Generation and characterisation of the Sheffield Living Biobank (SLB) of post-surgical residual Glioblastoma to identify novel therapeutic targets.

Glioblastoma remains the most common and aggressive CNS malignancy diagnosed accounting for ~5,300 deaths in the UK annually, and around 200,000 globally each year. Additionally, there has been no significant universal improvement to disease outcome or survival since the implementation of the chemotherapeutic Temozolomide (TMZ) in 2005. Numerous biological blockades perturbing the development of novel effective therapies have been identified which include extensive inter-patient heterogeneity, the presence of the blood-brain barrier (BBB) hindering the bioavailability of novel therapeutics, the highly invasive nature of the disease hindering extent of resection (EOR) safely possible, an immune suppressive tumour microenvironment (TME) and the presence of a treatment resistant subpopulations of cells, termed glioma stem cells (GSCs), which have been indicated as pivotal contributors to the currently inevitable recurrence of the disease post-“Stupp” regimen treatment and the death of the patient (median survival of 12-15 months). Recent reports have also now highlighted extensive intratumoural heterogeneity between spatially divergent populations of the same tumour (to a similar degree as inter-patient variation) and a lack of pre-clinical models capable of recapitulating this heterogeneity ultimately allowing for full characterisation of the disease. Here we display a wealth of data characterising a novel biobank of GSC populations isolated from the necrotic core and distal invasive edge of large en-bloc specimens resected within the Royal Hallamshire Hospital, Sheffield. These samples contain large portions of infiltrated brain (not possible in all resections) which allow for direct characterisation of the typically residual GSC populations left behind after surgery. This PhD study focussed on the characterisation of the DNA damage response (DDR) within these populations as these repair cascades directly reverse the damage induced by both standard- of-care (SoC) therapies (radiotherapy and TMZ) employed in the treatment of GBM. Additionally, therapies targeting these cascades have displayed unparalleled efficacy in a range of other cancers. This led us to pose the hypothesis that “GSCs typically left behind after surgery are inherently treatment resistant due to increased expression of both GSC stem markers and the DDR”. Initial biomolecular characterisation of these Core (resected) and Edge (residual) GSC populations (propagated in clinically relevant 3D AlvetexTM architecture) revealed differential expression and activation of several DDR mediators both endogenously and in response to SoC therapies. Additionally, we also recorded differential expression of several GSC markers suggestive of differential cell composition within these populations reflective of the previously established “Verhaak” GSC cell states. Further investigation also revealed differential DNA repair kinetics and chromatin order between Core and Edge GSC populations. We therefore believe that the differences identified within this study contribute to the differential TMZ sensitivity identified within these heterogeneity models. Interestingly, we did not record substantial differences in IR sensitivity between Core and Edge populations, however, we did observe differences in radio-sensitisation capacity of the clinically relevant ATR inhibitor (AZD6738) and ATM inhibitor (AZD1390), the latter of which is currently under investigation in clinical trials for GBM (NCT03523628). Finally, bulk RNA-seq datasets generated from these Core and Edge GSC populations revealed stark differences in the expression of all DNA repair cascades and stem markers again suggestive of differential GSC cell states between resected and residual tumour cells. This data also highlighted potential distinct preferences for different DNA repair pathways between these stem cell states therefore highlighting multimodal DDR interruptions as a viable treatment strategy for targeting entire tumour GSC populations. It is imperative to remember, all of the differences identified within this study largely followed a locational signature, however, there was significant exceptions to this ideology even within our small sample size leading us to reject our initial hypothesis. In conclusion, the unprecedented insight these models have provided suggests differential DNA repair capacity resulting in differential treatment sensitivity to SoC therapies, with further characterisation of these models employing single-cell technologies and the strategies outlined within the future work aiding in elucidating the viability of novel treatment strategies ideally leading to improved disease outcome. This has led us to pose a revised hypothesis which states: “Individual “Verhaak” GSC subtypes (Classical, Proneural and Mesenchymal) display distinct DDR expression and activation signatures, which with further characterisation, presents an exciting potential therapeutic window for the development of novel therapeutics capable of targeting the entire GBM tumour.”