Glioblastoma Patient Tumor Research Articles

Heterogeneous expression of the atypical chemokine receptor ACKR3 in glioblastoma patient-derived tissue samples and cell cultures

Glioblastoma (GBM) is the most aggressive glial tumor of the adult brain, associated with invariably fatal outcome, and a deeper understanding of the underlying malignant mechanisms is necessary to address the current therapeutic failure. We previously demonstrated the role of the CXCL12/CXCR4 axis in GBM cell migration and resistance to ionizing radiation. The atypical chemokine receptor ACKR3, responsible for CXCL12 scavenging, was previously suggested as additional important player in the context of GBM. Following validation of the detection tools, we observed that ACKR3 is expressed within GBM patient tumor tissue, distributed in diverse cell types. In contrast to CXCR4, ACKR3 expression in patient-derived stem-like cells (GSCs) remains however low, while ACKR3 gene expression by tumor cells appears to be modulated by the in-vivo environment. Using overexpression models, we also showed that in vitro ACKR3 had no significant direct effect on cell proliferation or invasion. Altogether, these results suggest that in vitro ACKR3 plays a minor role in malignant GBM cell biology and that its expression is possibly regulated by in-vivo influences. The subtle and multifaceted functions ACKR3 could exert in GBM should therefore only be tackled within a comprehensive tumor microenvironment considering tumoral but also non-tumoral cells.

Glioblastoma-instructed microglia transition to heterogeneous phenotypic states with phagocytic and dendritic cell-like features in patient tumors and patient-derived orthotopic xenografts

BackgroundA major contributing factor to glioblastoma (GBM) development and progression is its ability to evade the immune system by creating an immune-suppressive environment, where GBM-associated myeloid cells, including resident microglia and peripheral monocyte-derived macrophages, play critical pro-tumoral roles. However, it is unclear whether recruited myeloid cells are phenotypically and functionally identical in GBM patients and whether this heterogeneity is recapitulated in patient-derived orthotopic xenografts (PDOXs). A thorough understanding of the GBM ecosystem and its recapitulation in preclinical models is currently missing, leading to inaccurate results and failures of clinical trials.MethodsHere, we report systematic characterization of the tumor microenvironment (TME) in GBM PDOXs and patient tumors at the single-cell and spatial levels. We applied single-cell RNA sequencing, spatial transcriptomics, multicolor flow cytometry, immunohistochemistry, and functional studies to examine the heterogeneous TME instructed by GBM cells. GBM PDOXs representing different tumor phenotypes were compared to glioma mouse GL261 syngeneic model and patient tumors.ResultsWe show that GBM tumor cells reciprocally interact with host cells to create a GBM patient-specific TME in PDOXs. We detected the most prominent transcriptomic adaptations in myeloid cells, with brain-resident microglia representing the main population in the cellular tumor, while peripheral-derived myeloid cells infiltrated the brain at sites of blood–brain barrier disruption. More specifically, we show that GBM-educated microglia undergo transition to diverse phenotypic states across distinct GBM landscapes and tumor niches. GBM-educated microglia subsets display phagocytic and dendritic cell-like gene expression programs. Additionally, we found novel microglial states expressing cell cycle programs, astrocytic or endothelial markers. Lastly, we show that temozolomide treatment leads to transcriptomic plasticity and altered crosstalk between GBM tumor cells and adjacent TME components.ConclusionsOur data provide novel insights into the phenotypic adaptation of the heterogeneous TME instructed by GBM tumors. We show the key role of microglial phenotypic states in supporting GBM tumor growth and response to treatment. Our data place PDOXs as relevant models to assess the functionality of the TME and changes in the GBM ecosystem upon treatment.Graphical

BIOM-06. LONGITUDINAL CEREBROSPINAL FLUID LIQUID BIOPSY IN PHASE 0/2 TRIAL OF NIRAPARIB FOR NEWLY-DIAGNOSED GLIOBLASTOMA

Abstract INTRODUCTION Effective drug development in glioblastoma (GBM) requires visualization of tumor evolution in response to experimental therapy. Today, genetic and epigenetic mechanisms of tumor resistance are undetected in clinical trials. Here, we establish the feasibility of a liquid biopsy paradigm in tracking tumor evolution in response to a novel therapeutic using an intra-cavitary reservoir for serial cerebrospinal fluid (CSF) collection. METHODS We conducted a Phase 0/2 clinical trial of niraparib, a potent PARP1/2 inhibitor, in newly-diagnosed GBM patients (NCT05076513). Intra-cavitary reservoirs were placed in select patients following tumor resection. Monthly outpatient CSF samples were collected to longitudinally examine cell-free tumor DNA (cfDNA) using a 293-gene amplicon-based next-generation sequencing (NGS) panel and circulating tumor cells using single cell RNA-sequencing (scRNA-seq). Concordance analysis was performed using paired tumor tissue samples against multiple liquid biopsies. RESULTS Tumor-derived cfDNA and circulating tumor cells were successfully isolated from intra-cavitary reservoirs at regular intervals for the duration of the clinical trial. Twelve patients so far have been implanted with the reservoirs for CSF collection. Among 55 CSF samples, 53 (94%) has detectable cfDNA. Radiographic tumor burden is associated with concentration of cfDNA (ng/ml, p< 0.01). No adverse events were observed during liquid biopsy acquisition. CSF sequencing data was highly concordant with NGS analyses of tumor tissue. cfDNA and single-cell analyses from serial CSF samples reflected the disparate evolution of GBM patient tumors in the face of PARP1/2 inhibition. CONCLUSION This is the first reported experience of serial CSF liquid biopsy to monitor DNA- and RNA-based tumor evolution in the context of an experimental therapy. Our data demonstrate the feasibility of cfDNA NGS and scRNA-seq using longitudinal CSF sampling of clinical trial patients. This new strategy for serial tumor profiling will be an important component of future drug development efforts targeting the moving biology of glioblastoma.

The Impact of ioMRI on Glioblastoma Resection and Clinical Outcomes in a State-of-the-Art Neuro-Oncological Setup.

Intraoperative magnetic resonance imaging (ioMRI) aims to improve gross total resection (GTR) in glioblastoma (GBM) patients. Despite some older randomized data on safety and feasibility, ioMRI's actual impact in a modern neurosurgical setting utilizing a larger armamentarium of techniques has not been sufficiently investigated to date. We therefore aimed to analyze its effects on residual tumor, patient outcome, and progression-free survival (PFS) in GBM patients in a modern high-volume center. Patients undergoing ioMRI for resection of supratentorial GBM were enrolled between March 2018 and June 2020. ioMRI was performed in all cases at the end of resection when surgeons expected complete macroscopic tumor removal. Extent of resection (EOR) was performed by volumetric analysis, with GTR defined as an EOR ≥ 95%, respectively. Progression-free survival (PFS) was analyzed through univariate and multivariate Cox proportional regression analyses. In total, we enrolled 172 patients. Mean EOR increased from 93.9% to 98.3% (p < 0.0001) due to ioMRI, equaling an increase in GTR rates from 78.5% to 93.0% (p = 0.0002). Residual tumor volume decreased from 1.3 ± 4.2 cm3 to 0.6 ± 2.5 cm3 (p = 0.0037). Logistic regression revealed recurrent GBM as a risk factor leading to subtotal resection (STR) (odds ratio (OR) = 3.047, 95% confidence interval (CI) 1.165-7.974, p = 0.023). Additional resection after ioMRI led to equally long PFS compared to patients with complete tumor removal before ioMRI (hazard ratio (HR) = 0.898, 95%-CI 0.543-1.483, p = 0.67). ioMRI considerably reduces residual tumor volume and helps to achieve comparable PFS, even in patients with unexpected residual tumor after initial resection before ioMRI.

The benefit of complete resection of contrast enhancing tumor in glioblastoma patients: A population-based study.

New treatment modalities have not been widely adopted for patients with glioblastoma (GBM) after the addition of temozolomide to radiotherapy. We hypothesize that increased extent of resection (EOR) has resulted in improved survival for surgically treated patients with glioblastoma at the population level. Retrospective analysis of adult patients operated for glioblastoma in the population of South-Eastern Norway. Patients were stratified into Pre-temozolomide- (2003-2005), temozolomide- (2006-2012), and resection-focused period (2013-2019) and evaluated according to age and EOR. The study included 1657 adult patients operated on for supratentorial glioblastoma. The incidence of histologically confirmed glioblastoma increased from 3.7 in 2003 to 5.3 per 100 000 in 2019. The median survival was 11.4 months. Complete resection of contrast-enhancing tumor (CRCET) was achieved in 386 patients, and this fraction increased from 13% to 32% across the periods. Significant improvement in median survival was found between the first 2 periods and the last (10.5 and 10.6 vs. 12.3 months; P < .01), with a significant increase in 3- and 5-year survival probability to 12% and 6% (P < .01). Patients with CRCET survived longer than patients with non-CRCET (16.1 vs. 10.8 months; P < .001). The median survival doubled in patients ≥70 years and (12.1 months). Survival was similar between the time periods in patients where CRCET was achieved. We demonstrate an improved survival of GBM patients at the population level associated with an increased fraction of patients with CRCET. The data support the importance of CRCET to improve glioblastoma patient outcomes.

PIK3CB as a potential target to regulate chemosensitivity in glioblastoma.

e14051 Background: Glioblastoma (GBM) is the most common primary central nervous system malignancy but exhibits universally poor outcomes in part due to chemoresistance. Temozolomide (TMZ) can significantly improve overall survival, but resistance mechanisms limit its consistent efficacy. The phosphoinositide 3 kinase (PI3K) pathway is involved in multiple TMZ resistance mechanisms and is highly druggable; however, pan-inhibition of its four grossly homologous but functionally distinct catalytic subunits (PIK3CA/B/D/G encode p110α/β/δ/γ, respectively) was unsuccessful clinically. Therefore, by understanding how the individual PI3K subunits regulate TMZ sensitivity, a more selective treatment strategy can be developed to combat chemoresistance. Methods: Differential expression of PI3K subunit genes was assessed using RNA sequencing (RNAseq) results obtained from public databases (e.g., DepMap, The Cancer Tissue Genome Atlas) for GBM cell lines and patient tumors expressing high or low O6-methylguanine-DNA-methyltransferase (MGMT), a DNA repair enzyme highly implicated in TMZ resistance. PI3K protein levels were assessed using immunohistochemistry of gliomas available at The Human Protein Atlas (THPA). PI3K gene expression was correlated with PI3K pathway activation, quantified as protein kinase B phosphorylation (pAKT). PI3K mRNAs and pAKT were then correlated with the half maximal inhibitory concentrations (IC50s) of TMZ in GBM cell lines to evaluate for a relationship between PI3K gene expression, PI3K pathway activity, and TMZ sensitivity. Finally, hazard ratios were used to compare PI3K gene expression in patient tumors with survival and response to chemotherapy to determine if expression was predictive of prognosis and/or chemosensitivity. Results: PIK3CB was the only PI3K gene enriched in both GBM cell lines and tumors, demonstrating significantly higher expression than all other PI3K genes in GBM (P &lt; 0.0001), independent of MGMT levels. PIK3CB expression remained the highest even after controlling for GBM risk factors such as tumor recurrence, isocitrate dehydrogenase (IDH) mutations, molecular subtypes, or patient gender/age differences. Levels of the p110β protein were also significantly higher than those of other PI3K catalytic subunits (P &lt; 0.001). PIK3CB expression was positively correlated with pAKT levels (P &lt; 0.05), but only in cell lines and tumors with low MGMT activity. Both PIK3CB expression and pAKT levels correlated positively with TMZ IC50s (P &lt; 0.05) in MGMT-deficient GBMs, suggesting an inverse relationship between PIK3CB and TMZ sensitivity. Finally, better prognosis and significant survival benefit with chemotherapy were both associated with lower levels of PIK3CB in MGMT-deficient patients. Conclusions: PIK3CB may selectively regulate chemoresistance in GBM, suggesting a role for p110β selective inhibition to improve chemosensitivity.

Sex-specific T cell exhaustion drives differential immune responses in glioblastoma

Abstract Sex differences in glioblastoma (GBM) incidence and outcome are well recognized; males show higher incidence and worse prognosis than females. Emerging evidence suggests that these extend to genetic/epigenetic and cellular differences, including immune responses. However, the mechanisms driving immunological sex differences are not fully understood. Using orthotopic GBM animal models, we recapitulated these sex differences in immune-competent B6 mice but not in immune-deficient mouse strains (NSG and RAG1KO). By depleting CD8+ T cells, we further confirm that CD8+ T cells play a critical role in driving sex differences in survival. Flow cytometry analysis revealed that male CD8+ T cells from tumor expressed higher levels of exhaustion markers such as PD-1, CTLA-4 and Tox, with higher frequency of progenitor exhausted T cell subset whereas female tumor contained more effector-like CD8+ T cells expressing elevated cytokine production. Anti-PD-1 antibody treatment exclusively extended the survival of male mice via inducing decreased exhausted CD8+ T cell subsets and increased effector function and proliferation. Survival analysis on bone marrow chimera model and adoptive transfer model indicated that T cell-mediated tumor control was predominantly regulated in a cell-intrinsic manner, as the sex of donor cells was a critical contributing factor. Finally, we confirmed these observations in GBM patient tumor samples with male tumors having a higher frequency of progenitor exhausted T cells with increased TOX expression. Taken together, these findings demonstrate sex-specific pre-determined behavior of T cells is critical in inducing sex differences in GBM progression and immunotherapy responses. Supported by grants from NIH (R35 NS127083, P01 CA245705).

Abstract LB097: Armored bicistronic CAR T cells with dominant-negative TGF-β receptor II to alleviate antigenic heterogeneity and suppressive immune microenvironment in glioblastoma

Abstract Introduction: We have completed two CAR T cell clinical trials for glioblastoma (GBM) and have identified several key challenges to therapeutic efficacy, including the inherently heterogenous genomic landscape and the immunosuppressive tumor microenvironment (TME) found in GBM. Our previous study showed that EGFR variant III (EGFRvIII)-targeting monovalent CAR T cells reduced target-positive tumor cell populations, but tumor recurrence resulted from target-negative tumor cells, highlighting the limitation of single-target approaches in heterogenous tumors. With regards to the highly immunosuppressive TME in GBM, we found that transforming growth factor-β (TGFβ) was present in the GBM TME as a major driver of suppression of the anti-GBM response in clinical samples. TGFb is consistently highly expressed in both GBM tumor cell lines and patient tumor tissues. Methods: We used two parallel scFv constructs, independently targeting both IL13Rα2 and EGFRvIII, in combination with a truncated dominant negative (dn) TGFβ receptor II. This trivalent construct was designed to explore possible additive effects in both in vitro and in vivo GBM model systems to limit tumor escape and overcome the immunosuppressive GBM TME. The CART-EGFR-IL13Rα2-dnTGFb construct broadened the targeted tumor cell repertoire, blocked TGFβ signaling, and served as a sink for free TGFβ in the GBM TME to overcome the suppressive function of TGFβ. Results: The tri-modular CAR T construct had an enhanced proliferative response when compared with the CART-EGFR-IL13Rα2 construct, in vitro. In co-culture assays, this construct led to reduced PD-1 expression and increased central memory phenotype, when compared to the bicistronic CAR T construct, which suggested a lower fraction of exhausted T cells. Tri-modular CAR T cells blocked the suppressive pSmad2/3 signaling pathway, leading to the increased tumor killing activity in co-culture experiments with both adherent and suspension GBM cell lines. In an immunodeficient mouse model, tri-modular CAR T cells eradicated tumor cells efficiently and mice had a longer median survival when compared those treated with the bicistronic CART-EGFR-IL13Rα2 cells, lacking the dnTGFb receptor II. Conclusion: Overcoming the adaptive changes in the local TME and addressing antigen heterogeneity will be required to improve the clinical efficacy of CAR T-directed strategies. Our combination work showed that bicistronic CART constructs cooperate with truncated TGFβ receptor II efficiently. In summary, the dominant-negative TGFβ RII CART-EGFR-IL13Rα2 structure is a promising strategy to address the clinical challenges of antigenic heterogeneity and the immunosuppressive TME in GBM we have observed in our two GBM CART cell trials at UPenn. Citation Format: Nannan Li, Jesse Rodriguez, Zev Binder, Donald O’Rourke. Armored bicistronic CAR T cells with dominant-negative TGF-β receptor II to alleviate antigenic heterogeneity and suppressive immune microenvironment in glioblastoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 2 (Clinical Trials and Late-Breaking Research); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(8_Suppl):Abstract nr LB097.

OS05.5.A Glioblastoma-instructed microglia transit to heterogeneous phenotypic states with dendritic cell-like features in patient tumors and patient-derived orthotopic xenografts

Abstract Background A major contributing factor to Glioblastoma (GBM) development and progression is its ability to evade the immune system by creating an immune-suppressive tumor microenvironment (TME). GBM-associated myeloid cells, including resident microglia, macrophages and other peripheral immune cells are generally geared towards tumor-supportive roles. It is however unclear whether such recruited myeloid cells are phenotypically and functionally identical. Here, we aim to understand the heterogeneity of the GBM TME, using an unbiased, marker-free approach to systematically characterize cell type identities at the molecular and functional levels. Material and Methods We applied single-cell RNA-sequencing, multicolor flow cytometry, immunohistochemical analyses and functional studies to examine the heterogeneous TME instructed by GBM cells. GBM patient-derived orthotopic xenografts (PDOXs) representing different tumor phenotypes were compared to glioma mouse GL261 model and patient tumors. Results We show that PDOX models recapitulate major components of the TME found in human GBM. Human GBM cells reciprocally interact with mouse cells to create a GBM-specific TME. The most prominent transcriptomic adaptations are found in tumor-associated macrophages (TAMs), which are largely of microglial origin. We reveal inter-patient heterogeneity of TAMs and identify key signatures of distinct phenotypic states within the microglia-derived TAMs across distinct GBM landscapes. GBM-educated microglia adapt expression of genes involved in immunosuppression, migration, phagocytosis and antigen presentation, indicating functional cross-talk with GBM cells. We identify novel phenotypic states with astrocytic and endothelial-like features. Identified gene signatures and phenotypic states are confirmed in GBM patient tumor tissue. Finally we show that temozolomide treatment leads to transcriptomic adaptation of not only the GBM tumor cells but also adjacent TME components. Conclusion Our data provide insights into the phenotypic adaptation of the heterogeneous TME instructed by GBM tumor. We confirm a crucial role of microglia in supporting the immunosuppressive TME and show that PDOXs allow to monitor the highly plastic GBM ecosystem and its phenotypic adaptations upon treatment. This work further confirms the clinical relevance of PDOX avatars for testing novel therapeutics including modalities designed to target the myeloid compartment.

Long non-coding RNA (lncRNA) HOXD-AS2 promotes glioblastoma cell proliferation, migration and invasion by regulating the miR-3681-5p/MALT1 signaling pathway

ABSTRACT Glioblastoma (GBM) is the most lethal type of brain cancer. An increasing number of studies suggest that long non-coding RNAs (lncRNAs) are implicated in tumor progression. LncRNA HOXD‐AS2 was reported to be highly expressed in glioma and associated with glioma grade and poor prognosis. However, the molecular mechanism remains to be elucidated. In this study, we first analyzed differentially expressed lncRNAs in glioblastoma using RNA-seq dataset (156 GBM samples and 5 adjacent normal samples in TCGA (Cancer Genome Atlas) and GTEx (Genotype-Tissue Expression) database). HOXD-AS2 was found to be significantly up-regulated in GBM tissues, which was further confirmed in GBM patient tumor samples and GBM cell lines. Silencing HOXD-AS2 inhibited cell proliferation, migration and invasion, and promoted cell apoptosis. We further identified and validated miR-3681-5p as a target of HOXD-AS2, and miR-3681-5p was negatively regulated by HOXD-AS2. By negatively affecting miR-3681-5p, HOXD-AS2 could promote the expression of MALT1 to augment the aggressiveness of GBM cells. miR-3681-5p overexpression or MALT1 knockdown attenuated aggressiveness of GBM cells. Importantly, silencing HOXD-AS2 suppressed tumorigenesis of GBM cells in the xenograft mouse model. Collectively, our study clarified the role of miR-3681-5p/MALT1 axis underlying the oncogenic function of lncRNA HOXD-AS2 in GBM. Future work is required to study the mechanism by which HOXD-AS2 is upregulated in GBM cells, which can provide novel insights into therapeutic intervention for GBM treatment.

A Drosophila RNAi screen reveals conserved glioblastoma-related adhesion genes that regulate collective cell migration.

Migrating cell collectives are key to embryonic development but also contribute to invasion and metastasis of a variety of cancers. Cell collectives can invade deep into tissues, leading to tumor progression and resistance to therapies. Collective cell invasion is also observed in the lethal brain tumor glioblastoma (GBM), which infiltrates the surrounding brain parenchyma leading to tumor growth and poor patient outcomes. Drosophila border cells, which migrate as a small cell cluster in the developing ovary, are a well-studied and genetically accessible model used to identify general mechanisms that control collective cell migration within native tissue environments. Most cell collectives remain cohesive through a variety of cell–cell adhesion proteins during their migration through tissues and organs. In this study, we first identified cell adhesion, cell matrix, cell junction, and associated regulatory genes that are expressed in human brain tumors. We performed RNAi knockdown of the Drosophila orthologs in border cells to evaluate if migration and/or cohesion of the cluster was impaired. From this screen, we identified eight adhesion-related genes that disrupted border cell collective migration upon RNAi knockdown. Bioinformatics analyses further demonstrated that subsets of the orthologous genes were elevated in the margin and invasive edge of human GBM patient tumors. These data together show that conserved cell adhesion and adhesion regulatory proteins with potential roles in tumor invasion also modulate collective cell migration. This dual screening approach for adhesion genes linked to GBM and border cell migration thus may reveal conserved mechanisms that drive collective tumor cell invasion.

PL02.3. A Phenotypic heterogeneity and plasticity as resistance mechanisms in Glioblastoma

Abstract BACKGROUND Glioblastomas are among the most heterogeneous tumors, which hampers patient stratification and development of effective therapies. Glioblastomas create a dynamic ecosystem, where heterogeneous tumor cells interact with the tumor microenvironment to establish different niches. Upon tumor growth, Glioblastoma cells manifest remarkable plasticity and respond flexibly to selective pressures by transiting towards states favorable to the new tumor microenvitonment. How this phenotypic plasticity contributes to treatment resistance is currently less clear. The exact nature of treatment resistant, tolerant and sensitive Glioblastoma cells remains unresolved. Further studies at the single cell level are needed to reveal transient and long-term signatures of the resistant states. MATERIAL AND METHODS To investigate long-term phenotypic changes upon treatment at the single cell level we performed single cell RNA-seq (scRNA-seq) on the longitudinal patient-derived xenograft (PDOX) models derived from Glioblastoma patient tumors prior and after the standard-of-care treatment. In addition, direct treatment of PDOXs with temozolomide combined with scRNA-seq allowed revealing short-term transcriptomic changes both in tumor cells and in the mouse-derived cells forming tumor microenvironment. Advanced computational algorithms, including reference-free deconvolution methods, were applied to reveal treatment resistance signatures and master regulators of the identified treatment-resistant subpopulations. RESULTS We show that PDOX models recapitulate all the major cell types and transcriptional programs reported in Glioblastoma patient samples, providing clinically relevant models for investigating treatment resistance signatures of tumor cells and associated tumor microenvironment. Analysis of treatment naïve and treated Glioblastomas at the single cell level revealed presence of pre-existing treatment resistant states as well as newly established resistant subpopulatons. Certain transcriptomic changes are preserved long term, regardless of the lack of genetic evolution of the tumor cells. CONCLUSION Phenotypic plasticity is an important factor contributing to resistance mechanisms in Glioblastoma. Key molecular regulators of tumor cell plasticity towards treatment resistance states represent novel targets for future combinatory treatments.

ECOA-2. Immuno-sensitization of Glioblastoma to NY-ESO-1 Targeting via Promoter Demethylation

Abstract Introduction The immunotherapeutic targeting of New York-esophageal squamous cell carcinoma (NY-ESO-1) and other cancer/testis antigens (CTA) is an appealing strategy for the treatment of malignant gliomas because CTA are not expressed in most normal adult tissues and their expression can be induced in tumors for targeting by T-cells. Basally, NY-ESO-1 is often poorly expressed in glioblastoma (GBM), presumably through promoter methylation. Mechanisms governing the expression of CTA have been explored in other cancers; however, neither the prevalence of NY-ESO-1 downregulation in GBM patient tumors nor the presumed mechanism of downregulation by promoter methylation in GBM has been formally established. Methods We characterized baseline CpG methylation of NY-ESO-1 in 30 bulk patient GBM samples, 10 patient-derived gliomaspheres, and three established tumor cell lines using bisulfite sequencing. We induced NY-ESO-1 expression in vitro in U251 human GBM cells using the hypomethylating agent decitabine (DAC). We investigated the epigenetic response of DAC-treated U251 with bisulfite sequencing and NY-ESO-1 expression with quantitative real-time PCR. Lastly, we performed single-cell RNA sequencing on DAC-treated GBM U251 to evaluate tumor subpopulations that upregulate NY-ESO-1 and other co-expressed CTA after DAC treatment. Results Baseline NY-ESO-1 expression is associated with promoter methylation in the majority of GBM. Treatment of cells with 1 µM DAC every day for 4 days explicitly demethylated the promoter region of NY-ESO-1 and resulted in a 1000-fold increase in mRNA expression. DAC treatment upregulates NY-ESO-1 coordinately with other cancer/testis antigens CTAG2 and MAGEA4 as demonstrated by single-cell RNA sequencing. Conclusion Exposure of U251 to DAC results in promoter demethylation in NY-ESO-1 and increased expression of CTA. DAC treatment may therefore render GBM susceptible to targeting of these antigens by T-cells, revealing a feasible strategy of NY-ESO-1 and co-expressed CTA promoter demethylation to sensitize GBM to immunotherapy.

Understanding the Biological Basis of Glioblastoma Patient-derived Spheroids.

Resistance to glioblastoma (GB) therapy is attributed to the presence of glioblastoma stem cells (GSC). Here, we defined the behavior of GSC as it pertains to proliferation, migration, and angiogenesis. Human-derived GSC were isolated and cultured from GB patient tumors. Xenograft GSC were extracted from the xenograft tumors, and spheroids were created and compared with human GSC spheroids by flow cytometry, migration, proliferation, and angiogenesis assays. Oct3/4 and Sox2, GFAP, and Ku80 expression was assessed by immunoanalysis. The xenograft model showed the formation of two different tumors with distinct characteristics. Tumors formed at 2 weeks were less aggressive with well-defined margins, whereas tumors formed in 5 months were diffuse and aggressive. Expression of Oct3/4 and Sox2 was positive in both human and xenograft GSC. Positive Ku80 expression in xenograft GSC confirmed their human origin. Human and xenograft GSC migrated vigorously in collagen and Matrigel, respectively. Xenograft GSC displayed a higher rate of migration and invasion than human GSC. Human GSC were more aggressive in growth and proliferation than xenograft GSC, while xenograft GSC had increased invasion and migration compared to human GSC. A simple in vitro spheroid system for GSC provides a superior platform for the development of precision medicine in the treatment of GB.

Reduced EGFR and increased miR-221 is associated with increased resistance to temozolomide and radiotherapy in glioblastoma

Despite aggressive treatment with temozolomide and radiotherapy and extensive research into alternative therapies there has been little improvement in Glioblastoma patient survival. Median survival time remains between 12 and 15 months mainly due to treatment resistance and tumor recurrence. In this study, we aimed to explore the underlying mechanisms behind treatment resistance and the lack of success with anti-EGFR therapy in the clinic. After generating a number of treatment resistant Glioblastoma cell lines we observed that resistant cell lines lacked EGFR activation and expression. Furthermore, cell viability assays showed resistant cells were significantly less sensitive to the anti-EGFR agents when compared to parental cell lines. To further characterise the resistance mechanism in our cells microRNA prediction software identified miR-221 as a negative regulator of EGFR expression. miR-221 was up-regulated in our resistant cell lines, and this up-regulation led to a significant reduction in EGFR expression in both our cultured cell lines and a large cohort of glioblastoma patient tumor tissue.

IMMU-02. NEOANTIGENS ARISING FROM ALTERNATIVE SPLICING EVENTS MAY BE TARGETED BY TUMOR INFILTRATING LYMPHOCYTES IN GLIOBLASTOMAS

Abstract INTRODUCTION Alternative splicing, the cellular process that converts premature mRNA to mature mRNA and allows for single genes to produce multiple protein products, is frequently dysregulated in many cancers, including glioblastoma. However, along with non-synonymous mutations in the DNA, altered splicing mechanisms in cancers may produce novel antigens (so-called neoantigens) that distinguish cancer cells from healthy cells and can thus be targeted by the immune system. METHODS We developed a new computation pipeline (IRIS – Isoform peptides from RNA splicing for Immunotherapy targets Screening) that took bulk RNA-sequencing data from 23 glioblastoma patient tumor samples and predicted neoantigens that may arise from alternative splicing events. We prioritized predicted neoantigens that arose in HLA*A02:01 and HLA*A03:01 patients and selected 8 potential neoantigens to generate peptide:MHC Class 1 dextramers. We tested PBMCs and/or ex vivo expanded tumor infiltrating lymphocytes (TIL) from 6 of our glioblastoma patients against these dextramers, sorted for any neoantigen-reactive T cells, and performed single-cell RNAsequencing on the sorted population to determine the TCR sequence. RESULTS Among the 8 predicted neoantigens tested, 7 of the neoantigens were recognized by at least 1 patient’s T cells. 1 HLA*A03:01 epitope was recognized in 3 of the 4 HLA*A03:01 patients tested and this epitope was highly positive in an expanded TIL population, representing 1.7% of all CD3+ CD8+ cells. When we sorted for those neoantigen reactive T cells from the expanded TIL population and performed single-cell RNAsequencing, we found 325 unique T cell clonotypes, but the top 10 clonotypes represented 83.6% of all TCR clonotypes. The most frequent TCR clonotype represented 39.1% of the repertoire and suggests that clonal expansion of a select few TCR clones occurred within the tumor. CONCLUSIONS In total, our data indicates that neoantigens arising from alternative splicing events may represent a potential target for immunotherapy in glioblastoma.

Extracellular lipid loading augments hypoxic paracrine signaling and promotes glioma angiogenesis and macrophage infiltration

BackgroundPrimary brain tumors, in particular glioblastoma (GBM), remain among the most challenging cancers. Like most malignant tumors, GBM is characterized by hypoxic stress that triggers paracrine, adaptive responses, such as angiogenesis and macrophage recruitment, rescuing cancer cells from metabolic catastrophe and conventional oncological treatments. The unmet need of strategies to efficiently target tumor “stressness” represents a strong clinical motivation to better understand the underlying mechanisms of stress adaptation. Here, we have investigated how lipid loading may be involved in the paracrine crosstalk between cancer cells and the stromal compartment of the hypoxic tumor microenvironment.MethodsRegions from patient GBM tumors with or without the lipid loaded phenotype were isolated by laser capture microdissection and subjected to comparative gene expression analysis in parallel with cultured GBM cells with or without lipid loading. The potential involvement of extracellular lipids in the paracrine crosstalk with stromal cells was studied by immunoprofiling of the secretome and functional studies in vitro as well as in various orthotopic GBM mouse models, including hyperlipidemic ApoE−/− mice. Statistical analyses of quantitative experimental methodologies were performed using unpaired Student’s T test. For survival analyses of mouse experiments, log-rank test was used, whereas Kaplan-Meier was performed to analyze patient survival.ResultsWe show that the lipid loaded niche of GBM patient tumors exhibits an amplified hypoxic response and that the acquisition of extracellular lipids by GBM cells can reinforce paracrine activation of stromal cells and immune cells. At the functional level, we show that lipid loading augments the secretion of e.g. VEGF and HGF, and may potentiate the cross-activation of endothelial cells and macrophages. In line with these data, in vivo studies suggest that combined local tumor lipid loading and systemic hyperlipidemia of ApoE−/− mice receiving a high fat diet induces tumor vascularization and macrophage recruitment, and was shown to significantly decrease animal survival.ConclusionsTogether, these data identify extracellular lipid loading as a potentially targetable modulator of the paracrine adaptive response in the hypoxic tumor niche and suggest the contribution of the distinct lipid loaded phenotype in shaping the glioma microenvironment.

A Simple Three-dimensional Hydrogel Platform Enables Ex Vivo Cell Culture of Patient and PDX Tumors for Assaying Their Response to Clinically Relevant Therapies.

A cell culture platform that enables ex vivo tissue growth from patients or patient-derived xenograft (PDX) models and assesses sensitivity to approved therapies (e.g., temozolomide) in a clinically relevant time frame would be very useful in translational research and personalized medicine. Here, we present a novel three-dimensional (3D) ECM hydrogel system, VersaGel, for assaying ex vivo growth and therapeutic response with standard image microscopy. Specifically, multicellular spheroids deriving from either 5 patients with glioblastoma (GBM) or a renal cell carcinoma (RCC) PDX model were incorporated into VersaGel and treated with temozolomide and several other therapies, guided by the most recent advances in GBM treatment. RCC ex vivo tissue displayed invasive phenotypes in conditioned media. For the GBM patient tumor testing, all five clinical responses were predicted by the results of our 3D-temozolomide assay. In contrast, the MTT assay found no response to temozolomide regardless of the clinical outcome, and moreover, basement membrane extract failed to predict the 2 patient responders. Finally, 1 patient was tested with repurposed drugs currently being administered in GBM clinical trials. Interestingly, IC50s were lower than C max for crizotinib and chloroquine, but higher for sorafenib. In conclusion, a novel hydrogel platform, VersaGel, enables ex vivo tumor growth of patient and PDX tissue and offers insight into patient response to clinically relevant therapies. We propose a novel 3D hydrogel platform, VersaGel, to grow ex vivo tissue (patient and PDX) and assay therapeutic response using time-course image analysis.

RND1 regulates migration of human glioblastoma stem-like cells according to their anatomical localization and defines a prognostic signature in glioblastoma.

Despite post-operative radio-chemotherapy, glioblastoma systematically locally recurs. Tumors contacting the periventricular zone (PVZ) show earlier and more distant relapses than tumors not contacting the PVZ. Since glioblastoma stem-like cells (GSCs) have been proposed to play a major role in glioblastoma recurrence, we decided to test whether GSC migration properties could be different according to their anatomical location (PVZ+/PVZ–). For that purpose, we established paired cultures of GSCs from the cortical area (CT) and the PVZ of glioblastoma patient tumors. We demonstrated that PVZ GSCs possess higher migration and invasion capacities than CT GSCs. We highlighted specific transcriptomic profiles in PVZ versus CT populations and identified a down-regulation of the RhoGTPase, RND1 in PVZ GSCs compared to CT GSCs. Overexpression of RND1, dramatically inhibited PVZ GSC migration and conversely, downregulation of RND1 increased CT GSC migration. Additionally, transcriptomic analyses also revealed a down-regulation of RND1 in glioblastoma compared to normal brain. Using the glioblastoma TCGA database, low levels of RND1 were also shown to correlate with a decreased overall survival of patients. Finally, based on signaling pathways activated in patients with low levels of RND1, we identified an RND1low signature of six genes (MET, LAMC1, ITGA5, COL5A1, COL3A1, COL1A2) that is an independent prognostic factor in glioblastoma. These findings contribute to explain the shorter time to progression of patients with PVZ involvement and, point out genes that establish the RND1low signature as key targets genes to impede tumor relapse after treatment.

Glioblastoma cellular cross-talk converges on NF-κB to attenuate EGFR inhibitor sensitivity

In glioblastoma (GBM), heterogeneous expression of amplified and mutated epidermal growth factor receptor (EGFR) presents a substantial challenge for the effective use of EGFR-directed therapeutics. Here we demonstrate that heterogeneous expression of the wild-type receptor and its constitutively active mutant form, EGFRvIII, limits sensitivity to these therapies through an interclonal communication mechanism mediated by interleukin-6 (IL-6) cytokine secreted from EGFRvIII-positive tumor cells. IL-6 activates a NF-κB signaling axis in a paracrine and autocrine manner, leading to bromodomain protein 4 (BRD4)-dependent expression of the prosurvival protein survivin (BIRC5) and attenuation of sensitivity to EGFR tyrosine kinase inhibitors (TKIs). NF-κB and survivin are coordinately up-regulated in GBM patient tumors, and functional inhibition of either protein or BRD4 in in vitro and in vivo models restores sensitivity to EGFR TKIs. These results provide a rationale for improving anti-EGFR therapeutic efficacy through pharmacological uncoupling of a convergence point of NF-κB-mediated survival that is leveraged by an interclonal circuitry mechanism established by intratumoral mutational heterogeneity.