Glioblastoma is a very aggressive form of brain tumor particularly resistant to the standard therapies which include maximal surgical resection, followed by combined treatment with radiotherapy and chemotherapy [
6]. Alkylating agents, such as TMZ and CARB, and topoisomerase inhibitors, such as ETO and IRI, although effectively improve clinical outcomes when used alone or in combination with radiotherapy, display several adverse effects [
6] and their administration has not significantly changed the survival for GBM over the last years, making chemoresistance one of the biggest problems for GBM therapy [
37]. Thus, it is evident the need to identify new therapeutic strategies that can increase the life spans of patients affected by GBM.
Glioblastoma stem-like cells (GSCs) represent a subpopulation within the heterogeneous tumor mass of GBM, with a high similarity with neural stem cells [
6,
38]. They are characterized by elevated proliferative rate and tumorigenic capability in vivo [
39,
40] and are thought to be responsible for the resistance to standard therapies [
41,
42]. It has been demonstrated that the peritumor tissue is the site of tumor recurrence in 90% of the patients [
19]. This area shows complex changes such as edema, increased vascularization, abnormal gene expression and presence of numerous specialized cell types [
13‐
18], and is also the site in which a subpopulation of cancer stem-like cells (PCSCs) has been found [
12]. Although both GCSCs and PCSCs express stem cell markers, they have different characteristics in terms of self-renewal and tumorigenicity, suggesting that PCSCs may have high relevance in translational research [
12].
In this study, we investigated the response of IDH1-wildtype GCSCs and PCSCs derived from six patients affected by GBM to different chemotherapeutic drugs and verified the possibility to enhance their effect through the combined treatment with adjuvant molecules. The comparison of GCSCs and PCSCs behavior and the identification of the molecules involved in their differential response to the treatments may provide a further insight in the complexity of GBM-neighboring microenvironment, which plays a crucial role in tumor progression.
Since hematopoietic cells represent the primary target of chemotherapy-related adverse effects, here we first analyzed the efficacy of different chemotherapeutic drugs (TMZ, ETO, IRI and CARB) in Jurkat cells (Additional file
2: Figure S1A). As expected, all the tested antineoplastic agents dramatically decrease the proliferation of these cells. In particular, BrdU assay demonstrates that IRI and CARB have the stronger effect in decreasing proliferation while TMZ is less efficient. Western blot analysis of PCNA expression confirms that among the used compounds, TMZ exerts the lower anti-proliferative activity (Additional file
2: Figure S1B and C). This effect seems to be related to the activation of apoptotic pathway since all the chemotherapeutic drugs increased the activity of both pro- and effector-caspases (Additional file
3: Figure S2). We then investigated the effects of the different chemotherapeutic treatments on the proliferation rate of GCSCs and PCSCs. Both cell populations resulted more resistant than Jurkat cells to the different treatments and we noted that all the chemotherapeutic agents exerted significant effects only in the cells with higher values of BrdU incorporation (GCSCs and PCSCs derived from patients #1, #2 and #3), whereas none of the drugs affected significantly the cells with a lower proliferation rate (GCSCs and PCSCs derived from patients #4, #5 and #6). Moreover, although no significant differences have been found between GCSCs and PCSCs, TMZ is the drug with the lower efficacy in decreasing the proliferation of neurosphere clones derived from GBM of all the six patients. TMZ is a small molecule that is readily absorbed in the digestive tract and, because of its lipophilia, it is able to cross the blood–brain barrier. TMZ is the most widely chemotherapeutic drug used in patients with GBM, although the majority of the patients demonstrate de novo or acquired resistance, with subsequent tumor progression [
5,
43]. Thus, the identification of the mechanisms of resistance and the attempting to enhance its effect can represent a good therapeutic strategy. MGMT repairs cytotoxic DNA lesions generated by TMZ. Many studies have shown a mechanistic link between MGMT activity and TMZ resistance, with suppression of MGMT activity resulting in increased cytotoxicity and MGMT protein overexpression which lead to resistance [
31]. However, the role of MGMT in the evolution of acquired resistance is not well established as demonstrated by several papers revealing that MGMT expression is not always related to resistance to TMZ treatment [
44‐
46]. Since our results demonstrate that TMZ is the less efficient chemotherapeutic agent in decreasing the proliferation of both GCSCs and PCSCs, we have investigated the possibility of increasing its anti-tumor activity by means of the combined use with LEV. LEV is a relatively new non-enzyme inducing AED strongly recommended as a first line drug for patients with brain tumors [
47]. Recently, a growing body of evidence suggests that selected AEDs could lead to significant pharmaco-epigenetic interactions. It has been reported that LEV could act by favoring the recruitment of inhibitory complex, including HDAC, on the MGMT promoter, thus reducing its transcription [
25]. Here we show that in the GCSCs the combined treatment with LEV and TMZ decreases the MGMT expression levels and induces the nuclear translocation of HDAC4, suggesting, in agreement to what reported by Bobustuc et al. [
25], an HDAC4-dependent inhibitory role for MGMT transcription, thus increasing the sensitivity of these cells to TMZ treatment. In contrast, HDAC4 is expressed, although at low levels, in PCSCs, but none of the treatments modulated its expression and cellular localization. As GSCs show intrinsic deregulation in apoptotic cell death, we investigated whether in our system, the decreased proliferation rate observed in the presence of the combined treatment with LEV and TMZ could be associated with the activation of apoptotic pathway. Here we show an increased caspase 3 nuclear accumulation in GCSCs, while no change of its expression was observed in PCSCs. This result is supported by the analysis of pro-and effector-caspase expression evaluated in Jurkat cells after exposure to different chemotherapeutic agents showing that all the treatments significantly induced high levels of the activity of both classes of caspase compared to untreated cells (Additional file
3: Figure S2). All these data suggest that activation of apoptotic pathway is involved in the strong anti-proliferative effect exerted by the used antineoplastic drugs.
Taken together our results demonstrate that LEV enhances the TMZ effect on GCSCs by HDAC4-dependent downregulation of MGMT and by the activation of apoptotic pathways. PCSCs seem to be more resistant to the treatment, suggesting that the peritumoral microenvironment can favor the activation of survival mechanisms that make this therapeutic approach less effective. Our results are supported by data reported by Kim et al. demonstrating that the median progression-free survival and overall survival for patients who received LEV in combination with TMZ were significative longer than those for patients who did not receive LEV [
48]. In addition, a case report was published by Peddi et al. where a continuous regression of GBM was noted in a patient who received LEV and Dexamethasone without any cancer-targeted therapy, suggesting that the response may be related to Dexamethasone and/or LEV treatment [
49]. Although more studies are needed to better evaluate the role of LEV and its molecular mechanism of action, these papers together with our results strongly suggest the beneficial effect of LEV as a chemosensitizer agent.