238), with 3,3’-1524 diaminobenzidine (DAB) as the chromogen and counterstained with hematoxylin following the 1525 manufacturer’s protocol. After staining, slides were dehydrated and coverslipped using the 1526 Tissue-Tek Film® Automated Coverslipper (Sakura Finetek). Whole-slide images were acquired 1527 at 40× magnification using a Leica Aperio AT2 scanner (Leica Microsystems, 1528 RRID:SCR_021256). 1529 1530 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 80 Histology and immunofluorescence of murine PDX samples 1531 Mouse brain tissue for histology and immunofluorescence was dissected and fixed in 10% 1532 neutral buffered formalin (VWR 95042-908) and further processed for embedding and sectioning 1533 by the Histology Core at the University of Pittsburgh. H&E was performed by the Tissue 1534 Management and Shared Resource at the UT Southwestern Medical Center (UTSW), and the 1535 slides were imaged using a Hamamatsu Nanozoomer S60 (RRID:SCR_023762) at the UT 1536 Southwestern Whole Brain Microscopy Facility. 1537 1538 For immunofluorescence labeling, paraffin-embedded brain was sectioned at 4 µm and 1539 deparaffinized in xylene and re-hydrated using graded ethanol series and distilled water. Antigen 1540 retrieval was performed by boiling slides in 1 mM EDTA (pH 8.0) or citric acid buffer (pH 6.0) 1541 in a microwave oven for 25 minutes, followed by blocking (10% normal donkey serum, 0.4% 1542 Triton X-100 in PBS) at room temperature for 1 hour. Slides were incubated at 4 ºC overnight in 1543 the following primary antibodies: anti-KCC2 (1:500, Millipore Sigma C2366), anti-cFOS 1544 (1:1000, Synaptic System 226009, RRID: AB_2943525); anti-cFOS (1:250, Cell Signaling 1545 Technology 2250, RRID:AB_2247211) ; anti-KU80 (1:250, Cell Signaling Technology 2180, 1546 RRID: AB_2218736), and anti-vGLUT1 (1:500, Millipore Sigma AB5905, RRID: 1547 AB_2301751). The following day, slides were incubated at room temperature for 1 hour with 1548 secondary antibodies: 1:500 Alexa Fluor 488 AffiniPure donkey anti-rabbit IgG (H+L) (Jackson 1549 ImmunoResearch, 711-546-152, RRID:AB_2340619); 1:1000 Alexa Fluor 488 donkey anti-1550 chicken IgΥ (H+L) (Invitrogen, A78948); 1:1000 Alexa Fluor 594 goat anti-rabbit IgG (H+L) 1551 (Invitrogen, A11012, RRID:AB_2534079); 1:500 Alexa Fluor 647 AffiniPure donkey anti-guinea 1552 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 81 pig IgG (H+L) (Jackson ImmunoResearch, 706-606-148, RRID:AB_2340477), and 1:1000 1553 Alexa Fluor 647 donkey anti-rabbit IgG (H+L) (Invitrogen A32795). 1554 1555 Both primary and secondary antibodies were diluted in PBS containing 10% normal donkey 1556 serum and 0.2% Triton X-100. The nuclei and slides were stained and mounted by prolong gold 1557 antifade reagent with DAPI (Cayman Chemical 14285) followed by imaging using the Zeiss 1558 LSM880 with Airyscan (RRID:SCR_020925) Confocal Scanning Microscope in the Quantitative 1559 Light Microscopy Core (QLMC) at the UTSW or Zeiss Axioscan 7 Digital Slide Scanner 1560 (RRID:SCR_027284) at the Whole Brain Microscopy Facility at the UTSW. Quantification of 1561 immunofluorescent images was performed using ImageJ 1.43p (Wayne Rasband and 1562 contributors, National Institutes of Health, USA). 1563 1564 Acute brain slice preparation 1565 Acute hippocampal slices used in single-cell electrophysiological recordings in experiments in 1566 which slices were treated with GAA or creatine in the presence or absence of gabazine were 1567 prepared as previously described109 from C57BL/6J mice aged 5-7 weeks old. Briefly, mice were 1568 decapitated and the brain was rapidly extracted in oxygenated (95% O2 and 5% CO2) and ice-1569 cold sucrose-based artificial cerebrospinal fluid (sucrose-aCSF) at pH 7.4 and 300 mOsm 1570 containing: 185 mM sucrose (Sigma Aldrich S9378), 25 mM NaHCO3 (Thermo Fisher, BP328), 1571 2.5 mM KCl (Millipore Sigma, P9333), 25 mM glucose (Millipore Sigma, G8270), 1.25 mM 1572 NaH2PO4 (Sigma Aldrich S8282), 10 mM MgCl2 (Millipore Sigma, M9272), and 0.5 mM CaCl2 1573 (Millipore Sigma, 223506). The brain was dissected, glued to the specimen holder, and acute 1574 slices (300 µm) were prepared on a vibratome (Leica VT1200 S). Slices were transferred to 1575 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 82 oxygenated and heated (32°C) sucrose-aCSF for 30 minutes. After 30 minutes, the water bath 1576 temperature was turned off and slices were transferred to oxygenated recording aCSF at pH 7.4 1577 and 300 mOsm containing: 125 mM NaCl (Millipore Sigma S9888), 25 mM NaHCO3 (Thermo 1578 Fisher BP328), 2.5 mM KCl (Millipore Sigma P9333), 10 mM glucose (Millipore Sigma 1579 G8270), 2 mM CaCl2 (Millipore Sigma 223506), and 2 mM MgCl2 (Millipore Sigma M9272), 1580 and allowed to recover for an additional 30 minutes before recordings. Slices were maintained in 1581 oxygenated aCSF at room temperature for the rest of the day. 1582 1583 Brains for experiments measuring changes in input resistance in response to GAA in the presence 1584 or absence of inhibitors of glutamatergic and GABAergic signaling were prepared as described 1585 previously112. Briefly, wild-type male mice on postnatal days 18-23, were anesthetized by 1586 intraperitoneal injection of ketamine/xylazine. After toe pinch to confirm sufficient depth of 1587 anesthesia, mice were euthanized by rapid decapitation. Acute coronal slices (300 µm thickness) 1588 containing somatosensory barrel cortex were generated using a Leica VT1200S Vibratome in a 1589 semi-frozen 300 mOsm dissection buffer containing: 110 mM choline chloride (Millipore Sigma, 1590 C1879), 25 mM NaHCO3 (Millipore Sigma, S6014) 25 mM D-glucose (Millipore Sigma, 1591 G8270), 11.6 mM ascorbic acid (Millipore Sigma, A92902), 2.5 mM KCl (Millipore Sigma, 1592 P5405), 1.25 mM Na2H2PO4 (Millipore Sigma, S9638), 3.1 mM Na-pyruvate (Millipore Sigma, 1593 P5280), 7 mM MgCl2 (Millipore Sigma, M8266), 0.5 mM CaCl2 (Millipore Sigma, C7902), and 1594 1 mM kynurenic acid (Millipore Sigma, K3375). Brains and slices were continually aerated with 1595 95% O2 and CO2 prior to and during the slicing procedure. After slicing, slices were then 1596 transferred to 300mOsm normal artificial cerebrospinal fluid (ACSF) solution containing: 125 1597 mM NaCl (Millipore Sigma, S7653), 25 mM NaHCO3 (Millipore Sigma, S6014), 10 mM D-1598 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 83 glucose (Millipore Sigma, G8270), 2.5 mM KCl (Millipore Sigma, P5405), 1.25 mM Na2H2PO4 1599 (Millipore Sigma, S9638), 2 mM MgCl2 (Millipore Sigma, M8266), and 2 mM CaCl2 (Millipore 1600 Sigma, C7902) to recover at 37°C for 30 minutes, then at 21°C for 30 minutes prior to 1601 recording. 1602 1603 Acute brain slices for experiments assessing neuronal network activity in tumor-bearing and 1604 contralateral brain hemispheres were obtained from 14-week-old adult mice, 3 weeks following 1605 stereotactic xenograft injection. Slices were prepared at a thickness of 300 µm using a VT1000P 1606 vibratome (Leica) in oxygenated, ice-cold cutting buffer containing 205 mM sucrose (Thermo 1607 Fisher S5), 2.5 mM KCl (Thermo Fisher P330), 1.25 mM NaH/i1 PO/i1 (Thermo Fisher S397), 1608 26 mM NaHCO/i1 (Thermo Fisher P233), 10 mM glucose (Gibco 15023-021), 0.5 mM CaCl/i1 1609 (Thermo Fisher 012316.A1), and 5 mM MgSO/i1 (Thermo Fisher 033337.36). Following 1610 dissection, slices were incubated in recording buffer at 34°C for 40 minutes. 1611 1612 Single cell patch-clamp electrophysiological recordings 1613 Slices generated for electrophysiology experiments in which slices were treated with GAA or 1614 creatine in the presence or absence of gabazine were transferred under an upright microscope 1615 (Scientifica) and held under a harp in a chamber. Slices were perfused at 2 mL/minute with 1616 oxygenated aCSF heated to 32°C. The CA1 pyramidal cell layer was visually identified under a 1617 4× objective (NA = 0.1, Olympus) and visually targeted for whole-cell patch clamp recordings 1618 using a 40× objective (NA = 0.8, Olympus). Electrodes were made from borosilicate glass 1619 capillaries (BF150-86-10HP, Sutter Instrument) on a P-1000 micropipette puller (Sutter 1620 Instrument). Recording electrodes were filled using a pH 7.3 and 290 mOsm solution containing: 1621 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 84 130 mM Cs-methanesulfonate (Millipore Sigma C1426), 10 mM HEPES (Millipore Sigma 1622 H3375), 2 mM MgCl2 (Millipore Sigma M9272), 4 mM ATP (Millipore Sigma A9187), 0.3 mM 1623 GTP (Millipore Sigma G8877), 7 mM Phosphocreatine di(tris) (Millipore Sigma P7936), 0.6 1624 mM EGTA (Millipore Sigma E3889), and 5 mM KCl (Millipore Sigma P9333). The 1625 electrophysiological signal was amplified using a Multiclamp 700B (Molecular Devices) and 1626 digitized at 10 kHz with a Digidata 1550B (Molecular Devices) before being displayed and 1627 recorded on a personal computer using Clampex v11.3 (Molecular Devices). After establishing 1628 the whole cell configuration, neurons were assessed for electrophysiological parameters and 1629 voltage-clamped at 0 mV to begin experiments. Only neurons with a resting membrane potential 1630 more hyperpolarized than -55 mV and access resistance < 30 MΩ were included for analysis. 1631 The liquid junction potential was not corrected. Recording epochs consisted of a 3 minute 1632 baseline with aCSF, followed by 1 minute application of GAA (100 µM) or creatine (100 µM), 1633 and 6 minutes aCSF. GAA and creatine were dissolved in aCSF and prepared fresh every day. 10 1634 µM gabazine (EMD Millipore SR95531) was pre-applied and continuously applied in 1635 experiments where indicated. aCSF and all compounds were delivered using a gravity-driven 1636 perfusion system. Data from hippocampal patch-clamp slice recordings was analyzed in Clampfit 1637 v11.4 and Igor Pro v9.05. The area under curve was measured in Clampfit between recording 1638 time 4-10 minutes. 1639 1640 Recordings for experiments measuring changes in input resistance in response to GAA in the 1641 presence or absence of inhibitors of glutamatergic and gabaergic signaling were conducted at 1642 21°C in a submerged patch-clamp setting and perfused with oxygenated (95% O2/5% CO2) 1643 ACSF (1mL/minute) as described previously112. Depending on experiment, the ACSF contained 1644 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 85 the following inhibitors of neurotransmitter-gated ion channels: 10 µM CPP (Hello Bio 1645 HB0036), 20 µM DNQX (Tocris 2312), 100 µM PTX (Hello Bio HB0506), 2 µM CGP 52432 1646 (Tocris 1246), and/or 1 µM tetrodotoxin (Hello Bio HB1034). Layer V pyramidal neurons were 1647 visualized with differential interference contrast microscopy. Patch pipettes with 5-7 MΩ were 1648 pulled from borosilicate glass. Whole-cell recordings were performed on layer 5a pyramidal 1649 neurons using a Multiclamp 700A amplifier (Molecular Devices). Internal solution contained: 1650 130 mM K-gluconate (Millipore Sigma P1847), 0.2 mM EGTA (Millipore Sigma E3889), 6 mM 1651 KCl (Millipore Sigma P5405), 3 mM NaCl (Millipore Sigma S7653), 10 mM HEPES (Millipore 1652 Sigma H4034), 14 mM phosphocreatine-tris (Millipore Sigma P1937), 4 mM Mg-A TP (Millipore 1653 Sigma A9187) and 0.4 mM Na-GTP (Millipore Sigma G3776). Recordings were conducted in 1654 voltage clamp, holding the membrane potential at -65 mV . Neurons with resting membrane 1655 potential >-60mV and series resistance <40 MΩ were included in analysis. Input resistance in 1656 voltage clamp was measured with a -10 mV voltage step. For each recording, baseline 1657 electrophysiological measurements were obtained, followed by a 100 µM GAA wash-in for 5 1658 minutes. Following 5-minute GAA wash-in, post wash-in measurements were made. Data 1659 acquisition and analysis was performed using custom Labview 8.6 software (National 1660 Instruments). Input resistance following GAA treatment was normalized to pre-treatment input 1661 resistance averages for each neuron. 1662 1663 Multielectrode Array Recording 1664 Spike counts were recorded from acute slices using a 4096-electrode Complementary Metal–1665 Oxide Semiconductor (CMOS)-based HD-MEA (BioCam DupleX with CorePlate™ 3D, 1666 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 86 3Brain). Prior to use, chips were rinsed three times with deionized water, treated with 70% 1667 ethanol, re-rinsed, and preconditioned with PBS. 1668 1669 After incubation at 34°C, slices were incubated an additional 40-minute incubation at room 1670 temperature prior to recording. CLP257-treated slices were incubated in 30 μ M CLP257 (Bio-1671 Techne 5242/10) for 30 minutes immediately before recording. The recording buffer consisted of 1672 126 mM NaCl (Thermo Fisher P271), 3.5 mM KCl (Thermo Fisher P330), 1.25 mM NaH2PO4 1673 (Thermo Fisher S397), 26 mM NaHCO3 (Thermo Fisher P233), 10 mM glucose (Gibco 15023-1674 021), 1.6 mM CaCl2 (Thermo Fisher 012316.A1), and 1.25 mM MgSO4 (Thermo Fisher 1675 0333337.36). Spontaneous activity was sampled continuously for 5 minutes at 20 Hz across 1676 sequential pharmacological conditions: baseline aCSF and 20 µM GAA. Recordings were 1677 acquired under the following conditions (n = 3 slices per condition): Non-tumor-bearing left 1678 hemisphere, tumor-bearing right hemisphere, non-tumor-bearing left hemisphere plus 30 µM 1679 CLP-257 (Bio-Techne 5242/10), and tumor-bearing right hemisphere plus 30 µM CLP-257. 1680 1681 Spike detection was performed using the Precise Timing Spike Detection (PTSD)113 algorithm in 1682 BrainWave 5 software (3Brain), with a threshold set at 8 standard deviations. Spike counts were 1683 analyzed in R (v4.5.0); data were reshaped into long format using tidyverse100 tools and log-1684 transformed as log/i1/i1 (spike count + 1). 1685 1686 Cell growth assay 1687 TS516 AGA T WT and AGAT KO cells were plated in 6-well ultra-low adherence plates at 2 × 1688 105 cells per well in NeuroCult NS-A Basal Medium (Human) prepared as described above. Cell 1689 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 87 counts were obtained at 0, 24, 48, 72, 96, and 120 hours following dissociation in Accutase 1690 (StemCell 07922) using a Vi-CELL XR cell viability analyzer. Media on remaining wells was 1691 changed at 48 and 96 hours. Growth rates were compared using nonlinear regression to fit 1692 exponential growth curves in GraphPad Prism 10.4.1. 1693 1694 Xenograft studies of AGAT expression 1695 Mice were randomly assigned to receive intracranial injections of either TS516 AdCRISPR 1696 AGAT WT or TS516 AdCRISPR AGA T KO cells. Cells were injected in 5 µL NeuroCult NS-A 1697 Basal Medium (Human) 0.5-1 mm anterior and 2 mm lateral to the bregma, at a depth of 3 mm 1698 from the brain surface. Animals were monitored daily until the appearance of neurological 1699 symptoms, at which point they were euthanized. 1700 1701 Xenograft studies of dietary GAA modulation 1702 Mice were randomly assigned to receive one of the following diets: Amino Acid Diet (Inotiv 1703 TD.01084) or Arginine-Deficient, 0.665% Ornithine Diet (GAMT Deficiency Diet, Inotiv 1704 TD.240691). Cells were injected in 5 µL NeuroCult NS-A Basal Medium (Human) 0.5-1 mm 1705 anterior and 2 mm lateral to the bregma, at a depth of 3 mm from the brain surface. Mice were 1706 monitored daily for the development of neurological symptoms, at which point they were 1707 euthanized. 1708 1709 Secretion of GAA by human tissue explants 1710 Tumor tissue was collected and tissue explants generated as described previously48. Briefly, 1711 tissue collected from the operating room was suspended in Hibernate A (BrainBits HA) and 1712 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 88 transferred to RBC Lysis Buffer (Thermo Fisher 00433357) within 30 minutes of excision for 10 1713 minutes with rocking. Tissue was washed three times with Hibernate A supplemented with 2 mM 1714 GlutaMAX (Thermo Fisher 35050061), 100 U/mL and 100 μ g/mL (respectively) 1715 penicillin/streptomycin (Thermo Fisher 15140148), and 250 ng/mL amphotericin B (GeminiBio 1716 400104). Dissection scissors or scalpels were used to parcellate tissues into 1-2 mm3 pieces 1717 which were then plated in separate wells of a 24-well ultra-low adherence plate (Corning 3473) 1718 in 1 mL HPLM supplemented with 1× B27 supplement minus vitamin A (Thermo Fisher 1719 12587010), 1× N2 supplement (Thermo Fisher 17502048), 100 U/mL and 100 μ g/mL 1720 (respectively) penicillin/streptomycin (Thermo Fisher 15140148), 250 ng/mL plasmocin 1721 (InvivoGen ant-mpp), 55µM 2-mercaptoethanol (Thermo Fisher BP176100), and 2.375-2.875 1722 µg/mL insulin (Millipore Sigma I9278). Tissues were randomized to groups of 1 to 4 technical 1723 replicates per tracer from each biologic sample, depending on available tissue; remaining tissue 1724 was used for histopathology. After 30 minutes, media was exchanged for 1 mL HPLM as 1725 described previously48. Tissues were incubated for 18 hours in a 37°C incubator at ambient 1726 oxygen and 5% CO2 with shaking. After 18 hours, conditioned media was harvested, snap-1727 frozen in liquid nitrogen, and stored at -80°C until analysis. Tissues were then prepared for LC-1728 MS analysis and quantification of GAA secretion in media as described above. As in 1729 quantification of media GAA secretion by GSCs, a signal:noise threshold of GAA concentration 1730 in conditioned > 2.5 × unconditioned media was applied to all samples to discriminate GAA-1731 secreting samples. As no cell count information could be obtained for these explants, secreted 1732 GAA was rendered as the unconditioned-subtracted GAA content of conditioned media for those 1733 samples whose GAA content exceeded this threshold. Outliers within technical replicates were 1734 identified and excluded by ROUT test (Q=1%). 1735 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 89 1736 Histopathology of human explants 1737 Residual human tumor and brain tissues from explant preparation were fixed for 1 hour in 1738 neutral buffered 10% formalin solution (Millipore Sigma HT501128). Following fixation, tissues 1739 were washed and stored in 70% ethanol. Tissues were embedded in paraffin, sectioned, and 1740 stained with hematoxylin and eosin (H&E) by the University of Texas Southwestern Histo 1741 Pathology Core. H&E sections were reviewed by a board-certified neuropathologist (T.E.R). 1742 1743 QUANTIFICATION AND STATISTICAL ANALYSIS 1744 Analysis of nitrogen metabolism profiling platform data 1745 Analysis of nitrogen metabolism profiling platform data was performed as previously 1746 described48. Briefly, integrated peaks from LC-MS analysis of nitrogen metabolism profiling 1747 platform data were analyzed using R (RRID: SCR_001905). Metabolites which were not 1748 quantified in more than one sample, had a mean total pool size below a threshold of 1 × 106, had 1749 calculated total labeling across all conditions of less than 1%, had quantified labeling of >4% in 1750 any unlabeled sample, had calculated total labeling across all conditions of greater than 500% 1751 were filtered and removed from subsequent analysis. Differential labeling scores were calculated 1752 using the formula: /g3629 ln /g1858/g1870/g1853/g1855/g1872/g1861/g1867/g1866/g1853/g1864 /g1857/g1866/g1870/g1861/g1855/g1860/g1865/g1857/g1866/g1872 /g3030/g3042/g3041/g3031/g3036/g3047/g3036/g3042/g3041 /g2869 /g1858/g1870/g1853/g1855/g1872/g1861/g1867/g1866/g1853/g1864 /g1857/g1866/g1870/g1861/g1855/g1860/g1865/g1857/g1866/g1872 /g3030/g3042/g3041/g3031/g3036/g3047/g3036/g3042/g3041 /g2870 /g3415 /g3629 /g34001753 /g4666/g1858/g1870/g1853/g1855/g1872/g1861/g1867/g1866/g1853/g1864 /g1857/g1866/g1870/g1861/g1855/g1860/g1865/g1857/g1866/g1872 /g3030/g3042/g3041/g3031/g3036/g3047/g3036/g3042/g3041 /g2869 /g3398 /g1858/g1870/g1853/g1855/g1872/g1861/g1867/g1866/g1853/g1864 /g1857/g1866/g1870/g1861/g1855/g1860/g1865/g1857/g1866/g1872 /g3030/g3042/g3041/g3031/g3036/g3047/g3036/g3042/g3041 /g2870 /g4667 . For visualization, 1754 metabolites with less than 5% total labeling were filtered and removed from Sankey diagrams. 1755 1756 TCGA and GTEX RNA sequencing analysis 1757 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 90

are based in part upon data generated by the TCGA Research 1758 Network: https://www.cancer.gov/tcga74. RNA sequencing data for patient glioblastoma samples 1759 were downloaded from the TCGA portal (https://portal.gdc.cancer.gov/projects/TCGA-GBM), 1760 comprising 173 samples including 154 index tumor samples, 14 recurrent tumor samples, and 5 1761 normal controls sequenced by Illumina HiSeq 2000 RNA Sequencing. Of these, 151 primary 1762 GBM tissue samples with accompanying survival and prognostic information were selected for 1763 analysis. RNA sequencing data from non-malignant brain tissue samples were downloaded from 1764 the GTEx portal (https://gtexportal.org, file version: 2017-06-05_v8_brain_cortex, accession date 1765 August 26, 2019)75,76. Gene expression values for GATM, GAMT, CKB, CKMT1A, CKMT2A, and 1766 CKMT2B were first converted to Transcripts Per Million (TPM) to enable comparison across 1767 samples. A log/i1 (TPM + 1) transformation was then applied to stabilize variance and normalize 1768 the distribution. For each gene, Z-score transformation was performed across all samples to 1769 standardize expression levels and highlight relative up- or down-regulation. Samples were 1770 annotated according to associated histological subtype, IDH mutation status, CIMP status, and 1771 MGMT promoter methylation. 1772 1773 Other statistical analyses 1774 Information related to data presentation and statistical analysis for individual experiments can be 1775 found in the corresponding figure legends. Statistical analyses were carried out using GraphPad 1776 Prism software (version 10.4.1). Significance of all comparisons involving two groups was 1777 calculated by unpaired two-tailed t-test. For comparisons of two groups with significantly 1778 different variances, Welch’s t-test was used. For comparisons of two groups without significant 1779 differences in variances, Student’s t-test was used. Mann-Whitney tests or Kolmogorov-Smirnoff 1780 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 91 tests were used for comparisons between groups in which assumptions of normality were not 1781 appropriate. Significance of all comparisons involving three or more groups was calculated by 1782 one-way ANOV A. Survival data from mouse xenograft studies was analyzed and statistical 1783 significance was determined using log-rank tests. For all tests, two-tailed p-values < 0.05 were 1784 considered statistically significant. 1785 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 92 SUPPLEMENTAL INFORMATION 1786 1787 SUPPLEMENTAL FIGURE LEGENDS 1788 Figure S1. Composition of metabolomics dataset, related to Figure 1. (A) Pie chart depicting 1789 composition of human tissue sample cohort by tissue type. LGG = lower grade glioma. HGG = 1790 high grade glioma. (B) Heatmap depicting gene expression analyses and transcriptional subtype 1791 clustering of human HGG (n=39) and LGG (n=25) tissue samples. (C-D) Partial least-squares 1792 discriminant analysis (PLS-DA) of IDH mutant tumors by (C) tumor histology and (D) tumor 1793 grade. (E) PLS-DA of GBM tissue samples categorized by transcriptional subtype. (F) Peak 1794 intensities of 2-hydroxyglutarate (2HG) in IDH-mutant (IDHmut) and IDH-wild-type (IDHwt) 1795 glioma samples and non-malignant brain samples. (G-K) Peak intensities of (G) L-DOPA, (H) 1796 aminoadipate, (I) orotate, (J) deoxyguanosine, and (K) ratio of peak intensities of N-1797 acetylaspartate to creatine in HGG, LGG, and non-malignant brain samples. Data are presented 1798 as means. n.s.= not significant, *p<0.05, **p<0.01, ***p<0.001 (Welch’s t-test). 1799 1800 Figure S2. Comparison of creatine metabolism pathway intermediates by tumor subtype, 1801 related to Figure 2. (A-D) Peak intensities for (A) GAA, (B) creatine, (C) phosphocreatine, and 1802 (D) creatinine in tissue samples from index (n=15) and recurrent (n=10) LGG tumors. Data are 1803 means; *p<0.05 (Unpaired t-test for creatine; Welch’s t-test for all others). (E-H) Peak intensities 1804 for (E) GAA, (F) creatine, (G) phosphocreatine, and (H) creatinine in tissue samples from both 1805 index and recurrent HGG tumors and non-malignant brain specimens, stratified by diagnostic 1806 and transcriptomic subtypes (Astro, IDHmut, Gr4; n=4) (GBM; n=35) (non-malignant brain; 1807 n=9). GBMs were classified as classical (n=17), mesenchymal (n=12), or proneural (n=6) 1808 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 93 transcriptomic subtypes. (I-L) Peak intensities for (I) GAA, (J) creatine, (K) phosphocreatine, 1809 and (L) creatinine in tissue samples from squamous cell lung carcinoma (n=35) or tumor-1810 adjacent non-malignant lung (n=35). Peak intensities are reanalyzed from Moreno et al. (M-P) 1811 ROC analyses of GAA content as a discriminant marker of (M) HGG (n=39) compared to non-1812 malignant brain (n=9), (N) HGG (n=39) compared to LGG (n=25), or LGG (n=25) compared to 1813 non-malignant brain (n=9) including (O) both index and recurrent tumor resection samples, and 1814 (P) index LGGs only (n=15) compared to non-malignant brain (n=9). Data are presented as 1815 means. n.s.=not significant, *p<0.05, ***p<0.001 (Welch’s t-test in A, C-D, I, and K-L; unpaired 1816 t-test in B and J; one-way ANOVA in E-H). 1817 1818 Figure S3. Arginine tracer accumulation in non-malignant and glioma cell lines, related to 1819 Figure 3. Fractional enrichment of label from 15N4-arginine in intracellular arginine pools 1820 following 18 hour incubation of non-malignant cells (n=4), IDH-mutant GSCs (n=6), and IDH-1821 wild-type GSCs (n=7) in tracer-containing HPLM (n=3 per cell line). Data are means ± SEM. 1822 1823 Figure S4. Transcriptional regulation of creatine synthesis pathway enzymes, related to 1824 Figure 4. (A-C) Expression of (A) AGAT, (B) GAMT, and (C) ratio of AGA T:GAMT expression 1825 by cell type in human GBM by single-cell RNA sequencing. Data are reanalyzed from Nomura 1826 et al78. (D) GAA secretion by HMC3 microglia, NHA Donor #1, and TS516 cells (n=3 per line). 1827 n.s.= not significant, ***p<0.0001. (E) Diagram of mRNA splice isoforms expressed from the 1828 GATM gene. (F) Heatmap of RNA sequencing of non-malignant (NHA Donor #1), AGATlow 1829 (MGG152, BT260), and AGAThigh (all other) GSC cell lines comparing relative abundance of 1830 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 94 GATM mRNA splice isoforms (n=2 per line). Data are fragments per kilobase transcript per 1831 million mapped reads (FPKM). n.s.= not significant, ***p<0.001 (unpaired t-test). 1832 1833 Figure S5. Arginine tracer accumulation in NHA stable lines, related to Figure 5. (A) 1834 Fractional enrichment of label from 15N4-arginine in intracellular arginine pools following 18 1835 hour incubation in tracer-containing HPLM of NHA Donor #1 stable lines expressing either 1836 empty vector (EV), wild-type AGA T cDNA (AGAT WT), or catalytically-dead AGA T cDNA 1837 (AGAT C407A) and Cas9 with either an sgRNA targeting the AAVS1 safe-harbor locus control or 1838 one of two sgRNAs targeting GAMT. (B) Fractional enrichment of label from 15N4-arginine in 1839 intracellular arginine pools following 18 hour incubation in tracer-containing HPLM of NHA 1840 Donor #1 stable lines expressing either empty vector (EV), wild-type AGAT cDNA (AGA T WT), 1841 or catalytically-dead AGA T cDNA (AGA T C407A) and EV , wild-type GAMT cDNA (GAMT 1842 WT), or catalytically-dead GAMT cDNA (GAMT E45S) (n=3 per line). Data are means ± SEM. 1843 1844 Figure S6. Investigation of creatine synthesis pathway enzyme expression and GAA-1845 induced hyperexcitability in the HGG microenvironment, related to Figure 6. (A-B) 1846 Immunohistochemistry staining of kidney (positive control) and skeletal muscle (negative 1847 control) tissues for (A) AGAT and (B) GAMT. Scale bars=100 µm. (C) Neuronal activity in 1848 tumor-infiltrated brain slices before (0-5 minutes) and after (5-10 minutes) treatment with 20 µM 1849 GAA as measured by MEA assay. Representative neuronal spiking is shown in aggregate (top 1850 panels) and across electrodes (bottom panels). 1851 1852 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 95 Figure S7. Investigation of GAA-induced hyperexcitability in vivo, related to Figure 7. (A-1853 B) Standard curves with quality control samples (n=2 per concentration) for absolute 1854 quantification of (A) GAA and (B) creatine in a custom LC-MS assay. (C) Representative 1855 immunofluorescence analysis of cFOS and NeuN markers in tumor-adjacent neocortex regions 1856 of AGA T WT and AGA T KO TS516 xenografted mouse brains. (D) Quantitation of peritumoral 1857 cortical cFOS-positive cells co-expressing NeuN in AGAT WT TS516 xenografted mouse brains 1858 (n= 3 mice). (E-F) Immunofluorescence analysis of vGLUT1 and Ku80 markers in (E) AGAT 1859 WT and (F) AGAT KO TS516 xenografts. ROIs denote regions highlighted in Figures 7K and 1860 7L. Scale bars = 1 mm (wide view) and 100 µm (close view insets). (G) Relative body weights 1861 and (H) diet consumption for mice receiving standard diet or GAMT deficiency diet (n=3 mice 1862 per arm). Data are means ±SEM. 1863 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 96

1864 1. Louis, D.N., Perry, A., Wesseling, P., Brat, D.J., Cree, I.A., Figarella-Branger, D., Hawkins, 1865 C., Ng, H.K., Pfister, S.M., Reifenberger, G., et al. (2021). The 2021 WHO Classification of 1866 Tumors of the Central Nervous System: a summary. Neuro Oncol 23, 1231–1251. 1867 https://doi.org/10.1093/neuonc/noab106. 1868 2. Turcan, S., Rohle, D., Goenka, A., Walsh, L.A., Fang, F., Yilmaz, E., Campos, C., Fabius, 1869 A.W., Lu, C., Ward, P.S., et al. (2012). IDH1 mutation is sufficient to establish the glioma 1870 hypermethylator phenotype. Nature 483, 479–483. https://doi.org/10.1038/nature10866. 1871 3. Spitzer, A., Gritsch, S., Weisman, H.R., Gonzalez Castro, L.N., Nomura, M., Druck, N., 1872 Chanoch-Myers, R., Lee, C.K., Nagashima, H., Miller, J.J., et al. (2021). Mutant IDH 1873 Inhibitors Induce Lineage Differentiation in IDH-mutant Oligodendroglioma (Oncology) 1874 https://doi.org/10.1101/2021.11.16.21266364. 1875 4. Flavahan, W.A., Drier, Y ., Liau, B.B., Gillespie, S.M., Venteicher, A.S., Stemmer-1876 Rachamimov, A.O., Suva, M.L., and Bernstein, B.E. (2016). Insulator dysfunction and 1877 oncogene activation in IDH mutant gliomas. Nature 529, 110–114. 1878 https://doi.org/10.1038/nature16490. 1879 5. Rahme, G.J., Javed, N.M., Puorro, K.L., Xin, S., Hovestadt, V ., Johnstone, S.E., and 1880 Bernstein, B.E. (2023). Modeling epigenetic lesions that cause gliomas. Cell 186, 3674-1881 3685.e14. https://doi.org/10.1016/j.cell.2023.06.022. 1882 6. Koivunen, P., Lee, S., Duncan, C.G., Lopez, G., Lu, G., Ramkissoon, S., Losman, J.A., 1883 Joensuu, P ., Bergmann, U., Gross, S., et al. (2012). Transformation by the (R)-enantiomer of 1884 2-hydroxyglutarate linked to EGLN activation. Nature 483, 484–488. 1885 https://doi.org/10.1038/nature10898. 1886 7. Wu, M.-J., Kondo, H., Kammula, A.V ., Shi, L., Xiao, Y ., Dhiab, S., Xu, Q., Slater, C.J., Avila, 1887 O.I., Merritt, J., et al. (2024). Mutant IDH1 inhibition induces dsDNA sensing to activate 1888 tumor immunity. Science 385, eadl6173. https://doi.org/10.1126/science.adl6173. 1889 8. Xiao, Y ., Shi, D.D., Guo, L., Neumann, E., Levitt, M.M., Kaphle, P., Shipman, T., Li, H., Cai, 1890 F., Ramirez, D.M.O., et al. (2025). Mutant IDH silences GSX2 to reprogram neural progenitor 1891 cell fate and promote gliomagenesis. bioRxiv, 2025.08.12.667486. 1892 https://doi.org/10.1101/2025.08.12.667486. 1893 9. Lu, C., Ward, P.S., Kapoor, G.S., Rohle, D., Turcan, S., Abdel-Wahab, O., Edwards, C.R., 1894 Khanin, R., Figueroa, M.E., Melnick, A., et al. (2012). IDH mutation impairs histone 1895 demethylation and results in a block to cell differentiation. Nature 483, 474–478. 1896 https://doi.org/10.1038/nature10860. 1897 10. Yanchus, C., Drucker, K.L., Kollmeyer, T.M., Tsai, R., Winick-Ng, W., Liang, M., Malik, 1898 A., Pawling, J., De Lorenzo, S.B., Ali, A., et al. (2022). A noncoding single-nucleotide 1899 polymorphism at 8q24 drives IDH1-mutant glioma formation. Science 378, 68–78. 1900 https://doi.org/10.1126/science.abj2890. 1901 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 97 11. Shi, D.D., Savani, M.R., Levitt, M.M., Wang, A.C., Endress, J.E., Bird, C.E., Buehler, J., 1902 Stopka, S.A., Regan, M.S., Lin, Y .F., et al. (2022). De novo pyrimidine synthesis is a 1903 targetable vulnerability in IDH mutant glioma. Cancer Cell 40, 939-956.e16. 1904 https://doi.org/10.1016/j.ccell.2022.07.011. 1905 12. Philip, B., Yu, D.X., Silvis, M.R., Shin, C.H., Robinson, J.P., Robinson, G.L., Welker, 1906 A.E., Angel, S.N., Tripp, S.R., Sonnen, J.A., et al. (2018). Mutant IDH1 Promotes Glioma 1907 Formation In Vivo. Cell reports 23, 1553–1564. https://doi.org/10.1016/j.celrep.2018.03.133. 1908 13. Mellinghoff, I.K., van den Bent, M.J., Blumenthal, D.T., Touat, M., Peters, K.B., Clarke, 1909 J., Mendez, J., Y ust-Katz, S., Welsh, L., Mason, W.P ., et al. (2023). V orasidenib in IDH1- or 1910 IDH2-Mutant Low-Grade Glioma. N Engl J Med 389, 589–601. 1911 https://doi.org/10.1056/NEJMoa2304194. 1912 14. Scott, A.J., Correa, L.O., Edwards, D., Sun, Y ., Ravikumar, V ., Andren, A.C., Zhang, L., 1913 Srinivasan, S., Jairath, N., Verbal, K., et al. (2023). Metabolomic profiles of human glioma 1914 inform patient survival. Antioxid Redox Signal. https://doi.org/10.1089/ars.2022.0085. 1915 15. Chinnaiyan, P., Kensicki, E., Bloom, G., Prabhu, A., Sarcar, B., Kahali, S., Eschrich, S., 1916 Qu, X., Forsyth, P ., and Gillies, R. (2012). The metabolomic signature of malignant glioma 1917 reflects accelerated anabolic metabolism. Cancer Res 72, 5878–5888. 1918 https://doi.org/10.1158/0008-5472.CAN-12-1572-T. 1919 16. Prabhu, A.H., Kant, S., Kesarwani, P., Ahmed, K., Forsyth, P., Nakano, I., and 1920 Chinnaiyan, P. (2019). Integrative cross-platform analyses identify enhanced heterotrophy as a 1921 metabolic hallmark in glioblastoma. Neuro Oncol 21, 337–347. 1922 https://doi.org/10.1093/neuonc/noy185. 1923 17. Wang, L.-B., Karpova, A., Gritsenko, M.A., Kyle, J.E., Cao, S., Li, Y ., Rykunov, D., 1924 Colaprico, A., Rothstein, J.H., Hong, R., et al. (2021). Proteogenomic and metabolomic 1925 characterization of human glioblastoma. Cancer Cell 39, 509-528.e20. 1926 https://doi.org/10.1016/j.ccell.2021.01.006. 1927 18. Jeun, S.-S., Kim, M.-C., Kim, B.-S., Lee, J.-M., Chung, S.-T., Oh, C.-H., Lee, S.-Y ., and 1928 Choe, B.-Y . (2005). Assessment of malignancy in gliomas by 3T 1H MR spectroscopy. Clin 1929 Imaging 29, 10–15. https://doi.org/10.1016/j.clinimag.2004.03.004. 1930 19. Shimizu, H., Kumabe, T., Tominaga, T., Kayama, T., Hara, K., Ono, Y ., Sato, K., Arai, N., 1931 Fujiwara, S., and Yoshimoto, T. (1996). Noninvasive evaluation of malignancy of brain 1932 tumors with proton MR spectroscopy. AJNR Am J Neuroradiol 17, 737–747. 1933 20. Wang, X., Yang, K., Wu, Q., Kim, L.J.Y ., Morton, A.R., Gimple, R.C., Prager, B.C., Shi, 1934 Y ., Zhou, W., Bhargava, S., et al. (2019). Targeting pyrimidine synthesis accentuates 1935 molecular therapy response in glioblastoma stem cells. Sci Transl Med 11, eaau4972. 1936 https://doi.org/10.1126/scitranslmed.aau4972. 1937 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 98 21. Wang, X., Yang, K., Xie, Q., Wu, Q., Mack, S.C., Shi, Y ., Kim, L.J.Y ., Prager, B.C., 1938 Flavahan, W.A., Liu, X., et al. (2017). Purine synthesis promotes maintenance of brain tumor 1939 initiating cells in glioma. Nat Neurosci 20, 661–673. https://doi.org/10.1038/nn.4537. 1940 22. Yuan, H., Wu, X., Wu, Q., Chatoff, A., Megill, E., Gao, J., Huang, T., Duan, T., Yang, K., 1941 Jin, C., et al. (2023). Lysine catabolism reprograms tumour immunity through histone 1942 crotonylation. Nature 617, 818–826. https://doi.org/10.1038/s41586-023-06061-0. 1943 23. Neftel, C., Laffy, J., Filbin, M.G., Hara, T., Shore, M.E., Rahme, G.J., Richman, A.R., 1944 Silverbush, D., Shaw, M.L., Hebert, C.M., et al. (2019). An Integrative Model of Cellular 1945 States, Plasticity, and Genetics for Glioblastoma. Cell 178, 835-849.e21. 1946 https://doi.org/10.1016/j.cell.2019.06.024. 1947 24. Brennan, C.W., Verhaak, R.G., McKenna, A., Campos, B., Noushmehr, H., Salama, S.R., 1948 Zheng, S., Chakravarty, D., Sanborn, J.Z., Berman, S.H., et al. (2013). The somatic genomic 1949 landscape of glioblastoma. Cell 155, 462–477. https://doi.org/10.1016/j.cell.2013.09.034. 1950 25. Verhaak, R.G., Hoadley, K.A., Purdom, E., Wang, V ., Qi, Y ., Wilkerson, M.D., Miller, 1951 C.R., Ding, L., Golub, T., Mesirov, J.P., et al. (2010). Integrated genomic analysis identifies 1952 clinically relevant subtypes of glioblastoma characterized by abnormalities in PDGFRA, 1953 IDH1, EGFR, and NF1. Cancer Cell 17, 98–110. https://doi.org/10.1016/j.ccr.2009.12.020. 1954 26. Venkatesh, H.S., Johung, T.B., Caretti, V ., Noll, A., Tang, Y ., Nagaraja, S., Gibson, E.M., 1955 Mount, C.W., Polepalli, J., Mitra, S.S., et al. (2015). Neuronal Activity Promotes Glioma 1956 Growth through Neuroligin-3 Secretion. Cell 161, 803–816. 1957 https://doi.org/10.1016/j.cell.2015.04.012. 1958 27. Venkatesh, H.S., Tam, L.T., Woo, P.J., Lennon, J., Nagaraja, S., Gillespie, S.M., Ni, J., 1959 Duveau, D.Y ., Morris, P.J., Zhao, J.J., et al. (2017). Targeting neuronal activity-regulated 1960 neuroligin-3 dependency in high-grade glioma. Nature 549, 533–537. 1961 https://doi.org/10.1038/nature24014. 1962 28. Venkataramani, V ., Tanev, D.I., Strahle, C., Studier-Fischer, A., Fankhauser, L., Kessler, 1963 T., Körber, C., Kardorff, M., Ratliff, M., Xie, R., et al. (2019). Glutamatergic synaptic input to 1964 glioma cells drives brain tumour progression. Nature 573, 532–538. 1965 https://doi.org/10.1038/s41586-019-1564-x. 1966 29. Taylor, K.R., Barron, T., Hui, A., Spitzer, A., Yalçin, B., Ivec, A.E., Geraghty, A.C., 1967 Hartmann, G.G., Arzt, M., Gillespie, S.M., et al. (2023). Glioma synapses recruit mechanisms 1968 of adaptive plasticity. Nature 623, 366–374. https://doi.org/10.1038/s41586-023-06678-1. 1969 30. Pan, Y ., Hysinger, J.D., Barron, T., Schindler, N.F., Cobb, O., Guo, X., Yalçın, B., 1970 Anastasaki, C., Mulinyawe, S.B., Ponnuswami, A., et al. (2021). NF1 mutation drives 1971 neuronal activity-dependent initiation of optic glioma. Nature 594, 277–282. 1972 https://doi.org/10.1038/s41586-021-03580-6. 1973 31. Venkatesh, H.S., Morishita, W., Geraghty, A.C., Silverbush, D., Gillespie, S.M., Arzt, M., 1974 Tam, L.T., Espenel, C., Ponnuswami, A., Ni, L., et al. (2019). Electrical and synaptic 1975 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 99 integration of glioma into neural circuits. Nature 573, 539–545. 1976 https://doi.org/10.1038/s41586-019-1563-y. 1977 32. Huang-Hobbs, E., Cheng, Y .-T., Ko, Y ., Luna-Figueroa, E., Lozzi, B., Taylor, K.R., 1978 McDonald, M., He, P., Chen, H.-C., Yang, Y ., et al. (2023). Remote neuronal activity drives 1979 glioma progression through SEMA4F. Nature 619, 844–850. https://doi.org/10.1038/s41586-1980 023-06267-2. 1981 33. Venkataramani, V ., Yang, Y ., Schubert, M.C., Reyhan, E., Tetzlaff, S.K., Wißmann, N., 1982 Botz, M., Soyka, S.J., Beretta, C.A., Pramatarov, R.L., et al. (2022). Glioblastoma hijacks 1983 neuronal mechanisms for brain invasion. Cell 185, 2899-2917.e31. 1984 https://doi.org/10.1016/j.cell.2022.06.054. 1985 34. Campbell, S.L., Buckingham, S.C., and Sontheimer, H. (2012). Human glioma cells 1986 induce hyperexcitability in cortical networks. Epilepsia 53, 1360–1370. 1987 https://doi.org/10.1111/j.1528-1167.2012.03557.x. 1988 35. Buckingham, S.C., Campbell, S.L., Haas, B.R., Montana, V ., Robel, S., Ogunrinu, T., and 1989 Sontheimer, H. (2011). Glutamate release by primary brain tumors induces epileptic activity. 1990 Nat Med 17, 1269–1274. https://doi.org/10.1038/nm.2453. 1991 36. Yu, K., Lin, C.-C.J., Hatcher, A., Lozzi, B., Kong, K., Huang-Hobbs, E., Cheng, Y .-T., 1992 Beechar, V .B., Zhu, W., Zhang, Y ., et al. (2020). PIK3CA variants selectively initiate brain 1993 hyperactivity during gliomagenesis. Nature 578, 166–171. https://doi.org/10.1038/s41586-1994 020-1952-2. 1995 37. John Lin, C.-C., Y u, K., Hatcher, A., Huang, T.-W., Lee, H.K., Carlson, J., Weston, M.C., 1996 Chen, F., Zhang, Y ., Zhu, W., et al. (2017). Identification of diverse astrocyte populations and 1997 their malignant analogs. Nat Neurosci 20, 396–405. https://doi.org/10.1038/nn.4493. 1998 38. Scott, A.J., Mittal, A., Meghdadi, B., Palavalasa, S., Achreja, A., O’Brien, A., Kothari, 1999 A.U., Zhou, W., Xu, J., Lin, A., et al. (2023). Rewiring of cortical glucose metabolism fuels 2000 human brain cancer growth. medRxiv, 2023.10.24.23297489. 2001 https://doi.org/10.1101/2023.10.24.23297489. 2002 39. McBrayer, S.K., Mayers, J.R., DiNatale, G.J., Shi, D.D., Khanal, J., Chakraborty, A.A., 2003 Sarosiek, K.A., Briggs, K.J., Robbins, A.K., Sewastianik, T., et al. (2018). Transaminase 2004 Inhibition by 2-Hydroxyglutarate Impairs Glutamate Biosynthesis and Redox Homeostasis in 2005 Glioma. Cell 175, 101-116.e25. https://doi.org/10.1016/j.cell.2018.08.038. 2006 40. Pope, W.B., Prins, R.M., Albert Thomas, M., Nagarajan, R., Yen, K.E., Bittinger, M.A., 2007 Salamon, N., Chou, A.P., Yong, W.H., Soto, H., et al. (2012). Non-invasive detection of 2-2008 hydroxyglutarate and other metabolites in IDH1 mutant glioma patients using magnetic 2009 resonance spectroscopy. J Neurooncol 107, 197–205. https://doi.org/10.1007/s11060-011-2010 0737-8. 2011 41. Andronesi, O.C., Arrillaga-Romany, I.C., Ly, K.I., Bogner, W., Ratai, E.M., Reitz, K., 2012 Iafrate, A.J., Dietrich, J., Gerstner, E.R., Chi, A.S., et al. (2018). Pharmacodynamics of 2013 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 100 mutant-IDH1 inhibitors in glioma patients probed by in vivo 3D MRS imaging of 2-2014 hydroxyglutarate. Nat Commun 9, 1474. https://doi.org/10.1038/s41467-018-03905-6. 2015 42. Dang, L., White, D.W., Gross, S., Bennett, B.D., Bittinger, M.A., Driggers, E.M., Fantin, 2016 V .R., Jang, H.G., Jin, S., Keenan, M.C., et al. (2009). Cancer-associated IDH1 mutations 2017 produce 2-hydroxyglutarate. Nature 462, 739–744. 2018 43. Choi, C., Ganji, S.K., DeBerardinis, R.J., Hatanpaa, K.J., Rakheja, D., Kovacs, Z., Yang, 2019 X.L., Mashimo, T., Raisanen, J.M., Marin-V alencia, I., et al. (2012). 2-hydroxyglutarate 2020 detection by magnetic resonance spectroscopy in IDH-mutated patients with gliomas. Nat 2021 Med 18, 624–629. https://doi.org/10.1038/nm.2682. 2022 44. Andronesi, O.C., Kim, G.S., Gerstner, E., Batchelor, T., Tzika, A.A., Fantin, V .R., V ander 2023 Heiden, M.G., and Sorensen, A.G. (2012). Detection of 2-hydroxyglutarate in IDH-mutated 2024 glioma patients by in vivo spectral-editing and 2D correlation magnetic resonance 2025 spectroscopy. Sci Transl Med 4, 116ra4. https://doi.org/10.1126/scitranslmed.3002693. 2026 45. Choi, C., Raisanen, J.M., Ganji, S.K., Zhang, S., McNeil, S.S., An, Z., Madan, A., 2027 Hatanpaa, K.J., Vemireddy, V ., Sheppard, C.A., et al. (2016). Prospective Longitudinal 2028 Analysis of 2-Hydroxyglutarate Magnetic Resonance Spectroscopy Identifies Broad Clinical 2029 Utility for the Management of Patients With IDH-Mutant Glioma. J Clin Oncol 34, 4030–2030 4039. https://doi.org/10.1200/JCO.2016.67.1222. 2031 46. Moreno, P., Jiménez-Jiménez, C., Garrido-Rodríguez, M., Calderón-Santiago, M., 2032 Molina, S., Lara-Chica, M., Priego-Capote, F., Salvatierra, Á., Muñoz, E., and Calzado, M.A. 2033 (2018). Metabolomic profiling of human lung tumor tissues - nucleotide metabolism as a 2034 candidate for therapeutic interventions and biomarkers. Mol Oncol 12, 1778–1796. 2035 https://doi.org/10.1002/1878-0261.12369. 2036 47. Riviere-Cazaux, C., Carlstrom, L.P., Rajani, K., Munoz-Casabella, A., Rahman, M., 2037 Gharibi-Loron, A., Brown, D.A., Miller, K.J., White, J.J., Himes, B.T., et al. (2023). Blood-2038 brain barrier disruption defines the extracellular metabolome of live human high-grade 2039 gliomas. Commun Biol 6, 653. https://doi.org/10.1038/s42003-023-05035-2. 2040 48. Savani, M.R., Smith, B.C., Gu, W., Xiao, Y ., Baquer, G., Li, B., Oken, S.S., Manoj, N., 2041 Zacharias, L.G., Puliyappadamba, V .T., et al. (2025). Nitrogen metabolism profiling reveals 2042 cell state-specific pyrimidine synthesis pathway choice. Preprint at Cell Biology, 2043 https://doi.org/10.1101/2025.07.23.666448 https://doi.org/10.1101/2025.07.23.666448. 2044 49. Cantor, J.R., Abu-Remaileh, M., Kanarek, N., Freinkman, E., Gao, X., Louissaint, A., 2045 Lewis, C.A., and Sabatini, D.M. (2017). Physiologic Medium Rewires Cellular Metabolism 2046 and Reveals Uric Acid as an Endogenous Inhibitor of UMP Synthase. Cell 169, 258-272.e17. 2047 https://doi.org/10.1016/j.cell.2017.03.023. 2048 50. Savani, M.R., El Shami, M., Gattie, L.C., Smith, B.C., Hicks, W.H., Oken, S.S., 2049 Zacharias, L.G., Martin-Sandoval, M.S., Montgomery, E.Y ., Xiao, Y ., et al. (2023). Stable 2050 isotope tracing in human plasma-like medium reveals metabolic and immune modulation of 2051 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 101 the glioblastoma microenvironment. Preprint at Cancer Biology, 2052 https://doi.org/10.1101/2023.05.29.542774 https://doi.org/10.1101/2023.05.29.542774. 2053 51. Abdullah, K.G., Bird, C.E., Buehler, J.D., Gattie, L.C., Savani, M.R., Sternisha, A.C., 2054 Xiao, Y ., Levitt, M.M., Hicks, W.H., Li, W., et al. (2022). Establishment of patient-derived 2055 organoid models of lower-grade glioma. Neuro Oncol 24, 612–623. 2056 https://doi.org/10.1093/neuonc/noab273. 2057 52. Jacob, F., Salinas, R.D., Zhang, D.Y ., Nguyen, P.T.T., Schnoll, J.G., Wong, S.Z.H., 2058 Thokala, R., Sheikh, S., Saxena, D., Prokop, S., et al. (2020). A Patient-Derived Glioblastoma 2059 Organoid Model and Biobank Recapitulates Inter- and Intra-tumoral Heterogeneity. Cell 180, 2060 188-204 e22. https://doi.org/10.1016/j.cell.2019.11.036. 2061 53. Hicks, W.H., Gattie, L.C., Traylor, J.I., Davar, D., Najjar, Y .G., Richardson DO, T.E., 2062 McBrayer, S.K., and Abdullah, K.G. (2024). Matched three-dimensional organoids and two-2063 dimensional cell lines of melanoma brain metastases mirror response to targeted molecular 2064 therapy. bioRxiv, 2024.01.18.576318. https://doi.org/10.1101/2024.01.18.576318. 2065 54. Rashidi, A., Billingham, L.K., Zolp, A., Chia, T.-Y ., Silvers, C., Katz, J.L., Park, C.H., 2066 Delay, S., Boland, L., Geng, Y ., et al. (2024). Myeloid cell-derived creatine in the hypoxic 2067 niche promotes glioblastoma growth. Cell Metab 36, 62-77.e8. 2068 https://doi.org/10.1016/j.cmet.2023.11.013. 2069 55. Humm, A., Fritsche, E., Mann, K., Göhl, M., and Huber, R. (1997). Recombinant 2070 expression and isolation of human L-arginine:glycine amidinotransferase and identification of 2071 its active-site cysteine residue. Biochem J 322 ( Pt 3), 771–776. 2072 https://doi.org/10.1042/bj3220771. 2073 56. Komoto, J., Yamada, T., Takata, Y ., Konishi, K., Ogawa, H., Gomi, T., Fujioka, M., and 2074 Takusagawa, F. (2004). Catalytic mechanism of guanidinoacetate methyltransferase: crystal 2075 structures of guanidinoacetate methyltransferase ternary complexes. Biochemistry 43, 14385–2076 14394. https://doi.org/10.1021/bi0486785. 2077 57. Schulze, A. (2003). Creatine deficiency syndromes. Mol Cell Biochem 244, 143–150. 2078 58. Neu, A., Neuhoff, H., Trube, G., Fehr, S., Ullrich, K., Roeper, J., and Isbrandt, D. (2002). 2079 Activation of GABA(A) receptors by guanidinoacetate: a novel pathophysiological 2080 mechanism. Neurobiol Dis 11, 298–307. https://doi.org/10.1006/nbdi.2002.0547. 2081 59. Meera, P., Uusi-Oukari, M., Wallner, M., and Lipshutz, G.S. (2023). Guanidinoacetate 2082 (GAA) is a potent GABAA receptor GABA mimetic: Implications for neurological disease 2083 pathology. J Neurochem 165, 445–454. https://doi.org/10.1111/jnc.15774. 2084 60. Zugno, A.I., Oliveira, D.L., Scherer, E.B.S., Wajner, M., Wofchuk, S., and Wyse, A.T.S. 2085 (2007). Guanidinoacetate inhibits glutamate uptake in rat striatum of rats at different ages. 2086 Neurochem Res 32, 959–964. https://doi.org/10.1007/s11064-006-9245-1. 2087 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 102 61. Zugno, A.I., Scherer, E.B.S., Schuck, P.F., Oliveira, D.L., Wofchuk, S., Wannmacher, 2088 C.M.D., Wajner, M., and Wyse, A.T.S. (2006). Intrastriatal administration of guanidinoacetate 2089 inhibits Na+, K+-ATPase and creatine kinase activities in rat striatum. Metab Brain Dis 21, 2090 41–50. https://doi.org/10.1007/s11011-006-9003-8. 2091 62. Krishna, S., Choudhury, A., Keough, M.B., Seo, K., Ni, L., Kakaizada, S., Lee, A., 2092 Aabedi, A., Popova, G., Lipkin, B., et al. (2023). Glioblastoma remodelling of human neural 2093 circuits decreases survival. Nature. https://doi.org/10.1038/s41586-023-06036-1. 2094 63. Osswald, M., Jung, E., Sahm, F., Solecki, G., Venkataramani, V ., Blaes, J., Weil, S., 2095 Horstmann, H., Wiestler, B., Syed, M., et al. (2015). Brain tumour cells interconnect to a 2096 functional and resistant network. Nature 528, 93–98. https://doi.org/10.1038/nature16071. 2097 64. Sontheimer, H. (2008). A role for glutamate in growth and invasion of primary brain 2098 tumors. J Neurochem 105, 287–295. https://doi.org/10.1111/j.1471-4159.2008.05301.x. 2099 65. Campbell, S.L., Robel, S., Cuddapah, V .A., Robert, S., Buckingham, S.C., Kahle, K.T., 2100 and Sontheimer, H. (2015). GABAergic disinhibition and impaired KCC2 cotransporter 2101 activity underlie tumor-associated epilepsy. Glia 63, 23–36. 2102 https://doi.org/10.1002/glia.22730. 2103 66. MacKenzie, G., O’Toole, K.K., Moss, S.J., and Maguire, J. (2016). Compromised 2104 GABAergic inhibition contributes to tumor-associated epilepsy. Epilepsy Res 126, 185–196. 2105 https://doi.org/10.1016/j.eplepsyres.2016.07.010. 2106 67. Stockler-Ipsiroglu, S., van Karnebeek, C., Longo, N., Korenke, G.C., Mercimek-2107 Mahmutoglu, S., Marquart, I., Barshop, B., Grolik, C., Schlune, A., Angle, B., et al. (2014). 2108 Guanidinoacetate methyltransferase (GAMT) deficiency: outcomes in 48 individuals and 2109 recommendations for diagnosis, treatment and monitoring. Mol Genet Metab 111, 16–25. 2110 https://doi.org/10.1016/j.ymgme.2013.10.018. 2111 68. Schulze, A., Ebinger, F., Rating, D., and Mayatepek, E. (2001). Improving treatment of 2112 guanidinoacetate methyltransferase deficiency: reduction of guanidinoacetic acid in body 2113 fluids by arginine restriction and ornithine supplementation. Mol Genet Metab 74, 413–419. 2114 https://doi.org/10.1006/mgme.2001.3257. 2115 69. Santagata, S., Eberlin, L.S., Norton, I., Calligaris, D., Feldman, D.R., Ide, J.L., Liu, X., 2116 Wiley, J.S., Vestal, M.L., Ramkissoon, S.H., et al. (2014). Intraoperative mass spectrometry 2117 mapping of an onco-metabolite to guide brain tumor surgery. Proc Natl Acad Sci U S A 111, 2118 11121–11126. https://doi.org/10.1073/pnas.1404724111. 2119 70. Eberlin, L.S., Norton, I., Orringer, D., Dunn, I.F., Liu, X., Ide, J.L., Jarmusch, A.K., 2120 Ligon, K.L., Jolesz, F.A., Golby, A.J., et al. (2013). Ambient mass spectrometry for the 2121 intraoperative molecular diagnosis of human brain tumors. Proc Natl Acad Sci U S A 110, 2122 1611–1616. https://doi.org/10.1073/pnas.1215687110. 2123 71. Pirro, V ., Alfaro, C.M., Jarmusch, A.K., Hattab, E.M., Cohen-Gadol, A.A., and Cooks, 2124 R.G. (2017). Intraoperative assessment of tumor margins during glioma resection by 2125 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 103 desorption electrospray ionization-mass spectrometry. Proc Natl Acad Sci U S A 114, 6700–2126 6705. https://doi.org/10.1073/pnas.1706459114. 2127 72. Marques, E.P., and Wyse, A.T.S. (2016). Guanidinoacetate Methyltransferase Deficiency: 2128 A Review of Guanidinoacetate Neurotoxicity. Journal of Inborn Errors of Metabolism and 2129 Screening 4, 232640981666937. https://doi.org/10.1177/2326409816669371. 2130 73. Barron, T., Yalçın, B., Su, M., Byun, Y .G., Gavish, A., Shamardani, K., Xu, H., Ni, L., 2131 Soni, N., Mehta, V ., et al. (2025). GABAergic neuron-to-glioma synapses in diffuse midline 2132 gliomas. Nature 639, 1060–1068. https://doi.org/10.1038/s41586-024-08579-3. 2133 74. Scarpace, L., Mikkelsen, T., Cha, S., Rao, S., Tekchandani, S., Gutman, D., Saltz, J.H., 2134 Erickson, B.J., Pedano, N., Flanders, A.E., et al. (2016). The Cancer Genome Atlas 2135 Glioblastoma Multiforme Collection (TCGA-GBM). Version 4 (The Cancer Imaging 2136 Archive). https://doi.org/10.7937/K9/TCIA.2016.RNYFUYE9 2137 https://doi.org/10.7937/K9/TCIA.2016.RNYFUYE9. 2138 75. Lonsdale, J., Thomas, J., Salvatore, M., Phillips, R., Lo, E., Shad, S., Hasz, R., Walters, 2139 G., Garcia, F., Young, N., et al. (2013). The Genotype-Tissue Expression (GTEx) project. Nat 2140 Genet 45, 580–585. https://doi.org/10.1038/ng.2653. 2141 76. The GTEx Consortium, Aguet, F., Anand, S., Ardlie, K.G., Gabriel, S., Getz, G.A., 2142 Graubert, A., Hadley, K., Handsaker, R.E., Huang, K.H., et al. (2020). The GTEx Consortium 2143 atlas of genetic regulatory effects across human tissues. Science 369, 1318–1330. 2144 https://doi.org/10.1126/science.aaz1776. 2145 77. Schmidt, A., Marescau, B., Boehm, E.A., Renema, W.K.J., Peco, R., Das, A., Steinfeld, 2146 R., Chan, S., Wallis, J., Davidoff, M., et al. (2004). Severely altered guanidino compound 2147 levels, disturbed body weight homeostasis and impaired fertility in a mouse model of 2148 guanidinoacetate N-methyltransferase (GAMT) deficiency. Hum Mol Genet 13, 905–921. 2149 https://doi.org/10.1093/hmg/ddh112. 2150 78. Nomura, M., Spitzer, A., Johnson, K.C., Garofano, L., Nehar-Belaid, D., Galili Darnell, 2151 N., Greenwald, A.C., Bussema, L., Oh, Y .T., V arn, F.S., et al. (2025). The multilayered 2152 transcriptional architecture of glioblastoma ecosystems. Nat Genet 57, 1155–1167. 2153 https://doi.org/10.1038/s41588-025-02167-5. 2154 79. Sonoda, Y ., Ozawa, T., Hirose, Y ., Aldape, K.D., McMahon, M., Berger, M.S., and Pieper, 2155 R.O. (2001). Formation of intracranial tumors by genetically modified human astrocytes 2156 defines four pathways critical in the development of human anaplastic astrocytoma. Cancer 2157 Res 61, 4956–4960. 2158 80. Kelly, J.J.P., Blough, M.D., Stechishin, O.D.M., Chan, J.A.W., Beauchamp, D., Perizzolo, 2159 M., Demetrick, D.J., Steele, L., Auer, R.N., Hader, W.J., et al. (2010). Oligodendroglioma cell 2160 lines containing t(1;19)(q10;p10). Neuro-Oncology 12, 745–755. 2161 https://doi.org/10.1093/neuonc/noq031. 2162 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 104 81. Laks, D.R., Crisman, T.J., Shih, M.Y .S., Mott ahedeh, J., Gao, F., Sperry, J., Garrett, M.C., 2163 Yong, W.H., Cloughesy, T.F., Liau, L.M., et al. (2016). Large-scale assessment of the 2164 gliomasphere model system. NEUONC 18, 1367–1378. 2165 https://doi.org/10.1093/neuonc/now045. 2166 82. Rohle, D., Popovici-Muller, J., Palaskas, N., Turcan, S., Grommes, C., Campos, C., Tsoi, 2167 J., Clark, O., Oldrini, B., Komisopoulou, E., et al. (2013). An Inhibitor of Mutant IDH1 2168 Delays Growth and Promotes Differentiation of Glioma Cells. Science 340, 626–630. 2169 https://doi.org/doi:10.1126/science.1236062. 2170 83. Nguyen, T.P., Wang, W., Sternisha, A.C., Corley, C.D., Wang, H.L., Wang, X., Ortiz, F., 2171 Lim, S.K., Abdullah, K.G., Parada, L.F., et al. (2023). Selective and brain-penetrant lanosterol 2172 synthase inhibitors target glioma stem-like cells by inducing 24(S),25-epoxycholesterol 2173 production. Cell Chem Biol 30, 214-229.e18. https://doi.org/10.1016/j.chembiol.2023.01.005. 2174 84. Wang, X., Prager, B.C., Wu, Q., Kim, L.J.Y ., Gimple, R.C., Shi, Y ., Yang, K., Morton, 2175 A.R., Zhou, W., Zhu, Z., et al. (2018). Reciprocal Signaling between Glioblastoma Stem Cells 2176 and Differentiated Tumor Cells Promotes Malignant Progression. Cell Stem Cell 22, 514-2177 528.e5. https://doi.org/10.1016/j.stem.2018.03.011. 2178 85. Wakimoto, H., Tanaka, S., Curry, W.T., Loebel, F., Zhao, D., Tateishi, K., Chen, J., 2179 Klofas, L.K., Lelic, N., Kim, J.C., et al. (2014). Targetable Signaling Pathway Mutations Are 2180 Associated with Malignant Phenotype in IDH -Mutant Gliomas. Clinical Cancer Research 20, 2181 2898–2909. https://doi.org/10.1158/1078-0432.CCR-13-3052. 2182 86. Chen, C.-H., Xiao, T., Xu, H., Jiang, P ., Meyer, C.A., Li, W., Brown, M., and Liu, X.S. 2183 (2018). Improved design and analysis of CRISPR knockout screens. Bioinformatics 34, 4095–2184 4101. https://doi.org/10.1093/bioinformatics/bty450. 2185 87. Su, X., Lu, W., and Rabinowitz, J.D. (2017). Metabolite Spectral Accuracy on Orbitraps. 2186 Anal. Chem. 89, 5940–5948. https://doi.org/10.1021/acs.analchem.7b00396. 2187 88. Schneider, C.A., Rasband, W.S., and Eliceiri, K.W. (2012). NIH Image to ImageJ: 25 2188 years of image analysis. Nat Methods 9, 671–675. https://doi.org/10.1038/nmeth.2089. 2189 89. Krueger, F., James, F., Ewels, P., Afyounian, E., Weinstein, M., Schuster-Boeckler, B., 2190 Hulselmans, G., and Sclamons (2023). FelixKrueger/TrimGalore: v0.6.10 - add default 2191 decompression path. Version 0.6.10 (Zenodo). https://doi.org/10.5281/ZENODO.7598955 2192 https://doi.org/10.5281/ZENODO.7598955. 2193 90. Li, H. (2013). Aligning sequence reads, clone sequences and assembly contigs with 2194 BWA-MEM. Preprint at arXiv, https://doi.org/10.48550/arXiv.1303.3997 2195 https://doi.org/10.48550/arXiv.1303.3997. 2196 91. Tarasov, A., Vilella, A.J., Cuppen, E., Nijman, I.J., and Prins, P. (2015). Sambamba: fast 2197 processing of NGS alignment formats. Bioinformatics 31, 2032–2034. 2198 https://doi.org/10.1093/bioinformatics/btv098. 2199 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 105 92. Danecek, P., Bonfield, J.K., Liddle, J., Marshall, J., Ohan, V ., Pollard, M.O., Whitwham, 2200 A., Keane, T., McCarthy, S.A., Davies, R.M., et al. (2021). Twelve years of SAMtools and 2201 BCFtools. GigaScience 10, giab008. https://doi.org/10.1093/gigascience/giab008. 2202 93. Quinlan, A.R., and Hall, I.M. (2010). BEDTools: a flexible suite of utilities for 2203 comparing genomic features. Bioinformatics 26, 841–842. 2204 https://doi.org/10.1093/bioinformatics/btq033. 2205 94. Kharchenko, P.V ., Tolstorukov, M.Y ., and Park, P.J. (2008). Design and analysis of ChIP-2206 seq experiments for DNA-binding proteins. Nat Biotechnol 26, 1351–1359. 2207 https://doi.org/10.1038/nbt.1508. 2208 95. Zhang, Y ., Liu, T., Meyer, C.A., Eeckhoute, J., Johnson, D.S., Bernstein, B.E., Nusbaum, 2209 C., Myers, R.M., Brown, M., Li, W., et al. (2008). Model-based Analysis of ChIP-Seq 2210 (MACS). Genome Biol 9, R137. https://doi.org/10.1186/gb-2008-9-9-r137. 2211 96. Satija, R., Farrell, J.A., Gennert, D., Schier, A.F., and Regev, A. (2015). Spatial 2212 reconstruction of single-cell gene expression data. Nat Biotechnol 33, 495–502. 2213 https://doi.org/10.1038/nbt.3192. 2214 97. Dobin, A., Davis, C.A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., Batut, P., 2215 Chaisson, M., and Gingeras, T.R. (2013). STAR: ultrafast universal RNA-seq aligner. 2216 Bioinformatics 29, 15–21. https://doi.org/10.1093/bioinformatics/bts635. 2217 98. Liao, Y ., Smyth, G.K., and Shi, W. (2013). The Subread aligner: fast, accurate and 2218 scalable read mapping by seed-and-vote. Nucleic Acids Research 41, e108–e108. 2219 https://doi.org/10.1093/nar/gkt214. 2220 99. Hänzelmann, S., Castelo, R., and Guinney, J. (2013). GSV A: gene set variation analysis 2221 for microarray and RNA-Seq data. BMC Bioinformatics 14, 7. https://doi.org/10.1186/1471-2222 2105-14-7. 2223 100. Wickham, H., Averick, M., Bryan, J., Chang, W., McGowan, L., François, R., 2224 Grolemund, G., Hayes, A., Henry, L., Hester, J., et al. (2019). Welcome to the Tidyverse. 2225 JOSS 4, 1686. https://doi.org/10.21105/joss.01686. 2226 101. Xiao, Y ., Clima, R., Busch, J., Rabien, A., Kilic, E., Villegas, S.L., Timmermann, B., 2227 Attimonelli, M., Jung, K., and Meierhofer, D. (2020). Decreased Mitochondrial DNA Content 2228 Drives OXPHOS Dysregulation in Chromophobe Renal Cell Carcinoma. Cancer Res. 80, 2229 3830–3840. https://doi.org/10.1158/0008-5472.CAN-20-0754. 2230 102. Aurora, A.B., Khivansara, V ., Leach, A., Gill, J.G., Martin-Sandoval, M., Y ang, C., 2231 Kasitinon, S.Y ., Bezwada, D., Tasdogan, A., Gu, W., et al. (2022). Loss of glucose 6-2232 phosphate dehydrogenase function increases oxidative stress and glutaminolysis in 2233 metastasizing melanoma cells. Proc Natl Acad Sci U S A 119, e2120617119. 2234 https://doi.org/10.1073/pnas.2120617119. 2235 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 106 103. DeVilbiss, A.W., Zhao, Z., Martin-Sandoval, M.S., Ubellacker, J.M., Tasdogan, A., 2236 Agathocleous, M., Mathews, T.P., and Morrison, S.J. (2021). Metabolomic profiling of rare 2237 cell populations isolated by flow cytometry from tissues. Elife 10, e61980. 2238 https://doi.org/10.7554/eLife.61980. 2239 104. Faubert, B., Tasdogan, A., Morrison, S.J., Mathews, T.P., and DeBerardinis, R.J. (2021). 2240 Stable isotope tracing to assess tumor metabolism in vivo. Nat Protoc 16, 5123–5145. 2241 https://doi.org/10.1038/s41596-021-00605-2. 2242 105. Dieterle, F., Ross, A., Schlotterbeck, G., and Senn, H. (2006). Probabilistic Quotient 2243 Normalization as Robust Method to Account for Dilution of Complex Biological Mixtures. 2244 Application in1 H NMR Metabonomics. Anal. Chem. 78, 4281–4290. 2245 https://doi.org/10.1021/ac051632c. 2246 106. Bothe, H.-W., Bodsch, W., and Hossmann, K.-A. (1984). Relationship between specific 2247 gravity, water content, and serum protein extravasation in various types of vasogenic brain 2248 edema. Acta Neuropathol 64, 37–42. https://doi.org/10.1007/BF00695604. 2249 107. Food and Drug Administration (2018). Bioanalytical method validation: Guidance for 2250 Industry. 2251 108. Corces, M.R., Trevino, A.E., Hamilton, E.G., Greenside, P.G., Sinnott-Armstrong, N.A., 2252 Vesuna, S., Satpathy, A.T., Rubin, A.J., Montine, K.S., Wu, B., et al. (2017). An improved 2253 ATAC-seq protocol reduces background and enables interrogation of frozen tissues. Nat 2254

14, 959–962. https://doi.org/10.1038/nmeth.4396. 2255 109. Krueger, F., James, F., Ewels, P., Afyounian, E., Weinstein, M., Schuster-Boeckler, B., 2256 Hulselmans, G., and Sclamons (2023). FelixKrueger/TrimGalore: v0.6.10 - add default 2257 decompression path. Version 0.6.10 (Zenodo). https://doi.org/10.5281/ZENODO.7598955 2258 https://doi.org/10.5281/ZENODO.7598955. 2259 110. Landt, S.G., Marinov, G.K., Kundaje, A., Kheradpour, P ., Pauli, F., Batzoglou, S., 2260 Bernstein, B.E., Bickel, P ., Brown, J.B., Cayting, P ., et al. (2012). ChIP-seq guidelines and 2261 practices of the ENCODE and modENCODE consortia. Genome Res. 22, 1813–1831. 2262 https://doi.org/10.1101/gr.136184.111. 2263 111. Varn, F.S., Johnson, K.C., Martinek, J., Huse, J.T., Nasrallah, M.P., Wesseling, P ., Cooper, 2264 L.A.D., Malta, T.M., Wade, T.E., Sabedot, T.S., et al. (2022). Glioma progression is shaped by 2265 genetic evolution and microenvironment interactions. Cell 185, 2184-2199.e16. 2266 https://doi.org/10.1016/j.cell.2022.04.038. 2267 112. Rajkovich, K.E., Loerwald, K.W., Hale, C.F., Hess, C.T., Gibson, J.R., and Huber, K.M. 2268 (2017). Experience-Dependent and Differential Regulation of Local and Long-Range 2269 Excitatory Neocortical Circuits by Postsynaptic Mef2c. Neuron 93, 48–56. 2270 https://doi.org/10.1016/j.neuron.2016.11.022. 2271 113. Maccione, A., Gandolfo, M., Massobrio, P., Novellino, A., Martinoia, S., and 2272 Chiappalone, M. (2009). A novel algorithm for precise identification of spikes in 2273 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint 107 extracellularly recorded neuronal signals. J Neurosci Methods 177, 241–249. 2274 https://doi.org/10.1016/j.jneumeth.2008.09.026. 2275 2276 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted September 18, 2025. ; https://doi.org/10.1101/2025.09.15.676412doi: bioRxiv preprint