Perry, Christina Joy
(2016)
Targeting DNA repair for personalisation of glioblastoma therapy.
PhD thesis, University of Nottingham.
Abstract
Background
Glioblastomas (GBM) are the most common primary brain tumour in adults in the UK and are associated with a poor prognosis. High expression of MGMT (a direct DNA repair protein) has been shown to decrease response to temozolomide and radiotherapy in GBM. Little, however, is known of the effects of other DNA repair genes on survival and response to treatment in GBM.
PTEN is a tumour suppressor gene commonly mutated in high grade glioma. Tumours deficient in PTEN have been shown in some, but not all, studies to have defective DNA double strand break repair. PTEN deficient tumours may therefore be amenable to treatment targeting with a DNA repair inhibitor, exploiting a strategy known as synthetic lethality.
In this study the prognostic significance of DNA repair status in high grade glioma will be explored, alongside the effects of ATM inhibitor treatment in PTEN proficient and deficient glioblastoma cell lines.
Methods
Kaplan Meier survival analysis was performed for the mRNA expression levels of 248 probes (representing 157 DNA repair genes) in two adult glioblastoma datasets (Test dataset [n=191] and TCGA dataset [n=508]). After adjustment for multiple comparisons, multivariate analysis was performed, including genes significantly associated with survival on Kaplan Meier analysis. A DNA repair gene prognostic index was generated incorporating DNA repair genes whose expression was associated with survival on multivariate analysis in both datasets. Artificial neural network analysis (ANN) was conducted in the TCGA dataset to investigate global gene interactions of the five key DNA repair genes included in the prognostic index. The prognostic significance of key genes included in the prognostic index was explored at the protein level in a cohort of 61 adult high grade gliomas (Nottingham cohort) by immunohistochemistry.
To explore synthetic lethality in PTEN proficient/deficient glioblastoma, Kaplan Meier survival analysis was performed for both PTEN and ATM mRNA (TCGA cohort) and protein expression (Nottingham cohort). Differences in DNA repair gene expression between PTEN proficient/deficient glioblastoma cell lines was assessed using the RT2 DNA repair PCR array. Sensitivity of these cell lines to the ATM inhibitor KU55933 was assessed using a cell proliferation (MTS) assay. The functional consequences of this treatment were evaluated using the neutral comet assay and flow cytometric analysis of γH2AX accumulation, apoptosis and cell cycle progression. To further validate these findings, doxycycline-inducible PTEN knock-down glioblastoma cell lines were generated using lentiviral delivery of shRNA against PTEN.
Results
In both the Test and TCGA cohorts, a 14 DNA repair gene expression panel was associated with poor survival. On cox multivariate regression analysis APE1, NBN, PMS2, MGMT and PTEN were found to be independently associated with poor prognosis. A DNA repair prognostic index incorporating APE1, NBN, PMS2, MGMT and PTEN stratified patients into three distinct prognostic sub-groups with worsening survival (p<0.000001). ANN analysis of APE1, NBN, PMS2, MGMT and PTEN revealed interactions with genes involved in the response to hypoxia (HIF1AN), fatty acid metabolism (ACACB, ACSL4), transcriptional response (FOXG1, THRA), and nucleotide metabolism (HPRT). At the protein level, low PTEN (p=0.042) and low APE1 (p=0.031) retained a significant association with poor prognosis.
In the subgroup of GBMs with high expression of PTEN, high expression of ATM mRNA (probe 208442_s_at p=0.03, probe 210858_x_at p=0.025) and protein (p=0.022) was associated with worse overall survival than in patients with low ATM expression. In cell lines studies PTEN deficiency was not associated with sensitivity to an ATM inhibitor; however the PTEN proficient LN18 cell line exhibited increased sensitivity to KU55933 treatment. Higher rates of apoptosis and G1 cell cycle arrest were demonstrated in the LN18 cell line in response to 10µM KU55933. KU55933 treatment did not result in an increase in DNA damage (including double strand breaks), as assessed by the neutral comet assay and γH2AX accumulation. Knock-down of PTEN did not significantly influence sensitivity to KU55933.
Conclusion
DNA repair status clearly has prognostic significance in high grade glioma. There is, however, no evidence to support the use of ATM inhibitors in PTEN deficient GBM. Further work is required to identify why the LN18 PTEN proficient glioblastoma cell line is selectively sensitive to ATM inhibitor treatment.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[img]](/style/images/fileicons/application_pdf.png)