Despite extensive research, little progress has been made in glioblastoma therapy, owing in part to a lack of adequate preclinical in vivo models to study this disease. To mitigate this, primary patient-derived cell lines, which maintain their specific stem-like phenotypes, have replaced established glioblastoma cell lines. However, due to heterogenous tumour growth inherent in glioblastoma, the use of primary cells for orthotopic in vivo studies often requires large experimental group sizes. Therefore, when using intracranial patient-derived xenograft (PDX) approaches, it is advantageous to deploy imaging techniques to monitor tumour growth and allow stratification of mice. Here we show that stable expression of near-infrared fluorescent protein (iRFP) in patient-derived glioblastoma cells enables rapid, direct non-invasive monitoring of tumour development without compromising tumour stemness or tumorigenicity. Moreover, as this approach does not depend on the use of agents like luciferin, which can cause variability due to changing bioavailability, it can be used for quantitative longitudinal monitoring of tumour growth. Notably, we show that this technique also allows quantitative assessment of tumour burden in highly invasive models spreading throughout the brain. Thus, iRFP transduction of primary patient-derived glioblastoma cells is a reliable, cost- and time-effective way to monitor heterogenous orthotopic PDX growth.

25Despite extensive research, little progress has been made in glioblastoma therapy, 26 owing in part to a lack of adequate preclinical in vivo models to study this disease. To 27 mitigate this, primary patient-derived cell lines, which maintain their specific stem-like 28 phenotypes, have replaced established glioblastoma cell lines. However, due to 29 heterogenous tumour growth inherent in glioblastoma, the use of primary cells for 30 orthotopic in vivo studies often requires large experimental group sizes. Therefore, 31 when using intracranial patient-derived xenograft (PDX) approaches, it is 32 advantageous to deploy imaging techniques to monitor tumour growth and allow 33 stratification of mice. Here we show that stable expression of near-infrared 34 fluorescent protein (iRFP) in patient-derived glioblastoma cells enables rapid direct 35 non-invasive monitoring of tumour development without compromising tumour 36 stemness or tumorigenicity. Moreover, as this approach does not depend on the use 37 of agents like luciferin, which can cause variability due to changing bioavailability, it 38 can be used for quantitative longitudinal monitoring of tumour growth. Notably, we 39 show that this technique also allows quantitative assessment of tumour burden in 40 highly invasive models spreading throughout the brain. Thus, iRFP transduction of 41 primary patient-derived glioblastoma cells is a reliable, cost-and time-effective way 42 to monitor heterogenous orthotopic PDX growth.68 xenografts (PDX) 7 . However, due to their intrinsic heterogeneity, orthotopic PDXs 69 display a much greater range of growth rates between animal subjects compared to 70 established glioblastoma cell lines 4,10 . Although this variability is a good reflection of 71 the heterogeneity of the parental tumour, differences in growth characteristics, and 72 therefore PDX size and the degree of dissemination at defined time points, lead to 73 difficulties in quantifying and interpreting animal studies. Thus, large cohorts of 74 animals are needed to achieve the necessary statistical power in these preclinical 75 studies. 76 77 Currently, intra-vital bioluminescent imaging (BLI) with luciferase is commonly 78 used to monitor xenograft growth 11 . Yet, in vivo luciferase assays are limited to 79 descriptive analysis since factors such as fluctuating substrate access to the target 128 129 Cell proliferation assay 130 GSC were seeded in three wells of a 6-well plate at a density of 5x10 4 cells 131 per well. Cell number for each well was assessed using a haemocytometer 1, 3 and 132 6 7 days after seeding. A median was calculated for each time point. Cell count was 133 normalised to day 1 (n=3 independent repeats per cell line). 135Neurosphere formation assay 136 GSC were seeded in uncoated 96-well plates at a density of 10 cells per well.was first described in 2011, it has been shown to be a powerful and reliable tool to 193 measure cell proliferation in vitro and in vivo 11,14 . We deployed a retroviral 194 transduction approach to express iRFP in...

Glioblastoma (GBM) is the most prevalent malignant primary brain tumour in adults. GBM typically has a poor prognosis, mainly due to a lack of effective treatment options leading to tumour persistence or recurrence. We investigated the therapeutic potential of targeting anti-apoptotic BCL-2 proteins in GBM. Levels of anti-apoptotic BCL-xL and MCL-1 were consistently increased in GBM compared with non-malignant cells and tissue. Moreover, we found that relative to their differentiated counterparts, patient-derived GBM stem-like cells also displayed higher expression of anti-apoptotic BCL-2 family members. High anti-apoptotic BCL-xL and MCL-1 expression correlated with heightened susceptibility of GBM to BCL-2 family protein-targeting BH3-mimetics. This is indicative of increased apoptotic priming. Indeed, GBM displayed an obligate requirement for MCL-1 expression in both tumour development and maintenance. Investigating this apoptotic sensitivity, we found that sequential inhibition of BCL-xL and MCL-1 led to robust anti-tumour responses in vivo, in the absence of overt toxicity. These data demonstrate that BCL-xL and MCL-1 pro-survival function is a fundamental prerequisite for GBM survival that can be therapeutically exploited by BH3-mimetics.