PARP1 expression and its correlation with survival is tumour molecular subtype dependent in glioblastoma

Abstract

Overexpression of PARP1 exists in various cancers, including glioblastoma (GBM). Although PARP1 inhibition is a promising therapeutic target, no comprehensive study has addressed PARP1's expression characteristics and prognostic role regarding molecular heterogeneity in astrocytomas including GBM. Our aim was to evaluate PARP1's associations with survival, WHO grade, lineage specific markers, and GBM transcriptomic subtypes. We collected genomic and clinical data from the latest glioma datasets of The Cancer Genome Atlas and performed PARP1, ATRX, IDH1, and p53 immunohistochemistry on GBM tissue samples. We demonstrated that PARP1 gain and increased mRNA expression are characteristics of high-grade astrocytomas, particularly of Proneural and Classical GBM subtypes. Additionally, higher PARP1 levels exhibited an inverse correlation with patient survival (p<0.005) in the Classical subgroup. ATRX (p=0.006), and TP53 (p=0.015) mutations were associated with increased PARP1 expression and PARP1 protein level correlated with ATRX loss and p53 overexpression. Furthermore, higher PARP1 expression together with wildtype TP53 indicated shorter survival (p=0.039). Therefore, due to subtype specificity, PARP1 expression level and TP53 mutation status are reliable marker candidates to distinguish Proneural and Classical subtypes, with prognostic and therapeutic implications in GBM.

Keywords: PARP1; glioblastoma; glioma; p53.

Figure 1. The overview of the analysis strategy for PARP1 characterisation used in this study

Figure 2. PARP1 copy number and expression levels in GBM subtypes and glioma WHO grades

Association between copy number alterations (CNAs) and mRNA expressions of PARP1 in GBM (A). PARP1 mRNA expression level z-scores in different grades of astrocytomas (B) and in GBM subtypes (C). The TCGA glioblastoma multiforme (WHO grade IV) & brain lower grade glioma (WHO grade II & III) TCGA datasets were used for the analysis.

Figure 3. Association between TP53 gene status and PARP1 mRNA expression in glioblastoma

PARP1 mRNA expression was higher in mutated TP53 cases (A), but only missense TP53 mutations showed significant difference compared to wild-type TP53 (B). The TCGA glioblastoma multiforme, provisional dataset was used for the analysis.

Figure 4. Immunohistochemical staining pattern for various markers in a clinical cohort of glioblastoma

Immunohistochemical positive and negative staining for IDH1^R132H, positive and negative staining patterns of ATRX and p53, high and low nuclear expression of the PARP1. Original magnification: x200. Scale bar: 100 μm.

Figure 5

Kaplan–Meier plot of overall survival (OS) stratified by high and low PARP1 mRNA expression in GBM (A) and in Classical GBM subgroup (B). The TCGA glioblastoma multiforme, provisional dataset was used for the analysis. (Significance was assessed by the log-rank method.)

Figure 6

Kaplan–Meier plot of overall survival (OS) stratified by the combination of high and low PARP1 mRNA expression and TP53 mutation status in GBM. The TCGA glioblastoma multiforme, provisional dataset was used for the analysis. (Significance was assessed by the log-rank method.)

Figure 7. Association between PARP1 and p53 pathway in glioblastoma

A gene-gene interaction network presenting the correlation between PARP1 and the general members of p53 pathway in GBM according to mRNA expressions (A). The associations of PARP1 mRNA expression levels with CDKN2A and MDM4 copy number alterations in GBM (B). The TCGA glioblastoma multiforme, provisional dataset was used for the analysis.

Figure 8. The impact of PARP1 levels on p53 pathway-related genes in glioblastoma

The mRNA expression of TP53 and TP53BP1 was increased in PARP1-high group, whereas CDKN1A (p21) mRNA expression was higher in PARP1-low group. The TCGA glioblastoma multiforme, provisional dataset was used for the analysis. Level of significance: * (p < 0.05), ** (p < 0.01), and *** (p < 0.001).