The WYZ-1 mouse glioblastoma stem cell model: A novel preclinical tool for tumor and therapy studies Authors: Qingfeng Xue, Juan Ye, Samik Chakraborty, Xueyu Sang, Mitchell Sun, Yaping Zhang, Shuran Chen, Wanrun Lin, Jeeva Munasinghe, Chunzhang Yang, Herui Wang, Zhengping Zhuang Published: 2026 ABSTRACT We established a novel syngeneic glioblastoma stem cell (GSC) line from PDGFB-driven RCAS/tv-a glioblastomas and subsequently generated a Pten-deficient derivative, WYZ-1. In vitro, WYZ-1 cells exhibited robust self-renewal, high expression of GSC markers (CD133, Nestin, SOX2), and multipotent differentiation into neuronal and astrocytic lineages. In vivo, intracranial implantation in immunocompetent mice produced highly infiltrative and vascularized tumors with marked proliferation, necrosis, and white matter invasion, closely recapitulating key histopathological features of human glioblastoma. Immunohistochemical analysis confirmed the retention of oncogenic, stem-like, and mesenchymal-associated markers. Importantly, WYZ-1 tumors displayed an immune-excluded (“cold”) tumor microenvironment characterized by dense infiltration of immunosuppressive macrophages/microglial cells within the tumor core and restricted localization of exhausted T cells to the tumor margins. Consistent with these features, WYZ-1 tumors are resistant to temozolomide and anti-PD-1 monotherapy, with only a modest survival benefit observed following CTLA-4 blockade. Collectively, these findings establish WYZ-1 as a highly aggressive, stem-like, and immunocompetent glioblastoma model that mirrors the therapeutic resistance of human disease and provides a valuable platform for investigating tumor biology and evaluating novel immunotherapeutic strategies. Introduction Glioblastoma (GBM) remains the most aggressive primary malignant brain tumor in adults, with a median survival of approximately 14 months despite maximal surgical resection followed by radiation and temozolomide chemotherapy [1,2]. A primary factor contributing to poor outcomes is the highly immunosuppressive tumor microenvironment (TME), which confers resistance to checkpoint inhibitors and severely limits therapeutic efficacy [3]. The development of effective immunotherapies has been hindered by the scarcity of physiologically relevant syngeneic GBM models in immunocompetent mice, which are essential for accurately evaluating immune responses and treatment efficacy [4]. While numerous preclinical GBM models have been established, they all exhibit critical shortcomings. Patient-derived xenografts and human GSC lines require implantation into immunodeficient mice, precluding meaningful assessment of immunotherapies and often exhibiting variable engraftment efficiency and substantial inter-animal heterogeneity, which affects reproducibility [5]. Commonly used syngeneic GBM models, such as GL261 and SB28, also incompletely represent human disease. GL261 is highly immunogenic, with a high mutational burden and strong responsiveness to checkpoint blockade, resulting in exaggerated therapeutic effects. Conversely, SB28 more closely recapitulates the immunosuppressive microenvironment of GBM, characterized by limited mutational diversity, poor T-cell infiltration, and resistance to immunotherapy. Neither model recapitulates hallmark features of human GBM, including diffuse infiltration and cellular heterogeneity [6, 7]. Increasing evidence further suggests that stem-like, developmentally plastic cellular states play critical roles in gliomagenesis and tumor persistence [8,9]. GBM contains a population of glioblastoma stem-like cells (GSCs) that drive tumor initiation, therapy resistance, and recurrence [10,11]. These cells display self-renewal, multilineage differentiation, and immunoevasive properties, making them essential targets for understanding tumor biology and therapeutic response [12]. While several patient-derived GSC lines exist, their reliance on immunodeficient hosts restricts their utility for immunotherapy research [13,14]. A well-characterized, syngeneic GSC line would therefore fill a major gap by enabling investigation of tumor-immune interactions and therapeutic resistance mechanisms in an immunocompetent setting. The RCAS-TVA system, a traditional syngeneic GSC mouse model, enables somatic gene transfer through Replication-Competent Avian Sarcoma-Leukosis virus (RCAS) vectors delivered to cells expressing the TVA receptor, allowing for the modeling of gliomagenesis in an immunocompetent host [15]. Although powerful for studying tumor initiation and progression, the RCAS-TVA system is limited by the need for repeated stereotactic injections, variable tumor penetrance, and challenges in performing long-term or high-throughput studies, particularly because tumor-bearing mice often decline rapidly [16]. Establishing cell lines derived from RCAS-TVA tumors would provide renewable, genetically defined, and immunocompetent tools for mechanistic and translational research. In this study, we established a novel syngeneic glioblastoma stem cell line derived from tumors generated in the PDGFB-driven RCAS-TVA glioma model that recapitulates key molecular and histopathological features of human GBM [4,17]. The resulting line demonstrates hallmark GSC properties, including sustained self-renewal, expression of CD133 and Nestin, and multipotent differentiation into astrocytic and neuronal lineages. To more accurately model Pten-deficient GBM, we generated a Pten-knockout derivative, hereafter referred to as WYZ-1. When implanted intracranially, WYZ-1 cells form highly infiltrative gliomas in immunocompetent mice, closely mirroring the invasive behavior and treatment resistance characteristic of human disease. The establishment of WYZ-1 provides a robust platform for mechanistic, immunological, and translational studies of GBM. Materials and methods Cell lines and materials: The Ntv-a, Ink4a-Arf −/−, LPTEN mice were kindly provided by Eric C. Holland [16], and tumor induction was performed as previously described [18]. The RCAS-TVA-derived glioma stem cell line (RCAS-TVA-GSC01; RTG01) was generated from the tumor tissue isolated from an RCAS-TVA mouse, dissociated using the Mouse Tumor Dissociation Kit and GentleMACS Dissociator (Miltenyi Biotec), and cultured in DMEM/F12 (Gibco) supplemented with 2% B27 (Invitrogen), 20 ng/mL bFGF, and EGF (R&D Systems) for ten passages to establish a stable line. RTG01 cells were then infected with Cre-expressing lentivirus (Creative Biogene) to delete Pten, followed by puromycin selection to generate the stable Pten-knockout line WYZ-1. Accutase (Sigma-Aldrich) was used for routine passaging, and the cells were confirmed mycoplasma-free using the MycoAlert® Mycoplasma Detection Kit (Lonza, LT07–218). Animals: Healthy female C57BL/6 mice (6–8 weeks old) were obtained from Charles River Laboratories (Wilmington, MA). All animal experiments were conducted under the approval of the NCI Animal Care and Use Committee (ACUC), following federal regulatory guidelines and standards. The NCI-CCR-affiliated staff carried out all procedures in compliance with these regulations. The intramural NIH Animal Care and Use (ACU) program is fully accredited by AAALAC International (Frederick, MD). Syngeneic GBM Models: Syngeneic GBM models were generated using 6–8-week-old female C57BL/6 mice. WYZ-1 cells (30,000 cells in 2 μL HBSS; Crystalgen) were stereotactically injected into the right striatum at the following coordinates: 0 mm posterior to bregma, 3 mm lateral to midline, and 3.5 mm depth. Mice were monitored every three days after implantation and daily beginning two weeks post-injection to assess tumor progression and overall health. Glioma-bearing mice were euthanized at endpoint, and brains were collected, fixed in 10% neutral buffered formalin for at least 48 h, and paraffin-embedded. Tissue blocks were sectioned at 4 μm for downstream histological analyses. Genotyping PCR: Genotyping PCR reactions were set up with 2x Hieff PCR Master Mix (Arcegen, 10102ES08). The sequences of primers are listed in Table 1. PCR products were analyzed via 1–2% agarose gel electrophoresis in 1× TAE buffer at 120 V for 20 min, visualized with SYBR Safe, and imaged. Western blot: Lysates were prepared in RIPA buffer (ThermoFisher Scientific) containing 1 mM PMSF (Cell Signaling Technology) and 1x phosphatase inhibitor (Cell Signaling Technology), then incubated on ice for 30 min. Following quantification via BCA assay (ThermoFisher Scientific), samples were normalized, mixed with 4× loading buffer, and denatured at 98°C for 10 min. Samples were separated by electrophoresis on NuPAGE 4–12% Bis-Tris gels (Invitrogen) and transferred to nitrocellulose membranes, which were then blocked with 5% non-fat milk. Primary antibodies: PTEN (Cell Signaling Technology, #9559) and HA (Cell Signaling Technology, #3724). Secondary antibodies: Anti-Rabbit IgG HRP-Linked antibody (Cell Signaling Technology, #7074). IHC staining: Paraffin-embedded tissue sections were deparaffinized, rehydrated, and subjected to heat-induced antigen retrieval in citrate buffer (pH 6.0) (Sigma, #C9999) or Tris-EDTA (pH 9.0) (Abcam, ab93684), as appropriate for each antibody. Endogenous peroxidase activity was quenched with 3% H2O2, and non-specific binding was blocked with 4% goat serum. Sections were incubated with primary antibodies for 2 h at room temperature or overnight at 4°C, followed by MACH4 HRP-Polymer and ABC reagent. Immunoreactivity was visualized using DAB substrate (Sigma), and sections were counterstained with hematoxylin, dehydrated, cleared, and mounted with Permount (Fisher Chemical, SP15–100). The following primary antibodies were used: Ki67 (ThermoFisher Scientific, MA5–14520), p-STAT3 (Cell Signaling Technology, #9145), Cyclin D1 (Cell Signaling Technology, #2978), Integrin β1 (Cell Signaling Technology, D6S1W), PDGF Receptor alpha (Cell Signaling Technology, #3174), Vimentin (Cell Signaling Technology, #5741), F4/80 (Cell Signaling Technology, #70076), CD31 (Cell Signaling Technology, #77699), Nestin (Abcam, ab316018), GFAP (Abcam, ab7260), CD4 (Abcam, ab183685), CD8 (Abcam, ab209775), Olig2 (Abcam, ab109186), CD44 (Cell Signaling Technology, #37259), Fibronectin (Cell Signaling Technology, #63779), ZEB1 (Cell Signaling Technology, #70512). Multiplex Immunohistochemistry: For spatial microenvironmental phenotyping, paraffin-embedded tissue sections were subjected to multiplex immunohistochemistry (mIHC). Following deparaffinization in xylene and rehydration through a graded ethanol series, slides were heated in a microwave with citrate buffer (pH 6.0; Sigma-Aldrich, #C9999) for antigen retrieval. The slides were then incubated with primary antibodies overnight at 4°C, followed by incubation with SignalStain® Boost IHC Detection Reagent (HRP, Rabbit; Cell Signaling Technology, #8114) for 30 min at room temperature. Subsequently, signal amplification was performed for 10 min at room temperature using Opal dyes (Akoya Biosciences). Sequential rounds of antigen retrieval, antibody staining and signal amplification on the same tissue sections were performed, enabling spatial characterization of multiple immune markers within the WYZ-1 tumor microenvironment. Finally, after DAPI counterstaining, the slides were mounted using an anti-fade mounting medium before imaging. Detailed antibody and dye specifications are provided in Table 2. IF staining: RTG01 or WYZ-1 cells (2000 cells per well) were seeded onto 8-chamber slides and cultured overnight. Cells were fixed with 4% paraformaldehyde (PFA), permeabilized with Triton X-100 for intracellular staining, and blocked with 2% goat serum. Slides were incubated with Nestin (Cell Signaling Technology, #89529), Sox2 (Cell Signaling Technology, #4900), β3-tubulin (Cell Signaling Technology, #5568), and GFAP (Cell Signaling Technology, #3670) overnight at 4°C, followed by Goat anti-Mouse IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 488 (ThermoFisher Scientific, A-11001), Goat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 594 (ThermoFisher Scientific, A-11012), Goat anti-Rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 488 (ThermoFisher Scientific, A-11008), and Goat anti-Mouse IgG (H + L) Cross-Adsorbed Secondary Antibody, Alexa Fluor™ 647 (ThermoFisher Scientific, A-21235) for 1 h at room temperature. Nuclei were counterstained with DAPI, and coverslips were mounted using an anti-fade mounting medium before imaging. Gene Set Variation Analysis (GSVA): Single-sample pathway enrichment was assessed via Gene Set Variation Analysis (GSVA). The Transcripts Per Million (TPM) expression matrix was converted into an enrichment score matrix utilizing immune-related gene sets. Immune signature variability across the three WYZ-1 biological replicates was subsequently displayed as a bubble plot. Magnetic resonance (MR) imaging: Mice were anesthetized with 1.5% isoflurane, placed in a stereotaxic holder, and monitored for respiration (60–80 bpm) and body temperature (37°C). The head was positioned in a cryo-cooled 2×2 receive coil within a 9.4 T Bruker MRI scanner (Bruker Biospin, Inc., Billerica, MD). A scout image acquired in three orthogonal directions was used to specify and position twelve 0.75-mm-thick axial slices encompassing the region from the olfactory bulb to the cerebellum. Quantitative T2 (echo time (TE) = 15 ms, number of echoes = 16, repetition time (TR) = 3000 ms, in-plane resolution = 100 μm, number of averages (NA) = 2) and diffusion-weighted (TE/TR = 20/2000 ms, NA = 2, gradient duration/gradient separation = 4/10 ms, gradient strength = 326 sec-1/cm-1) images were acquired, followed by T1-weighted (TR/TE = 4/250 ms, averages 8, flip angle = 30°, in-plane resolution = 75 μm) images obtained before and after administration of the contrast agent Gadavist (0.3 ml/kg). All images were processed using in-house MATLAB software (Mathworks Inc, MA). Statistical analysis: Kaplan-Meier survival curves were evaluated using the Log-rank (Mantel-Cox) test. All statistical analyses were performed using GraphPad Prism (Version 10.2.0, GraphPad Software, CA, USA). Data are presented as mean ± SD, with P-values less than 0.05 considered statistically significant. Results Establishment and genetic characterization of the WYZ-1 glioblastoma stem cell line To establish a stable syngeneic mouse glioblastoma stem cell (GSC) line, glioblastoma tissue was isolated from tumors arising in the RCAS/tv-a GBM mouse model, which harbors Nestin-driven tv-a (Ntv-a), HA-tagged PDGFB expression, homozygous deletion of the Ink4a-Arf (Cdkn2a) locus (resulting in loss of p16INK4a and p19ARF), and homozygous floxed Pten alleles (Ptenf/f). Following tumor dissociation and culture in GSC medium (DMEM/F12 supplemented with B27), a stable cell line was established and designated RCAS-TVA-GSC01 (RTG01). Genotyping of RTG01 cells confirmed retention of the parental genotype, including the Ntv-a transgene, Ink4a–Arf deletion, and intact floxed Pten alleles (Fig. 1A). Consistent with the animal model, Western blot analysis demonstrated robust expression of HA-tagged PDGFB in RTG01 cells (Fig. 1B). Because PTEN loss occurs in approximately 25–30% of human glioblastomas, we next generated a Pten-deficient derivative to better model this clinically relevant alteration. RTG01 cells were transduced with a Cre-expressing lentivirus to excise the floxed Pten alleles, yielding the derivative line WYZ-1. Successful Pten deletion in WYZ-1 cells was confirmed by agarose gel electrophoresis and Western blot analysis (Fig. 1C and D). To further characterize the molecular landscape of WYZ-1 cells, we performed whole-genome sequencing (WGS) and whole-genome bisulfite sequencing (WGBS). Genome-wide circos visualization demonstrated widespread genomic alterations and structural variation across multiple chromosomes (Supplementary Fig. 1A). WGBS analysis showed that CpG methylation represented the predominant methylation context in WYZ-1 cells, whereas CHG and CHH methylation levels remained low (Supplementary Fig. 1B). Analysis of CpG methylation distribution revealed a broad range of methylation states across the genome, with enrichment of highly methylated CpG regions (Supplementary Fig. 1C). Collectively, these data provide additional genomic and epigenomic characterization of the WYZ-1 model. Characterization of the WYZ-1 GBM GSC cell line: sphere formation, stemness, and differentiation WYZ-1 cells exhibited robust neurosphere-forming potential when cultured in glioblastoma stem cell (GSC) media. Longitudinal observation revealed a progressive shift in sphere kinetics; initial single-cell suspensions at day 0 transitioned into nascent aggregates by day 3, followed by significant volumetric expansion by day 5, culminating in the formation of mature, cohesive spheres by day 7, a phenotype consistent with sustained self-renewal capacity (Fig. 2A). In contrast, WYZ-1 cells cultured in differentiation medium containing 5% FBS failed to form spheres and instead adopted a flattened, differentiated morphology (Fig. 2B). To rigorously evaluate the stemness of the WYZ-1 line, we performed an in vitro mathematical limiting dilution analysis (Supplementary Fig. 2A). WYZ-1 cells exhibited strong sphere-forming efficiency, achieving an estimated stem cell/sphere-initiating frequency of 1 in 3.1 cells (95% CI: 2.05–4.68). The log-dose linear regression was tightly aligned with a single-hit Poisson model (Likelihood ratio test, χ2 = 1.87, DF = 1, P = 0.171), proving that individual spheres arose successfully from clonal projection rather than random cell aggregation. Microscopic evaluation at day 14 validated the presence of well-demarcated, spherical single-cell clones growing in isolation (Supplementary Fig. 2B). The identity of these GSCs was validated via immunofluorescence, which showed positive co-staining for the stemness markers Nestin and SOX2 (Fig. 2C). Flow cytometric analysis further revealed high CD133 expression in WYZ-1 cells relative to unstained and secondary antibody-only controls, supporting their glioblastoma stem cell identity (Fig. 2D). Upon culture in differentiation medium for 7 days, WYZ-1 cells expressed lineage-specific markers of neuronal (β3-tubulin) and astrocytic (GFAP) differentiation (Fig. 2E), indicating multilineage differentiation potential. Together, these results demonstrate that WYZ-1 cells maintain stemness under supportive conditions while retaining the capacity to differentiate into multiple neural lineages upon induction. Temporal progression of WYZ-1 tumor development in vivo The temporal progression of WYZ-1 tumor growth following intracranial implantation was evaluated by histology and magnetic resonance imaging (MRI) over a 3-week period (Fig. 3). H&E staining revealed no detectable tumor at week 1, with preserved normal brain architecture. By week 2, tumors exhibited emerging hypercellularity and infiltrative growth with indistinct boundaries, consistent with early tumor establishment. At week 3, a well-developed tumor mass was evident, characterized by dense hypercellularity and extensive infiltration into adjacent white matter (Fig. 3A). MRI findings were consistent with histological observations. Quantitative T2-weighted maps showed no abnormal signal at week 1, a subtle signal increase at week 2, and a pronounced, widespread signal at week 3, reflecting tumor expansion (Fig. 3B). Similar temporal patterns were observed in diffusion coefficient maps and T1-weighted contrast-enhanced images. Histopathological features of WYZ-1 tumors Further histopathological analysis was performed on WYZ-1 tumors at the study endpoint (approximately 4 weeks after intracranial implantation). H&E staining revealed diffuse infiltration of tumor cells into the surrounding brain parenchyma, particularly along white matter tracts, as evidenced by indistinct tumor-normal tissue boundaries at higher magnification (Fig. 4A). Areas of intratumoral necrosis were also observed, accompanied by prominent vascular structures consistent with tumor-associated neoangiogenesis (Fig. 4A). Ki67 immunohistochemical staining demonstrated high proliferative activity within the tumor core, with increased Ki67-positive cells surrounding necrotic regions, indicative of active tumor growth (Fig. 4B). At the tumor margins, sporadic Ki67 positivity suggested ongoing infiltration into adjacent tissue. Notably, Ki67-positive tumor cells were detected extending from the implanted right hemisphere across the midline into the contralateral hemisphere, recapitulating the highly infiltrative behavior characteristic of human glioblastoma. Immunohistochemical characterization of WYZ-1 glioblastoma The immunohistochemical (IHC) analysis of brain tissue from WYZ-1-bearing mice at the experimental endpoint revealed distinct expression patterns of selected markers critical for understanding glioblastoma stem cell (GSC) biology (Fig. 5 and Supplementary Fig. 3). Phosphorylated signal transducer and activator of transcription 3 (p-STAT3) positivity indicates active signaling pathways often linked to tumor proliferation and survival [19]. Nestin, a neural stem cell marker, and GFAP, an astrocytic marker, suggest a mixed population of stem-like and differentiated cells, reflecting the tumor's heterogeneity [20,21]. Cyclin D1, involved in cell cycle progression, and Integrin β1, associated with cell adhesion and invasion, highlight proliferative and invasive potentials [22]. Olig2, a transcription factor essential for glioma stemness, and PDGFRA, a growth factor receptor driving tumorigenesis, indicate a highly aggressive phenotype [23,24]. Vimentin, Integrin β1, and ZEB1, mesenchymal markers, reflect a partial mesenchymal-like phenotype associated with invasiveness. However, canonical mesenchymal markers CD44 and Fibronectin were not detected [25]. Together, these immunohistochemical profiles demonstrate that WYZ-1 tumors recapitulate key molecular features of aggressive, stem-like human glioblastoma, including intratumoral heterogeneity, high proliferative capacity, and invasive potential. Immunotherapeutic response of the WYZ-1 glioblastoma model To assess therapeutic sensitivity, WYZ-1 tumor-bearing mice were treated with standard-of-care chemotherapy (TMZ) or immune checkpoint inhibitors (anti-PD-1, anti-CTLA-4) (Fig. 6A). While anti-CTLA-4 therapy provided a modest improvement in median survival (32.5 days vs. 27.5 days in controls), neither TMZ nor anti-PD-1 treatment (median survival: 28 days for both) yielded a significant survival benefit. To further characterize the tumor microenvironment of this model, we performed RNA sequencing on three independent mouse tumor samples. Gene Set Variation Analysis (GSVA) of immune-related pathways revealed that the model is characterized by a predominant infiltration of myeloid lineage cells, with a paucity of adaptive immune cells (T cells, B cells) and NK cells (Fig. 6B). Furthermore, we performed IHC analysis using markers for T cells (CD4, CD8), macrophages/microglial (F4/80), and blood vessels (CD31). The results demonstrated that T cells were confined to the tumor periphery and failed to infiltrate the tumor core, which was instead densely populated by macrophages/microglial and blood vessels (Fig. 6C). To further dissect the spatial constraints of the WYZ-1 “cold” phenotype, we performed mIHC and parallel conventional IHC on consecutive serial tissue sections. Spatial profiling revealed a striking architectural compartmentalization of the immune infiltrate: CD4+ and CD8+ T cells were excluded from the tumor proper, remaining confined to the peripheral invasive margins and perivascular stroma. In sharp contrast, the tumor core was densely packed with highly infiltrative myeloid cell populations concurrently positive for F4/80 and the M2-polarization marker CD204 (Fig. 6D and E, and Supplementary Fig. 4A). In addition, this distinct architectural compartmentalization was further substantiated by flow cytometric analysis of WYZ-1 tumors 3 weeks post-implantation (n = 4) (Supplementary Fig. 4B). Phenotypic characterization of the sparse tumor-associated lymphocytic populations revealed a highly suppressed immune state: CD8+ T cells and CD4+ T cells exhibited profound checkpoint exhaustion, with PD-1 expression reaching 91% and 78%, respectively. Crucially, a dominant portion of these cells co-expressed the secondary inhibitory receptor LAG-3 (75% in CD8+; 41% in CD4+), pointing to terminal exhaustion (Supplementary Fig. 4C). Additionally, Foxp3+ regulatory T cells (Tregs) comprised 19% of the total tumor-associated CD4+ population (Supplementary Fig. 4C), confirming active immunosuppression within the tissue. Collectively, this limited therapeutic efficacy and immune-excluded (“cold”) phenotype highlights the aggressive, treatment-resistant nature of WYZ-1 tumors, supporting the utility of this model for studying GBM progression and resistance mechanisms in an immunocompetent setting. Discussion The clinical translation of effective immunotherapies for glioblastoma (GBM) has been critically impeded by a lack of physiologically relevant, immunocompetent preclinical models. Many commonly used syngeneic GBM models, such as GL261 and SB28, fail to fully recapitulate key features of human GBM. GL261 acts as an immunologically “hot” tumor with a high mutational burden that overestimates checkpoint blockade efficacy, whereas SB28 is poorly immunogenic yet lacks the molecular profile and histopathology characteristic of the human disease [7,26]. In this study, we address this translational gap by establishing WYZ-1, a novel syngeneic glioblastoma stem cell (GSC) line derived from the PDGFB-driven RCAS/tv-a mouse model with Pten deletion. The WYZ-1 cell line provides several advantages for neuro-oncology research compared with conventional genetically engineered mouse model (GEMM)-derived glioma models. Derived from the PDGFB-driven RCAS/tv-a platform, WYZ-1 preserves the biological context of somatic gliomagenesis in an immunocompetent setting while overcoming limitations such as variable tumor penetrance, long latency, and limited scalability. The combination of PDGFB activation, Ink4a/Arf loss, and Pten deletion generates highly aggressive glioblastoma cells with strong stem-like properties, cellular plasticity, and invasive behavior. As a stable and transplantable glioblastoma stem-like cell line, WYZ-1 enables rapid, reproducible, and medium-throughput studies in immunocompetent hosts. Importantly, WYZ-1 recapitulates key features of human glioblastoma, including diffuse invasion, necrosis, vascularization, stem-like phenotypes, and an immune-excluded (“cold”) tumor microenvironment, making it a valuable and experimentally accessible platform for translational glioblastoma research. In vitro, WYZ-1 cells maintained robust self-renewal capacity and high expression of GSC markers CD133, Nestin, and SOX2. Importantly, these cells exhibited multipotency, differentiating into astrocytic and neuronal lineages upon serum exposure. This plasticity mirrors the developmental hierarchy driving tumor recurrence in patients [27]. In vivo, this stem-like phenotype translates into highly aggressive behavior. By week 3, WYZ-1 tumors display prominent vascularization, necrosis, and diffuse infiltration along white matter tracts, extending into the contralateral hemisphere as demonstrated by H&E and Ki-67 staining. This contrasts sharply with the circumscribed growth often seen in other syngeneic lines, confirming that WYZ-1 successfully recapitulates the invasive “gliomatosis” patterns that make surgical cure impossible in humans [28]. Immunohistochemical profiling of WYZ-1 tumors at the study endpoint reveals key molecular features that underscore the model's clinical relevance. The elevated expression of Olig2, Nestin, GFAP, and PDGFRA indicates an aggressive, proneural-like phenotype driven by aberrant growth factor signaling [29]. In parallel, tumors expressed mesenchymal/plasticity-associated markers, including Integrin β1, p-STAT3, Vimentin, and ZEB1. However, canonical mesenchymal markers CD44 and Fibronectin were negative. These findings suggest that WYZ-1 tumors do not represent a fully mesenchymal subtype, but rather exhibit a mixed proneural/neural-lineage state with partial mesenchymal-like plasticity [30,31]. Furthermore, the activation of p-STAT3 reflects the engagement of oncogenic signaling pathways known to mediate immunosuppression [32]. In all, this intratumoral heterogeneity, encompassing stem-like, proliferative, and differentiated populations, faithfully recapitulates the cellular complexity of human glioblastoma and contributes to therapeutic resistance [33,34]. Immunologically, WYZ-1 represents an “immune-excluded” or “cold” tumor phenotype, which is the dominant landscape in human GBM [35]. Our transcriptomic and immunohistochemical analyses revealed a microenvironment heavily infiltrated by myeloid lineage cells (macrophages/microglial) but largely devoid of adaptive immune effectors (T cells) within the tumor core. Complementing these spatial observations, multi-parameter flow cytometry and mIHC analyses demonstrated that the sparse T cell populations restricted to the tumor periphery exist in a state of terminal exhaustion, marked by profound expression of PD-1 and LAG-3 alongside a substantial infiltration of immunosuppressive Foxp3+ regulatory T cells. Conversely, the heavily infiltrated tumor core is dominated by macrophages/microglial cells exhibiting an M2-polarized, immunosuppressive phenotype characterized by co-expression of F4/80 and CD204. This physical sequestration of exhausted T cells at the parenchyma boundaries, paired with an intensely suppressive myeloid core shield, provides a definitive mechanical explanation for the absolute resistance to anti-PD-1 monotherapy observed in our survival studies. Unlike GL261, which responds robustly to PD-1 blockade, the WYZ-1 model's resistance to PD-1 and standard-of-care temozolomide (median survival: 28 days vs. 27.5 days control) aligns more closely with the disappointing results of recent clinical trials [36]. Interestingly, we observed a modest but distinct survival benefit with anti-CTLA-4 treatment (median survival 32.5 days). Collectively, the WYZ-1 model fills a critical niche in neuro-oncology research. By combining the genetic precision of the RCAS/tv-a system with the experimental flexibility of a transplantable cell line, WYZ-1 offers a renewable platform for high-throughput studies. Its aggressive growth, immune-excluded microenvironment, and resistance to standard checkpoint inhibitors position it as a rigorous tool for testing next-generation combinatorial therapies, particularly those designed to overcome myeloid-mediated immunosuppression or enhance T-cell infiltration. Conclusions In summary, we have established and characterized WYZ-1, a novel Pten-deficient glioblastoma stem cell line that overcomes the limitations of existing syngeneic models. WYZ-1 tumors faithfully recapitulate the hallmarks of human glioblastoma, including stem-cell plasticity, diffuse invasion, and prominent necrosis. Crucially, the model mimics the “cold” immune microenvironment of human GBM, characterized by myeloid infiltration and T-cell exclusion, leading to resistance against anti-PD-1 and temozolomide therapy. We propose WYZ-1 as a robust, clinically relevant platform for dissecting the mechanisms of therapeutic resistance and accelerating the development of effective immunotherapies for glioblastoma. Supplementary Material