Temozolomide (TMZ) is the first line of standard therapy in glioblastoma (GBM). However, relapse occurs due to TMZ resistance. We attempted to establish an acquired TMZ resistance model that recapitulates the TMZ resistance phenotype and the relevant gene signature. Two GBM cell lines received two cycles of TMZ (150 µM) treatment for 72 h each. Regrown cells (RG2) were defined as TMZ resistant cells. MTT assay revealed significantly less susceptibility and sustained growth of RG2 compared with parental cells after TMZ challenge. TMZ-induced DNA damage significantly decreased in 53BP1-foci reporter transduced-RG2 cells compared with parental cells, associated with downregulation of MSH2 and MSH6. Flow cytometry revealed reduced G2/M arrest, increased EdU incorporation and suppressed apoptosis in RG2 cells after TMZ treatment. Colony formation and neurosphere assay demonstrated enhanced clonogenicity and neurosphere formation capacity in RG2 cells, accompanied by upregulation of stem markers. Collectively, we established an acute TMZ resistance model that recapitulated key features of TMZ resistance involving impaired mismatch repair, redistribution of cell cycle phases, increased DNA replication, reduced apoptosis and enhanced self-renewal. Therefore, this model may serve as a promising research tool for studying mechanisms of TMZ resistance and for defining therapeutic approaches to GBM in the future.

Hyper-angiogenesis is a typical feature of glioblastoma (GBM), the most aggressive brain tumor. We have reported the expression of aldehyde dehydrogenase 1A3 (ALDH1A3) in proliferating vasculature in GBM patients. We hypothesized that ALDH1A3 may act as an angiogenesis promoter in GBM. Two GBM cell lines were lentivirally transduced with either ALDH1A3 (ox) or an empty vector (ev). The angiogenesis phenotype was studied in indirect and direct co-culture of endothelial cells (ECs) with oxGBM cells (oxGBMs) and in an angiogenesis model in vivo. Angiogenesis array was performed in oxGBMs. RT²-PCR, Western blot, and double-immunofluorescence staining were performed to confirm the expression of targets identified from the array. A significantly activated angiogenesis phenotype was observed in ECs indirectly and directly co-cultured with oxGBMs and in vivo. Overexpression of ALDH1A3 (oxALDH1A3) led to a marked upregulation of PAI-1 and IL-8 mRNA and protein and a consequential increased release of both proteins. Moreover, oxALDH1A3-induced angiogenesis was abolished by the treatment of the specific inhibitors, respectively, of PAI-1 and IL-8 receptors, CXCR1/2. This study defined ALDH1A3 as a novel angiogenesis promoter. oxALDH1A3 in GBM cells stimulated EC angiogenesis via paracrine upregulation of PAI-1 and IL-8, suggesting ALDH1A3-PAI-1/IL-8 as a novel signaling for future anti-angiogenesis therapy in GBM.

Glioblastoma (GBM) is an extremely invasive and incurable tumor. We previously reported predominant ALDH1A3 expression at the invasive front of GBM tumors, which was associated with shorter patient survival, and further showed that ALDH1A3 promoted tumor angiogenesis involving plasminogen activator inhibitor-1 (PAI-1). Here, we investigated whether ALDH1A3 drives cell invasion through retinoic acid (RA) and PAI-1 signaling. Analysis of the TCGA-GBM dataset revealed a positive association between ALDH1A3 and PAI-1 (SERPINE1) expression. Overexpression of ALDH1A3 in GBM cells markedly increased PAI-1 mRNA and protein levels, with cellular colocalization of both proteins, accompanied by robust migration and invasion. These effects were reversed by treatment with a pan-RA receptor (RAR) antagonist AGN193109 (AGN), with a specific PAI-1 inhibitor tiplaxtinin (Tip) or by CRISPR/Cas9-mediated knockout of PAI-1. In a chick chorioallantoic membrane (CAM) model, ALDH1A3-overexpressing cells showed increased invasion, which was reduced by tiplaxtinin (Tip) treatment or PAI-1 knockout. Mechanistically, ChIP-qPCR demonstrated that RA treatment or ALDH1A3 overexpression increased RARα occupancy at the PAI-1 regulatory region, accompanied by increased PAI-1 expression, both of which were diminished by AGN. Collectively, the present study defines an ALDH1A3-RA-PAI-1 signaling axis that contributes to GBM cell motility and invasion.

Glioblastoma (GBM) is characterized by extensive neo-angiogenesis, which drives rapid tumor growth and therapeutic resistance. We previously identified PDCD10 as a tumor suppressor in GBM. Here, we investigated whether PDCD10 loss promotes neo-angiogenesis through paracrine signaling and GBM cell plasticity. PDCD10 knockdown (shPDCD10) enhanced endothelial angiogenic activity after treatment with conditioned medium (CM) from shPDCD10 cells and in a direct co-culture model. Moreover, application of CM to the chicken chorioallantoic membrane increased vascular branching in an in vivo angiogenesis model. Antibody array analysis detected elevated levels of multiple pro-angiogenic factors following PDCD10 depletion. In a GBM mouse model, shPDCD10 tumors exhibited pronounced hypervascularity and stromal expansion. Indeed, immunofluorescence revealed colocalization of CD31 with the tumor cell reporter RFP in a subset of implanted shPDCD10 GBM cells, suggesting that these tumor cells acquired an endothelial molecular signature. PDCD10 loss also promoted a glioma stem cell (GSC)-like phenotype, characterized by enhanced clonogenicity, sphere formation, and upregulation of Nestin, KLF4, and SOX2. Under endothelial induction conditions, shPDCD10 sphere-derived cells exhibited endothelial-like characters, showing greater tube-forming capacity and increased Ac-LDL uptake. Taken together, these findings demonstrate that PDCD10 loss promotes GBM neo-angiogenesis involving complementary paracrine and GSC-like plasticity mechanisms, highlighting PDCD10 as a potential therapeutic target to suppress neo-angiogenesis in GBM.