Targeting the Fanconi anaemia DNA damage repair pathway to establish effective combinatorial strategies in glioblastoma

Introduction: Glioblastoma represent around half of newly diagnosed glial tumours and continue to portend poor survival (median 15.6 months) despite maximal safe surgical resection followed by DNA-damaging adjuvant therapies. Anomalies within the cellular DNA damage response (DDR) contribute to the treatment refractory nature of this devastating brain tumour with difficult-to-treat glioblastoma stem cell (GSC) subpopulations reported to exhibit heightened activity in various DDR processes. The Fanconi anaemia (FA) pathway represents a key DDR process which remains relatively inactive in normal brain, but is re-activated in glioblastoma, raising its value as a foundational target for cancer-specific treatment. However, similar to other difficult-to-treat diseases such as HIV and multi-drug resistant TB, novel drug combinations may be required to overcome the extensive genetic heterogeneity and resistance mechanisms observed in glioblastoma. This thesis therefore, considers whether combined targeting of the FA pathway and interconnected DDR processes could form a basis for new, effective multimodal therapies.

Methods: Established cell lines and more clinically-relevant 2D and 3D primary, patient- derived glioblastoma stem cell (GSC) models were used in conjunction with viability and clonogenic survival assays to assess therapeutic combinations targeting the DDR. Inhibitors of the FA pathway (FAPi), including novel specific FAPi under development (nFAPi), were validated and assessed alongside other key DDR targets including PARP1, ATR and ATM, and a range of techniques were used to assess FA pathway activity and the effects of combinatorial strategies on DNA damage and responses in GSCs.

Results: Elevated FA pathway gene expression in gliomas is associated with poor survival. Furthermore, patient-derived GSC populations, which drive therapeutic resistance, display high FA pathway expression relative to paired bulk tumour cell populations. We further show that inhibition of a single DDR process (FA pathway, PARP, ATR or ATM) increases the susceptibility of glioblastoma cell lines and patient-derived GSCs to current adjuvant therapy. Importantly, clinically approved PARP inhibitor (PARPi) monotherapy stimulates robust FANCD2 monoubiquitination, supporting a role of FA pathway activation in response to current DDR-targeted therapy. In clinically-relevant 3D GSC models, simultaneous inhibition of the FA pathway (FAPi) and PARP or ATR enhanced temozolomide sensitisation compared to a single DDR inhibitor (DDRi). Furthermore, combined PARPi + FAPi consistently conferred radiosensitisation whilst combined ATRi + FAPi led to a profoundly radiosensitising effect. Furthermore, comparison of a/b ratio enhancement suggests dual-DDRi strategies fundamentally alter the response of GSCs, whilst single cell gel electrophoresis and immunofluorescence studies suggest FA pathway-based DDRi combinations substantially delay the resolution of IR-induced DNA strand breaks at 6 hours post-treatment, with increased persistent DNA double strand breaks at 24 hours. In contrast to PARPi + FAPi, treatment with the ATRi + FAPi combination portended marked generation of RNA:DNA hybrids, however the substantial efficacy of both combinations over single DDRi strategies appeared to be independent of their cell cycle modulatory effects.

Finally, during these studies a ‘living biobank’ of 24 clinically- and surgically-relevant GSC models has been established. These include novel resected and residual disease models based on careful macrodissection of rare en-bloc partial lobectomy specimens to liberate parallel GSC lines from the tumour core and adjacent infiltrated brain, which represent cells removed and those typically left behind after surgery. Relative to resected disease models generated from the same patient, biological characterisation reveals residual disease models display increased connecting tumour microtube formation, stem marker expression and early evidence of elevated DDR marker expression including BRCA2, ATM and ATR. These are all consistent with enhanced therapeutic resistance in disease normally left behind after surgery.

Conclusion: Simultaneously targeting the FA pathway and interconnected DDR processes represents an appealing therapeutic strategy. Novel preclinical models which recapitulate spatiofunctional intratumoural heterogeneity have been developed and will support future evaluation of combinatorial strategies against recalcitrant post-surgical disease. Additionally, constitutive lack of FA pathway function in some tumours, could serve as a novel predictive biomarker for patient response to PARPi and ATRi which are currently in clinical trials.