Identification of transcriptome signature for predicting clinical response to bevacizumab in recurrent glioblastoma

Abstract

Glioblastomas are among the most fatal brain tumors. Although no effective treatment option is available for recurrent glioblastomas (GBMs), a subset of patients evidently derived clinical benefit from bevacizumab, a monoclonal antibody against vascular endothelial growth factor. We retrospectively reviewed patients with recurrent GBM who received bevacizumab to identify biomarkers for predicting clinical response to bevacizumab. Following defined criteria, the patients were categorized into two clinical response groups, and their genetic and transcriptomic results were compared. Angiogenesis-related gene sets were upregulated in both responders and nonresponders, whereas genes for each corresponding angiogenesis pathway were distinct from one another. Two gene sets were made, namely, the nonresponder angiogenesis gene set (NAG) and responder angiogenesis gene set (RAG), and then implemented in independent GBM cohort to validate our dataset. A similar association between the corresponding gene set and survival was observed. In NAG, COL4A2 was associated with a poor clinical outcome in bevacizumab-treated patients. This study demonstrates that angiogenesis-associated gene sets are composed of distinct subsets with diverse biological roles and they represent different clinical responses to anti-angiogenic therapy. Enrichment of a distinct angiogenesis pathway may serve as a biomarker to predict patients who will derive a clinical benefit from bevacizumab.

Keywords: Angiogenesis; bevacizumab; biomarkers; gene expression signatures; glioblastoma.

Figure 1

MRIs following BEZ treatment. (A) G364 demonstrated remarkable decrease in contrast enhancement and T2 infiltration accompanied with edema following bevacizumab, while both contrast enhancement and T2 infiltration of G074 (B) had been maintained and somewhat increased despite bevacizumab. MRI, magnetic resonance image; BEZ, bevacizumab; pre‐BEZ, before treatment of bevacizumab; post‐BEZ#2, after 2 cycles of bevacizumab treatment.

Figure 2

Upregulated genes and associated GO pathways in responder versus nonresponder group. (A) Differentially expressed genes between responder and nonresponder groups are plotted as a volcano plot. Genes within the same genetic pathways are represented using the same color scale. Dots in the right side of the graph depict genes that are upregulated in nonresponders while dots in the left side demonstrate genes upregulated in responders. (B) Upregulated genes in each group are functionally annotated with GO terminology, and top pathways with statistical significance are illustrated. ECM, extracellular matrix.

Figure 3

Responders and nonresponders have distinct angiogenesis‐related gene sets. Two angiogenesis‐associated gene sets regarding the clinical prognosis to bevacizumab therapy are made and termed as responder angiogenesis gene set (RAGs) and nonresponder angiogenesis gene set (NAGs). (A) We applied these gene sets to the independent glioblastoma cohort from the BELOB trial. The ssGSEA scores of each gene set are represented as Z‐scores. (B) RAGs expression was upregulated in responders (long‐term survivors), whereas NAGs expression was upregulated in nonresponders (short‐term survivors) of the BELOB trial. Short‐ and long‐term survivors are defined as follows: short‐term survivors denote patients who survived for less than 25% of the survival period of the entire cohort. Long‐term survivors are patients who survived for more than 75% of the overall survival of entire cohort. Error bars mean standard deviation of ssGSEA Z‐scores and n.s refer statistical insignificance from Wilcoxon test. ssGSEA, single‐sample gene set enrichment analysis; RAG, responder angiogenesis gene set; NAG, nonresponder angiogenesis gene set.

Figure 4

Good‐prognosis angiogenesis genes (GPAGs) and poor‐prognosis angiogenesis genes (PPAGs) scores in the SMC glioblastoma cohort. Tian et al. identified two distinct subsets related to clinical prognosis within the angiogenesis‐associated genes and defined them as GPAG and PPAG. These two gene sets are applied to our cohort 26. (A) GPAGs and PPAGs are implemented to our glioblastoma cohort and showed similar association between prognosis and corresponding gene sets. (B) A few genes are shared between GPAG/PPAG and NAG/RAGs. SMC, Samsung Medical Center; ssGSEA, single‐sample gene set enrichment analysis; GPAG, good‐prognosis angiogenesis genes; PPAG, poor‐prognosis angiogenesis genes; RAG, responder angiogenesis gene set; NAG, nonresponder angiogenesis gene set.

Figure 5

COL4A2 is associated with poor prognosis of glioblastoma with BEZ treatment. Public data from AVAglio+RTOG0825 (Gene Expression Omnibus; Access number: GSE84010 (https://www.ncbi.nlm.nih.gov/geo)) are used to validate the prognostic significance of COL4A2 in BEZ‐treated patients with GBM. (A) COL4A2 expression and overall survival of the patients with GBM treated with BEZ. Patients treated with BEZ are stratified according to COL4A2 mRNA expression. (B) COL4A2 expression and overall survival of the total glioblastoma dataset regardless of BEZ treatment. Patients from the entire cohort stratified according to COL4A2 mRNA expression. BEZ, bevacizumab; GBM, glioblastoma.