Warrington, A.P.

Abstract BACKGROUND Cerebrospinal fluid (CSF) is emerging as an abundant source of candidate biomarkers, including proteins, for gliomas. Longitudinal CSF access may enable insights into treatment response and resistance. To that end, we have established a growing biobank of CSF samples, many of which have been obtained longitudinally via CSF access devices, such as Ommaya reservoirs, through our ongoing biomarker discovery clinical trials (NCT04692337, 04692324). We have previously presented our initial data demonstrating a conserved proteomic signature of HGG that decreased with resection and increased with recurrence. Since then, samples have been acquired throughout individual patients’ treatments toward our goal of identifying CSF-derived biomarkers of treatment response. MATERIAL AND METHODS CSF samples obtained prior to, during, and throughout treatment underwent global proteomics analysis via the SomaScan® assay by Somalogic, which uses slow off-rate modified aptamers to sensitively detect over 7,000 proteins. Fold-changes for each protein were calculated between different timepoints for each patient and ranked based on the comparison of interest. Comparisons included prior to versus after resection, chemoradiation or adjuvant temozolomide (each with or without pembrolizumab), bevacizumab, and low versus high blood contamination. RESULTS Treatment with pembrolizumab alone resulted in increased IL-23 (5.01x), IL-12 (4.25x), lymphocyte activation gene 3 protein (LAG3) (3.26x), and granzyme A (2.93x). Chemoradiation, with or without pembrolizumab, consistently downregulated NAD kinase, acetyl CoA transferase, and Bcl-2 and upregulated proteins related to neuronal signaling, such as neurexins and cerebellins, in addition to LAG-3, suggesting potential for synergy between chemoradiation and anti-LAG-3 antibodies currently in clinical trials. Interestingly, pembrolizumab and chemoradiation decreased MMP-12 and MMP-9 abundance as compared to chemoradiation alone (0.25x vs. 18.5x; 0.54x vs. 4.99x). Finally, bevacizumab treatment consistently increased VEGF (7.32x), GFAP (3.38x), and MMP-3 (2.99x) abundance, while decreasing the presence of our previously reported plasma-associated protein signature in CSF, suggesting decreased blood-brain barrier disruption. CONCLUSION The global CSF proteome can be longitudinally accessed to evaluate response to standard-of-care and experimental therapies. Further studies are ongoing to identify signatures of therapeutic resistance and to assess the impact of novel immunotherapeutic combinations on the CSF proteome.

Abstract BACKGROUND D-2-hydroxyglutarate (D-2-HG) is a well-established oncometabolite of isocitrate dehydrogenase (IDH) mutant gliomas. While prior studies have demonstrated that D-2-HG is elevated in the cerebrospinal fluid (CSF) of patients with IDH-mutant gliomas, no study has evaluated the ability of CSF D-2-HG to quantitatively monitor tumor burden and therapeutic response. Toward this goal, we are routinely obtaining CSF samples from patients prior to, during, and after treatment via intra-operative collection and CSF-access devices. MATERIAL AND METHODS Patients were enrolled into our Ommaya reservoir study (NCT04692337), which enables longitudinal access to CSF. The catheter for the implanted reservoir was placed within the surgical resection cavity following a standard-of-care resection. Intra-operatively collected CSF samples were also biobanked through Mayo Clinic’s Neuro-Oncology biorepository. CSF samples were also acquired via clinically indicated ventriculoperitoneal shunts (VP shunts) or external ventricular drains via our ongoing CSF biomarkers study (NCT04692324). D-and-L-2-HG were assayed from each CSF sample via gas chromatography-mass spectrometry (GC-MS) at Mayo Clinic’s Biochemical Genetics Laboratory. RESULTS CSF was collected prior to and after resection in patients via Ommaya reservoirs or VP shunts, all of whom underwent near gross total resections (>95% of tumor volume). D-2-HG and the D/L ratio fell on average by 57.9% (range: 15.1%-98.6%) and 68.8% (range: 49.3%-98.9%), respectively, following resection. In one patient with a largely intraventricular grade 4 IDH-mutant astrocytoma, D-and-L-2-HG fell by 99.1% and 99.3% by post-operative day 2, consistent with the impact of ventricular contact on baseline CSF D-2-HG levels. In patients for whom CSF was collected throughout chemoradiation, with or without immunotherapy, D-2-HG and the D/L ratio continued to decrease over time, despite concern for pseudoprogression versus progression on imaging. CONCLUSION Our early experience with this strategy demonstrates the potential for intracranial CSF D-2-HG as a monitoring biomarker for IDH-mutant gliomas. Work is ongoing to evaluate the reproducibility of these findings in a larger cohort of patients with low or high-grade gliomas, in addition to the impact of tumor recurrence on 2-HG. Future studies will evaluate the impact of IDH inhibitors and other experimental therapies on CSF 2-HG levels for pharmacodynamic monitoring of treatment response.

Abstract BACKGROUND Few methods exist via which to assess treatment response with high fidelity in gliomas. Cell-free DNA (cfDNA) is of growing interest as a liquid biomarker in plasma and CSF and has been used in other cancer types to optimize therapeutic regimens for individual patients. While CSF has previously been used as a source of cfDNA for gliomas, it remains unknown how cfDNA abundance and sequencing change throughout treatment and inevitable recurrence. To that end, we leveraged our growing biobank of longitudinal CSF specimens, acquired via CSF access devices, to determine the potential utility of CSF cfDNA as a monitoring tool for treatment response in patients with gliomas. MATERIAL AND METHODS Longitudinal CSF specimens were acquired via Ommaya reservoirs (NCT04692337) or ventriculoperitoneal shunts (NCT04692324) in patients with brain tumors, including gliomas. cfDNA was extracted from 1-5 mL of CSF based on available sample volumes and quantified via Qubit. Next-Generation sequencing (PredicineCARE) or low-pass whole genome sequencing (LPWGS) was performed by Predicine, Inc based on the amount of cfDNA available. RESULTS CSF acquired prior to versus after resection demonstrated an increase in quantified cfDNA (2.97x, range: 1.58-5.26x), consistent with the impact of increased parenchymal disruption and closer contact with CSF. Thereafter, CSF cfDNA continued to decrease throughout chemoradiation and adjuvant treatment and increased at recurrence, including in patients co-enrolled in immunotherapy trials for IL-7 agonists and pembrolizumab. Interestingly, copy number burden (CNB) increased by the time of recurrence in three patients with known disease progression, and continued to decrease throughout immunotherapy treatment in one patient despite concerns of pseudoprogression. In one patient with a hypermutated glioblastoma, CSF cfDNA revealed over 59 genomic alterations, including MAP2K1, KIT, and PDGFRA, and over 100 more variants of unknown significance, each of which had been detected in sequencing of the tissue acquired at resection and subsequently decreased with chemoradiation. In another patient with a known progressing EGFR-amplified GBM, over 200 new variants were identified by the patient’s final sample and EGFR copy number increased from 2 to over 30. CONCLUSION Analysis of cfDNA over a patient’s disease course, including resection and treatment with standard-of-care and experimental therapies, may be of use for disease monitoring and treatment response. Further work is needed to evaluate these findings in a larger cohort of patients and to determine the sensitivity of changes in mutations, CNB, and cfDNA quantity for glioma disease burden.

Abstract BACKGROUND The microenvironment of glioblastoma is complex and defined in part by protumorigenic metabolic changes. Within this microenvironment, blood-brain barrier disruption is often associated with the most aggressive portions of the tumor. To date, the metabolic tumor microenvironments of enhancing, blood-brain barrier (BBB) disrupted portions of gliomas, as compared to BBB-intact tumor, remain poorly understood. Specifically, it is unclear if the associated blood-brain barrier disruption merely reflects or functionally supports tumor aggressiveness. METHODS Intra-operative high-molecular weight microdialysis was utilized under an investigational device exemption (NCT NCT04047264) to sample the extracellular metabolome of radiographically diverse regions during fifteen neurosurgical resections. The global extracellular metabolome of recovered microdialysate from contrast-enhancing and non-enhancing tumor, as well as brain adjacent to tumor, was evaluated via ultra-performance liquid chromatography tandem mass spectrometry (UPLC-MS/MS) on the Metabolon platform and assessed using enrichment and correlation analyses. RESULTS Among 162 named metabolites identified via ultra-performance liquid chromatography tandem mass spectrometry, a novel glioma-associated metabolite, guanidinoacetate (GAA), was 126.32x higher in enhancing tumor than in adjacent brain and 13.63x higher in non-enhancing tumor than brain. Forty-eight additional metabolites were 2.05-10.18x more abundant in enhancing tumor than brain, including carnitine family members and all nine large neutral amino acids. Except for GAA and 2-hydroxyglutarate in IDH-mutant gliomas, differences between non-enhancing tumor and brain microdialysate were comparatively modest and less consistent, suggesting heterogeneity in the glioma metabolic microenvironment of blood-brain barrier (BBB) intact areas as compared to BBB-disrupted portions of gliomas. Interestingly, the enhancing but not the non-enhancing glioma metabolome was significantly enriched for plasma-associated metabolites, largely comprising amino acids and carnitines. CONCLUSION Our patient-derived data suggest that the BBB-disrupted metabolome of high-grade gliomas is characterized by metabolite diffusion from circulation into the tumor microenvironment. Further studies are ongoing to determine how this alteration in the glioma extracellular metabolome impacts glioma behavior and how it could be therapeutically targeted via BBB-modulating drugs.