Patient-derived Glioma Stem Cell Organoids

Patients with glioblastoma (GM) have a median overall survival of approximately 15 months.Standard therapy for GM encompasses maximum surgical resection followed by radiation and chemotherapy using temozolomide (TMZ) (1). Regardless of initial tumor response, tumor recurrence is inevitable, after which survival drops to less than 6 months. GM tailored approaches targeting oncogenes that might drive the growth of the bulk of primary tumors, have been unsuccessful so far in clinical trials(2), creating a large unmet need warranting new approaches to overcome intrinsic and acquired resistance to current treatment schedules. The objective of this research is to establish primary patient derived organoid cultures from GM to study mechanisms that contribute to aggressive tumor growth and treatment resistance in primary and recurrent GM.

1. Inter-and intratumoral heterogeneity in glioblastoma. Tumor tailored approaches for GM are being hampered by inter-and intratumoral heterogeneity of both microenvironment and genomic alterations in GM cells. It has been shown that tumors are composed of multiple clones harboring distinct genetic alterations (3-7). The clonal evolution model posits that tumor formation is initiated in a cell of origin and is followed by the subsequent accumulation of multiple genetic and epigenetic alterations, leading to tumor cell survival and growth advantage (8). Divergent genetic alterations in early transformed cells give rise to a variety of clones under the selective pressure of the tumor microenvironment (3-7). An important microenvironmental stressor is intratumoral hypoxia, which is frequent in GM and a negative prognostic and predictive factor associated with reduced survival (9,10). Emerging evidence implicates a subpopulation of tumor cells with characteristics of normal stem cells so-called glioma stem cells (GSC) in intrinsic and acquired treatment resistance. GSC are endowed with specific properties including high tumor initiating ability, unlimited self-renewal potential and capacity for multipotent differentiation, generating a diverse progeny(11). GSC are marked by common stem cell markers including CD133+, SOX2, Olig1, and have been shown to reside in the perivascular region as well as in hypoxic areas. GSC are expanded under hypoxia12depend on glycolysis (13,14). Combined with their low proliferation, increased DNA repair, high anti-oxidant activity and among others, make GCS more resistant to conventional treatments (radiation and temozolomide) then non-GSC (15,16) {Jamal, 2012 #52}. This implies that GSCs form an important driver of GM recurrence after chemoradiation. There are currently no effective treatments to eliminate glioma stem cells. Blocking hypoxia signalling in tumors (inhibiting self-renewal and survival of GSC cells)(12,17) and blocking the NOTCH stem cell pathway (rendering GSCs sensitive to radiation(18) and TMZ(19-21)), seem to promising but drugs interfering with these pathways have not passed beyond early phase clinical trials yet(22). Current standard of care of newly diagnosed glioblastoma is multimodal and consists of surgery, radiation therapy and TMZ, an alkylating agent modifying purine bases of DNA (O6-guanine; N7-guanine and N3-adenine). The addition of TMZ to radiotherapy has increased the overall survival of GM patients significantly, but only up to 14.6 months(1). Intratumoral hypoxia has been shown to decrease therapeutic efficacy of RT and chemotherapy(23). Hypoxic GM cells are genetically unstable and show increased MGMT expression and thereby resistance to alkylating TMZ chemotherapy(24). In non-GSCs, MGMT promoter methylation is a predictive marker for response to TMZ treatment(25,26). However the interpretation of the MGMT methylation assay is complex, since the extent of MGMT promoter methylation in GM is heterogeneous and the level of heterogeneity is underestimated, since only GM biopsies or fragmented GM tissue is being analysed(27,28). Importantly, MGMT is also expressed in the normal brain endothelial cells and in immune cells including tumor infiltrating cells(29). Thus depending on the extent of normal tissue contamination in GM biopsies the level MGMT methylation may also differ. Since intra-tumoral heterogeneity in MGMT expression cannot be objectivated using currently available diagnostics and since undertreatment of patients should be prevented at all times, at present most patients diagnosed with ´de novo´ GM, receive TMZ; although there seems to be little benefit of TMZ in GMs with MGMT unmethylated glioma cells. Designing new treatment protocols for newly diagnosed glioblastoma is quite complex, since diagnostic tools, predicting inter-and intra-GM heterogeneity of MGMT promoter methylation status (or level of MGMT expression), are lacking. Additionally, the level of MGMT expression, needed for significant TMZ response and improvement of clinical outcome, is not known. Thus since most GMs are defined by both MGMT methylated and MGMT unmethylated tumor clones, combining TMZ with drugs that target non-methylated glioma cells/glioma stem cells and/or microenvironment might be necessary.
2. Glioblastoma recurrence Even if a macroscopically complete resection of a GM can be achieved, tumor cells remain at the resection site. It has been shown that GM cells have a high capacity for dissemination. Invading tumor cells escape at the periphery of the tumor mass and diffusely infiltrate the normal brain parenchyma. Deeply infiltrated tumor cells are more likely to escape surgery and what is not known is whether infiltration is a property of a more resilient cell population that initiates and drives tumor recurrence. Even after postoperative chemotherapy and radiation outside of the field of surgery to reduce infiltrative tumor cells, almost all GM recur, mostly around the resection cavity. If a gross total resection cannot be performed; primary radiation therapy and chemotherapy are also able to reduce clonal diversity, but are thus insufficient to prevent recurrence.

This implies that tumor cells resistant to multiple therapies persist in the brain parenchyma around the tumor cavity after gross total resection or in the remaining tumor after chemoradiation which are responsible for tumor repopulation, making them a critical target to overcome tumor recurrence. Genomic analysis of GM has shown that dominant clones in recurrent tumors are composed of clones representative of the primary tumor as well as new clones with little resemblance to the original tumor(3,30,31,32,33). Consequently targets identified based on the analysis of the primary tumor may not be informative to identify the best molecular targets to prevent recurrence(5).

After failure of first line GM treatment (radiation, TMZ), recurrence seems to be accompanied with a phenotypic shift of GSCs to the mesenchymal (MES) subtype (loss of CD133; gain of BMI1, SOX2 and CD44)(34-38). Such cells are more aggressive, invasive and angiogenic than primary tumor derived GSCs, mostly the proneural (PN) subtype (CD133+, CD15+)(39-41). MES GSC also show a higher expression of the inflammasome genes, such as IL6, IL8, IL1B1C and CXCL2, reinforcing the notion that interplay with the microenvironment and plays an important role in recurrence and progression(42,43). GM specific clinical trials are being developed, using immune modulators. It has been shown that MSH (mutS homologue) mutations correlate with TMZ resistance, as they are neither found in pre-treatment GM nor radiotherapy post-treatment GM, but were detected in approximately half of recurrent GM patients treated with TMZ and radiotherapy. This strongly indicates that MSH6 alterations are associated with resistance to alkylating agent therapy. Thus GM patients who initially respond to TMZ may acquire MMR defective hypermutator phenotype(44-46). It is well established that an intact mismatch repair and base excision repair (BER) contribute to effective TMZ cytotoxicity. Pharmacological inhibition of BER using PARP inhibitors either alone or in combination with TMZ has shown promise in clinical trials. However, acquired resistance to PARP inhibitors is observed through up-regulation of base excision repair and homologous recombination repair to compensate for diminished BER (47,48).

In conclusion, understanding the therapy induced genetic and epigenetic alterations of the remaining tumor cells and GSCs and also the impact of standard of care to the microenvironment /niche, in which recurrence occurs, will guide the development of new treatments.

In this project we will use patient derived glioma stem cell organoids(49), mimicking de novo GM and its intratumoral heterogeneity. Also these patient-derived organoids (PDO) organoids will be used to study acquired temozolomide resistance and thus can be used for identifying novel targeted agents.

Tumor cell heterogeneity in MGMT expression and its relevance for TMZ response will also be addressed using these organoids, using single cell proteomics and IHC for MGMT, GSC markers and exome sequencing (for common driver mutations) to identify those populations that clonally expand under TMZ (RT) selection and those that disappear.

Another interesting feature we will explore using PDO is the possibility to analyze the supernatants for circulating tumor DNA (ctDNA) or exosomes (secreted vesicles which contain RNA, small non-coding RNA, proteins as well as DNA) before during after acquired TMZ resistance which might lead to the identification of pharmacological response and predictive biomarkers. Such biomarkers 'liquid biopsy' may become useful to measure treatment response in blood or lumbar fluid of GM patients and enable dose-modification (escalation or deintensification or termination).

The development of the PDOs (costs of materials) will be funded (607061 PI M. Vooijs, MAASTRO and KWF grant Alpe D'Huzes PI M.Vooijs, MAASTRO. There are no additional costs associated with obtaining the tumor tissue.