Targeted Proteomics to Assess the Response to Anti-Angiogenic Treatment in Human Glioblastoma (GBM)

Glioblastoma (GBM) is a highly aggressive primary brain tumor with dismal outcome for affected patients. Because of the significant neo-angiogenesis exhibited by GBMs, anti-angiogenic therapies have been intensively evaluated during the past years. Recent clinical studies were however disappointing, although a subpopulation of patients may benefit from such treatment. We have previously shown that anti-angiogenic targeting in GBM increases hypoxia and leads to a metabolic adaptation toward glycolysis, suggesting that combination treatments also targeting the glycolytic phenotype may be effective in GBM patients. The aim of this study was to identify marker proteins that are altered by treatment and may serve as a short term readout of anti-angiogenic therapy. Ultimately such proteins could be tested as markers of efficacy able to identify patient subpopulations responsive to the treatment. We applied a proteomics approach based on selected reaction monitoring (SRM) to precisely quantify targeted protein candidates, selected from pathways related to metabolism, apoptosis and angiogenesis. The workflow was developed in the context of patient-derived intracranial GBM xenografts developed in rodents and ensured the specific identification of human tumor versus rodent stroma-derived proteins. Quality control experiments were applied to assess sample heterogeneity and reproducibility of SRM assays at different levels. The data demonstrate that tumor specific proteins can be precisely quantified within complex biological samples, reliably identifying small concentration differences induced by the treatment. In line with previous work, we identified decreased levels of TCA cycle enzymes, including isocitrate dehydrogenase, whereas malectin, calnexin, and lactate dehydrogenase A were augmented after treatment. We propose the most responsive proteins of our subset as potential novel biomarkers to assess treatment response after anti-angiogenic therapy that warrant future analysis in clinical GBM samples.In the context of glioblastoma (GBM)¹, the quest for effective biomarkers is vital given that GBM is the most aggressive primary brain tumor in adults and no curative treatment is currently available (1). GBM is characterized by extensive invasion into the brain parenchyma, a high proliferation rate, neo-angiogenesis and significant cellular and molecular heterogeneity. Current treatment involves neurosurgery, radiotherapy and chemotherapy, yet the median life expectancy of affected patients is less than fifteen months. Recent efforts have focused on targeting the vascular endothelial growth factor (VEGF) system which is critical for tumor angiogenesis, however GBM quickly develop escape mechanisms leading to tumor progression (2, 3). Previous work from our group demonstrated that GBMs adapt to anti-VEGF treatment via a metabolic switch in tumor cells toward increased glycolysis (4, 5). This was accompanied by increased hypoxia and tumor cell invasion, with little or no effect on tumor growth (4). In agreement with these preclinical studies, two large scale clinical trials addressing the impact of bevacizumab, a VEGF targeting antibody, in newly diagnosed GBM patients reported disappointing results: although progression free survival appeared to be improved, no effect on overall survival was observed (6, 7). The evaluation of such studies are complicated by the fact that anti-angiogenic agents affect blood vessel permeability thereby directly modulating neuroimaging parameters used to determine treatment effects (8, 9). Thus there is a need for molecular biomarkers to adequately determine treatment response to anti-angiogenic agents.MS-based proteomics (10, 11) is widely used in the field of cancer research in particular in the context of biomarker development including discovery and verification. The application of the selected reaction monitoring (SRM) approach to proteomics reinforced the importance of MS in biomarker development (12–14). Indeed, SRM is a targeted proteomics approach that allows a precise and absolute quantification of previously selected marker candidates (15, 16). Moreover it can be applied in a supervised discovery phase for potential biomarkers (17, 18), i.e. the precise quantification of a wider range of selected biomarkers of interest by the use of stable isotope labeled (SIL) peptides in crude quality. Because of its high selectivity, sensitivity and accuracy, SRM, also named multiple reaction monitoring (MRM), is currently the reference method in targeted proteomics (14, 19).The aim of this study was to identify proteins that are altered by anti-angiogenic treatment, thereby providing biomolecular signatures of tumor response in GBM. Ultimately such protein markers could be evaluated for their utility as markers of efficacy that allow to discriminate responders from nonresponders. The study was focused on target proteins that may exhibit significant differences in protein expression reflecting the metabolic switch exhibited during anti-angiogenic therapy. An SRM workflow designed on a triple quadrupole platform (20), was developed and optimized in the context of GBM xenografts treated with bevacizumab in order to perform, in a supervised manner, a precise relative quantification of target proteins. We have previously shown that patient derived GBM xenografts developed in rodents faithfully reflect human pathology and allow a detailed analysis of the tumor and stromal compartments (4, 21–24). Furthermore xenograft models facilitate the access to control samples as well as the possibility of controlled interventional studies (25). The results presented herein demonstrate the feasibility of SRM to precisely quantify small changes in protein concentration after treatment. We highlight the importance of peptide selection, data normalization and consideration of the variability of target proteins within complex biological samples before assessing their concentration changes in subsequent comparative studies. From an initial set of 100 candidates, we screened 74 proteins and identified 32 responsive to anti-angiogenic treatment. We propose malectin, calnexin, lactate dehydrogenase A (LDHA), and isocitrate dehydrogenase (IDH) as novel response markers to anti-angiogenic therapy.