Treatment Options in Recurrent GBM Research

Treatment Options in Recurrent GBM Research Strategies for clinical applications The multi-omics data may also reveal important leads for therapeutic applications. A very recent review on GBM, reported outcomes of clinical trials investigating current treatment options in recurrent GBM, including anti-angiogenic, signaling pathway blockade and immunotherapy based approaches (1). However the genetic and cellular heterogeneity reflects in the modest results obtained so far. This necessitates identification and validation of better therapeutic targets and active strategies to combat GBM. Some novel strategies are showing promise in Phase II trials and preliminary data is becoming available, such as, EGFRvIII peptide vaccine, Rindopepimut; CD95 targeted monoclonal antibody, APG100 and multi-targeted tyrosine kinase inhibitor cabozantinib (1). A multi-pronged approach targeting a panel of proteins may thus hold the key to eliciting a synergistic response and prove more beneficial than current treatment modalities targeting individual markers. When it comes to circulat ory or plasma-based biomarkers, in view of the technical limitations encountered in deep and direct plasma analysis as discussed earlier, alternate methods which would allow prediction of tumor related molecules and their targeted exploration would be highly useful. One of the outcomes of the study was the identification of effective strategies for data analysis and integration, facilitated by the bioinformatics tools available today. It shows experimental identification of proteins passed through the screen to ensure analytical rigor and functional relevance as above (Stage 1). Biologically important and potential tumor specific proteins identified in expression studies are then assessed for their secretory potential based on computational prediction algorithms for signal peptide and transmembrane domain containing proteins, such as, SignalP and TMHMM, respectively and via non-classical secretory mechanisms using SecretomeP. These proteins are further prioritized based on their de tectability and occurrence in proteomic data for secretome, CSF and plasma analysis (normal or patient) (Stage 2). The potential secretory candidates are then explored in plasma in a targeted manner (Stage 3). Interestingly, some of these proteins were identified in analysis of plasma or CSF from GBM patients (2, 3). Once bioinformatically scrutinized as above and compiled, the candidate biomarker panels, can be subjected to validation and experimentation in cohorts of tissue sections, blood plasma/serum specimens from patients (Stage 4). We believe construction of such high confidence protein panels would be a valuable paradigm for studies in larger cohorts in clinical experimental designs. High confident lead candidates for experimental application GBM Secreted proteins Secreted proteins have an integral role in GBM tumorigenesis through cell growth, migration, invasion, and angiogenesis besides being important in normal physiological processes and thus instrumental to the discovery of cancer biomarkers. Besides being useful as markers for typing the tumor, their presence in easily accessible body fluids makes them useful for monitoring the disease progression or treatment response and recurrence. A thorough survey of all available literature was done to identify the several candidate biomarkers have been reported in serum or plasma of GBM patients and these are shown in Table 1 in Chapter 1. However, such potential and promising new biomarkers are yet to be rigorously evaluated for application against this unmet need. Non-invasive methods based on circulatory biomarkers would be useful for monitoring not only GBM patients but also for lower grades Grade II and III tumors that exhibit longer survival periods. Further, some new reports on circulating tumor DNA (ctDNA) that have identified in the plasma of GBM patients such as mutated IDH1 DNA (4), methylated MGMT DNA (5) and EGFRvIII mutant DNA (6). The highly sensitive sequencing based methods for detection of circulatory tumor DNA (ctDNA) in patients plasma are under progress (7). These ctDNA markers shed by dead tumor cells may surface in future to be reasonable indicators for tumor diagnostics. Kinases in GBM Identification of GnRH signaling pathway using an alternate approach As mentioned in Chapter 2, I used alternate approaches to enhance pathway views by targeting specific protein families, i.e. kinases. Protein kinases (PKs) are well known therapeutic targets in different cancers and a family of proteins that are major components of signal transduction pathways acting as membrane receptors (RTKs) or as intracellular signaling mediators (non-receptor PKs) and several protein kinases have been implicated in gliomagenesis (8, 9). Several studies have also shown altered expression of protein kinases in GBM and targeted therapies directed towards RTKs using kinase inhibitors are in clinical trials (10, 11). There is renewed optimism in the use of kinase inhibitors to treat GBM (12). New therapeutic strategies have emerged that use multi-targeted kinase inhibitors to simultaneously disrupt multiple kinases (13). The GBM data was found to be enriched with several kinases. A total of 102 kinases were present in GBM datasets; 77 different kinases in transcript omics data and 30 kinases in proteomics data with 26 in common between them.   Pathway analysis using these kinases revealed GnRH signaling as the top pathway that has still not investigated in the context of GBM. We observe an overall enrichment of about 129 entities from omics datasets of which 26 kinases and 57 non-kinase members are coming from the concordant (n=711) transcriptome and proteome dataset. The 26 concordant kinases along with their fold changes are shown in the Figure 48 below. A large proportion of GnRH pathway entities include kinases (MAPKs, CAMKs, and RTKs) that enabled its identification as a top pathway using this approach. A targeted search of other non-kinase members of the pathway resulted in additional members of the pathway in omics datasets that further increased its significance value. In GBM, it has been shown that human GnRH receptors are expressed in tumor cells and receptor activation affects apoptosis, adhesion and angiogenesis to promote tumorigenesis. GnRH signalling as a possible therapeutic target in cancer has already been suggested and put together with my observations it strongly supports this possibility in the context of GBM. The expanded hand-curated map of GnRH signaling is a valuable resource for the scientific community. Expression of GnRH and GnRH receptor has been reported in GBM cell lines and tissue samples at both mRNA and protein levels concordant with clinical data obtained using GBM tumor tissues and treatment with GnRH agonists resulted in anti-proliferative activity (14-16).There is also evidence that the analogues can cross the blood-brain barrier, indicating suitability for treatment of malignant glioblastomas (17). Given the significance of this pathway in cancers and GBM, further understanding the molecular interplay involving GnRH signalling pathway in light of my findings will reveal is use as a potential molecular and therapeutic target.      Ã‚   Glioma Amplicon and Risk Regions The protein coding genes implicated in Glioma and other cancers were clustered based on their chromosomal locations using Gene Set Enrichment Analysis tools to compute overlaps with positional gene sets from Molecular Signatures database and further clustered based on proximity to other known oncogenes from Atlas of Genetics and Cytogenetics in Oncology and Haematology data resource, to identify colocalized gene clusters on Chr. 12 and other chromosomes as shown in Chapter 3. An important finding was that larger number of overexpressed differential regulated genes in glioma datasets mapped to two significant regions the glioma amplicon (n=37) in 12q13-15 region and the glioma susceptibility (n=16) in the 12p13 region implicated as a major risk region in patients with a family history of gliomas. The discovery of these two clusters of overexpressed genes provides a biological validation of mass-spectrometry derived data. Apart from these two essential regions, several genes from the glioma dataset were found to cluster around amplicons on other chromosomes and other known cancer associated genes that were not identified in GBM datasets but present in close proximity to them. These can be investigated in a more targeted manner in glioma.   Many studies have been done to understand the biological significance of these amplicon regions in gliomas that indicate that these amplifications are more frequent in gliomas than previously thought and have different distribution patterns in low grade versus high grade tumors (18, 19). Overall, a relative high degree of amplifications and deletions are seen in GBM that have implications on the expression of the genes involved and contribute to relevant pathogenic genes (20). Novel genes and isoforms Alternative splicing increases the repertoire of protein functionality and heterogeneity and aberrant splicing events have been frequently seen in several cancers, including GBM and increasing evidence now points to their important role in tumor initiation and progression. The concept of proteogenomics has emerged rapidly as a valuable approach to integrate mass spectrometry (MS)-derived proteomic data with transcriptomic data to identify novel splice variants. However, the role of alternative splicing in GBM is still nascent and needs to be explored as potential biomarkers or molecular targets. As detailed in Chapter 4, the identification of a novel variant of NCAM1, using a proteogenomics approach with 5 peptide evidences from MS data spanning a novel exonic region, is very significant finding in GBM. NCAMs are well characterized glycoproteins that mediate cell-cell or cell-matrix adhesion among neurons and between neurons and muscle. Several splice variants of NCAM1 have been identified (21, 22) and alterations in these have been found in serum and tissues of brain tumors (23, 24). NCAM1 has 5 known isoforms and also exhibits glycoforms as it can be post-translationally modified by the addition of polysialic acid (PSA), which is thought to abrogate its homophilic binding properties and affect the adhesive properties of NCAM (25). Further, PSA conjugated NCAM, was shown to potentiate migration via FGFR signaling distinct from its adhesion capability (26).   The following observations may be noted with respect to this novel variant: The observation is supported in transcriptomics data in 18 out of 25 RNAseq samples. Multiple gene modelling software such as Augustus, GenScan, AceView and Ensemble support the presence of this novel exon in their gene models and a high degree of conservation was seen as expected for an exonic region. This variant was also separately identified in MS-derived Human Proteome and IvyGAP RNAseq datasets NCAM1 is upregulated in several cancers; however, in GBM both transcript and protein data support its down regulation.   We observed two known forms of NCAM1 as well as the novel form to be down regulated. It is interesting to note that the miRNA (hsa-mir-30a-5p) that regulates NCAM1 is upregulated in GBM indicating the deregulation of a putative oncogenic cascade. In summary, our findings demonstrate the usefulness of combining omics approaches to identify novel putative candidates in GBM. Although, it is not clear if the novel splice variant represents a major or minor form of NCAM1. At the transcript level, it seems to be a minor component; however, preliminary assessment at the protein level is suggestive of it being a predominant form. Regardless, it would be interesting to explore the biological significance of the novel splice variant of NCAM1 and examine its role in GBM tumorigenesis. Hence, in the light of this observation my identification of novel NCAM1 splice variant through proteogenomics analysis using GBM RNAseq data is a very important finding in GBM. The effect of this novel variant on cell-cell adhesion and migration in GBM needs to be further investigated in a targeted manner. Disease implications and targeted analysis Studies suggest that gliomas constitute a rapidly progressing neurodegenerative disease caused by the malignant growth of glial cells that nourish neurons, resulting in a loss of brain function. Glutamate excitotoxicity is observed in several neurological diseases, which is also utilized by gliomas to gain growth advantage (27). My observations that neurological conditions like Alzhiemers and Parkinsons disease share many common genes with gliomas possible indicate shared molecular mechanisms inducing neurodegeneration. Further, the chromosomal mapping of glioma differentials revealed two clusters; one around 12p13 implicated as a glioma risk region and another around 12q13-15 region harboring a glioma amplicon with several overexpressed and amplified genes. Hence, extracting gene/disease associations and generation of a glioma-centric functional and diseasome network is important for understanding GBM tumorigenesis. Further, this region was found to be enriched in several cancers in cluding other brain neoplasms and neurological diseases that may share disease genes and processes with gliomas. Only 22 of the observed 108 disease genes in the diseasome network were identified in our proteomic analysis. The other 86 disease genes implicated in gliomas but not identified in our dataset can be investigated in a more targeted manner in gliomas, providing a global view of linkages between disease phenotypes. Additionally, the finding that chromosomal proximity of genes may have an impact on their functions can be used to explore the functions of missing proteins mapping within functional cassettes of related protein/genes. Such investigations offer newer paradigms that may be valuable to investigate and present clinically important targets. Future Scope Metabolomic data integration and potential Compared to the genome and proteome, metabolome represents the phenotypic changes more closely and has already been investigated for malignancies such as breast, ovarian, colon, prostrate and esophageal cancers. This line of investigation has been extended to gliomas albeit on a smaller scale, revealing novel insights into the role of metabolites in GBM tumorigenesis (reviewed in ref. (28)). Previous studies have revealed how mutations can lead to generation of oncometabolites such as 2-hydroxyglutarate (2-HG) specifically in IDH1 mutated gliomas (29). The discovery by Otto Warburg that cancer cells prefer to metabolize glucose through a seemingly inefficient process of aerobic glycolysis   led to the application of 18-FDG-PET imaging to predict the histological grade of gliomas.   Using this technique we could now distinguish low grade gliomas that have low specific uptake (SUV) values from grade III and IV that have higher SUVs.   One study performed global metabolic profilin g using mass-spectrometry coupled to liquid/gas chromatography on patient derived tumor samples and found increased levels of glutathione, tryptophan and metabolites associated with phentose phosphate and nucleotide synthesis and glycolytic intermediates such as phosphoenolpyruvate (PEP) and 3-phosphoglycerate (30). These studies have collectively provided a window of opportunity for further investigation and integrating these changes with the changes at proteomic, transcriptomic and genomic levels will be the next big step in to study the underlying biology of these tumors. Improving pathway analysis with phosphoproteomics data Protein phosphorylation plays a central role in transmitting the signal from outside the cell through a cascade effect into an intracellular signal to control the biochemical pathways in all living cells. This mechanism of activation or deactivation can be orchestrated by protein kinases via phosphorylation and phosphatases via dephosphorylation. Modifications to these signaling networks via mutations or abnormal protein expression or post-translational modifications may underlie both development and progression of tumorigenesis. Glioma Repository In order to facilitate annotation of key terms and manage the collection of high-throughput data coming from different omics technologies and platforms and make it easier to store and retrieve large amounts of information, I proposed to a schema for data annotation, collection and deposition. The data will be stored in the backend, in separate tables in a relational database (RDBMS), to enable effortless retrieval of key information for particular candidates of interest and also allow for complex querying. The outline for the schema is given below. Figure 49: Schema for development of a glioma repository