Rahman, Masum

Abstract BACKGROUND Progress for gliomas is slowed in part by the paucity of mechanistic feedback during treatment with experimental therapies. Access to extracellular tumor pharmacodynamic biomarkers could provide an avenue to accelerate progress. We have initiated a program of intra-operative microdialysis to accelerate biomarker discovery and to identify candidate outcome measures for translational therapies. METHODS Intraoperative microdialysis was performed with M-dialysis 100kDA catheters and 107 variable rate pumps under an IDE. Four IDH-mutant and two IDH-WT lesions were studied intraoperatively with 3 divergently placed catheters. Microperfusate (artifical CSF+ 3% dextran) was perfused at 2uL/min and collected in 20 min increments. Paired CSF was also obtained when accessible. A parallel cohort of nude mice bearing human IDH-mutant, IDH-WT, or sham intracranial xenografts (n=6-12) underwent intratumoral microdialysis. A pilot murine study of intracranial drug delivery was performed via concurrent microdialysis during convection-enhanced delivery (CED) of saline or the IDH-inhibitor AG120. RESULTS Microdialysate from IDH-mutant intracranial xenografts revealed >100 differentially abundant metabolites compared to sham or IDH-WT tumors, including D2-HG (21x) and MTA(18x), p< 10^-5. The most significantly abundant metabolite was DMA (4x, p< 10^-10). 15-1000uM D2HG was recovered from intra-operative human IDH-mutant tumors and 1-2uM from normal brain adjacent to IDH-WT gliomas and < 1uM in all IDH-WT samples. Forty metabolites differentiated enhancing tumor from adjacent brain in 3/3 paired human samples including upregulated Aminoacyl-tRNA biosynthesis and downregulated purine metabolism. Serial aliquots of microdialysate during saline CED yielded steady D2-HG levels whereas CED with AG120 yielded undetectable D2-HG within 6 hours. CONCLUSION The extracellular metabolic landscape of glioma is diverse, dynamic and reflects tumor biology and response to therapy. Collectively, these studies suggest that intra-tumoral drug testing should be feasible with realistic expectation of gaining metabolic feedback within a short timeframe. Leveraging this paradigm can provide opportunities to accelerate therapeutic translation for gliomas.

Abstract BACKGROUND NAD+ is required for cell metabolism and DNA repair. It is generated from nicotinic acid (NA) by NAPRT and from Nicotinamide (NAM) by NAMPT. D2HG in IDH-mutant tumors methylates and inactivates NAPRT, increasing dependence on NAMPT. Toxic side effects of NAMPT inhibition can be prevented by NA supplementation in healthy cells without NAPRT methylation. A1326133 is a recently described CNS-penetrant NAMPT inhibitor hypothesized to selectively eliminate IDH-mutant NAPRT-methylated gliomas, likely in combination with other therapies. Our group is looking for biomarkers of drug efficacy to augment individualized therapies. To that end, we sought to identify GBM cell lines with varying sensitivity to NAMPT inhibition. METHODS Human non-immortalized astrocytes and human GBM cell lines were utilized from the Mayo Clinic Glioma patient-derived xenograft resource, including IDH-R132Hmutant lines (GBM164, 196) and IDH-WT lines (GBM6, 12, 76). Cell viability was analyzed after 4days incubation with the NAMPT inhibitor, A1326133 +/- Temozolomide (TMZ) or NA. IC50 for A1326133 was estimated based on intracellular ATP using Cell-Titer-Glo. RESULTS Marked heterogeneity between lines was observed in response to A1326133 +/- NA or TMZ. Sensitive and resistant lines were identified among both IDH-mutant and IDH-WT cell lines. IC50s: GBM164, 12, 6, 196 and 76 were 5.6, 9.3, 39.2, 910, and 9455nM, respectively. NA partially rescued GBM164 by NA (IC50 increased to 20.8nM) but not GBM6 nor 12. IC50 for Human astrocytes was 221.7nM, but >10,000nM with NA. Addition of TMZ did not improve A1326133 efficacy. CONCLUSION Our data illustrate the potential utility of NAMPT inhibition to kill a subset of IDH-WT and IDH mutant lines, but conflict with previously reported TMZ synergy and correlation with mutant IDH. NA may increase safety but could decrease efficacy in certain lines. Ongoing studies seek metabolic biomarkers of therapeutic efficacy to guide individualized therapy with NAMPT inhibitors.

Abstract Glioblastoma multiforme (GBM) is the most aggressive CNS neoplasm with a mere 15 month progression-free survival following current standards-of-care. Furthermore, conventional therapies have been effective, in part, by inducing a senescent-like phenotype in at least a proportion of glioma, as well as within the tumor-adjacent, healthy brain cells. Through an N-glyco-capture large-scale screening method, we were able to identify 25 genes exclusively overexpressed on glioma cell surfaces. Of those 25 targets, CD38 was an attractive target due to pre-existing FDA approved therapeutics. Recent studies have shown, however, that senescent cells, such as those induced via chemo- and radiotherapies, secrete a pro-inflammatory array of cytokines, known collectively as senescence associated secretory phenotype (SASP), that are capable of increasing CD38 expression in monocytes. CD38 functions as an ectoenzyme to convert extracellular NAD+ to nicotinamide (required for glioma cell salvage pathway NAD+ synthesis) and ADPR/cADPR (a cellular proliferation signal). To examine the effects of radiation on human glioma tissue, we performed reverse cyclase assays (RCA) on paired primary and recurrent human glioma tissue samples and found an increase in CD38 activity in post-irradiated tissues (recurrent glioma) compared to pre-irradiated (primary glioma). Through proliferation assays, we also found an increase in glioma cell growth following treatment with cADPR compared to untreated. Our results demonstrate that CD38 activity is tumorgenic, and furthermore that conventional chemo- and radiotherapies increase this CD38 activity. This indicates that treating CD38 with previously FDA approved therapeutics may provide hope for increasing progression free survival in our GBM patient population, and moreover emphasize the importance of leveraging novel large scale screening methods for identifying additional opportunities for treatment.

Abstract BACKGROUND Radiotherapy, combined with surgical resection and chemotherapy, remains a first-line treatment for infiltrative gliomas. However, these tumors are not surgically curable, and often recur, even within the prior radiation field, and may demonstrate a more aggressive phenotype. Importantly, high grade gliomas display diverse molecular phenotypes, and whether this genetic variability leads to divergent behaviour in the radiated tumor microenvironment is unknown. Herein, we characterize the effects of the irradiated brain microenvinroment on nine additional unique GBM cell lines to better understand the nuances of how tumor molecular phenotypes influence cellular dynamics. METHODS Female athymic nude mice were randomly divided into cranial radiation (15 Gy) and non-radiated groups. Mice then underwent intracranial implantation with one of the selected patient-derived xenograft (PDX) GBM cell lines (GBM 6, 10, 12, 39, 46, 76, 123, 164, 196; total n=8-15, per group, per line). Kaplan-Meyer (K-M) and log-rank tests were performed to compare the survival between irradiated and non-irradiated groups. RESULT Of nine previously untested human GBM lines, we found that five demonstrated shorter survival in the pre-radiated brain (GBM 6, 46, 76, 164, 196). However, two lines yielded prolonged survival in the pre-radiated brain (GBM 10, 12); GBM 39, 123 whose rate of growth was not impacted by the radiated brain. CONCLUSION These results highlight the likely critical impact of the irradiated microenvironment on tumor behaviour, yet illustrate that different tumors may exhibit opposing responses. Although further evaluation will be needed to understand mechanisms of divergent behavior, our data suggest the increased rate of growth in the radiated microenvironment may not apply to the fastest-growing tumor lines, which could instead demonstrate a paradoxical response.

Abstract INTRODUCTION Laser Interstitial Thermal Therapy (LITT) is a novel treatment modality that has been used for an array of intracranial pathology. In the current manuscript, we sought to conduct a systematic review and meta-analysis to summarize all available literature to date, on outcomes of patients with recurrent GBM (r-GBM) undergoing LITT, pooling together quantitatively the overall survival and progression-free survival data. METHODS A comprehensive literature search was performed to retrieve all studies investigating overall survival, post-procedure survival, and progression-free survival outcomes of patients with r-GBM undergoing LITT. All statistics were pooled together by the meta-analysis of the mean using a weighted random-effects (RE) or fixed-effect (FE) model. RESULTS Eleven studies were included in the final cohort, representing a total of 134 patients with rGBM. The pooled mean age of the cohort at the time of recurrence diagnosis was found to be 56.7 ± 4.56 while 41% of the cohort were females. For the delivery of LITT, two studies utilized neodymium-yttrium aluminum-garnet laser (Neodp-YAG Laser), three studies utilized the Visualase system, five studies utilized the Neuroblate system, and one study used both the Neuroblate and the Visualase system. A total of eight studies with 107 patients had available data for overall median survival. The pooled overall survival was found to be 18.6 months (95%CI 16.2-21.1). A total of six studies with 93 patients had available data for post-LITT survival. The pooled post-LITT survival was found to be 10.1 months (95%CI 8.8-11.6). A total of eight studies with 119 patients had available data for progression-free survival. The pooled progression-free survival was found to be 6 months (95%CI 5.3-6.7). CONCLUSION LITT is a novel minimally invasive procedure which, when used with optimal adjuvant therapy, may confer survival benefit for patients with r-GBM.

Abstract BACKGROUND Glioblastoma (GBM) is an infiltrative, uniformly fatal brain tumor, treated with surgery, radiation, and Temozolomide (TMZ). Chemoradiation induces a senescent-like phenotype, which contributes to disease recurrence. We recently found that, radiated GBM cells can be eliminated by inhibition of BCL-XL. However, it remains unknown whether p53 is involved in this process and whether the elimination of senescent cells by BCL-XL inhibition could be augmented by MDM2 inhibition, a negative p53 regulator. METHODS p53-mutant (GBM6/GBM123) and p53-WT (GBM39/GBM76) human GBM cells were treated with 5Gy or TMZ following 48 hours of MDM2 inhibitor (AMG232) or vehicle treatment and maintained for seven days to establish a senescent-like phenotype. We evaluated the IC50 for BCL-XL inhibitor (A1331852) in radiated vs. non-radiated cells with or without MDM2 inhibitor pre-treatment. RESULTS MDM2-inhibitor treatment prior to radiation increased the expression of p21 and lead to increased cell death when combined with BCL-XL inhibition in p53-WT GBM cells. IC50 of BCL-XL inhibitor after prior MDM2 pretreatment and radiation in GBM76 was 4.5□M compared to 33.5□M, 18.1□M, and 32.3□M in vehicle without radiation, vehicle with radiation, and MDM2 inhibitor-alone treatment groups, respectively(p=0.0036). The IC50 of BCL-XL inhibitor without MDM2-inhibition in non-irradiated and radiated, as well as pre-administration of MDM2 inhibitor in non-irradiated and radiated GBM39 cells was 5618nM, 16117nM, 3926nM and 276.5nM, respectively(p=0.0003). Conversely, MDM2 co-inhibition with BCL-XL did not lead to increased rates of cell death in p53-mutant cell(p>0.05). CONCLUSION We previously identified that BCL-XL inhibition promotes cell death in senescent GBM cells. We build upon that work, demonstrating the increased rates of cell death can be augmented by MDM2 inhibition, but only in p53-WT cells. These findings highlight a novel therapeutic target for treating latent GBM tumors prior to recurrence, with the additional of MDM2 inhibition greatly increasing the efficacy of BCL-XL targeting agents.

Abstract BACKGROUND Glioblastoma is a fatal infiltrative primary brain tumor, and standard care includes maximal safe surgical resection followed by radiation and Temozolomide (TMZ). Therapy-resistant residual cells persist in a latent state a long time before inevitable recurrence. Conventional radiation and Temozolomide (TMZ) treatment cause oxidative stress and DNA damage resulting senescent-like state of cell-cycle arrest. However, increasing evidence demonstrates escaping senescence leads to tumor recurrence. Thus, the ablation of senescent tumor cells after chemoradiation may be an avenue to limit tumor recurrence. METHODS 100uM TMZ for 7days or 10-20Gy radiation (cesium gamma radiator) was used for senescence induction in human glioblastoma in vitro and confirmed by SA-Beta gal staining and PCR. Replication arrest assessed by automated quantification of cellular confluence (Thermo Scientific Series 8000 WJ Incubator). We evaluated the IC50 for several senolytics targeting multiple SCAPs, including Dasatinib, Quercetin, AMG-232, Fisetin, Onalespib, Navitoclax, and A1331852, and in senescent vs. proliferating cells. RESULTS Among the senolytic tested, the Bcl-XL inhibitors A1331852 and Navitoclax both shown senolytic effect by selectively killing radiated, senescent tumor cells at lower concentrations as compared to 0Gy treated non-senescent cells. Across 12 GBM cell lines, IC50 for senescent cells was 6–500 times lower than non-senescent GBM(p< 0.005). Such differential sensitivity to Bcl-XL inhibition after radiation has also observed by BCL-XL knockdown in radiated glioma. CONCLUSION These findings suggest the potential to harness radiation-induced biology to ablate surviving quiescent cells and demonstrate Bcl-XL dependency as a potential vulnerability of surviving tumor cells after exposure to chemoradiation.

Abstract INTRODUCTION Gliomas are the most common and deadly adult brain cancers with a median survival time of less than 18 months. Many therapeutics have failed translating into effective therapies due to incomplete understanding of the disease and heterogeneity of tumors between patients. There is a need for methods that allow for continual access and analysis of glioma biomarkers. We sought to establish a reliable method for quantifying the onco-metabolite, D2HG. After optimizing this method, we used microdialysate and microperfusate to test our hypothesis that analysis of D2HG in response to therapy can indicate therapeutic efficacy sooner than current methods permit. METHODS Microdialysate and microperfusate were collected from the tumor centers and lateral ventricles of athymic nude mice implanted with IDH1 mutant PDX line GBM196. We performed microdialysis and microperfusion on consecutive days to determine baseline D2HG concentration. Following a five day period of TMZ treatment, we collected another round of microdialysis and microperfusion one week after treatment and again two weeks after treatment. We then used our optimized D2HG enzymatic assay to quantify the changes in D2HG in response to therapy. RESULTS Using our enzymatic D2HG assay, we were able to quantify D2HG as low as 30nM (30fmol/uL). We found a ~20% decrease in D2HG concentration in the tumor center and right lateral ventricle after 1 week of TMZ treatment (n=2), and ~36% decrease after 2 weeks of TMZ treatment (n=2). CONCLUSION Using a D2HG quantification assay to monitor changes in concentration of D2HG in microdialysate/microperfusate can predict sensitivity to therapy.

Abstract BACKGROUND Despite best current therapies, median survival for glioblastoma remains little more than a year. There remains an urgent need for improved treatment strategies. Current therapies, including radiation and temozolomide, are known to induce senescence in tumor cells. Advances in aging research have identified drugs with “senolytic” properties to kill senescent cells. This work tests the hypothesis that induction of senescence pathways following radiation can provide a window of opportunity to eliminate residual glioblastoma cells using drugs that block senescent cell-associated pathways (SCAPs). METHODS Human glioblastoma cells were radiated with 10, 15, or 20 Gy and maintained in culture for 3 to 4 weeks to facilitate maturation of a senescent phenotype. Given existence of multiple SCAPs, we evaluated the IC50 for several senolytic candidates including Dasatinib, Quercetin, Fisetin, AMG-232, Onalespib, A1331852, and Navitoclax in radiated vs non-radiated cells. RESULTS Among these, the Bcl-XL inhibitors A1331852 and Navitoclax both selectively killed multiple lines of radiated, senescent tumor cells at lower concentrations as compared to unirradiated, non-senescent tumor cells. Across multiple GBM cell line, the average IC50 for senescent cells was 12.88 nM (95%CI= 9.068–18.58 nM), as compared to the IC50 of non-senescent cells of 103.2 nM (95%CI= 74.24–143.9 nM; p< 0.005). Such differential sensitivity to Bcl-XL inhibition after radiation was not observed in the non-GBM control cells evaluated to date. CONCLUSION Our findings suggest that blocking the anti-apoptotic pathway Bcl-XL in GBM cells rendered at least transiently senescent following radiation may provide an avenue to selectively ablate residual glioblastoma cells following radiation. Our data provide proof of principle that senolytic therapy may represent a previously untapped window of opportunity to increase extent of tumor ablation following completion of standard cytotoxic therapies.