Sen Peng

Radiogenomics to characterize regional genetic heterogeneity in glioblastoma

Background Glioblastoma (GBM) exhibits profound intratumoral genetic heterogeneity. Each tumor comprises multiple genetically distinct clonal populations with different therapeutic sensitivities. This has implications for targeted therapy and genetically informed paradigms. Contrast-enhanced (CE)-MRI and conventional sampling techniques have failed to resolve this heterogeneity, particularly for nonenhancing tumor populations. This study explores the feasibility of using multiparametric MRI and texture analysis to characterize regional genetic heterogeneity throughout MRI-enhancing and nonenhancing tumor segments. Methods We collected multiple image-guided biopsies from primary GBM patients throughout regions of enhancement (ENH) and nonenhancing parenchyma (so called brain-around-tumor, [BAT]). For each biopsy, we analyzed DNA copy number variants for core GBM driver genes reported by The Cancer Genome Atlas. We co-registered biopsy locations with MRI and texture maps to correlate regional genetic status with spatially matched imaging measurements. We also built multivariate predictive decision-tree models for each GBM driver gene and validated accuracies using leave-one-out-cross-validation (LOOCV). Results We collected 48 biopsies (13 tumors) and identified significant imaging correlations (univariate analysis) for 6 driver genes: EGFR, PDGFRA, PTEN, CDKN2A, RB1, and TP53. Predictive model accuracies (on LOOCV) varied by driver gene of interest. Highest accuracies were observed for PDGFRA (77.1%), EGFR (75%), CDKN2A (87.5%), and RB1 (87.5%), while lowest accuracy was observed in TP53 (37.5%). Models for 4 driver genes (EGFR, RB1, CDKN2A, and PTEN) showed higher accuracy in BAT samples (n = 16) compared with those from ENH segments (n = 32). Conclusion MRI and texture analysis can help characterize regional genetic heterogeneity, which offers potential diagnostic value under the paradigm of individualized oncology.

A Novel Signaling Complex between TROY and EGFR Mediates Glioblastoma Cell Invasion

Glioblastoma is the most frequent primary brain tumor in adults and a highly lethal malignancy with a median survival of about 15 months. The aggressive invasion of the surrounding normal brain makes complete surgical resection impossible, increases the resistance to radiation and chemotherapy, and assures tumor recurrence. Thus, there is an urgent need to develop innovative therapeutics to target the invasive tumor cells for improved treatment outcomes of this disease. Expression of TROY (TNFRSF19), a member of the tumor necrosis factor (TNF) receptor family, increases with increasing glial tumor grade and inversely correlates with patient survival. Increased expression of TROY stimulates glioblastoma cell invasion in vitro and in vivo and increases resistance to temozolomide and radiation therapy. Conversely, silencing TROY expression inhibits glioblastoma cell invasion, increases temozolomide sensitivity, and prolongs survival in an intracranial xenograft model. Here, a novel complex is identified between TROY and EGFR, which is mediated predominantly by the cysteine-rich CRD3 domain of TROY. Glioblastoma tumors with elevated TROY expression have a statistically positive correlation with increased EGFR expression. TROY expression significantly increases the capacity of EGF to stimulate glioblastoma cell invasion, whereas depletion of TROY expression blocks EGF stimulation of glioblastoma cell invasion. Mechanistically, TROY expression modulates EGFR signaling by facilitating EGFR activation and delaying EGFR receptor internalization. Moreover, the association of EGFR with TROY increases TROY-induced NF-κB activation. These findings substantiate a critical role for the TROY–EGFR complex in regulation of glioblastoma cell invasion. Implications: The TROY–EGFR signaling complex emerges as a potential therapeutic target to inhibit glioblastoma cell invasion. Mol Cancer Res; 16(2); 322–32. ©2017 AACR.

EGFRvIII-Stat5 Signaling Enhances Glioblastoma Cell Migration and Survival

Glioblastoma multiforme (GBM) is the most common brain malignancies in adults. Most GBM patients succumb to the disease less than 1 year after diagnosis due to the highly invasive nature of the tumor, which prevents complete surgical resection and gives rise to tumor recurrence. The invasive phenotype also confers radioresistant and chemoresistant properties to the tumor cells; therefore, there is a critical need to develop new therapeutics that target drivers of GBM invasion. Amplification of EGFR is observed in over 50% of GBM tumors, of which half concurrently overexpress the variant EGFRvIII, and expression of both receptors confers a worse prognosis. EGFR and EGFRvIII cooperate to promote tumor progression and invasion, in part, through activation of the Stat signaling pathway. Here, it is reported that EGFRvIII activates Stat5 and GBM invasion by inducing the expression of a previously established mediator of glioma cell invasion and survival: fibroblast growth factor-inducible 14 (Fn14). EGFRvIII-mediated induction of Fn14 expression is Stat5 dependent and requires activation of Src, whereas EGFR regulation of Fn14 is dependent upon Src–MEK/ERK–Stat3 activation. Notably, treatment of EGFRvIII-expressing GBM cells with the FDA-approved Stat5 inhibitor pimozide blocked Stat5 phosphorylation, Fn14 expression, and cell migration and survival. Because EGFR inhibitors display limited therapeutic efficacy in GBM patients, the EGFRvIII–Stat5–Fn14 signaling pathway represents a node of vulnerability in the invasive GBM cell populations. Implications: Targeting critical effectors in the EGFRvIII–Stat5–Fn14 pathway may limit GBM tumor dispersion, mitigate therapeutic resistance, and increase survival. Mol Cancer Res; 16(7); 1185–95. ©2018 AACR.

Differential expression of the TWEAK receptor Fn14 in IDH1 wild-type and mutant gliomas

The TNF receptor superfamily member Fn14 is overexpressed by many solid tumor types, including glioblastoma (GBM), the most common and lethal form of adult brain cancer. GBM is notable for a highly infiltrative growth pattern and several groups have reported that high Fn14 expression levels can increase tumor cell invasiveness. We reported previously that the mesenchymal and proneural GBM transcriptomic subtypes expressed the highest and lowest levels of Fn14 mRNA, respectively. Given the recent histopathological re-classification of human gliomas by the World Health Organization based on isocitrate dehydrogenase 1 (IDH1) gene mutation status, we extended this work by comparing Fn14 gene expression in IDH1 wild-type (WT) and mutant (R132H) gliomas and in cell lines engineered to overexpress the IDH1 R132H enzyme. We found that both low-grade and high-grade (i.e., GBM) IDH1 R132H gliomas exhibit low Fn14 mRNA and protein levels compared to IDH1 WT gliomas. Forced overexpression of the IDH1 R132H protein in glioma cells reduced Fn14 expression, while treatment of IDH1 R132H-overexpressing cells with the IDH1 R132H inhibitor AGI-5198 or the DNA demethylating agent 5-aza-2′-deoxycytidine increased Fn14 expression. These results support a role for Fn14 in the more aggressive and invasive phenotype associated with IDH1 WT tumors and indicate that the low levels of Fn14 gene expression noted in IDH1 R132H mutant gliomas may be due to epigenetic regulation via changes in DNA methylation.

The TNF receptor family member Fn14 is highly expressed in recurrent glioblastoma and in GBM patient-derived xenografts with acquired temozolomide resistance

Background Glioblastoma (GBM) is a difficult to treat brain cancer that nearly uniformly recurs, and recurrent tumors are largely therapy resistant. Our prior work has demonstrated an important role for the tumor necrosis factor-like weak inducer of apoptosis (TWEAK) receptor fibroblast growth factor-inducible 14 (Fn14) in GBM pathobiology. In this study, we investigated Fn14 expression in recurrent GBM and in the setting of temozolomide (TMZ) resistance. Methods Fn14 mRNA expression levels in nonneoplastic brain, primary (newly diagnosed) GBM, and recurrent GBM (post-chemotherapy and radiation) specimens were obtained from The Cancer Genome Atlas data portal. Immunohistochemistry was performed using nonneoplastic brain, patient-matched primary and recurrent GBM, and gliosarcoma (GSM) specimens to examine Fn14 protein levels. Western blot analysis was used to compare Fn14 expression in parental TMZ-sensitive or matched TMZ-resistant patient-derived xenografts (PDXs) established from primary or recurrent tumor samples. The migratory capacity of control and Fn14-depleted TMZ-resistant GBM cells was assessed using the transwell migration assay. Results We found that Fn14 is more highly expressed in recurrent GBM tumors than their matched primary GBM counterparts. Fn14 expression is also significantly elevated in GSM tumors. GBM PDX cells with acquired TMZ resistance have higher Fn14 levels and greater migratory capacity than their corresponding parental TMZ-sensitive cells, and the migratory difference is due, at least in part, to Fn14 expression in the TMZ-resistant cells. Conclusions This study demonstrates that the Fn14 gene is highly expressed in recurrent GBM, GSM, and TMZ-resistant GBM PDX tumors. These findings suggest that Fn14 may be a valuable therapeutic target or drug delivery portal for treatment of recurrent GBM and GSM patients.

Abstract 289: Probing the non-enhancing component of glioblastoma: Targeting what is left behind

While genomic profiling and therapeutic selection support individualized GBM treatment, such therapeutic decision-making is usually made with reference to tumor obtained from the enhancing core region. GBM is known to be heterogeneous and exhibits a high resistance to standard therapies. To address whether non-enhancing tumor (representing the majority of tumor left behind after surgery) shows distinct genomic characteristics and therapeutic targets compared to the enhancing tumor core, we performed genome-wide exome-sequencing and RNA-sequencing for 12 patients with matched enhancing region and at least one non-enhancing region. Non-enhancing biopsies show a surprisingly high level of tumor content, with a median of 28% tumor cells and 6 of the 22 samples having greater than 50% tumor cells. Cognate non-enhancing and enhancing specimens demonstrated overall concordance in therapeutically actionable alterations (single nucleotide variants) and copy number alterations. However, non-enhancing regions were not genetically identical and did reveal additional and distinct variants compared to enhancing cores. For example, the non-enhancing region of patient 1 showed two nonsense NF1 mutations (R1534X; R2517X) while the enhancing region showed an NF1 frameshift mutation (F1247fs). Clonality analysis by LumosVar also indicated that 7 out of 12 patients harbored dissimilar cellular prevalence patterns between enhancing and non-enhancing regions. In addition, comparison of alternative polyadenylation between enhancing and non-enhancing regions uncovered distinct 39 UTR usage: e.g. SGMS2 and TOB1 tended to have longer 39 UTR in enhancing regions whereas longer 39 UTR of SYNPO and NOS1AP were prevalent in non-enhancing regions. We posit that the enhancing component of glioblastoma probably underrepresents the genomic alterations in patients9 tumors. Given non-enhancing tumor is left behind after surgical debulking, genomic profiling of this region would potentially reveal more accurate tumor vulnerabilities and lead to more effective therapy. Supported by a grant from the Ben & Catherine Ivy Foundation. Citation Format: Sen Peng, Rebecca Halperin, Harshil Dhruv, Sara Byron, Christophe Legendre, Joanna Phillips, Michael Prados, Michael Berens, Nhan Tran. Probing the non-enhancing component of glioblastoma: Targeting what is left behind [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 289.

Abstract 1142: Novel target discovery for glioblastoma using chemical biology fingerprinting

The most common adult brain tumor is Glioblastoma Multiforme (GBM), an extremely aggressive cancer with only scant treatment options. Even with standard of care most patients present with a recurrence and the median survival is only circa 15 months. The need, therefore, for new therapeutic targets and treatment options is pressing. Here we describe here a multipronged approach to identifying said targets. We present an established methodology for the isolation and culture of patient derived GBM samples that retain the “stem-like” fraction thought to underlie resistance and recurrence. Furthermore we show genomically that these samples represent specific subtypes of the disease yet still form distinct groups in unbiased clustering analysis. Thus we have multiple representative patient derived cultures that are suitable for our drug discovery and chemical biology analyses. Using a process we term Chemical Biology Fingerprinting (CBF) we utilize small focused, and clinically relevant, chemical collections in order to identify patterns of chemovulnerabilities across multiple samples. This allows an unbiased yet cancer relevant sub-stratification and the identification of agents, and therefore targets, which may be relevant for GBM patient subtypes. Indeed our use of the highly annotated NCI CTD2 Informer Set of chemicals allows ready drug-to-target mapping and facilitates data sharing across the CTD2 network. Moreover, already defined subgroups can be clustered to find agents, or groups of agents, that show selective activity against traditional classifications (e.g. proneural, mesenchymal etc.). Finally our strategy is permissive for the identification of “exceptional responders”. That is, individual patient samples that respond to a specific drug whilst most samples are refractory. In sum we demonstrate generation of patient derived models and identify specific, and novel, drugs that may be relevant for specific GBM subtypes. Supported by NIH U01CA168397 Citation Format: Darren Finlay, Pedro Aza-Blanc, Harshil Dhruv, Alexey Eroshkin, Craig Hauser, Jeff Kiefer, Seungchan Kim, Tao Long, Robert G. Oshima, Sen Peng, Gil Speyer, Michael Berens, Kristiina Vuori. Novel target discovery for glioblastoma using chemical biology fingerprinting [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1142. doi:10.1158/1538-7445.AM2017-1142

Abstract 1886: mGluR1 drives invasion of proneural subtype of glioblastoma cells

A major cause for the therapeutic failure and subsequent morbidity and mortality of glioblastoma (GBM) is the aggressive invasion of malignant cells into the surrounding normal brain that effectively renders complete surgical resection impossible and virtually assures recurrent tumor growth. Multi-omic profiling of GBM led to their molecular sub-classification into two distinct molecular subtypes: proneural (PN) and mesenchymal (MES). However, very little is known about shared or distinct invasion processes of cells in these genomically different subtypes. Using Microarray gene expression profiles of microdissected paired stationary core and invasive rim samples from 19 patients, we demonstrated that invasive gene signature of MES subtypes differs from PN. Specifically, using three orthogonal but intersecting bioinformatic approaches, i.e. gene set variation analysis, causal network analysis, and iRegulon analysis; we discovered that genes differentially expressed between PN core and rim could, to a meaningful degree, be accounted for based on their annotation as “regulated by the transcription factor REST”. REST functions as a repressor of gene expression. Of the genes repressed by REST, mGluR1 (Metabotropic Glutamate Receptor 1) was significantly overexpressed at the rim of PN glioma cells as compared to the core. Finally, we also investigated the role of mGluR1 in glutamate induced glioma cell migration; our results show that glutamate stimulates migration of proneural-like glioma cells (A172) as compare to non-proneural-like glioma cell (T98G). mGluR1 activation by glutamate has shown to induce activation of Pyk2 and Src in astrocytes; knockout of mGluR1 is not embryonic lethal. In summary, our data demonstrate that glutamate-induced glioma cell migration of PN subtype of GBM is dependent on mGluR1 and thus raises the prospect that therapeutic targeting of mGluR1 may be a novel approach to controlling the invasion of this deadly disease. Citation Format: Alena Gladwin, Sen Peng, Jeff Kiefer, Seugchan Kim, Michael Berens, Harshil D. Dhruv. mGluR1 drives invasion of proneural subtype of glioblastoma cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1886. doi:10.1158/1538-7445.AM2017-1886

Abstract 4671: Identification of novel drugs for glioblastoma using chemical biology fingerprinting

Proceedings: AACR 106th Annual Meeting 2015; April 18-22, 2015; Philadelphia, PA Glioblastoma Multiforme (GBM) is an aggressive brain tumor with very poor prognosis and extremely limited therapeutic options. GBM is the most common malignant brain tumor and the search for novel targets and/ or the repurposing of already extant drugs to treat the disease is therefore of utmost importance. We describe here a comprehensive multidisciplinary approach to identifying said targets and ergo potential therapies. We have applied a novel analytical strategy to The Cancer Genome Atlas (TCGA) GBM expression data to stratify GBM into novel subtypes we call molecular contexts, or mCs. Subsequently, a panel of patient-derived GBM xenografts was ascribed to our novel mCs. Utilizing a technique we term Chemical Biology Fingerprinting, or CBF, short-term cultures derived from these clinically-relevant preclinical models were screened for chemosensitivity with a deeply annotated, yet clinically relevant, chemical library. Agents that were statistically more toxic to one context than another were then re-tested in true drug dose response experiments to confirm sensitivity. Preliminary data demonstrated that mC14, characterized by mutant p53 and transcriptionally similar to the GBM proneural subtype, showed distinct vulnerability to Arsenic Trioxide (ATO) as compared to mC4, enriched for NF1 mutations and with transcriptional patterns similar to the GBM mesenchymal subtype. To validate the ATO vulnerability signature in GBM, we acquired 20 treatment naive archival patient samples, that were part of a Phase I/II clinical trial to study the efficacy of ATO and Temozolomide in combination with ionizing radiation ([NCT00275067][1]). Participants in the trial exhibited varied survival with ATO treatment (91 days to >1000 days) and the clinical samples were subtyped into our molecular contexts using RNAseq data. In summary, we demonstrate a subclassification of GBM into novel molecular contexts (mCs) and show that these contexts are differentially sensitive to clinically relevant drugs. Citation Format: Darren Finlay, Harshil Dhruv, Lisa Evers, Sen Peng, Jeff Kiefer, Seungchan Kim, Jeffrey Raizer, Michael Berens, Kristiina Vuori. Identification of novel drugs for glioblastoma using chemical biology fingerprinting. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4671. doi:10.1158/1538-7445.AM2015-4671 [1]: /lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT00275067&atom=%2Fcanres%2F75%2F15_Supplement%2F4671.atom

Abstract 1468: Characterization of patient-derived xenograft (PDX) models to evaluate clinical and therapeutic responses of glioblastoma multiforme

BACKGROUND/PURPOSE The Mayo Clinic has developed a large panel of patient-derived xenografts (PDX) from patients with glioblastoma multiforme (GBM). Here we report initial molecular profiling of the Mayo PDX panel and a comparison of patient and PDX molecular profile and response to therapy. METHODS & MATERIALS Clinical data was retrieved by retrospective chart review. Animal data from all PDX investigations conducted between 2004-2014 were retrieved from experimental logs and subsequently consolidated into a database for analysis. RESULTS From 1999 to 2014, 182 patient tumor samples of varying histological grades were attempted for xenograft with 73 resulting viable PDX lines. Viable xenografts were only produced from WHO grade IV tumor specimens, yielding an overall success rate of 49% for these tumors. GBM patients that produced viable xenografts compared to those that did not exhibited a trend for decreased overall survival (P = 0.18). There was no significant association between successful xenografting and whether tissue was from newly diagnosed (45/93%) or recurrent (20/40; 50%) tumors. Patient age > 45 at diagnosis was correlated with increased PDX viability in GBMs (p = 0.05). Of the viable PDX models analyzed, EGFR mutation was identified in 17 lines, TERT mutation was found in 13 lines, IDH mutation in 1 line, and MGMT hypermethylation in 25 lines. RNAseq was performed on orthotopic tumor samples from 53 PDX models. After excluding contaminating murine sequence reads, expression analysis demonstrated 32 models with a mesenchymal phenotype. Array comparative genomic hybridization was performed on 9 patient samples and derivative early, mid and late passage PDX tumors. Using unbiased hierarchical clustering, there was a high concordance between patient and xenograft models. Within the PDX panel, fractionated radiation (RT) alone and RT combined with temozolomide (TMZ) was tested in orthotopic tumors in 38 lines. The overall median survival benefit (ratio of median survival for treated vs. placebo) in PDX lines treated with RT only was 1.6 (range: 0.9-2.5) and with RT/TMZ was 2.5 (range: 1.1 - 8.9). There was a positive association between observed patient survival and the corresponding survival benefit in the PDX for subjects treated with RT/TMZ (r = 0.2; n = 17). Response to adjuvant TMZ was evaluated in 42 tumor lines, and response to bevacizumab was tested in 33 tumor lines, and correlations with clinical treatment response are being evaluated. CONCLUSIONS: The Mayo GBM PDX panel is widely used in the neuro-oncology community. The initial molecular analysis suggests a good correlation between patient and PDX at the level of genotypic characteristics and therapeutic sensitivity. Citation Format: Dioval A. Remonde, Brett L. Carlson, Mark A. Schroeder, Brock Armstrong, Sen Peng, Lisa Evers, Paul A. Decker, Jeanette Eckel Passow, Michael E. Berens, Nhan L. Tran, Robert B. Jenkins, Jann N. Sarkaria. Characterization of patient-derived xenograft (PDX) models to evaluate clinical and therapeutic responses of glioblastoma multiforme. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1468. doi:10.1158/1538-7445.AM2015-1468