Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Oct 7.
Published before final editing as: Neuro Oncol. 2026 Aug 10:noag185. doi: 10.1093/neuonc/noag185

Potent antitumor preclinical efficacy of B7-H3 chimeric antigen receptor (CAR) T cells in H3G34-mutant, diffuse hemispheric glioma

David Akhavan 1,†, Monika Yadav 1,†, Siddharth Subham 1, John Jeppson 1, Omar Tinoco 1, Scott Chin 1, Nadine Santana-Magal 1, Jacqueline Babai 1, John Szarejko 1, Garrett Gibson 1, Kirill Sabitov 1, Daniel B Zamler 1, Gary Doug Myers 1, Brooke L Fridley 1, Stefan H Bossmann 1, Kalyani Pyaram 1, Viveka Nand Yadav 1
PMCID: PMC13640953  NIHMSID: NIHMS2210348  PMID: 42573529

Abstract

Background.

Diffuse hemispheric glioma (DHG), H3 G34-mutant, is a highly aggressive and fatal pediatric brain tumor, with a median survival of only ~18 months and no effective targeted therapies. While chimeric antigen receptor (CAR) T cell therapy has shown success in B-cell leukemia and promise in other pediatric central nervous system tumors, it has not yet been evaluated explicitly in preclinical or clinical models of DHG. B7-H3 (CD276), a tumor-associated antigen overexpressed in pediatric brain tumors, represents a promising target for immunotherapeutic intervention in DHG. Here, we present the first preclinical evaluation of B7-H3-directed CAR T cell therapy in DHG.

Methods.

We evaluated B7-H3 expression in specimens from DHG patients and patient-derived cell lines. Two B7-H3-targeted CAR constructs were generated and evaluated for antigen-specific activation, cytokine production, exhaustion, and cytotoxicity in vitro. Therapeutic efficacy was assessed in multiple orthotopic DHG xenograft models, followed by multiplex immunofluorescence and serum cytokine profiling.

Results.

B7-H3 was highly expressed in DHG tumors and cell lines, but at minimal levels in normal brain tissue. B7-H3 CAR T cells demonstrated potent, antigen-specific cytotoxicity against DHG cells, with negligible activity against B7-H3-knockout DHG cells. In vivo, the intratumoral administration of B7-H3 CAR T cells led to significant tumor regression and a substantial extension of survival in 3 DHG mouse models, with durable tumor eradication.

Conclusion.

This study provides preclinical evidence supporting B7-H3 CAR T cell therapy as a highly effective immunotherapeutic strategy for DHG. Our findings support B7-H3 as a compelling immunotherapeutic target for DHG, laying the groundwork for clinical translation.

Keywords: B7-H3, CAR T cells, DHG, H3G34-mutant, immunotherapy, pediatric high-grade gliomas

Lay Summary

Diffuse hemispheric glioma (DHG) is an aggressive childhood brain tumor with limited treatment options and poor survival. In this study, we explored an immunotherapy approach using engineered immune cells, called chimeric antigen receptor (CAR) T cells, that specifically target a protein, B7-H3, found on tumor cells. We found that B7-H3 is highly present in DHG tumors but not in normal brain tissue. In brain tumor animal models, these CAR T cells effectively killed tumor cells, shrank tumors, and significantly extended survival. These findings offer promising evidence for a potential treatment option and support moving this approach toward clinical trials for children with DHG.

Pediatric high-grade gliomas (pHGGs) are among the most lethal primary brain tumors and represent the leading cause of cancer-related mortality in children and adolescents.1 These tumors are biologically and clinically heterogeneous and are now stratified into molecular subgroups based on recurrent driver mutations in Histone H3.3 (H3.3A) or Histone H3.1 (H3C2).2–4 Approximately 50%-60% of pHGGs harbor mutually exclusive somatic point mutations, such as the presence of methionine at amino acid 27 instead of lysine (K27M) or the presence of arginine or valine at amino acid 34 instead of glycine (G34R/V) within the histone H3 tail.5–7 Pediatric high-grade gliomas with H3K27M mutation, characterized as diffuse midline glioma (DMG), typically arise in midline brain structures, such as the pons, thalamus, or spinal cord, and predominantly affect young children. In contrast, pHGGs with H3G34R/V mutation are classified as diffuse hemispheric glioma (DHG) and are usually located in the cerebral hemispheres, particularly the cortex and subcortical white matter, and occur more frequently in adolescents and young adults.2,6,8,9 These recurrent somatic mutations disrupt local chromatin structure and transcriptional regulation, promoting epigenetic reprogramming and oncogenesis. Despite aggressive multimodal therapy, including maximal surgical resection, radiation, and chemotherapy, DHGs remain uniformly fatal, with a median survival of 12-18 months and 5-year survival rates below 20%.8–10 Unlike their DMG counterparts, DHG remain understudied, and currently, no targeted or immunotherapeutic approaches are approved for these patients, highlighting a critical and unmet need for innovative therapeutic strategies.

Chimeric antigen receptor (CAR)-based immunotherapy has demonstrated remarkable success in treating childhood B-cell acute lymphoblastic leukemia,11–13 suggesting its potential as a promising therapeutic avenue to explore for the treatment of DHG. High-grade gliomas pose significant barriers to CAR T cell therapy due to their infiltrative nature, molecular heterogeneity, and localization within the immune-privileged central nervous system (CNS). Effective immunotherapy hinges on the identification of tumor-restricted antigens that enable selective tumor targeting while minimizing the risk of off-tumor neurotoxicity.14,15 Despite these challenges, several preclinical and early phase clinical studies have shown encouraging results using CAR T cells directed against glioma-associated antigens, including IL-13Rα2, EGFRvIII,16,17 and HER2 in glioblastoma (NCT02208362, NCT05168423, NCT03500991)18–20 as well as GD2 and B7-H3 in DMG (NCT04196413, NCT04185038).21–26 Notably, recent phase I trials targeting GD2 and B7-H3 have demonstrated safety and feasibility, with some patients showing extended survival, indicating promising therapeutic efficacy and marking important advancements in CAR T cell-based strategies for pediatric brain tumors. B7-H3 (CD276) has emerged as a highly promising immunotherapeutic target due to its broad and consistent overexpression across a diverse array of solid tumors, including prostate, lung, and breast cancers, as well as high-grade gliomas (pHGGs) such as DMG, medulloblastoma, atypical teratoid/rhabdoid tumors, and ependymoma.27–30 Importantly, its expression remains limited in normal tissues, making it an attractive CAR T cell-based therapeutic target.25,27–29,31,32 Despite significant advancements in CAR T cell therapy for other malignancies, no preclinical or clinical trials have yet evaluated efficacy of CAR T cell therapy in H3G34R/V mutated DHG. Given the tumor-specific overexpression and immunomodulatory role of B7-H3, this antigen emerges as a promising target for CAR T cell-based immunotherapy in DHG. However, the expression of B7-H3 in H3G34R/V-mutated DHG and its therapeutic implications remain largely unexplored.

To address this gap, we conducted the first investigation of CAR T cell therapy targeting B7-H3 in DHG. First, we analyzed and found robust B7-H3 expression in tumor specimens and multiple patient-derived DHG cell lines but not in normal brain tissues. We developed a second-generation CAR construct utilizing a humanized single-chain variable fragment (scFv) derived from 8H9 and MGA271 targeting B7-H324,29,30 Functional assays revealed that the B7-H3 CAR T cells exhibited potent antigen-specific cytotoxicity against B7-H3 positive DHG cell lines in vitro. Furthermore, intracranial tumor (ICT) administration of B7-H3 CAR T cells in an NSG orthotopic mouse model significantly inhibited tumor growth and extended survival. These findings establish B7-H3 targeted CAR T cell therapy as a viable and innovative immunotherapeutic strategy for treating DHG with H3G34R/V mutations. Our study highlights the promising potential of B7-H3 CAR T cells as a therapeutic strategy, paving the way for their future human trials targeting this high-risk pediatric DHG population.

Methods

Cell Lines and Culture Conditions

Human DHG cell lines KNS42 (H3.3G34V) were obtained from Dr Carl Koschmann. SJ-HGGX42 (H3.3G34R) was kindly provided by Dr Suzanne Baker, St. Jude Children’s Research Hospital, USA. OPBG-GBM001(H3.3G34R) and HSJD-GBM002 (H3.3G34R) were kindly provided by Dr Chris Jones, from the Institute of Cancer Research, UK (see Supplementary Material S1 for details).

In Silico Comparative RNA-Seq Analysis of B7-H3 Expression in DHG and Normal Brain Tissues

For B7-H3 RNA-sequencing (RNA-seq) analysis, the data for DHG patients were acquired from PedcBioPortal (PBTA-all), and for normal brain tissue samples from the Genotype-Tissue Expression (GTEx) project, version 8 (see Supplementary Material S1 for details).

Construction of B7-H3 CAR Vector

Lentivirus-based second-generation B7-H3 CAR (LV-B7-H3) vector was designed to encompass an scFv sequence against the human B7-H3 antigen derived from the 8H9 and MGA271 antibodies (see Supplementary Material S1 for details).

Lentivirus Production

LentiX-293T cells were plated in a 6-well plate at 1 million cells/well a day before transfection. When 70%-80% confluency was achieved, transfection was done using TransIT-LT1 Transfection Reagent (MirusBio; MIR 2300), LV-MAX Lentiviral Packaging Mix (Gibco; A43237), and respective plasmid constructs as per the manufacturer’s protocol(see Supplementary Material S1 for details).

Human B7-H3 CAR T Cell Production

Human PBMCs (ZenBio; SER-PBMC-F) were thawed and washed with LymphoONE T-Cell Expansion Xeno-Free media (Takara; WK552S) supplemented with 10% fetal bovine serum (FBS) (T cell media). PBMCs were cultured at a concentration of 2 × 106 cells/well in 1 mL of T cell media augmented with 50 U/mL IL2 (Peprotech; 200-02), 0.5 ng/mL IL15 (Peprotech; 200-15), and 20 μL T Cell Trans-Act (Miltenyi Biotec; 130-128-758) in a 24-well culture plate (see Supplementary Material S1 for details).

CRISPR/Cas9-Based B7-H3 Knockout in DHG Cell Lines

Lentivirus containing a CRISPR/Cas9-based vector containing gRNA CACCGACCCAGTGGTGGCACTGGT against human B7-H3 (VectorBuilder; VB231128-1328tgf) was transduced in DHG cell lines in media containing 8 μg/mL Polybrene (Sigma-Aldrich; TR-1003-G) (see Supplementary Material S1 for details).

Flow Cytometry

The surface expression of B7-H3 on DHG cell lines was detected with anti-B7-H3 antibody (Biolegend; 351011). All samples were collected using a BD Fortessa X-20 flow cytometer, and the data were analyzed with FlowJo software v10.9.0 (see Supplementary Material S1 for details).

Western Blot

DHG cell lines were harvested and lysed with RIPA lysing and extraction buffer (Gibco; 89900) containing 1X Halt Protease and Phosphatase Inhibitor Cocktail (Invitrogen; PI78440), and total protein was extracted (see Supplementary Material for details).

Cytotoxicity Assay

Cytotoxic efficacy of B7-H3 CAR-T cells was assessed using different detection systems. Tumor cells were CFSE-labeled and plated in a 96-well plate in triplicate and incubated for a minimum of 2 h at 37°C, 5% CO2. Effector cells (mock T, CD19 CAR-T, and B7-H3 CAR-T) were added at desired ratios in the same plate and incubated for 72 h (see Supplementary Material S1 for details).

Cytokine Secretion Analysis

KNS42, SJ-HGGX42, and KNS42 B7-H3KO cells were cocultured with effector cells (mock T cells or B7-H3 CAR T cells) at an E:T of 1:1 in triplicate wells of a 96-well plate for 24 h. Then, the plate was centrifuged briefly to let the cells settle down. Culture media was collected and aliquoted in a fresh 96-well plate and frozen at −80°C until used for analysis (see Supplementary Material S1 for details).

Animal Studies

All the mouse experiments were performed in accordance with the Institutional Animal Care and Use Committee of the University of Kansas Medical Center (KUMC). Male and female NSG (NOD-scid IL2Rgammanull) mice aged 8-12 weeks were purchased from the Jackson Laboratory. Tumor cells were intracranially transplanted by drilling a burr hole on the skull at 2.5 mm posterior, then 2.0 mm mediolateral to the bregma; 1 × 106 KNS42 and 1.5 × 106 SJ-HGGX42 cells expressing luciferase were resuspended in 4 μL of phosphate buffered saline (PBS) and were injected in 1 μL increments at 3.5 mm, 2.5 mm, and 2 μL in 1.5 mm from the depth of burr hole (see Supplementary Material S1 for details).

Patient Tumor Tissue Specimens and Immunohistochemical Analysis

Brain tumor tissue sections of 2 DHG patient samples were procured from the Biospecimen Repository Core Facility of the KUMC. Sequencing of tumor samples confirmed the presence of the H3.3G34R mutation in the H3F3A gene in both DHG patients. Dr Nelli S. Lakis, MD (Neuropathologist), confirmed the H&E-stained sections for the malignant and matched nonmalignant tissue from the same patients (normal brain tissue), specifically in the supratentorial region of the brain (see Supplementary Material S1 for details).

IHC Staining of the Mouse Brains

Mice were euthanized at experimental end-time points and were perfused with PBS, then with 4% paraformaldehyde (PFA). Brain tissues were coronally sliced into 2, placed cut sections facing downward in the cassettes, and processed for paraffin embedding (see Supplementary Material S1 for details).

IHC Quantitation

For quantitation of the area of B7-H3 and H3G34R expression, and the number of Ki67 positive cells, 4-5 sections from a 20× image of the respective mouse brains were captured and analyzed using ImageJ software.

Secretome Analysis

Multiplex analysis was performed on the Isoplexis CodePlex Secretome platform (Bruker Cellular Analysis) using serum collected from 4 CD19 or B7-H3 CAR-treated mice on the day of euthanasia (n = 2, per group) (see Supplementary Material S1 for details).

Multiplex Immunofluorescence and Workflow for Multiplex Image Analysis

The imaging was performed using the Akoya PhenoCycler-Fusion platform following standard histological preparation, antigen retrieval, and antibody panel staining33–35 (see Supplementary Material S1 for details).

Statistical Analysis

All the experimental data were analyzed for statistics using GraphPad Prism software (v10). Data are represented as mean ± SEM. RNA-seq for B7-H3 expression in DHG patient samples was analyzed using Mann-Whitney U tests, nonparametric (see Supplementary Material S1 for details).

Results

B7-H3 Is Highly Expressed in DHG Patient Tumor Specimens and Patient-Derived Cell Lines

We explored the potential of B7-H3 to be utilized as the target for the CAR T cell-based immunotherapy in DHG. First, we assessed the expression of the B7-H3 antigen in DHG tumor tissues from patients. Immunohistochemistry (IHC) of brain tumor tissue from 2 DHG patients revealed high expression of B7-H3 (Figure 1A). In contrast, the matched nontumor brain tissues had almost undetectable expression of B7-H3. Importantly, IHC analysis confirmed the presence of the DHG-specific H3G34R mutation in both the patient tumor tissues, along with elevated Ki-67 expression (Figure S1). Further, in silico analysis of bulk RNA-seq from 31 H3G34R/V DHG patient (PBTA-cohort) tumors revealed significant overexpression of B7-H3 expression compared to the normal brain (GTEXV8 dataset, n = 2655, ****P < .0001) (Figure 1B). Next, we also confirmed B7-H3 expression in the 4 DHG patient-derived cell lines with H3G34V mutation (KNS42), and H3G34R mutation (SJ-HGGX42, OPBG-GBM001, HSJD-GBM002) using immunoblot staining and flow cytometry. Western blot analysis revealed high levels of B7-H3 expression (Figure 1C) and flow cytometry confirmed the high surface expression of B7-H3 in all the patient-derived cell lines (Figure 1D). These results indicate B7-H3’s tumor-specific expression in DHG, making it a potential target for developing novel CAR T cell-based therapeutic strategies.

Figure 1.

Figure 1.

Open in a new tab

Significantly elevated B7-H3 expression in tumor tissue and patient-derived cell lines of H3G34R/V mutated diffuse hemispheric gliomas (DHGs). (A) Immunohistochemical analysis for B7-H3 expression in 2 DHG patient tissue samples harboring H3G34R mutation. Images for tumor and nontumor tissue from the same section for B7-H3 have been represented (scale bar: 2× × 5 mm, and 20× × 200 μm). (B) Violin plots showing the relative RNA expression of B7-H3 in DHG patient samples (n = 31) from the PedcBioPortal (PBTA-all) compared to normal brain tissue samples (n = 2655) RNA-seq from the Genotype-Tissue Expression (GTEx) project, version 8. Mann-Whitney U tests, nonparametric (****P < .0001). (C) Western blot showing variable expression of B7-H3 in different patient-derived DHG cell lines with H3G34R/V mutation. (D) Flow cytometry staining for B7-H3 expression on the surface of DHG cell lines.

Generation of B7-H3 Targeted CAR-T Cells

We designed a second-generation B7-H3 CAR construct using the B7-H3 scFv sequence derived from monoclonal antibodies 8H929,30 and MGA271.24,25 To generate B7-H3-directed CAR T cells, a humanized anti-B7-H3 (scFv), IgG4 hinge with EQ mutation, CD4 transmembrane domain,36,37 4-1BB costimulatory domain, and CD3ζ signaling domain (Figure 2A), was cloned into a lentiviral vector.37 The use of a 4-1BB costimulatory domain supports sustained CAR T cell persistence and reduced exhaustion, as previously demonstrated to promote memory formation and oxidative metabolism compared to CD28-based designs.38–41 This is particularly important for achieving durable responses in aggressive solid tumors such as DHG. Further, the selection of the IgG4 hinge/spacer was guided by comparative optimization studies that demonstrated enhanced antitumor activity and reduced exhaustion relative to CD8-based constructs (Figure S2). This highlights the importance of CAR structural design in balancing efficacy and T cell fitness. Human CD3+ T cells were activated with CD3/CD28 and then transduced with lentivirus containing the B7-H3 CAR construct, referred to as LV-B7-H3 CAR-T. The transduced T cells were checked for CAR expression by flow cytometry using antihuman IgG (Fcg fragment specific) after gating on the lymphocyte population, single and live cells (Figure 2B). We achieved ~45% B7-H3-positive CAR expression (Figure 2C). We next evaluated the basal exhaustion status of transduced T cells prior to antigen encounter. Surface expression of exhaustion markers 2B4, PD-1, Tim-3, and Lag-3 in B7-H3 CAR T cells was comparable to mock-transduced controls, indicating that CAR transduction alone did not induce premature T cell exhaustion (Figure 2D).

Figure 2.

Figure 2.

Open in a new tab

Design, generation, and phenotypic characterization of B7-H3 CAR T cells. (A) Schematic of the second-generation B7-H3 CAR construct consisting of a humanized anti-B7-H3 single-chain variable fragment (scFv) derived from 8H9 monoclonal antibody, IgG4 hinge (EQ mutation), CD4 transmembrane domain, 4-1BB co-stimulatory domain, and CD3ζ signaling domain. (B) Representative gating strategy for flow cytometric analysis of transduced human T cells: lymphocytes were gated based on FSC-A vs SSC-A (left), followed by selection of single cells (middle), and live cells (right) using Ghost Red 780 viability dye. (C) CAR expression was quantified by staining with an antihuman IgG antibody targeting the Fc domain. Approximately 44.6% of live T cells expressed the B7-H3 CAR construct. (D) Surface expression of T cell exhaustion markers (2B4, PD-1, Tim-3, and Lag-3) was analyzed by flow cytometry. Histograms show overlays of B7-H3 CAR T cells and mock-transduced T cells. No major upregulation of exhaustion markers was observed in CAR T cells prior to antigen stimulation.

While prior preclinical and phase I/II clinical studies in diffuse intrinsic pontine glioma (DIPG) and pHGG models have predominantly utilized MG271-based B7-H3 CAR T cells,24,25 we sought to determine whether the choice of antigen-binding domain influences therapeutic efficacy in our DHG model. To address this, we evaluated the cytotoxicity activity of an alternative 8H9-based B7-H3 CAR in parallel with the established MGA271 CAR in a DHG model (Figure S3A). Our data demonstrate that both constructs confer comparable cytotoxic activity, indicating that the 8H9 binder is equally effective (Figure S3B). The 8H9-based CAR exhibited a less exhausted phenotype than the MGA271-based CAR following antigen stimulation (Figure S3C). Given prior validation of MG271-based CAR T cells in DIPG models, our study advances the field by prioritizing the 8H9-based construct for all in vitro and in vivo experiments, thereby expanding the translational applicability of B7-H3-targeted therapies in the DHG model.

Following identification of 8H9 as the preferred antigen-binding domain based on its comparable antitumor activity and lower exhaustion profile relative to MGA271, we evaluated an alternative 8H9-based CAR configuration as a subsequent optimization strategy. This retrovirus-based second-generation B7-H3 CAR construct (RV-B7-H3 CAR T cells) incorporated the same 8H9 scFv but utilized a distinct backbone comprising a CD8α transmembrane region, a CD28 costimulatory domain, and a CD3α signaling domain (Figure S4A).30 Following CD3/CD28 stimulation, T cells exhibited activation marker expression (CD69, CD25, HLA-DR) and were successfully transduced with the RV-B7-H3 CAR construct, achieving ~60% surface CAR expression (Figure S4B and C). Proliferation assays on days 0, 4, and 8 showed no significant difference between CAR T and mock T cells, indicating that CAR expression did not impair T cell expansion (Figure S4D). Although both the 8H9 CAR constructs yielded similar killing efficacy (Figure S4E), CAR with CD4 transmembrane domain, 4-1BB costimulatory domain exhibited less exhaustion (Figure S4F). Collectively, these in vitro optimization studies established that CAR architecture, including the choice of scFv, transmembrane region, and costimulatory domain, substantially influences T cell fitness. Among the 3 B7-H3 CAR constructs evaluated, the 8H9-based CAR incorporating the CD4 transmembrane domain, 4-1BB costimulatory domain, and CD3ζ signaling domain exhibited the most favorable balance of cytotoxicity and reduced exhaustion. Therefore, this optimized construct was selected for all subsequent in vitro and in vivo studies in DHG model.

B7-H3 CAR T Cells Demonstrate Effective In Vitro Killing Against DHG Cell Lines

We investigated the in vitro cytotoxic effect of LV-B7-H3 CAR-T cells against 2 DHG cell lines: KNS42 (H3G34V-mutated)-adherent cells and H3G34R-mutated (SJ-HGGX42) neurosphere cells. We employed different cytotoxicity assessment methods, that is, fluorescence-based assay using Celigo and Incucyte, and luminescence-based assay using a spectrophotometer. KNS42 and SJ-HGGX42 cells were cocultured with mock and B7-H3 CAR T cells at the indicated effector-to-target (B7-H3 CAR T: tumor cells) ratios for 72 h. In vitro killing revealed that B7-H3 CAR T cells had robust cytotoxicity, achieving near-complete eradication of tumor cells compared to mock T cells (Figure 3A–E). KNS42 cells (Figure 3A and B) demonstrated greater sensitivity to CAR T cell-mediated killing than SJ-HGGX42 (Figure 3D and E), supporting that CAR T cell efficacy correlates with B7-H3 expression on target tumor cells. Live-cell imaging revealed a marked reduction in viable KNS42 tumor cells following treatment with B7-H3 CAR T cells compared to mock T cells. At increasing E:T ratios (2:1 and 4:1), a substantial decrease in GFP-positive tumor cells was observed, consistent with enhanced tumor clearance. In contrast, mock T cells showed minimal cytotoxic effect, with abundant GFP-positive cells persisting across all conditions (Figure 3C).

Figure 3.

Figure 3.

Open in a new tab

Cytotoxic potential of B7-H3 CAR T cells against DHG cell lines. Mock and B7-H3 CAR T cells with ~30% of surface CAR expression were cocultured with the DHG cell lines at different effector-to-target ratios (E:T ratio). Cytotoxicity was assessed after 72 h to determine the percentage of viable tumor cells. (A) Bar graph representing killing in KNS42 cell lines. (B) Line graph for kinetics of CAR T cytotoxic activity in KNS42. (C) Representative fluorescence image of cocultures of KNS42 cells with mock and B7-H3 CAR T cells at 2:1 and 4:1 E:T ratio (scale bar: 10× × 1 mm). (D) Bar graph for cytotoxicity in SJ-HGGX42. (E) Kinetics of CAR-T cytotoxic activity in SJ-HGGX42. (F) Cytokine secretion profiles of B7-H3 CAR-T cells after 24-h coculture with DHG cell lines (KNS42 and SJ-HGGX42), measured by multiplex bead-based immunoassay by flow cytometry. (G) Flow cytometric analysis of exhaustion markers 2B4, PD1, TIM3, and LAG3 following 24 h of antigen stimulation. Cytotoxicity, cytokine, and exhaustion data are represented as mean ± SEM and were analyzed by ordinary 1-way ANOVA with Tukey’s multiple comparison. Killing kinetics data are represented as mean ± SEM and were analyzed by 2-way ANOVA with Tukey’s multiple comparison. ****P < .0001; ***P < .001; **P < .01; *P < .05; ns > 0.05.

We comprehensively evaluated the kinetics of B7-H3 CAR T cell-mediated cytotoxicity against KNS42 tumor cells across multiple effector-to-target (E:T) ratios (1:8, 1:4, 1:2, 1:1, 2:1, and 4:1) taking readings at every 8-h interval. A time and dose-dependent increase in tumor cell killing was observed compared to mock and CD19 CAR T cells at the 4:1 ratio. Higher E:T ratios (4:1) demonstrated rapid and near 80% tumor clearance beginning at 48-72 h postcoculture, whereas lower E:T ratios (1:8 and 1:4) showed slower and less efficient killing, with partial tumor cell killing even at 72 h (Figure S5A and B). These results indicate that both time and effector cell concentration critically influence the efficacy of B7-H3 CAR T cell-mediated cytotoxicity. Additionally, to further investigate if B7-H3 CAR T cell cytotoxicity is dependent on antigen density on the target cells, we compared the cytotoxicity susceptibility of 2 neurosphere-forming glioma cell lines with differential B7-H3 expression: SJ-HGGX42 (low B7-H3) and OPBG-GBM001 (high B7-H3) (Figure 1C and D). At similar effector-to-target (E:T) ratios, we observed significantly greater cytotoxic activity of B7-H3 CAR T cells against OPBG-GBM001 relative to SJ-HGGX42 (Figure S6). Our data provide strong functional evidence that CAR T cell efficacy scales with B7-H3 expression on tumor cells. Similarly, 8H9-based retrovirus-B7-H3 CAR T cells also demonstrated significant killing potential against 2 DHG cell lines (KNS42 and SJ-HGGX42) (Figure S7A and B), further demonstrating the potential of B7-H3 CAR T cells as an effective strategy.

B7-H3 CAR T Cells Activation, Exhaustion, and Cytokine Profiles Following Co-culture With DHG Cell Lines (KNS42 and SJ-HGGX42)

We evaluated CAR T cell activation and effector cytokine production following 24-h coculture of B7-H3 CAR T cells with DHG cell lines KNS42 and SJ-HGGX42 using a bead-based 13-plex cytokine detection kit employing a flow cytometer. B7-H3 CAR T cells secreted significantly higher levels of IL-2, IL-6, TNF-α, IFN-γ, Granzyme A, Granzyme B, Perforin, IL-10, IL-17A, IL-4, sFas, sFasL, and Granulysin compared to mock T cells (Figure 3F, Figure S8). These findings confirm that B7-H3 CAR T cells undergo robust antigen-specific activation and mount a strong effector cytokine response upon engagement with DHG tumor targets. We also checked the status of T cell exhaustion after exposure of mock and CAR T cells to the target antigen in KNS42 and SJ-HGGX42 cell lines. After antigen stimulation, B7-H3 CAR T cells displayed significantly higher surface expression of PD1, TIM3, and LAG3 in both DHG models than mock T cells (Figure 3G), consistent with activation-induced exhaustion. These results demonstrate that B7-H3 CAR T cells are capable of robust, dose-dependent cytotoxicity and effector cytokine release upon target recognition, indicative of potent, but sustained T cell activation in the DHG microenvironment.

B7-H3 CAR T Cell Did Not Affect the Viability of B7-H3 Negative Cells

To directly assess the antigen specificity of B7-H3 CAR T cell effector functions, we performed CRISPR-Cas9-based knockout (KO) of B7-H3 antigen in KNS42 and SJ-HGGX42 DHG cell lines and generated respective B7-H3KO cell lines. Flow cytometry and immunoblotting confirmed complete loss of B7-H3 surface and total protein expression in B7-H3KO cells (Figure 4A and B, Figure S9A–C). We observed a considerable loss of CAR T cells’ cytolytic activity against KNS42 B7-H3KO and SJ-HGGX42 B7-H3KO cells (Figure 4C, Figure S9D and E). B7-H3 CAR T cells showed no significant cytotoxicity against B7-H3KO KNS42 cells across E:T ratios up to 4:1, with tumor cell viability comparable to mock and CD19 CAR T cell controls, confirming target-specific activity (Figure 4D). Fluorescence imaging further confirmed the lack of B7-H3 CAR-mediated tumor killing in B7-H3KO cells, even at high E:T ratios (Figure 4E). Comparison of cytokine release after antigen stimulation revealed significant reduction by B7-H3KO cells compared to control cells (Figure 4F). Time-course analysis of B7-H3 CAR T cell cytotoxicity revealed that B7-H3 CAR T cells exhibited negligible cytotoxic activity against KNS42 B7-H3KO cells across all tested E:T ratios (1:8, 1:4, 1:2, 1:1, 2:1, and 4:1) over 72 h unlike parent KNS42 cells. Tumor cell viability remained unchanged throughout the assay, confirming the antigen-specific nature of B7-H3 CAR T cell-mediated killing and the requirement for B7-H3 expression on target cells (Figure S10). Further, to assess the dependence of CAR T cell activity on antigen availability/density, we performed a density-dependent cytotoxicity assay using mixed populations of B7-H3-positive and B7-H3KO KNS42 cell lines. B7-H3 CAR T cells demonstrated robust, antigen-dependent cytotoxicity that progressively decreased with declining proportions of B7-H3-positive targets and was completely abolished under antigen-negative conditions, whereas mock T cells showed no cytotoxicity across all conditions (Figure S11A). Lastly, we evaluated B7-H3 CAR T cell killing of B7-H3 negative TNBC cell line MDA-MB-453 and did not see any cytotoxicity (Figure S11B). Together, these results establish that B7-H3 CAR-T cell activation, cytotoxicity, and effector cytokine responses are strictly dependent on tumor-intrinsic B7-H3 expression, underscoring the high specificity and precision of this therapeutic strategy.

Figure 4.

Figure 4.

Open in a new tab

Inhibition of cytolytic activity by B7-H3 CAR T in B7-H3 negative cell lines. Flow cytometry and immunoblot validation of B7-H3 knockout (B7-H3KO) in KNS42 cells using CRISPR/Cas9. (A) Flow cytometry for B7-H3 surface expression in KNS42 B7-H3KO cells. (B) Immunoblot to verify the absence of B7-H3 in KNS42 B7-H3KO cell line. (C) Cytotoxicity assays show that B7-H3 CAR T cells fail to kill B7-H3KO tumor cells. (D) Cytotoxicity killing kinetic assessment at multiple E:T ratios shows no significant difference between B7-H3 CAR, CD19 CAR, and mock T-mediated killing in KNS42 B7-H3KO cell line. At the time of plating for the cytotoxicity assay, both CD19 and B7-H3 CAR T cell groups exhibited approximately 40% surface CAR expression. (E) Representative fluorescence image of cocultures of KNS42 B7-H3KO cells with mock and B7-H3 CAR T cells at 2:1 and 4:1 E:T ratio (scale bar: 10× × 1 mm). (F) Comparison of cytokine release in KNS42 and KNS42 B7-H3KO cell lines by B7-H3 CAR-T cells. Cytotoxicity, and cytokine data are represented as mean ± SEM and were analyzed by ordinary 1-way ANOVA with Tukey’s multiple comparison. Kinetics data are represented as mean ± SEM and were analyzed by 2-way ANOVA with Tukey’s multiple comparison. ****P < .0001; ***P < .001; **P < .01; *P < .05; ns > 0.05.

B7-H3 CAR T Cells Eradicate Tumor Burden and Prolong Survival in an Orthotopic DHG Mouse Model With H3G34R Mutation

After B7-H3 CAR T cells’ efficient and specific cytotoxic effect against B7-H3 positive DHG cell lines in vitro, we evaluated their antitumor efficacy in vivo. First, we established an orthotopic xenograft model using SJ-HGGX42 cells implanted intracranially in the cortex of NSG mice. Following confirmation of tumor engraftment via bioluminescence imaging (BLI), mice were randomized into B7-H3 CAR or CD19 CAR treatment arms and administered CAR T cell therapy intratumorally on day 8 postimplantation (Figure 5A). Longitudinal BLI revealed a striking and sustained reduction in tumor burden of B7-H3 CAR T cell-treated mice, whereas CD19 CAR-treated controls exhibited progressive tumor growth (Figure 5B and C). Quantitative analysis of total flux confirmed a significant decline in tumor signal in the B7-H3 CAR cohort, beginning shortly after treatment and sustaining through 300 days (Figure 5C and D). Survival analysis demonstrated a significant survival advantage in B7-H3 CAR T cell-treated mice compared to CD19 CAR controls. The median survival for CD19 CAR-treated mice was 136 days, whereas a median survival was not reached in the B7-H3 CAR T cell-treated cohort (n = 9), with 4 mice remaining alive beyond 300 days at the time of study termination (P = .0004; Figure 5D). H&E staining of brains revealed the absence of tumor in B7-H3 CAR T cell-treated mice, while the control mice had highly dense and diffuse tumoral growth (Figure 5E, Figure S12). Immunohistochemistry confirmed the absence of B7-H3 and H3G34R staining in the B7-H3 CAR T celll-treated group, accompanied by near-complete ablation of the proliferation marker Ki-67 (Figure 5E and F, Figure S12). To evaluate potential neurotoxicity and treatment-associated inflammation, we performed IHC for Iba1 on brain sections from CD19 CAR and B7-H3 CAR-treated mice. Quantitative analysis revealed a significant reduction in Iba1+ cells in B7-H3 CAR-treated animals (Figure S13), consistent with decreased microglial activation following effective tumor clearance. Additionally, we utilized multiplex immunofluorescence imaging in CD19 CAR and B7-H3 CAR-treated mice brains to analyze tumor, microglial, and T-cell markers in SJ-HGGX42 tumors following CAR-T treatment. CD19 CAR-T-treated mice showed invasive tumor regions (Vimentin+), proliferating tumor cells (Ki67+), microglia (Iba1+), whereas B7-H3 CAR-T-treated mice brain displayed absence of tumor related phenotype with minimal expression of Iba1+, Vimentin+ and Ki67+ (Figure 5G and H, Figure S14). Analysis of T-cell markers (CD3e and CD8) in CD19- and B7-H3 CAR T cell-treated mice revealed an absence of detectable T cells at the time point analyzed, suggesting either limited persistence or clearance of infused CAR T cells.

Figure 5.

Figure 5.

Open in a new tab

B7-H3 CAR T cells mediate robust tumor regression and extend survival in an orthotopic H3G34R-mutant DHG mouse model. (A) Experimental schema: NSG mice were intracranially implanted at 2.5 mm posterior and 2.0 mm lateral to bregma with 1.5 × 106 SJ-HGGX42 cells and treated with intratumoral injection of 1 × 106 total B7-H3 CAR or CD19 CAR-T cells. The CD19 and B7-H3 CAR T cell groups were normalized based on surface CAR expression (~40%) to ensure equivalent numbers of functional CAR + cells were administered. Tumor burden was monitored via bioluminescence imaging (BLI) at indicated time points. (B) Representative longitudinal BLI images showing tumor regression in B7-H3 CAR-treated mice vs progressive tumor growth in CD19 CAR-treated controls. (C) Quantification of tumor burden over time based on total flux (photons/s) from BLI. B7-H3 CAR T cell-treated mice exhibited a significant and sustained reduction in tumor signal compared to controls. (D) Kaplan-Meier survival analysis of mice harboring H3.3G34R mutated SJ-HGGX42 tumors treated with CD19 (control, blue, n = 9 with median survival = 136 days, B7-H3-treated animals, red, n = 9) median survival did not reached as 4 mice were alive past 300 days at the time of experiment termination (P = .0004) Log-rank (Mantel-Cox) test. (E) Representative H&E and Immunohistochemistry (IHC) images from B7-H3 (n = 9) and CD19 treatment groups (n = 9) for expression of B7-H3, H3G34R, and Ki67 in brains of CD19 and B7-H3 CAR-treated mice groups (scale bar: 1.2× × 2 mm, and 20× × 200 μm). (F) IHC quantitation for the B7-H3, H3G34R, and Ki67. Bar plot represents IHC quantification of 4-5 randomly selected regions per mouse (n = 3), per treatment condition. IHC images were quantified using ImageJ software. Data were analyzed using unpaired Student’s t tests. (G) Representative multiplex immunofluorescence image panel comparing expression of CD3e (cyan), CD8 (magenta), Iba1 (yellow), Ki67 (red), and Vimentin (white) with DAP (blue) in B7-H3 and CD19 treatment groups (screenshots captured at 10× magnification). (H) Bar plots showing the percent abundance of Iba1, Ki67, and Vimentin out of all segmented cells per brain sample for respective groups. ****P < .0001.

To further strengthen the translational relevance and robustness of our findings, we performed a repeat in vivo experiment using a more aggressive SJ-HGGX42 model, which has a median survival of 50 days. Additionally, CAR T cells are administered at a later time point (3 weeks postimplantation), when tumors are more established. This design mimics advanced disease with a higher tumor burden, which is clinically more challenging to treat. Importantly, even under these conditions, B7-H3 CAR T cell therapy resulted in tumor clearance in the majority of treated mice, with long-term survival observed compared to CD19 CAR T control mice at the end of the study (Figure S15A–C). To evaluate potential systemic toxicity and cytokine release associated with CAR T cell therapy, we performed multiplex serum cytokine profiling at the end of the experiment, which revealed elevations of IL-6 and IL-8, consistent with CAR T cell activation, without significant increases in IFN-γ, TNF-α, or IL-10. These findings suggest an absence of severe cytokine release or systemic toxicity (Figure S15D). Notably, serum was collected after tumor clearance, which may account for the lack of significant elevation of pro-inflammatory cytokines.

Collectively, these results demonstrate that the antitumor efficacy of B7-H3 CAR T cells is both robust, reproducible, and remains effective even in aggressive and late-stage tumor settings. These results establish that B7-H3 CAR T cells mediate potent and selective tumor regression in vivo and significantly extend survival in a clinically relevant model of H3G34R-mutated diffuse hemispheric glioma.

Potent AntiTumor Activity of B7-H3 CAR T Cells in an In Vivo Orthotopic H3G34V Mutated DHG Mouse Model

To further validate the antitumor efficacy of B7-H3 CAR T cells, we employed an orthotopic model using the H3G34V-mutant DHG cell line KNS42. NSG mice were intracranially injected with 1 × 106 million KNS42 cells and monitored by BLI to confirm tumor establishment before treatment allocation (Figure 6A). Mice were randomized based on tumor burden and received intratumoral injections of either B7-H3 CAR or CD19 CAR T cells. Longitudinal BLI revealed rapid tumor regression in B7-H3 CAR T treated mice across both models, whereas CD19 CAR-treated controls exhibited continued tumor progression (Figure 6B and C). Quantitative analysis demonstrated a profound and sustained reduction in tumor burden in the B7-H3 CAR T treated groups, with durable responses persisting for over 290 days in KNS42 models. Survival analysis confirmed a significant extension of survival in B7-H3 CAR T cell-treated mice compared to controls (P = .0095 for KNS42; Figure 6D). Histological examination revealed complete tumor clearance in B7-H3 CAR T cell-treated mice by H&E staining, in contrast to dense, infiltrative tumors in controls (Figure 6E, Figure S16). Immunohistochemistry showed a marked absence of B7-H3 expression and depletion of Ki-67+ proliferative cells in B7-H3 CAR T cell-treated mice brains (Figure 6E and , F Figure S16). Further, body weight monitoring of mice treated with B7-H3 CAR T cells and CD19 CAR T cells showed no significant differences over time, indicating that B7-H3 CAR T cell therapy was well tolerated and did not induce systemic toxicity (Figure S17). Collectively, these findings demonstrate that B7-H3 CAR T cells mediate potent and durable tumor eradication, significantly prolonging survival across distinct genetic subtypes of DHG in orthotopic preclinical models.

Figure 6.

Figure 6.

Open in a new tab

In vivo efficacy of B7-H3 CAR in KNS transplanted mice. (A) Schematic of the in vivo study design. 8-10 weeks old NSG mice were intracranially implanted at 2.5 mm posterior and 2.0 mm lateral to bregma with 1 × 106 KNS42 cells (H3G34V-mutant) and treated with intratumoral injection of 1 × 106 total B7-H3 CAR or CD19 CAR T cells on day 17. The CD19 and B7-H3 CAR T cell groups were normalized based on surface CAR expression (~40%) to ensure equivalent numbers of functional CAR + cells were administered. Bioluminescence imaging (BLI) was performed at the indicated time points to monitor tumor progression. (B) Representative longitudinal BLI images showing sustained tumor regression in B7-H3 CAR-treated mice, whereas CD19 CAR-treated mice exhibited rapid tumor expansion. (C) Quantification of tumor burden based on total photon flux over time. B7-H3 CAR-treated mice exhibited significantly reduced tumor burden compared to controls, with signal suppression maintained up to 294 days in a few mice posttreatment. (D) Kaplan-Meier survival analysis demonstrated significantly prolonged survival in B7-H3 CAR-treated mice (n = 7, median survival = 244 days) vs CD19 CAR controls (n = 7, median survival = 68 days (P = .0095), Log-rank (Mantel-Cox) test. (E) Histological and immunohistochemical analysis of brain sections at the endpoint. Representative images from B7-H3 (n = 7) and CD19 treatment groups (n = 7) showed an absence of tumor in B7-H3 CAR-treated brains, while CD19 controls displayed dense tumor infiltration. Immunohistochemistry (IHC) revealed near-complete loss of B7-H3 and Ki67 expression in B7-H3 CAR-treated tumors (scale bar: 1.2× × 2 mm, and 20× × 200 μm). (F) IHC quantitation for the B7-H3, and Ki67. Bar plot represents IHC quantification of 4-5 randomly selected regions per mouse (n = 2 for CD19 and n = 3 for B7-H3) treatment condition. IHC images were quantified using ImageJ software. Data were analyzed using unpaired Student’s t tests. ****P < .0001; **P < .01 .

Discussion

CAR T cell therapy has shown remarkable efficacy in hematologic malignancies; however, translating this approach to solid tumors, particularly CNS malignancies, remains challenging due to barriers such as antigen heterogeneity, limited trafficking, and immune suppression within the tumor microenvironment. Pediatric high-grade gliomas, including DHG, exhibit pronounced molecular and phenotypic heterogeneity, aggressive infiltration into eloquent brain regions, and reside within the immunoprivileged CNS, compounding the difficulty of achieving effective and durable immune-mediated tumor clearance. Moreover, the restrictive nature of the blood-brain barrier (BBB) poses an additional obstacle for systemically administered cellular therapies.

In this context, our study provides the first preclinical evidence supporting the therapeutic efficacy of B7-H3 targeted CAR T cell therapy in DHGs harboring H3G34R/V mutations. We demonstrate that B7-H3-targeted CAR T cells mediate potent and selective cytotoxicity against DHG cells, both in vitro and in vivo. The 2 orthotopic models used in this study, derived from H3G34R-(SJ-HGGX42) and H3G34V-(KNS42) mutated cell lines, represent distinct yet clinically relevant subtypes of DHG. Intratumoral delivery of B7-H3 CAR T cells in both models resulted in robust tumor regression, marked reductions in proliferation markers (Ki67), and significantly prolonged survival compared to CD19 CAR T treated controls (Figures 5 and 6). These outcomes suggest that direct intracerebral administration of CAR T cells may circumvent BBB-associated limitations and achieve therapeutic levels at the tumor site, offering a viable strategy for CNS tumor immunotherapy. Although a small number of tumor-free mice in the B7-H3 CAR T-cell group died (Figure 6) during long-term follow-up, no residual tumor was detected histologically, suggesting that the deaths were likely attributable to treatment-or host-related complications inherent to the NSG xenograft model rather than to tumor progression. Importantly, our study extends beyond expression profiling to demonstrate functional in vitro and in vivo efficacy of B7-H3 CAR T cells in DHG. These findings establish a proof-of-concept and address a critical gap in immunotherapeutic strategies for this high-risk DHG model.

Our data show that B7-H3 is robustly and selectively expressed in DHG patient tumor tissues and across multiple H3G34R/V-mutated DHG cell lines, whereas it is minimally expressed in normal brain tissues. Identifying tumor-associated antigens that are highly enriched and uniformly expressed is critical for achieving effective and comprehensive tumor eradication.28,29,42 Consistent and high-level B7-H3 expression is crucial for the effectiveness of CAR T cell therapy, as intratumoral antigen heterogeneity remains a significant barrier to the successful treatment of brain tumors.27,43 Importantly, our findings indicate that CAR T cell efficacy is primarily driven by antigen density rather than tumor-intrinsic differences. Under uniform conditions, cytotoxicity correlated with B7-H3 expression levels, consistent with threshold-dependent CAR T activation, highlighting antigen heterogeneity as a key determinant of therapeutic response in DHG (Figure 3, Figure S6). Several studies have reported high expression of B7-H3 in pediatric brain tumors.21,26,29,44 Haydar et al conducted a comprehensive analysis of 5 CAR targets in tumors from 49 pediatric patient-derived orthotopic xenografts (PDOXs), identifying B7-H3 as the most highly expressed antigen, followed by GD2, IL-13Rα2, HER2, and EphA2.28 Notably, in the SJ-HGGX42 (PDOX) model of DHG, more than 96% of tumor cells coexpressed both B7-H3 and GD2, highlighting their strong potential as dual immunotherapeutic targets.28

The selective cytotoxicity of B7-H3 CAR T cells may be attributed to a threshold-dependent activation mechanism, wherein the relatively low antigen density on normal tissues remains insufficient to trigger CAR T cell activation, thereby preserving healthy cells while targeting tumor cells with high B7-H3 expression.24,25,27–29 Our data showed minimal expression of B7-H3 in nonmalignant brain tissue from DHG patients (Figure 1A). Furthermore, our experiments using B7-H3KO cell lines confirmed the antigen specificity of the therapeutic effect, as B7-H3 CAR T cells failed to elicit cytotoxicity or cytokine production in the absence of B7-H3 expression (Figure 4, Figure S11).This is a clinically relevant observation that underscores the critical importance of selecting tumor-restricted targets to minimize off-tumor effects and enhance the safety profile of CAR T cell therapies in the CNS. Most of the B7-H3 targeted CARs have used an scFv sequence derived from either MGA271 or 8H9 monoclonal antibodies.44,45 Notably, we performed a direct comparison of 2 clinically relevant B7-H3 CAR constructs (8H9 and MGA271), demonstrating comparable efficacy with distinct functional profiles (Figure S3). Given the clinical use of MGA271 in DIPG, these findings provide translational insight and support the selection of 8H9 for subsequent in vitro and in vivo studies. The 8H9 monoclonal antibody has been extensively studied and safely administered in pediatric patients for over a decade.45 Delivered directly into the CNS as a radio-conjugate, 8H9 has shown clinical promise in treating B7-H3-expressing malignancies.45 Its specificity and safety profile underscore its translational potential in patients with DHG. In parallel, MGA271 exhibits minimal binding to normal tissues and has demonstrated a favorable safety profile in early phase clinical trials.46 Indeed, recently, Vitanza et al showed a favorable safety profile and clinical activity of B7-H3 CAR T cells in early phase trials, in DIPG and DMG patients.24,25

B7-H3 CAR T cells were well tolerated, as we did not observe any evidence of toxicity after ICT delivery of B7-H3 CAR T cells in both DHG in vivo models (Figure S17). Our findings support the therapeutic advantages of local-regional intracranial delivery of B7-H3 CAR T cells, a strategy that bypasses the BBB and potentially minimizes systemic toxicity. Thus, preclinical studies suggest superior efficacy of locoregional delivery for B7-H3 CAR T cells and have demonstrated tumor regression in some models.47,48 Recent pediatric trials, including incorporated locoregional delivery, showed enhanced tumor penetration and reduced risks of cytokine release syndrome and neurotoxicity, especially in DMG patients with preexisting neurological deficits.22,24 While locoregional delivery may reduce the incidence of systemic cytokine release syndrome, it does not eliminate the risk of neurotoxicity. Tumor inflammation-associated neurotoxicity and localized inflammatory effects remain important considerations following intracranial CAR T cell administration. In our models, comprehensive toxicity monitoring revealed no behavioral or neurological deficits following treatment, and histologic analysis demonstrated reduced microglial activation, consistent with resolution of tumor-associated inflammation rather than treatment-induced neurotoxicity (Figure 5G and H, Figure S14). In addition, serum cytokine profiling showed no evidence of severe systemic inflammatory responses (Figure S15D), supporting a favorable safety profile of B7-H3 CAR T therapy with locoregional delivery.19,20,28,49 The optimal route of CAR T cell delivery may depend on both anatomical tumor location and spatial distribution of the target antigen. Tumors located in surgically accessible cortical regions may be well suited for locoregional administration, whereas deep-seated or multifocal lesions may require alternative strategies. Importantly, antigen heterogeneity across primary and disseminated tumor sites can influence CAR T engagement and distribution, thereby impacting the effectiveness of localized vs systemic delivery approaches. Although we did not directly compare systemic vs intracranial administration, prior preclinical and clinical studies have demonstrated superior tumor control and CNS persistence with locoregional CAR T delivery in brain tumors.19,28 Given the cortical accessibility of DHG and emerging clinical evidence supporting intracranial approaches, we prioritized locoregional administration to enhance translational relevance. Consistent with prior preclinical and clinical studies,19,20,28,49 our findings support the therapeutic advantages of local-regional intracranial delivery of B7-H3 CAR T cells by enabling direct tumor targeting while bypassing the BBB. This approach achieved durable tumor clearance with minimal toxicity, likely driven by enhanced local effector activity and limited systemic exposure. Collaborative efforts are needed to optimize delivery routes and dosing regimens for B7-H3 targeted CAR T cell immunotherapies in pediatric CNS tumors.

Antigen heterogeneity in brain tumors like DIPG and DHG contributes to CAR T cell resistance and tumor escape. Importantly, antigen escape was not observed in our DHG model, as durable tumor clearance occurred without recurrence, and no residual B7-H3-positive tumor was detected. Although transient antigen modulation was seen in vitro (Figure S18B). Notably, in the few nonresponding animals, mice 10 and 15 showed persistent B7-H3 expression, indicating that resistance was unlikely due to antigen loss but may instead reflect technical variability in intratumoral CAR T cell delivery (Figure S18A). To counter the antigen heterogeneity, multiplexed CAR T strategies targeting EGFR, HER2, and IL13Rα2 are being developed. Previous studies have shown high expression of B7-H3 and GD2 that dual or bispecific “OR”-gated CARs enhance antigen coverage but may increase off-tumor toxicity risks, whereas “AND”-gated designs, such as SynNotch, improve specificity and enhance CAR T cell persistence and trafficking.50

While B7-H3 CAR T cells demonstrated strong activation and cytokine production upon target engagement, they also exhibited upregulation of exhaustion markers (PD-1,TIM-3, LAG-3), suggesting that repetitive antigen stimulation within the immunosuppressive tumor microenvironment may lead to functional T cell exhaustion. Future studies should explore combination strategies to enhance CAR T cell persistence and function, such as checkpoint blockade, cytokine support, or targeting the immunosuppressive myeloid-rich microenvironment (eg CD14+/CD163+ macrophages) may further improve efficacy and reduce relapse.

Our findings align with and extend recent clinical efforts investigating GD2 and B7-H3-directed therapies in pediatric brain tumors, such as DIPG, where early phase trials have demonstrated feasibility and preliminary efficacy. However, DHGs remain underrepresented in preclinical and clinical studies, and our work addresses this critical gap. By demonstrating B7-H3 expression in DHG and the efficacy of B7-H3 CAR T cells, we provide a strong preclinical rationale for extending CAR T cell-based immunotherapies to this high-risk patient population.

In conclusion, this study establishes B7-H3 as a compelling immunotherapeutic target in DHG and demonstrates that B7-H3 CAR T cells elicit potent, antigen-specific antitumor responses in both in vitro and in vivo models. These results lay the foundation for clinical translation of B7-H3-directed CAR T cell therapies in H3G34R/V-mutated DHGs and support the development of early phase clinical trials to evaluate their safety and efficacy in pediatric patients with these otherwise intractable tumors.

Supplementary Material

Supplemental figures

Supplementary material is available online at Neuro-Oncology (https://academic.oup.com/neuro-oncology).

Key Points.

  • B7-H3 is overexpressed in diffuse hemispheric glioma (DHG), highlighting its potential as a selective target for chimeric antigen receptor (CAR) T cell immunotherapy.

  • Anti-B7-H3 CAR-T cells demonstrate robust antitumor activity, resulting in a significant improvement in overall survival in multiple in vivo DHG models.

Importance of the Study.

This study represents an advancement in the treatment of diffuse hemispheric gliomas (DHGs) with H3G34R/V mutations, an aggressive pediatric brain tumor with no curative therapies. For the first time, chimeric antigen receptor (CAR) T-cell therapy was evaluated against H3G34R/V DHG tumors. Key findings include significant B7-H3 expression on DHG tumor models and the development of anti-B7-H3-specific CAR T cells, which demonstrated antigen-specific cytokine release, immune activation, and cytotoxicity in vitro. In multiple orthotopic mouse models of H3G34R/V-mutated DHG, these CAR T cells reduced tumor growth and significantly improved mouse survival. This study highlights the potential of B7-H3 as a promising therapeutic target. It provides a strong preclinical rationale for initiating the first anti-B7-H3 CAR T cell-mediated phase 1 clinical trial for DHG patients. It sets the stage for translating these preclinical improvements into clinical applications, offering hope for a previously untreatable disease.

Funding

This work was supported by the Noah’s Bandage project grant (AWD00000727) (to V.N.Y.), the National Institute of Health (NIH)/National Cancer Institute (NCI) 1R01CA284065 (to D.A. and S.H.B.), the Kansas Institute of Precision Medicine (to D.A.), Head for the Cure Foundation (to D.A.), (NIH)/(NCI)P30CA168524 (PI: Roy Jensen, MD) (to D.A., S.H.B., and K.P.), the NIGMS Kansas Institute of Precision Medicine P20 GM130423 (to K.P.), the Nora’s Sparkle of Love Project (to K.P. and V.N.Y.), the Braden’s Hope Research grant (AWD00000907) (to K.P. and V.N.Y.), and the Masonic Cancer Alliance (AWD00000851) (to V.N.Y.).

Footnotes

Conflict of Interest Statement

The authors declare that they have no potential conflicts of interest.

Data Availability

The data, reagents, and resources described in this manuscript will be made available to the research community by the corresponding author upon reasonable request. Transfers of research materials will be facilitated under a Materials Transfer Agreement (MTA). Figure 1B was generated by reanalyzing publicly available RNA-seq data from the Pediatric Brain Tumor Atlas via PedcBioPortal and normal brain tissue from the (GTEx) project v8. (See Supplementary Material S1 for details of the analysis.)

References

  • 1.Ostrom QT, Price M, Ryan K, et al. CBTRUS statistical report: pediatric brain tumor foundation childhood and adolescent primary brain and other central nervous system tumors diagnosed in the United States in 2014-2018. Neuro Oncol. 2022;24:iii1–iii38. 10.1093/neuonc/noac161 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Jones C, Karajannis MA, Jones DTW, et al. Pediatric high-grade glioma: biologically and clinically in need of new thinking. Neuro Oncol. 2017;19:153–161. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Louis DN, Perry A, Wesseling P, et al. The 2021 WHO classification of tumors of the central nervous system: a summary. Neuro Oncol. 2021;23:1231–1251. 10.1093/neuonc/noab106 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Mackay A, Burford A, Carvalho D, et al. Integrated molecular meta-analysis of 1,000 pediatric high-grade and diffuse intrinsic pontine glioma. Cancer Cell. 2017;32:520–537.e525. 10.1016/j.ccell.2017.08.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Buczkowicz P, Hoeman C, Rakopoulos P, et al. Genomic analysis of diffuse intrinsic pontine gliomas identifies three molecular subgroups and recurrent activating ACVR1 mutations. Nat Genet. 2014;46:451–456. 10.1038/ng.2936 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Funato K, Smith RC, Saito Y, Tabar V. Dissecting the impact of regional identity and the oncogenic role of human-specific NOTCH2NL in an hESC model of H3.3G34R-mutant glioma. Cell Stem Cell. 2021;28:894–905. e897. 10.1016/j.stem.2021.02.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Schwartzentruber J, Korshunov A, Liu XY, et al. Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma. Nature. 2012;482:226–231. 10.1038/nature10833 [DOI] [PubMed] [Google Scholar]
  • 8.Chen CCL, Deshmukh S, Jessa S, et al. Histone H3.3G34-mutant interneuron progenitors co-opt PDGFRA for gliomagenesis. Cell. 2020;183:1617–1633.e1622. 10.1016/j.cell.2020.11.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Haase S, Banerjee K, Mujeeb AA, et al. H3.3-G34 mutations impair DNA repair and promote cGAS/STING-mediated immune responses in pediatric high-grade glioma models. J Clin Invest. 2022;132:e154229. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Sweha SR, Chung C, Natarajan SK, et al. Epigenetically defined therapeutic targeting in H3.3G34R/V high-grade gliomas. Sci Transl Med. 2021;13:eabf7860. 10.1126/scitranslmed.abf7860 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Maude SL, Laetsch TW, Buechner J, et al. Tisagenlecleucel in children and young adults with B-cell lymphoblastic leukemia. N Engl J Med. 2018;378:439–448. 10.1056/NEJMoa1709866 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Porter DL, Hwang WT, Frey NV, et al. Chimeric antigen receptor T cells persist and induce sustained remissions in relapsed refractory chronic lymphocytic leukemia. Sci Transl Med. 2015;7:303ra139. 10.1126/scitranslmed.aac5415 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Schuster SJ, Svoboda J, Chong EA, et al. Chimeric antigen receptor T cells in refractory B-cell lymphomas. N Engl J Med. 2017;377:2545–2554. 10.1056/NEJMoa1708566 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Park S, Maus MV, Choi BD. CAR-T cell therapy for the treatment of adult high-grade gliomas. NPJ Precis Oncol. 2024;8:279. 10.1038/s41698-024-00753-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Thomas BC, Staudt DE, Douglas AM, Monje M, Vitanza NA, Dun MD. CAR T cell therapies for diffuse midline glioma. Trends Cancer. 2023;9:791–804. 10.1016/j.trecan.2023.07.007 [DOI] [PubMed] [Google Scholar]
  • 16.Gust J, Cole BL, Ronsley R, et al. Locoregional infusion of EGFR806-CAR T cells for recurrent or refractory pediatric CNS tumors: results of the completed BrainChild02 phase 1 clinical trial. Neuro Oncol. 2025;27:2170–2181. 10.1093/neuonc/noaf064 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Ravanpay AC, Gust J, Johnson AJ, et al. EGFR806-CAR T cells selectively target a tumor-restricted EGFR epitope in glioblastoma. Oncotarget. 2019;10:7080–7095. 10.18632/oncotarget.27389 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Bagley SJ, Logun M, Fraietta JA, et al. Intrathecal bivalent CAR T cells targeting EGFR and IL13Ralpha2 in recurrent glioblastoma: phase 1 trial interim results. Nat Med. 2024;30:1320–1329. 10.1038/s41591-024-02893-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Brown CE, Hibbard JC, Alizadeh D, et al. Locoregional delivery of IL-13Ralpha2-targeting CAR-T cells in recurrent high-grade glioma: a phase 1 trial. Nat Med. 2024;30:1001–1012. 10.1038/s41591-024-02875-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Vitanza NA, Johnson AJ, Wilson AL, et al. Locoregional infusion of HER2-specific CAR T cells in children and young adults with recurrent or refractory CNS tumors: an interim analysis. Nat Med. 2021;27:1544–1552. 10.1038/s41591-021-01404-8 [DOI] [PubMed] [Google Scholar]
  • 21.Majzner RG, Ramakrishna S, Yeom KW, et al. GD2-CAR T cell therapy for H3K27M-mutated diffuse midline gliomas. Nature. 2022;603:934–941. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Monje M, Mahdi J, Majzner R, et al. Intravenous and intracranial GD2-CAR T cells for H3K27M(+) diffuse midline gliomas. Nature. 2025;637:708–715. 10.1038/s41586-024-08171-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Mount CW, Majzner RG, Sundaresh S, et al. Potent antitumor efficacy of anti-GD2 CAR T cells in H3-K27M(+) diffuse midline gliomas. Nat Med. 2018;24:572–579. 10.1038/s41591-018-0006-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Vitanza NA, Ronsley R, Choe M, et al. Intracerebroventricular B7-H3-targeting CAR T cells for diffuse intrinsic pontine glioma: a phase 1 trial. Nat Med. 2025;31:861–868. 10.1038/s41591-024-03451-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Vitanza NA, Wilson AL, Huang W, et al. Intraventricular B7-H3 CAR T cells for diffuse intrinsic pontine glioma: preliminary first-in-human bioactivity and safety. Cancer Discov. 2023;13:114–131. 10.1158/2159-8290.CD-22-0750 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Zhou Z, Luther N, Ibrahim GM, et al. B7-H3, a potential therapeutic target, is expressed in diffuse intrinsic pontine glioma. J Neurooncol. 2013;111:257–264. 10.1007/s11060-012-1021-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Du H, Hirabayashi K, Ahn S, et al. Antitumor responses in the absence of toxicity in solid tumors by targeting B7-H3 via chimeric antigen receptor T cells. Cancer Cell. 2019;35:221–237.e228. 10.1016/j.ccell.2019.01.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Haydar D, Houke H, Chiang J, et al. Cell-surface antigen profiling of pediatric brain tumors: B7-H3 is consistently expressed and can be targeted via local or systemic CAR T-cell delivery. Neuro Oncol. 2021;23:999–1011. 10.1093/neuonc/noaa278 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Majzner RG, Theruvath JL, Nellan A, et al. CAR T cells targeting B7-H3, a pan-cancer antigen, demonstrate potent preclinical activity against pediatric solid tumors and brain tumors. Clin Cancer Res. 2019;25:2560–2574. 10.1158/1078-0432.CCR-18-0432 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Li S, Zhang M, Wang M, et al. B7-H3 specific CAR-T cells exhibit potent activity against prostate cancer. Cell Death Discov. 2023;9:147. 10.1038/s41420-023-01453-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Maachani UB, Tosi U, Pisapia DJ, et al. B7-H3 as a prognostic biomarker and therapeutic target in pediatric Central nervous system tumors. Transl Oncol. 2020;13:365–371. 10.1016/j.tranon.2019.11.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Picarda E, Ohaegbulam KC, Zang X. Molecular pathways: targeting B7-H3 (CD276) for human cancer immunotherapy. Clin Cancer Res. 2016;22:3425–3431. 10.1158/1078-0432.CCR-15-2428 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Bankhead P, Loughrey MB, Fernandez JA, et al. QuPath: open source software for digital pathology image analysis. Sci Rep. 2017;7:16878. 10.1038/s41598-017-17204-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Uwe Schmidt MW, Broaddus C, Myers G. Medical Image Computing and Computer Assisted Intervention – MICCAI 2018; 2018, 265–273. 10.1007/978-3-030-00934-2_30 [DOI] [Google Scholar]
  • 35.Xiong J, Kaur H, Heiser CN, et al. GammaGateR: semi-automated marker gating for single-cell multiplexed imaging. Bioinformatics. 2024;40:btae356. 10.1093/bioinformatics/btae356 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Jonnalagadda M, Mardiros A, Urak R, et al. Chimeric antigen receptors with mutated IgG4 Fc spacer avoid fc receptor binding and improve T cell persistence and antitumor efficacy. Mol Ther. 2015;23:757–768. 10.1038/mt.2014.208 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Subham S, Jeppson JD, Worcester C, et al. EGFR as a potent CAR T target in triple negative breast cancer brain metastases. Breast Cancer Res Treat. 2023;197:57–69. 10.1007/s10549-022-06783-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Kawalekar OU, O’Connor RS, Fraietta JA, et al. Distinct signaling of coreceptors regulates specific metabolism pathways and impacts memory development in CAR T cells. Immunity. 2016;44:380–390. 10.1016/j.immuni.2016.01.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Long AH, Haso WM, Shern JF, et al. 4-1BB costimulation ameliorates T cell exhaustion induced by tonic signaling of chimeric antigen receptors. Nat Med. 2015;21:581–590. 10.1038/nm.3838 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Milone MC, Fish JD, Carpenito C, et al. Chimeric receptors containing CD137 signal transduction domains mediate enhanced survival of T cells and increased antileukemic efficacy in vivo. Mol Ther. 2009;17:1453–1464. 10.1038/mt.2009.83 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Zhao X, Yang J, Zhang X, et al. Efficacy and safety of CD28-or 4-1BB-based CD19 CAR-T cells in B cell acute lymphoblastic leukemia. Mol Ther Oncolytics. 2020;18:272–281. 10.1016/j.omto.2020.06.016 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Caruso HG, Hurton LV, Najjar A, et al. Tuning sensitivity of CAR to EGFR density limits recognition of normal tissue while maintaining potent antitumor activity. Cancer Res. 2015;75:3505–3518. 10.1158/0008-5472.CAN-15-0139 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.O’Rourke DM, Nasrallah MP, Desai A, et al. A single dose of peripherally infused EGFRvIII-directed CAR T cells mediates antigen loss and induces adaptive resistance in patients with recurrent glioblastoma. Sci Transl Med. 2017;9:eaaa0984. 10.1126/scitranslmed.aaa0984 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Modak S, Kramer K, Gultekin SH, Guo HF, Cheung NK. Monoclonal antibody 8H9 targets a novel cell surface antigen expressed by a wide spectrum of human solid tumors. Cancer Res. 2001;61:4048–4054. [PubMed] [Google Scholar]
  • 45.Kramer K, Smith M, Souweidane MM. Safety profile of long-term intraventricular access devices in pediatric patients receiving radioimmunotherapy for central nervous system malignancies. Pediatr Blood Cancer. 2014;61:1590–1592. 10.1002/pbc.25080 [DOI] [PubMed] [Google Scholar]
  • 46.Aggarwal C, Prawira A, Antonia S, et al. Dual checkpoint targeting of B7-H3 and PD-1 with enoblituzumab and pembrolizumab in advanced solid tumors: interim results from a multicenter phase I/II trial. J Immunother Cancer. 2022;10:e004424. 10.1136/jitc-2021-004424 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Donovan LK, Delaidelli A, Joseph SK, et al. Locoregional delivery of CAR T cells to the cerebrospinal fluid for treatment of metastatic medulloblastoma and ependymoma. Nat Med. 2020;26:720–731. 10.1038/s41591-020-0827-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Theruvath J, Sotillo E, Mount CW, et al. Locoregionally administered B7-H3-targeted CAR T cells for treatment of atypical teratoid/rhabdoid tumors. Nat Med. 2020;26:712–719. 10.1038/s41591-020-0821-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Wang H, Xu T, Huang Q, Jin W, Chen J. Immunotherapy for malignant glioma: current status and future directions. Trends Pharmacol Sci. 2020;41:123–138. 10.1016/j.tips.2019.12.003 [DOI] [PubMed] [Google Scholar]
  • 50.Jin L, Tao H, Karachi A, et al. CXCR1- or CXCR2-modified CAR T cells co-opt IL-8 for maximal antitumor efficacy in solid tumors. Nat Commun. 2019;10:4016. 10.1038/s41467-019-11869-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental figures
NIHMS2210348-supplement-Supplemental_figures.docx (8.6MB, docx)

Data Availability Statement

The data, reagents, and resources described in this manuscript will be made available to the research community by the corresponding author upon reasonable request. Transfers of research materials will be facilitated under a Materials Transfer Agreement (MTA). Figure 1B was generated by reanalyzing publicly available RNA-seq data from the Pediatric Brain Tumor Atlas via PedcBioPortal and normal brain tissue from the (GTEx) project v8. (See Supplementary Material S1 for details of the analysis.)

RESOURCES