Mesenchymal glioblastoma is a subtype of glioblastoma multiforme (GBM) characterized by pronounced inflammatory features and resistance to conventional therapies. Proneural GBM acquires a mesenchymal phenotype through proneural-mesenchymal transition (PMT), in which NF-κB signaling plays a central role. Through RNA-sequencing analysis of glioma-initiating cells (GICs), we found that expression of herpes virus entry mediator (HVEM, also known as tumor necrosis factor (TNF) receptor superfamily member 14 or TNFRSF14) is highly expressed in mesenchymal GBM cells. Functional analyses revealed that HVEM promotes GIC proliferation, neurosphere formation, and invasive capacity in vitro, and enhances tumor formation following intracranial transplantation of GICs in mice. Among the TNF superfamily ligands, APRIL (a proliferation-inducing ligand, also known as TNF superfamily 13 or TNFSF13) binds to HVEM and activates NF-κB signaling, thereby inducing a mesenchymal phenotype in GBM cells. To therapeutically target this pathway, we have generated nanobodies from camelid-derived heavy-chain-only antibodies against human HVEM. An anti-human HVEM nanobody significantly inhibited the invasion of mesenchymal GICs in organotypic cultures and suppressed tumor growth in a mouse xenograft model. Furthermore, HVEM expression contributed to resistance to anticancer drugs, which was relieved by knockout of HVEM expression in the mesenchymal GICs. Collectively, these findings suggest that beyond its known role in immune evasion through interaction with BTLA (B and T lymphocyte attenuator), activation of HVEM on GBM cells promotes invasion and proliferation of mesenchymal GICs. Thus, HVEM represents a promising therapeutic target for the treatment of mesenchymal GBM. ### Competing Interest Statement R.T., B.W., C.-H. H, and K.M. hold patents (WO 2020/138503 and WO 2025/033553) related to this work, and hold stocks in Mesenkia Therapeutics AB (Uppsala, Sweden). Swedish Cancer Society, 100452, 222363, 243486 Swedish Research Council, 2024-03002 Ministry of Education, Culture, Sports, Science, and Technology of Japan, JP17H06326 Japan Society for the Promotion of Science, JP15H05774, JP23H05486 Japan Agency for Medical Research and Development, JP21ck0106705, JP24ck0106914
Glioblastoma (GBM) is an aggressive and incurable brain tumor, with treatment resistance and recurrence posing persistent clinical challenges. The transcription factor SOX9 has recently emerged as a key regulator of therapy resistance. Still, no specific SOX9 drugs exist, likely due to the inherent difficulty of targeting transcription factors. Here, we delineate a critical role for SOX9 in driving resistance to standard-of-care radiation and temozolomide (TMZ) therapy. Using a panel of 35 patient-derived GBM cell lines of varying subgroups and MGMT status we defined TMZ sensitivities and demonstrated that SOX9 is selectively activated in resistant lines following cytotoxic stress. Elevated SOX9 expression correlated with poor patient survival by attenuating TMZ-induced DNA damage. CRISPR-based depletion of SOX9 sensitized GBM cells to radiation and TMZ, underscoring its functional importance in mediating resistance. To therapeutically exploit this vulnerability, we developed a SOX9-responsive suicide gene therapy. This system couples the Herpes Simplex Virus Thymidine Kinase (HSV-TK) gene to a SOX9-inducible enhancer, enabling selective ablation of SOX9-positive GBM cells upon ganciclovir (GCV) treatment. Delivery via AAV2 or AAV9 vectors effectively eliminated resistant tumors and significantly extended survival in both immunodeficient and immunocompetent orthotopic GBM mouse models. SOX9-directed gene therapy synergized with radiation to enhance DNA damage and even led to long term survival due to a reprogrammed tumor immune microenvironment. We establish SOX9 as a central driver of therapeutic resistance in GBM and present a first-in-class strategy to selectively target resistant tumor populations. Clinical translation of this strategy holds promise for improving outcomes in patients with SOX9-driven tumor relapse.
Glioblastoma (GBM) is a highly malignant brain tumor with extensive cellular heterogeneity and plasticity. Bone morphogenetic protein 4 (BMP4) has shown potential as a therapeutic agent by promoting differentiation, but its effects are complex and context dependent. While BMP4’s role in differentiation is well established, its impact on senescence remains unclear. This study investigates BMP4’s ability to induce senescence in GBM cells. Primary GBM cultures were treated with BMP4 and analyzed for senescence markers, including cell enlargement, p21 expression, senescence-related gene enrichment, and senescence-associated-β-galactosidase activity. A p21 knockout model was used to determine its role in BMP4-induced senescence, and sensitivity to the senolytic agent navitoclax was evaluated. BMP4 induced senescence in the GBM cultures, particularly in mesenchymal (MES)-like GBM cells with high baseline p21 levels. The knockout of p21 nearly abolished BMP4-induced senescence, maintaining cell size and proliferation. Furthermore, navitoclax effectively eliminated BMP4-induced senescent cells through apoptosis, while sparing cells with normal p21 expression. Our findings highlight BMP4 as an inducer of p21-dependent senescence in GBM, particularly in MES-like cells. This study clarifies BMP4’s dual roles in differentiation and senescence, emphasizing their context dependence. Given the strong link between MES-like cells and therapy resistance, their heightened susceptibility to senescence may aid in developing targeted therapies for GBM and potentially other cancers with similar cellular dynamics.
Bone morphogenetic protein 4 (BMP4) has emerged as a potential glioblastoma therapy due to its anti- proliferative effect via SOX2 downregulation and differentiation promotion. However, BMP4 responses vary across and within tumors. Our previous data indicate that BMP4 induces transition to a mesenchymal-like cell state. Mesenchymal transition is associated with therapy-resistance and tumor recurrence, as is senescence in cancer. In this study, we investigated BMP4’s potential to induce senescence in primary glioblastoma cells, including proneural- and mesenchymal-like clones derived from the same tumor. BMP4 treatment induced senescence-associated genes and phenotypic changes such as cell enlargement, senescence- associated-β-gal expression, lamin B1 downregulation, and elevated p21 levels. The most robust senescence induction was observed in the mesenchymal-like clone, compared to its proneural counterpart. Notably, mesenchymal-like cells displayed high basal levels of p21 and other senescence- associated markers, suggesting a convergence of mesenchymal and senescent traits. p21 knockout abolished BMP4-induced senescence, maintaining proliferation and cell size despite SOX2 downregulation. Additionally, senolytic treatment effectively eliminated senescent cells through apoptosis, thereby favoring survival of cells retaining normal p21 levels. Our findings demonstrate BMP4’s ability to induce p21-dependent senescence in glioblastoma, particularly in therapy-resistant mesenchymal-like cells. These insights provide potential therapeutic strategies targeting senescence pathways in this challenging disease. ### Competing Interest Statement The authors have declared no competing interest. * ### List of abbreviations BMP4: : bone morphogenetic protein 4 Cas9: : CRISPR-associated protein 9 CDK: : cyclin-dependent kinase CDKN1A/2A: : cyclin-dependent kinase inhibitor 1A/2A CL: : classical CRISPR: : clustered regularly interspaced short palindromic repeats EGF: : epidermal growth factor FGF: : fibroblast growth factor G1: : gap1 (cell cycle phase) G2: : gap 2 (cell cycle phase) GBM: : glioblastoma KO: : knockout MAPK: : mitogen-activated protein kinase MES: : mesenchymal OLIG2: : oligodendrocyte transcription factor 2 PI3K: : phosphatidylinositol 3 kinase PN: : proneural RB: : retinoblastoma transcriptional corepressor SA-b-gal: : senescence-associated beta-galactosidase SASP: : senescence-associated secretory profile SMAD: : suppressor of mothers against decapentaplegic SOX2: : SRY-box transcription factor 2 TGF-b: : transforming growth factor beta
The migration of neural progenitor cells (NPCs) to their final destination during development follows well-defined pathways, such as along blood vessels. Cells originating from the highly malignant tumor glioblastoma (GBM) seem to exploit similar routes for infiltrating the brain parenchyma. In this report, we have examined the migration of GBM cells using three-dimensional high-resolution confocal microscopy in brain tumors derived from eight different human GBM cell lines xenografted into immunodeficient mice. The primary invasion routes identified were long-distance migration along white matter tracts and local migration along blood vessels. We found that GBM cells in the majority of tumors (6 out of 8) did not exhibit association with blood vessels. These tumors, derived from low lamin A/C expressing GBM cells, were comparatively highly diffusive and invasive. Conversely, in 2 out of 8 tumors, we noted perivascular invasion and displacement of astrocyte end-feet. These tumors exhibited less diffusive migration, grew as solid tumors, and were distinguished by elevated expression of lamin A/C. We conclude that the migration pattern of glioblastoma is distinctly tumor cell-specific. Furthermore, the ability to invade the confined spaces within white matter tracts may necessitate low expression of lamin A/C, contributing to increased nuclear plasticity. This study highlights the role of GBM heterogeneity in driving the aggressive growth of glioblastoma.
‘Of all the processes that the student of pathology is privileged to study, none is as intriguing, fascinating, and perplexing as neoplasia.’ This statement, a quote from William Boyd’s Textbook of Pathology (1961), remains as thought-provoking today as when it was first written. Over the decades since Boyd penned these words, our understanding of the molecular mechanisms underlying tumorigenesis and the development of novel therapeutic approaches have advanced at a revolutionary pace. This special issue of UJMS on cancer exemplifies the remarkable progress in cancer research, reaffirming that Boyd’s perspective on the enigmatic nature of cancer endures.
Glioblastomas are aggressive brain tumors that are largely immunotherapy resistant. This is associated with immunosuppression and a dysfunctional tumor vasculature, which hinder T cell infiltration. LIGHT/TNFSF14 can induce high endothelial venules (HEVs) and tertiary lymphoid structures (TLS), suggesting that its therapeutic expression could promote T cell recruitment. Here, we use a brain endothelial cell-targeted adeno-associated viral (AAV) vector to express LIGHT in the glioma vasculature (AAV-LIGHT). We found that systemic AAV-LIGHT treatment induces tumor-associated HEVs and T cell-rich TLS, prolonging survival in αPD-1-resistant murine glioma. AAV-LIGHT treatment reduces T cell exhaustion and promotes TCF1+CD8+ stem-like T cells, which reside in TLS and intratumoral antigen-presenting niches. Tumor regression upon AAV-LIGHT therapy correlates with tumor-specific cytotoxic/memory T cell responses. Our work reveals that altering vascular phenotype through vessel-targeted expression of LIGHT promotes efficient anti-tumor T cell responses and prolongs survival in glioma. These findings have broader implications for treatment of other immunotherapy-resistant cancers.
Table S1. Enriched GO terms for commonly upregulated genes in responsive cell lines and U3017MG harboring the control vector upon BMP4 treatment, according to the Database for Annotation, Visualization and Integrated Discovery (DAVID).
Abstract The malignant primary brain tumor glioblastoma (GBM) has a dismal prognosis and current treatment options are insufficient to cure patients. The stem cell transcription factor SOX2 has been indicated as vital for GBM stem cell maintenance in vitro and GBM tumor formation in vivo, rendering it a putative therapeutic target for GBM. However, a few studies, including previous work from our lab, suggest a dichotomy in SOX2 dependence between different patients. To elucidate whether SOX2 independent GBMs exist and if so, how they are distinguished, we have investigated how altered SOX2 expression affects a panel of primary human GBM lines cultured under stem cell conditions. We found that elevated SOX2 expression inhibited proliferation in a dose-dependent manner in three out of four GBM cell lines. Global gene expression analysis showed that cells representing proliferation-inhibited and refractory cell lines end up on different ends of the developmental- to injury response GBM stem cell axis, were resistance to SOX2 overexpression was connected to an injury response expression profile. CRISPR/Cas9 mediated SOX2 knockout revealed a SOX2 independence in the refractory cell line, where cells lacking SOX2 could be propagated both in vitro and in vivo, implicating SOX2 as a non-essential gene for injury response-driven GBM stem cells. In current work, SOX2-dependent cell lines are engineered to express inducible exogenous SOX2 with concomitant endogenous SOX2 knockout, providing a system where SOX2 can be turned off at any time point. This system will be used in further studies elucidating the nature of SOX2 dependence and independence in GBM stem cells.
There is an urgent need for simple and non-invasive identification of live neural stem/progenitor cells (NSPCs) in the developing and adult brain as well as in disease, such as in brain tumors, due to the potential clinical importance in prognosis, diagnosis, and treatment of diseases of the nervous system. Here, we report a luminescent conjugated oligothiophene (LCO), named p-HTMI, for non-invasive and non-amplified real-time detection of live human patient-derived glioblastoma (GBM) stem cell-like cells and NSPCs. While p-HTMI stained only a small fraction of other cell types investigated, the mere addition of p-HTMI to the cell culture resulted in efficient detection of NSPCs or GBM cells from rodents and humans within minutes. p-HTMI is functionalized with a methylated imidazole moiety resembling the side chain of histidine/histamine, and non-methylated analogues were not functional. Cell sorting experiments of human GBM cells demonstrated that p-HTMI labeled the same cell population as CD271, a proposed marker for stem cell-like cells and rapidly migrating cells in glioblastoma. Our results suggest that the LCO p-HTMI is a versatile tool for immediate and selective detection of neural and glioma stem and progenitor cells.
Bone morphogenetic protein 4 (BMP4) was initially suggested as a potential differentiation-inducing factor to be used in the therapy of glioblastoma. We and others have however demonstrated that the response is reversible, variable among patient samples, and heterogeneous within the same cell line. To deepen our knowledge on how BMP4 affects different types of glioblastoma cells, we treated phenotypically different clones from the same patient tumor—a multitherapy-sensitive/proneural-like (SENS/PN) clone and a multitherapy-resistant/mesenchymal-like (RES/MES) clone—with BMP4. In the SENS/PN clone, BMP4 turned on a mesenchymal-related gene program, whereas this response was less prominent in the RES/MES clone. Untreated SENS/PN cells were smaller than RES/MES cells, but both clones responded to BMP4 by cell size enlargement. Increase in cell size has been suggested to precede senescence; young cells are smaller than old cells that eventually enter replicative senescence. BMP4 induced a senescence-like phenotype in a subpopulation of cells, demonstrated by induction of senescence-associated (SA)-β-gal, p21 up-regulation, lamin B1 down-regulation, as well as increased lysosomal mass and granularity. This was more pronounced in the RES/MES clone than in the SENS/PN clone, and it was dependent on canonical SMAD signaling. Senolytic treatment ablated the SA-β-gal positive cells and reduced the p21 level at the population level. Targeted deletion of p21 abolished BMP4-induced SA-β-gal and increase in cell growth, while lamin B1 down-regulation remained, demonstrating that p21 signaling is crucial for one part of the senescence induction by BMP4. We are currently further investigating a connection between cell size, mesenchymality and senescence. We hypothesize that large mesenchymal-like cells are closer to senescence than smaller proneural cells within the cell culture. A combination of senescence-induction and senolytic treatment may open treatment opportunities to target therapy-resistant glioblastoma cells.
Glioblastoma (GBM) is an aggressive nervous system tumor with a mean survival time of 12-14 months. Cells with neural stem cell-like properties can be derived from GBM tumors. These cells seem to escape conventional irradiation treatment, chemotherapy, and surgery, and may play a crucial role for relapse. It is therefore urgent to develop novel approaches for reliable detection of neural stem cell-like cells in GBM. Here we report a luminescent conjugated oligothiophene (LCO), named GlioStem (p-HTMI), for non-invasive and non-amplified real-time detection of live human patient-derived GBM cells and embryonic neural stem/progenitor cells (NSPCs). Within a maximum of 10 minutes after administration of the molecule in vitro, in the existing media, fluorescence emission was observed without any modulation of the cells or additional vehicle, resulting in efficient detection of cytoplasmic luminescent signal in NSPCs or GBM cells from rodents and humans, detectable at Alexa488/GFP wavelength. GlioStem is functionalized with a methylated imidazole moiety resembling the side chain of histidine/histamine, and non-methylated analogues were not functional. In vitro, GlioStem was shown to identify fetal cortical NSPCs from rat (FGF2-expanded), embryonic stem cell-derived NSPCs from mouse (FGF2/EGF-expanded), and FGF2-exposed C6 glioma cell cultures from rat, but not any other cell types investigated. Cell sorting experiments of patient-derived, FGF2/EGF-expanded GBM cells demonstrated that GlioStem in addition to NSPC-markers like Nestin and Sox2 labeled the same population (overlap > 90%) of cells as CD271, a proposed marker for stem cell-like cells and rapidly migrating cells in glioblastoma. Our results suggest that the LCO GlioStem is a versatile tool for immediate and selective detection of subpopulations of neural stem and glioma cells.
Glioma‐initiating cells (GICs), a major source of glioblastoma recurrence, are characterized by the expression of neural stem cell markers and the ability to grow by forming nonadherent spheres under serum‐free conditions. Bone morphogenetic proteins (BMPs), members of the transforming growth factor‐β family, induce differentiation of GICs and suppress their tumorigenicity. However, the mechanisms underlying the BMP‐induced loss of GIC stemness have not been fully elucidated. Here, we show that paired related homeobox 1 (PRRX1) induced by BMPs decreases the CD133‐positive GIC population and inhibits tumorigenic activity of GICs in vivo. Of the two splice isoforms of PRRX1, the longer isoform, pmx‐1b, but not the shorter isoform, pmx‐1a, induces GIC differentiation. Upon BMP stimulation, pmx‐1b interacts with the DNA methyltransferase DNMT3A and induces promoter methylation of the PROM1 gene encoding CD133. Silencing DNMT3A maintains PROM1 expression and increases the CD133‐positive GIC population. Thus, pmx‐1b promotes loss of stem cell‐like properties of GICs through region‐specific epigenetic regulation of CD133 expression by recruiting DNMT3A, which is associated with decreased tumorigenicity of GICs.
Abstract Bone morphogenetic protein 4 (BMP4) was initially suggested as a potential differentiation-inducing factor to be used in the therapy of glioblastoma. We and others have however demonstrated that the response is reversible, variable among patient samples, and heterogeneous within the same cell line. To deepen our knowledge on how BMP4 affects different types of glioblastoma cells, we treated phenotypically different clones from the same patient tumor—a multitherapy-sensitive/proneural-like (SENS/PN) clone and a multitherapy-resistant/mesenchymal-like (RES/MES) clone—with BMP4. In the SENS/PN clone, BMP4 turned on a mesenchymal-related gene program, whereas this response was less prominent in the RES/MES clone. Untreated SENS/PN cells were smaller than RES/MES cells, but both clones responded to BMP4 by cell size enlargement. Increase in cell size has been suggested to precede senescence; young cells are smaller than old cells that eventually enter replicative senescence. BMP4 induced a senescence-like phenotype in a subpopulation of cells, demonstrated by induction of senescence-associated (SA)-β-gal, p21 up-regulation, lamin B1 down-regulation, as well as increased lysosomal mass and granularity. This was more pronounced in the RES/MES clone than in the SENS/PN clone, and it was dependent on canonical SMAD signaling. Senolytic treatment ablated the SA-β-gal positive cells and reduced the p21 level at the population level. Targeted deletion of p21 abolished BMP4-induced SA-β-gal and increase in cell growth, while lamin B1 down-regulation remained, demonstrating that p21 signaling is crucial for one part of the senescence induction by BMP4. We are currently further investigating a connection between cell size, mesenchymality and senescence. We hypothesize that large mesenchymal-like cells are closer to senescence than smaller proneural cells within the cell culture. A combination of senescence-induction and senolytic treatment may open treatment opportunities to target therapy-resistant glioblastoma cells.
Tumor cell heterogeneity is a crucial characteristic of malignant brain tumors and underpins phenomena such as therapy resistance and tumor recurrence. Advances in single-cell analysis have enabled the delineation of distinct cellular states of brain tumor cells, but the time-dependent changes in such states remain poorly understood. Here, we construct quantitative models of the time-dependent transcriptional variation of patient-derived glioblastoma (GBM) cells. We build the models by sampling and profiling barcoded GBM cells and their progeny over the course of 3 weeks and by fitting a mathematical model to estimate changes in GBM cell states and their growth rates. Our model suggests a hierarchical yet plastic organization of GBM, where the rates and patterns of cell state switching are partly patient-specific. Therapeutic interventions produce complex dynamic effects, including inhibition of specific states and altered differentiation. Our method provides a general strategy to uncover time-dependent changes in cancer cells and offers a way to evaluate and predict how therapy affects cell state composition.
The brain vasculature has several specific features, one of them being the blood-brain barrier (BBB), which supports and protects the brain by allowing for the passage of oxygen and nutrients, while at the same time preventing passage of pathogens and toxins. The BBB also prevents efficient delivery of drugs to the brain, e.g. for treatment of brain tumors. In the murine brain, perivascular fibroblasts were recently identified as a novel potential constituent of the BBB. Here we present the existence of human cells that could be the equivalent to the murine brain perivascular fibroblasts. Using RNA sequencing, we show a similar transcriptomic profile of cultured human brain cells and murine perivascular fibroblasts. These data open up a window for new hypotheses on cell types involved in human CNS diseases.
Abstract Glioblastoma multiforme continues to have a dismal prognosis. Even though detailed information on the genetic aberrations in cell signaling and cell-cycle checkpoint control is available, no effective targeted treatment has been developed. Despite the advanced molecular defects, glioblastoma cells may have remnants of normal growth-inhibitory pathways, such as the bone morphogenetic protein (BMP) signaling pathway. We have evaluated the growth-inhibitory effect of BMP4 across a broad spectrum of patient samples, using a panel of 40 human glioblastoma initiating cell (GIC) cultures. A wide range of responsiveness was observed. BMP4 sensitivity was positively correlated with a proneural mRNA expression profile, high SOX2 activity, and BMP4-dependent upregulation of genes associated with inhibition of the MAPK pathway, as demonstrated by gene set enrichment analysis. BMP4 response in sensitive cells was mediated by the canonical BMP receptor pathway involving SMAD1/5/9 phosphorylation and SMAD4 expression. SOX2 was consistently downregulated in BMP4-treated cells. Forced expression of SOX2 attenuated the BMP4 sensitivity including a reduced upregulation of MAPK-inhibitory genes, implying a functional relationship between SOX2 downregulation and sensitivity. The results show an extensive heterogeneity in BMP4 responsiveness among GICs and identify a BMP4-sensitive subgroup, in which SOX2 is a mediator of the response. Implications: Development of agonists targeting the BMP signaling pathway in glioblastoma is an attractive avenue toward a better treatment. Our study may help find biomarkers that predict the outcome of such treatment and enable stratification of patients.
Glioblastoma multiforme (GBM) is the most common primary malignant brain tumor in adults. Patients usually undergo surgery followed by aggressive radio- and chemotherapy with the alkylating agent temozolomide (TMZ). Still, median survival is only 12–15 months after diagnosis. Many human cancers including GBMs demonstrate addiction to MYC transcription factor signaling and can become susceptible to inhibition of MYC downstream genes. JQ1 is an effective inhibitor of BET Bromodomains, a class of epigenetic readers regulating expression of downstream MYC targets. Here, we show that BET inhibition decreases viability of patient-derived GBM cell lines. We propose a distinct expression signature of MYCN-elevated GBM cells that correlates with significant sensitivity to BET inhibition. In tumors showing JQ1 sensitivity, we found enrichment of pathways regulating cell cycle, DNA damage response and repair. As DNA repair leads to acquired chemoresistance to TMZ, JQ1 treatment in combination with TMZ synergistically inhibited proliferation of MYCN-elevated cells. Bioinformatic analyses further showed that the expression of MYCN correlates with Aurora Kinase A levels and Aurora Kinase inhibitors indeed showed synergistic efficacy in combination with BET inhibition. Collectively, our data suggest that BET inhibitors could potentiate the efficacy of either TMZ or Aurora Kinase inhibitors in GBM treatment.